Earth SciencesPROFESSIONAL SCIENCE SAMPLER

INCLUDING

Chapter 1: The Changing El Niño–Southern Oscillation and Associated Climate ExtremesFrom Climate Extremes: Patterns and Mechanisms, Geophysical Monograph 226, First Edition.Edited by S.-Y. Simon Wang, Jin-Ho Yoon, Christopher C. Funk, and Robert R. Gillies

Chapter 1: Microstructural, Geomechanical, and Petrophysical Characterization of Shale CaprocksFrom Geological Carbon Storage: Subsurface Seals and Caprock Integrity, Geophysical Monograph 238, First Edition.Edited by Stéphanie Vialle, Jonathan Ajo-Franklin, and J. William Carey

Chapter 1: The Need for the Nexus ApproachFrom Water-Energy-Food Nexus: Principles and Practices, Geophysical Monograph 229, First Edition.Edited by P. Abdul Salam, Sangam Shrestha, Vishnu Prasad Pandey, and Anil Kumar Anal

Chapter 1: On the Origin of the Upper Mantle Seismic DiscontinuitiesFrom Lithospheric Discontinuities, Geophysical Monograph 239, First Edition.Edited by Huaiyu Yuan and Barbara Romanowicz

Chapter 1: Bioenergy and Land Use Change: An OverviewFrom Bioenergy and Land Use Change, Geophysical Monograph 231, First Edition.Edited by Zhangcai Qin, Umakant Mishra, and Astley Hastings

Chapter 1: Chemostratigraphy as a Formal Stratigraphic MethodFrom Chemostratigraphy Across Major Chronological Boundaries, Geophysical Monograph 240, First Edition.Edited by Alcides N. Sial, Claudio Gaucher, Muthuvairavasamy Ramkumar, and Valderez Pinto Ferreira

CONTENTS

Chapter 1: The Changing El Niño–Southern Oscillation and Associated Climate ExtremesFrom Climate Extremes: Patterns and Mechanisms, Geophysical Monograph 226, First Edition.

Chapter 1: Microstructural, Geomechanical, and Petrophysical Characterization of Shale CaprocksFrom Geological Carbon Storage: Subsurface Seals and Caprock Integrity, Geophysical Monograph 238, First Edition.

Chapter 1: The Need for the Nexus Approach From Water-Energy-Food Nexus: Principles and Practices, Geophysical Monograph 229, First Edition.

Chapter 1: On the Origin of the Upper Mantle Seismic DiscontinuitiesFrom Lithospheric Discontinuities, Geophysical Monograph 239, First Edition.

Chapter 1: Bioenergy and Land Use Change: An OverviewFrom Bioenergy and Land Use Change, Geophysical Monograph 231, First Edition.

Chapter 1: Chemostratigraphy as a Formal Stratigraphic MethodFrom Chemostratigraphy Across Major Chronological Boundaries, Geophysical Monograph 240, First Edition.

3

39

67

75

105

121

3

Climate Extremes: Patterns and Mechanisms, Geophysical Monograph 226, First Edition. Edited by S.-Y. Simon Wang, Jin-Ho Yoon, Christopher C. Funk, and Robert R. Gillies. © 2017 American Geophysical Union. Published 2017 by John Wiley & Sons, Inc.

1.1. INTRODUCTION

El Niño–Southern Oscillation (ENSO) is a prominent climate phenomenon in the tropical Pacific that can dis­rupt global atmospheric and oceanic circulation patterns and exert profound impacts on global climate and socio­economic activities. Since Bjerknes [1969] first recognized that ocean‐atmosphere coupling was a fundamental

characteristic of ENSO, a tremendous amount of effort has been expended by the research community to describe and understand the complex nature of this phenomenon and its underlying generation mechanisms. By the 1990s, the effort had led to the developments of successful theo­retical frameworks that could explain the major features observed during ENSO events [Neelin et  al., 1998] and useful forecast systems had been formulated to predict ENSO evolution with significant lead times of up to three seasons [Latif et  al., 1998]. The typical ENSO impacts in  various parts of the globe had also been extensively examined and documented. Through these efforts, it was also recognized that ENSO properties were not the same among events and could change from one decade to another. The diversity of ENSO characteristics attracted renewed interest at the beginning of the 21st century (see Capotondi et al. [2015], for a summary of these efforts), when it became obvious that the central location of the

The Changing El Niño–Southern Oscillation and Associated Climate Extremes

Jin‐Yi Yu1, Xin Wang2, Song Yang3,4, Houk Paek1, and Mengyan Chen2

1

ABSTRACT

The El Niño–Southern Oscillation (ENSO) is one of the most powerful climate phenomena that produce p rofound global impacts. Extensive research since the 1970s has resulted in a theoretical framework capable of explaining the observed properties and impacts of the ENSO and predictive models. However, during the most recent two decades there have been significant changes observed in the properties of ENSO that suggest r evisions are required in the existing theoretical framework developed primarily for the canonical ENSO. The observed changes include a shift in the location of maximum sea surface temperature variability, an increased importance in the underlying dynamics of coupled ocean‐atmosphere process in the subtropical Pacific, and different remote atmospheric teleconnection patterns that give rise to distinct climate extremes. The causes of these recent changes in ENSO are still a matter of debate but have been attributed to both global warming and natural climate variability involving interactions between the Pacific and Atlantic oceans. The possible future changes of ENSO properties have also been suggested using climate model projections.

1 Department of Earth System Science, University of California, Irvine, California, USA

2 State Key Laboratory of Tropical Oceanography, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou, Guangdong, China

3 School of Atmospheric Sciences, Sun Yat‐sen University, GuangZhou, Guangdong, China

4 Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies

0003060439.INDD 3 6/5/2017 7:05:43 PM

COPY

RIGH

TED

MATER

IAL

4 FORCINGS OF CLIMATE EXTREMES

sea surface temperature (SST) anomalies associated with ENSO appeared to be moving from the tropical eastern Pacific near the South American coast to locations near the International Dateline in the tropical central Pacific. Most of the El Niño events that have occurred so far in the 21st century developed primarily in the central Pacific [Lee and McPhaden, 2010; Yu et  al., 2012b; Yu et  al., 2015a; Yeh et al., 2015]. The fact that El Niño can some­times occur in the eastern Pacific, sometimes in the central Pacific, and sometimes simultaneously in both portions of the Pacific has suggested that there may exist more than one type of ENSO, whose generation mecha­nisms and associated climate extremes may be different. The changes in ENSO observed during the recent d ecades have motivated the research community to revisit the conventional views of ENSO properties, dynamics, and global impacts.

1.2. CHANGES IN ENSO PROPERTIES

1.2.1. Flavors of ENSO

Slow or interdecadal changes have been observed in ENSO properties, including its intensity, period, and propagation direction [Gu and Philander, 1995; Wang and Wang, 1996; Torrence and Webster, 1999; An and Wang, 2000; Fedorov and Philander, 2000; Wang and An, 2001; Timmermann, 2003; An and Jin, 2004; and many others]. The amplitude of ENSO‐associated SST anomalies, for example, was found to be stronger at the beginning and the end of the twentieth century, but weaker in between [Gu and Philander, 1995; Wang and Wang, 1996]. The propagation direction of ENSO SST anomalies has also alternated between eastward, westward, and standing during the past few decades [Wang and An, 2001; An and Jin, 2004]. Its recurrence frequency changed from about 2 to 3 yr before 1975 to a longer frequency of 4 to 6 yr after­ward [An and Wang, 2000]. In recent decades, one of the most noticeable changes in ENSO properties has been the displacement of the central location of ENSO SST anomalies [Larkin and Harrison, 2005; Ashok et al., 2007; Kao and Yu, 2009; Kug et al., 2009]. ENSO is character­ized by interannual SST variations in the equatorial eastern and central Pacific. In the canonical ENSO events portrayed by Rasmusson and Carpenter [1982], SST anomalies first occur near the South American coast and then spread westward along the equator. However, ENSO events characterized by SST anomalies primarily in the equatorial central Pacific, and which spread eastward [An and Wang, 2000] also occur. Trenberth and Stepaniak [2001] recognized that the different SST evolutions of ENSO events cannot be fully described with a single index such as the Niño‐3 (5°S–5°N, 90°–150°W) SST index. They defined a trans‐Niño index, which measures

the SST gradient along the equator by taking the differ­ence between normalized Niño‐1 + 2 (10°S–0°, 80°–90°W) and Niño‐4 (5°S–5°N, 160°E–150°W) SST indices to help discriminate the different SST evolutions. Their study implies that different types of ENSO may be better iden­tified by contrasting SST anomalies between the eastern and central Pacific. A similar conclusion was reached by Yu and Kao [2007] in their analysis of the persistence b arrier of Niño indices in the central‐to‐eastern Pacific (i.e., Niño‐1 + 2, Niño‐3, Niño‐3.4 (5°S–5°N, 120°–170°W), and Niño‐4 SST indices). They found different decadal changes between the indices in the equatorial central and eastern Pacific. They also found that the decadal changes in ocean heat content variations along the equatorial Pacific coincide with the decadal changes in the SST per­sistence barriers in the eastern Pacific, but not with those in the central Pacific. This finding led them to suggest that there are two types of ENSO: An Eastern‐Pacific (EP) type located primarily in the tropical eastern Pacific and whose generation mechanism involves equatorial thermocline variations, and a Central‐Pacific (CP) type located in the central tropical Pacific and whose genera­tion is less sensitive to thermocline variations. A detailed comparison of these two types of ENSO was presented in Kao and Yu [2009], where they identified the spatial struc­ture, temporal evolution, and underlying dynamics of these two types of events. While this “two types of ENSO” point of view is adopted in this chapter, it should be noted that debate remains (for example, see the discus­sion in Capotondi et al. [2015]) as to whether these two types are really dynamically distinct.

Examples of the EP and CP types of ENSO are shown in Figure 1.1, which displays the SST anomaly patterns during the peak phase of the 1977–1978 and 1997–1998 El Niño events. During the 1997–1998 El Niño (Fig. 1.1a), SST anomalies were mostly located in the eastern part of the tropical Pacific, extending from the South American coast around 80°W to 160°W, covering the Niño‐1 + 2 and Niño‐3 regions. During the 1977–1978 El Niño (Fig. 1.1b), SST anomalies were mostly concentrated in the equatorial central Pacific from 160°E to 120°W, cov­ering the Niño‐3.4 and Niño‐4 regions. Some early stud­ies had already noticed the existence of a group of ENSO events that developed in the central Pacific around the International Dateline [e.g., Weare et al., 1976; Fu et al., 1986; Hoerling and Kumar, 2002; Larkin and Harrison, 2005a]. Larkin and Harrison [2005a] suggested that this group of ENSO events can produce different impacts on the US climate from conventional ENSO events and referred to them as the Dateline El Niño. Ashok et  al. [2007] also focused on this group of ENSO events and termed them El Niño Modoki. Furthermore, Wang and Wang [2013] classified El Niño Modoki into two sub­types: El Niño Modoki I and El Niño Modoki II because

0003060439.INDD 4 6/5/2017 7:05:43 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 5

they show significantly different impacts on rainfall in southern China and typhoon landfall activity. These two subtypes of El Niño Modoki have distinct origins and patterns of SST anomalies during their developing phase. Kug et al. [2009] also named this ENSO‐type warm pool El Niño and the other type cold tongue El Niño.

1.2.2. Distinct Characteristics of the Eastern‐Pacific (EP) and Central‐Pacific (CP) Types of ENSO

Significant differences were found between the struc­ture and evolution of the EP and CP types of ENSO [Kao and Yu, 2009; Kug et al., 2009]. The typical SST anomaly patterns associated with these two types of ENSO are shown in Figure 1.1c and d, which are obtained by apply­ing a regression‐EOF (empirical orthogonal function) analysis [Kao and Yu,2009; Yu and Kim, 2010a] to tropical Pacific SST anomalies. In this method, SST anomalies regressed on the Niño‐1 + 2 SST index (representing the

EP ENSO influence) were first removed from the original monthly SST anomalies, and then an EOF analysis was applied to the residual SST anomalies to obtain the spa­tial pattern of the CP ENSO. Similarly, SST anomalies regressed on the Niño‐4 index (representing the CP ENSO influence) were first removed from the original SST anomalies, and then an EOF analysis was applied to  identify the leading structure of the EP ENSO. The principal components (PCs) associated with these pat­terns are referred to, respectively, as the CP ENSO index and the EP ENSO index. The EOF patterns show that the EP ENSO‐related SST anomalies are mostly located in the eastern equatorial Pacific and connected to the coast of South America (Fig. 1.1c), while the CP ENSO has  most of its SST anomalies confined in the central Pacific with extension into the northeastern subtropical Pacific (Fig. 1.1d). Although other studies also proposed different identification methodologies for separating the two types of ENSO, including weighted averaged SST

120°E30°S

15°S

0°

15°N

30°N

150°E 150°W 120°W 90°W180°

Nov 1997–Jan 1998(a) (b)

(c) (d)Eastern–Pacific (EP) ENSO Central–Pacific (CP) ENSO

120°E30°S

15°S

0°

15°N

30°N

150°E 150°W 120°W 90°W180°

120°E30°S

15°S

0°

15°N

30°N

150°E 150°W 120°W 90°W180° 120°E30°S

15°S

0°

15°N

30°N

150°E 150°W 120°W 90°W180°

Nov 1977–Jan 1978

Figure 1.1 SST anomalies during the peak phase (November–January) of (a) the 1997–1998 El Niño event and (b)  the 1977–1978 El Niño event. The leading EOF (empirical orthogonal function) patterns obtained from a regression‐EOF analysis [Kao and Yu, 2009] are shown in (c) for the Eastern‐Pacific type of ENSO and in (d) for the Central‐Pacific type of ENSO. The contour intervals are 0.5 °C for (a), 0.3 °C for (b), and 0.1 °C for (c) and (d). The Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST) [Rayner et al., 2003] for the period 1948–2010 is used in these calculations.

0003060439.INDD 5 6/5/2017 7:05:44 PM

6 FORCINGS OF CLIMATE EXTREMES

anomalies in the tropical western, central, and eastern Pacific [Ashok et  al., 2007]; relative magnitudes of the Niño‐3 and Niño‐4 indices [Kug et al., 2009; Yeh et al., 2009]; a simple transformation of the Niño‐3 and Niño‐4 indices [Ren and Jin, 2011] or two PCs of tropical Pacific SST anomalies [Takahashi et al., 2011]; and mean upper ocean temperature anomalies in CP and EP regions [Yu et al., 2011], their corresponding SST anomaly patterns are similar to those shown in Figure 1.1c and d. The SST anomaly features revealed by the regression‐EOF analysis were observed in the 1977–1978 and 1997–1998 El Niño events as shown in Figure  1.1a and b. The meridional extension into the subtropics of the CP ENSO SST anomalies is an important feature, because it resembles the Pacific Meridional Mode (PMM) [Chiang and Vimont, 2004] that is the leading coupled variability mode of the subtropical Pacific. This connection indicates that the CP ENSO dynamic has a stronger tie to atmospheric and oceanic process outside the tropical Pacific than does the EP ENSO dynamic.

Below the ocean surface (Fig.  1.2), the EP ENSO is accompanied by significant subsurface temperature anomalies propagating across the Pacific basin, while the CP ENSO is associated only with subsurface ocean tem­perature anomalies that develop in situ in the central Pacific. The different subsurface evolution indicates that, in contrast to the EP ENSO, the underlying dynamics of the CP ENSO seem to be less dependent on thermocline variations [Kao and Yu, 2009; Kug et al., 2009], which is consistent with the suggestion of Yu and Kao [2007]. McPhaden [2008] already noticed that the variations in the equatorial Pacific Ocean heat content (i.e., the so‐called warm water volume; Meinen and McPhaden [2000]) tended to lead the ENSO SST anomalies by about two to three seasons in the 1980s and 1990s but by only one sea­son after 2000. McPhaden [2012] attributed this decrease to the shift of ENSO from the EP type to the CP type in recent decades. The decreased lead time implies a reduced impact of thermocline feedback to SST anomalies for the CP ENSO and may explain the reduced predictability observed in recent decades [Barnston et al., 2012] as this type of ENSO occurred more frequently. The weakened relationship between ENSO SST anomalies and the underlying thermocline variation indicates that the CP ENSO dynamic differs from the traditional ENSO dynamic, such as the delayed oscillator theory [Suarez and Schopf, 1988; Battisti and Hirst, 1989] or the recharged oscillator theory [e.g., Wyrtki 1975; Zebiak 1989; Jin 1997] that emphasize the subsurface ocean dynamics in governing ENSO SST evolution.

In the atmosphere, wind stress and precipitation anom­aly patterns associated with the two types of ENSO are also different. While the EP El Niño is associated with significant westerly anomalies over a large part of the

tropical Pacific, the westerly anomalies associated with the CP El Niño have a smaller spatial scale and are cen­tered in the equatorial central‐to‐western Pacific [Kao and Yu, 2009; Kug et al., 2009]. This more westward loca­tion of the westerly anomalies in the CP El Niño is con­sistent with the location of its SST anomalies. Significant easterly anomalies also appear over the tropical eastern Pacific during the CP El Niño. Theses anomalies were suggested to prevent the development of positive SST anomalies in this region during CP El Niño events [Ashok et al., 2007; Kug et al., 2009]. These differences in wind stress anomalies are important to the interpretation of the underlying dynamics of these two types of ENSO. Moreover, positive anomalies of precipitation associated with the EP El Niño typically extend from the equatorial eastern‐to‐central Pacific where the largest SST anoma­lies are located. For the CP El Niño, the precipitation anomalies are characterized by a dipole pattern, with positive anomalies located mainly in the western Pacific and negative anomalies in the eastern Pacific [Kao and Yu, 2009; Kug et  al., 2009]. The different precipitation patterns associated with these two types of ENSO imply that the locations of associated convective heating and the induced global teleconnections are different as well.

In addition to the central locations of their SST a nomalies, the two types of ENSO also differ in their evolution, amplitude, and recurrence frequency. The propagation of SST anomalies is weaker and less appar­ent in the CP type of ENSO than in the EP type of ENSO [Kao and Yu, 2009]. Also, there is no identifiable phase‐reversal signal in the composite analysis of the CP ENSO SST anomalies, implying that this type of ENSO does not behave like a cycle or oscillation as is often the case for the EP ENSO. Discussions of whether ENSO should be viewed as an event or a cycle [e.g., Kessler, 2002] might be aided by recognizing that there are two types of ENSO: one behaves more as an event and the other more as a cycle. Sun and Yu [2009] pointed out that the amplitude of the CP El Niño tends to be smaller than that of the EP El Niño, while the CP La Niña is often stronger than the EP La Niña. The reason for this difference in behavior between the El Niño and La Niña phases of these two types of ENSO has yet to be explained. Sun and Yu [2009], however, proposed an ENSO‐mean state inter­action mechanism to suggest that the amplitude asym­metries between the El Niño and La Niña phases of these two types of ENSO enable ENSO to modify the tropical Pacific mean state, which can give rise to a 10–15 yr m odulation cycle in ENSO amplitude. Strong El Niño events can occur every 10–15 yr due to this cycle, which is consistent with the time interval between the last three strongest El Niño events: the 1972–1973, 1982–1983, and 1997–1998 El Niño events. For the recurrence frequency, it is well known that ENSO events tend to occur every 2

0003060439.INDD 6 6/5/2017 7:05:44 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 7

120°E300

200

100

0

Dep

th (

m)

300

200

100

0

Dep

th (

m)

300

200

100

0

Dep

th (

m)

300

200

100

0

Dep

th (

m)

300

200

100

0

Dep

th (

m)

300

200

100

0

Dep

th (

m)

300

200

100

0EP CPLag (–6) Lag (–6)

Lag (–3)

0

0

0

0

2

2

22

2

2

0

0

00

0

0

0

0 0

00

0

00

02

2

4

0

0

0

2

24

2

2

4

42

2

2

2

24

4

2

2

20

4

4

6 2

2

2

0

00

0

4

4

2

2

2

2

2

2

2

2

2

2 2

2

2

22

2

24

4

4

4

4

2

2

2

2

Lag (–3)

Lag (0) Lag (0)

Lag (+3) Lag (+3)

Lag (+6) Lag (+6)

Lag (+9) Lag (+9)

Lag (+12) Lag (+12)

(a) (b)

Dep

th (

m)

150°E 150°W 120°W 90°W180° 120°E 150°E 150°W 120°W 90°W180°

Figure  1.2 Evolution of equatorial (5°S–5°N) subsurface ocean temperature anomalies associated with the (a) EP ENSO and (b) CP ENSO. Values shown are temperature anomalies regressed on the EP and CP ENSO i ndices from a lag of −6 mo to a lag of +12 mo. The contour interval is 0.2 °C. Ocean temperature assimilation from the German Estimating the Circulation and Climate of the Ocean (GECCO) [Kohl et al., 2006] for the period 1970–2001 is used.

0003060439.INDD 7 6/5/2017 7:05:46 PM

8 FORCINGS OF CLIMATE EXTREMES

to 7 yr. Spectral analyses have suggested that there are two dominant bands of ENSO variability: a quasibien­nial (~2 yr) band and a low‐frequency (3–7 yr) band [Rasmusson and Carpenter, 1982; Rasmusson et al., 1990; Barnett, 1991; Gu and Philander, 1995; Jiang et al., 1995; Wang and Wang, 1996]. The EP ENSO is dominated more by the low‐frequency band, while the CP ENSO is dominated by the quasibiennial band [Kao and Yu, 2009; Yu et al., 2010; Yu and Kim, 2010a]. As indicated above, El Niño events in the 21st century have been mostly of the CP type, and it is interesting to note that they occurred about every 2 yr in 2002, 2004, 2006, and 2009. In con­trast, the El Niño events observed during the 1970s and 1980s tended to occur every 4 yr. If model projections are correct about the more frequent occurrence of the CP ENSO in the future warmer world [Yeh et  al., 2009; Kim and Yu, 2012], we should expect El Niño events in the coming decades to become more frequent but weaker.

1.2.3. Possible Causes for the Change in ENSO Properties

The recent change in the ENSO type may be due to background state changes in the tropical Pacific that are either caused by anthropogenic warming [Yeh et al., 2009; Kim and Yu, 2012] or decadal/multidecadal natural vari­ability [McPhaden et al., 2011; Yeh et al., 2011]. Random fluctuations have also been suggested as possibly giving rise to the change in ENSO type on decadal timescales even without a change in the background state of the tropical Pacific [Newman et al., 2011].

1.2.3.1.  Anthropogenic Warming. Yeh et  al. [2009] and Kim and Yu [2012] used multiple model datasets of phases 3 and 5 of the Coupled Model Intercomparison Project (CMIP3 and CMIP5) [Meehl et al., 2007; Taylor et  al., 2012] to examine the influence of anthropogenic warming on the ENSO type. Yeh et  al. [2009] analyzed CMIP3 model projections to show that anthropogenic warming can flatten the equatorial Pacific thermocline (due to a weakened Walker circulation) resulting in a shift of the thermocline feedback mechanism from eastern to central Pacific and increasing the occurrence of ENSO in the central Pacific. They suggested that the same mecha­nism may be at work in recent decades to increase the occurrence frequency of the CP ENSO. Kim and Yu [2012] compared the amplitude ratio of these two types of ENSO in CMIP5 models. They found that the intensity of the CP ENSO increases steadily from the preindustrial to the his­torical simulations and the future projections, but the intensity of the EP ENSO decreases. The CP‐to‐EP ENSO intensity ratio, as a result, is projected to increase.

This global warming view has been challenged on two fronts. First, Lee and McPhaden [2010] found that not

only has the occurrence of the CP El Niño increased in recent decades but its intensity has also significantly increased. However, the intensity of CP La Niña events has not changed during the same period. As a result, they argued, the warming trend observed in the central Pacific is a result of the increasing frequency and intensity of the CP El Niño, rather than the other way around. Second, McPhaden et  al. [2011] found the Pacific background state during the most recent decade (2000–2010) has a steeper equatorial thermocline and stronger trade winds than during earlier decades (1980–1999), which are oppo­site from the background states changes expected by Yeh et al. [2009] should anthropogenic warming be the cause for the change in the ENSO type. This inconsistency implies that the change in ENSO type may be a manifes­tation of natural variability. McPhaden et al. [2011] also interpreted the Pacific background state during 2000–2010 to be a result, rather than a cause, of the shift of the ENSO from the EP type to the CP type. Their argument is that change in the central location of the ENSO SST anomalies can project onto the changes in the back­ground states.

1.2.3.2. ENSO–Mean State Interaction. Earlier studies have suggested that ENSO can force changes in the tropi­cal Pacific background state [e.g., Rodgers et  al., 2004; Schopf and Burgman, 2006; Sun and Yu, 2009; Choi et al., 2009]. Due to nonlinearities in ENSO dynamics, it is known that asymmetries exist in the intensity, frequency, or spatial structures of the warm (El Niño) and cold (La  Niña) phases of the cycle. For example, strong El Niño events tend to reach a larger intensity than strong La Niña events [Hannachi et al., 2003; An and Jin, 2004; Duan et al., 2008; Frauen and Dommenget, 2010; Su et al., 2010]. Also, strong El Niño events tend to be located more in the eastern Pacific while strong La Niña events tend to be located more in the central Pacific [Monahan, 2001; Hsieh, 2004; Rodgers et  al., 2004; Schopf and Burgman, 2006; Sun and Yu, 2009]. Because of these asymmetries, the positive and negative anomalies associ­ated with the El Niño and La Niña phases may produce a nonzero residual effect on the time‐mean state of the tropical Pacific. Sun and Zhang [2006] also conducted numerical experiments using a coupled model with and without ENSO to show that ENSO works as a basin‐scale mixer on the time‐mean thermal stratification in the upper equatorial Pacific. The mean state change associ­ated with decadal ENSO variability is one of the leading modes of tropical Pacific decadal variability in many coupled atmosphere‐ocean models [Rodgers et al., 2004; Yeh and Kirtman, 2004; Yu and Kim, 2011b].

Sun and Yu [2009] suggested an ENSO–mean state inter­action mechanism to explain how the ENSO type may shift between the EP and CP types on decadal timescales.

0003060439.INDD 8 6/5/2017 7:05:46 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 9

A key element in their hypothesized mechanism is that strong El Niño events tend to be of the EP type, whereas strong La Niña events tend to be of the CP type. In con­trast, weak El Niño events tend to be of the CP type, whereas weak La Niña events tend to be of the EP type. They verified this reversed spatial asymmetry characteris­tic using observed ENSO events during the 1880–2006 period in two reconstructed SST datasets. During strong ENSO periods or decades, strong El Niño events are located in the eastern tropical Pacific and strong La Niña events in the central tropical Pacific. This spatial asym­metry between El Niño and La Niña can gradually reduce the background east‐west SST gradient to weaken the ocean‐atmosphere coupling and to shift the background state toward one favoring a weak ENSO regime. Conversely, during the weak ENSO periods or decades, weak El Niño events are located in the eastern tropical Pacific and weak La Niña events in the central tropical Pacific, which can gradually increase the east‐west SST gradient, strengthen the ocean‐atmosphere coupling, and shift the background state toward one favoring a strong ENSO regime. The reversed spatial asymmetries between El Niño and La Niña enable ENSO to force the tropical Pacific mean state in opposite ways between the strong and weak ENSO regimes, acting as a restoring force to push the mean state back and forth between states that sustain strong and weak ENSO intensities. ENSO can also alternate between the EP and CP types on decadal timescales. Choi et al. [2012] have noted such an ENSO–mean state interaction mechanism and its associations with the ENSO types and Pacific background states in coupled model simulations.

1.2.3.3.  The Pacific Decadal Oscillation. Decadal changes in ENSO properties can also be associated with forcing external to the tropical Pacific, such as that related to slow SST oscillations in the North Pacific, Atlantic, or  Indian oceans [e.g., Gu and Philander, 1997; Behera and Yamagata, 2003; Verdon and Franks, 2006; Luo et al., 2005; Dong et  al., 2006]. One of the most well‐known forcing of this kind is related to the Pacific Decadal Oscillation (PDO) [Zhang et  al., 1997; Mantua et  al., 1997], which is a leading mode of decadal SST variability in the North Pacific. Earlier studies considered the PDO a physical mode of climate variability that results from atmosphere‐ocean coupling within the North Pacific [e.g., Latif and Barnet, 1994; Alexander and Deser, 1995]. More recent views consider it not a single physical mode but a combination of several processes operating on vari­ous timescales (see a review by Newman et al. [2015]). It has been suggested that the PDO is capable of modulat­ing the ENSO properties by influencing the tropical Pacific background states through ocean subduction, thermocline ventilation, or atmospheric teleconnections

[e.g., McCreary and Lu, 1994; Gu and Philander, 1995; Wang and Weisberg, 1998; Zhang et  al., 1998; Barnett et  al., 1999; Pierce et  al., 2000; Chang et  al., 2001; McPhaden and Zhang, 2002]. During the period of relia­ble instrumental records, the PDO shifted its phase twice: from a negative to a positive phase around 1976–1977 and back to a negative phase around 1999–2000. Along with the 1976–1977 phase shift, the background SSTs increased in the eastern and central tropical Pacific. This background state change was accompanied by noticeable changes in ENSO frequency and intensity [e.g., Trenberth and Hoar, 1996; Rajagopolan et  al., 1997]. Specifically, ENSO amplitude increased after the shift [An and Wang, 2000], and the ENSO frequency increased from being dominated by a quasi‐2 yr band to a quasi‐4 yr band [Wang and Wang, 1996]. In addition, ENSO SST anoma­lies were found to propagate westward before the shift but eastward after the shift [Wang and An, 2001], and there were more frequent El Niño events and fewer La Niña events after the shift [Trenberth and Hoar, 1996]. However, there was no noticeable change in the ENSO type across the 1976–1977 climate shift.

The 1999–2000 phase shift of the PDO is closer to the time that the occurrence frequency of the CP ENSO began to increase, which was suggested to be sometime around the 1980s [Ashok et al., 2007], the 1990s [Yu et al., 2012a], or the beginning of the 21st century [Lee and McPhaden, 2010; McPhaden et  al., 2011]. Associated with this PDO phase shift, a late‐1990s climate shift has also been suggested to occur in the Pacific Rim [e.g., Minobe, 2000; Peterson and Schwing, 2003]. Dramatic changes around 1999 were reported in the North Pacific sea level pressure (SLP), SST and sea level patterns [Minobe, 2000; Cummins et al., 2005], and in the north­eastern Pacific ecosystems [Peterson and Schwing, 2003]. In the western North Pacific, tropical cyclone activity abruptly increased after 2000 [Tu et al., 2009; He et al., 2015]. Over the neighboring continents, precipitation pat­terns changed after the late 1990s over North and South America [Huang et al., 2005] and over China [Zhu et al., 2011; Xu et al., 2015]. Surface temperatures in the Tibetan Plateau suddenly increased during the late 1990s [Xu et al., 2009]. A La Niña‐like background pattern appeared in the tropical Pacific [e.g., McPhaden et al., 2011; Chung and Li, 2013; Hu et al., 2013; Xiang et al., 2013; Kumar and Hu, 2014], characterized by a steeper equatorial ther­mocline slope [McPhaden et al., 2011] and strengthened Walker circulation [Chung and Li, 2013; Dong and Lu, 2013]. The breakdown of the subsurface ocean variabil­ity–ENSO relationship [McPhaden, 2008, 2012; Horri et al., 2012] and a coherent weakening of tropical ocean‐atmosphere variability [Hu et  al., 2013] after 2000 are likely associated with the decreases in ENSO forecast skill since 1999–2000 [Barnston et  al., 2012; Xue et  al.,

0003060439.INDD 9 6/5/2017 7:05:46 PM

10 FORCINGS OF CLIMATE EXTREMES

2013]. Beyond the Pacific, changes in the SST and sea ice variability over the Bering and Chukchi seas [Minobe, 2002; Yeo et  al., 2014] and in the southern ocean SST variability modes [Yeo and Kim, 2015] have also been found after the late 1990s.

1.2.3.4.  The Atlantic Multidecadal Oscillation. In addition to the PDO‐related late‐1990s climate shift, there is evidence that the Pacific also experienced a basin­wide climate shift in the early‐1990s that can also be responsible for the change in the type of ENSO. As shown in Figure 1.1, the EP ENSO has an SST anomaly pattern that spans the tropical eastern and central Pacific. During this type of ENSO, SST variations in the central Pacific (i.e., the Niño‐4 region) have an in‐phase association with the SST variations in the eastern Pacific (i.e., the Niño‐3 region). During the CP ENSO, the association between SST anomalies in the central Pacific and the eastern Pacific is weaker. Instead, the SST variations in the tropi­cal central Pacific are more related to SST variations in the northeastern subtropical Pacific that are induced by an interannual atmospheric variability mode, the so‐called North Pacific Oscillation (NPO) [Walker and Bliss, 1932; Rogers, 1981; Linkin and Nigam, 2008]. The NPO is the second leading mode of the North Pacific SLP varia­bility and is characterized by out‐of‐phase SLP fluctua­tions between the Aleutian low and the Pacific subtropical high. Based on these features, Yu et al. [2012a] used three oceanic and atmospheric indices and found that the SST variations in the tropical central Pacific (i.e., the Niño‐4 index) are strongly related to the SST variations in the tropical eastern Pacific (i.e., the Niño‐3 index) before the early 1990s but more closely related to the SLP variations associated with NPO (i.e., the NPO index) afterward. They concluded that the change in ENSO from the EP type to the CP type occurred in the early 1990s. This cli­mate shift resulted in a change in ENSO location from the eastern Pacific to the central Pacific. Yu et al. [2015a] further revealed that the early‐1990s climate shift was related to the Atlantic Multidecadal Oscillation (AMO) [Schlesinger and Ramankutty, 1994; Kerr, 2000], which changed from a negative phase to a positive phase in the early 1990s as well. The AMO is characterized by slow changes in SST in the North Atlantic and is believed to have a period of 60–80 yr. Although there is still ongoing debate on the causes of the AMO, it is agreed that the slow variability mode influences climate not only within the Atlantic basin but also across the entire Northern Hemisphere, including the Pacific basin [e.g., Dong et al., 2006; Zhang and Delworth, 2007]. The physical processes that link the NPO and AMO to the CP ENSO are described in detail in section 1.3.

Evidence of the early‐1990s climate shift in the Pacific has also been found in other climate phenomena of the

low‐latitude Pacific. In the western Pacific, upper ocean currents and sea level structure also changed around this time, with the north equatorial current and the north equatorial countercurrent migrating southward [Qiu and Chen, 2012] and sea level rise accelerating to a rate not seen in the preceding several decades [Merrifield, 2011]. These oceanic state changes were accompanied by changes in trade wind patterns [Qiu and Chen, 2012]. Sui et al. [2007] found that the interannual variability of the western Pacific subtropical high was dominated by fre­quencies of 3–5 yr before the early 1990s but 2–3 yr after­ward. The relationships between the Asian monsoon and the other monsoon systems in the Northern Hemisphere (e.g., the North American and North African monsoons) changed in such a way that the interannual rainfall vari­ability produced by these the monsoon systems decreased dramatically after 1993 [Lee et  al., 2014]. The relation­ships between the East Asian winter monsoon and ENSO also undergo low‐frequency oscillation [Zhou et  al., 2007]. Moreover, the typical drought pattern in China was replaced by a new pattern after the early 1990s [Qian et al., 2014]. These recent studies together indicate that the Pacific climate shift in the early 1990s affected the ENSO location, the monsoons, the subtropical high, and the subtropical ocean gyre.

1.2.3.5.  Random Forcing. Other mechanisms have also been put forth to explain the increased occurrence of the CP ENSO during the past few decades. Newman et al. [2011] used a linear inverse modeling technique to suggest that natural random variations in the atmosphere can project onto two particular initial SST anomaly patterns, each of which can eventually develop into either the EP or CP type of ENSO through thermocline and zonal advection feedbacks, respectively. They argued that such natural random variations can result in a slow alternation in the type of ENSO without the need for changes in the background state.

1.3. CHANGES IN ENSO DYNAMICS

1.3.1. The Different Generation Mechanisms for the Two Types of ENSO

The fundamental dynamics of ENSO have been tradi­tionally believed to reside mostly within the tropical Pacific [Neelin et al., 1998; Philander 1999; Wang et al., 2015; and the references therein]. The prevailing ENSO theories emphasize the delayed response of subsurface ocean to atmospheric forcing for the oscillation of ENSO. In this view, the delayed oceanic response results from the propagation and reflection of oceanic waves along the equator in the delayed oscillator theory [Suarez and Schopf 1988; Battisti and Hirst 1989], the adjustment of

0003060439.INDD 10 6/5/2017 7:05:46 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 11

the nonequilibrium between zonal‐mean thermocline depth and wind forcing in the recharged oscillator theory [e.g., Wyrtki 1975; Zebiak 1989; Jin 1997], or the lagged response of western Pacific wind stress to ENSO SST anomalies in the western Pacific oscillator theory [Weisberg and Wang 1997; Wang et al., 1999]. Variations in ENSO properties (such as its low‐frequency and qua­sibiennial periodicity) can be produced by these delayed atmosphere‐ocean coupling processes and their interac­tions with the annual cycle or by the nonlinear chaotic characteristics of these ENSO dynamics [Jin et al., 1994; Jin 1996, 1997; Tziperman et al., 1994, 1995; Chang et al., 1995, 1996; Mantua and Battisti 1995; An and Jin, 2001]. From this point of view, the various types of ENSO do not necessarily represent distinct ENSO dynamics. However, recent studies of the EP and CP types of ENSO have begun to suggest that these two types of ENSO are likely to have different underlying dynamics [e.g., Kao and Yu, 2009; Yu et al., 2011; Yu and Kim, 2011a; S.‐T. Kim et  al., 2012; Yu et  al., 2012a; Lin et  al., 2014]. In this framework, the EP ENSO is produced by the traditional ENSO dynamics, while the generation of the CP ENSO relies more on forcing from the subtropical Pacific and its  interaction with ocean mixed‐layer dynamics in the equatorial central Pacific.

As shown in Figure  1.2, the composite CP ENSO is associated with in situ subsurface ocean temperature anomalies that do not show the basinwide propagation characteristics typical of the delayed oscillator theory [Suarez and Schopf 1988; Battisti and Hirst 1989]. Yu and Kim [2010b] further examined the thermocline variations associated with nine major CP El Niño events since 1960. Three of the CP El Niño events (i.e., 1968–1969, 1990–1991, and 1991–1992) they examined undergo a pro­longed decaying phase (Fig. 1.3a) that decays slowly and is followed by a warm event in the eastern Pacific. Figure 1.3d shows that the mean thermocline depth (rep­resented by the 20 °C isotherm) for this group of events (i.e., Group 1) is deeper than normal in the equatorial Pacific (i.e., recharged state), which later gives rise to an EP El Niño to prolong the decay of these three CP El Niño events. The other three CP El Niño events (1963–1964, 1977–1978, and 1987–1988) occurred in the p resence of shallow‐than‐normal thermoclines (i.e., discharged state; Group 2 of Fig. 1.3d). This group of CP El Niño events is followed by a La Niña event in the east­ern Pacific, resulting in an abrupt termination of the CP El Niño (Fig.  1.3b). The last three events (1994–1995, 2002–2003, and 2004–2005) occurred with near‐normal thermocline depths (i.e., neutral state; Group 3 of Fig. 1.3d). This group of events undergoes a symmetric‐decaying pattern, whose SST anomalies grow and decay symmetrically with respect to a peak phase (Fig.  1.3c). Therefore, CP El Niño events can occur when the equatorial

Pacific thermocline is above, below, or near the normal depth, as also suggested by a recent modeling study [Fedorov et al., 2014]. This finding further confirms that the generation of CP ENSO does not depend on thermo­cline dynamics. However, these different thermocline structures do affect whether a warming, cooling, or neu­tral event may occur in the eastern Pacific as the CP event decays. Depending on the thermocline structure, the east­ern Pacific warming, or cooling may interact with the SST evolution of the CP event and give rise to the three distinct evolution patterns of CP El Niño. Although the thermocline variations are not crucial to the generation and the dynamics of the CP type of El Niño, the state of the thermocline at the peak phase of a CP ENSO event can potentially be used to predict when and how the CP ENSO event will terminate. It is the interplay between the dynamics of the CP and EP types of ENSO that makes the thermocline information useful for the prediction of both types of SST variations.

1.3.2. The Central‐Pacific ENSO: Extratropical Atmospheric Forcing Mechanism

As mentioned, for the CP El Niño the equatorial westerly anomalies appear to the west of the positive SST anomalies in the central Pacific and the equatorial east­erly anomalies to the east. Ashok et al. [2007] argued that the thermocline variations induced by this wind anomaly pattern are responsible for the generation of the CP ENSO. The equatorial westerly anomalies induce down­welling Kelvin waves propagating eastward and the equatorial easterly anomalies induce downwelling Rossby waves propagating westward and, together, they deepen the thermocline in the central Pacific to produce the CP El Niño. Kug et al. [2009] emphasized the fact that the equatorial easterly anomalies can suppress warming in the eastern Pacific during a CP El Niño event by enhanc­ing upwelling and surface evaporation. However, they also argued that the mean depth of thermocline in the central Pacific is relatively deep and the wind‐induced thermocline variations may not be efficient in producing the CP SST anomalies. Instead, they suggested that ocean advection is responsible for the development of the cen­tral Pacific warming. Yu et al. [2010] also concluded that the SST anomalies of the CP ENSO undergo rapid inten­sification through ocean advection processes. However, they showed that the initial establishment of the CP SST anomalies is related to forcing from the extratropical atmosphere and subsequent atmosphere‐ocean coupling in the subtropics.

Yu et  al. [2010] performed an analysis of the near‐s urface ocean temperature budget to identify the physi­cal processes involved in the development of the CP ENSO. They showed that the SST anomalies associated

0003060439.INDD 11 6/5/2017 7:05:46 PM

12 FORCINGS OF CLIMATE EXTREMES

with the CP ENSO undergo rapid intensification in the equatorial central Pacific through local atmosphere‐ocean interactions, which involve only mixed layer ocean dynamics and are similar to those described for the slow‐SST mode of ENSO [Neelin 1991; Neelin and Jin 1993]. This mode requires surface wind anomalies to induce anomalous air‐sea fluxes and ocean advection processes to intensify the initial SST perturbation. Yu et al. [2010] suggested that the initial SST anomalies in the central equatorial Pacific were, however, established by forcing from the extratropical atmosphere. This s uggestion is supported by regressions of the Pacific SST anomalies onto the CP ENSO index. As shown in Figure 1.4, the regressed positive SST anomalies appear first in the northeastern subtropical Pacific and later

spread toward the central equatorial Pacific to initiate ENSO event in the central Pacific.

The possibility of subtropical precursors in triggering ENSO events has long been recognized. However, recent studies on CP ENSO dynamics have begun to emphasize that these subtropical processes are not just precursors of ENSO but an essential element in the generation of CP ENSO events. As revealed by Figure 1.4, the CP ENSO is preceded by SST anomalies in the northeastern Pacific, which are characterized by a band of SST anomalies extending typically from Baja California toward the equatorial central Pacific. As the subtropical SST anom­alies approach the equatorial Pacific, El Niño events develop and begin to grow. Several recent studies [e.g., Vimont et al., 2003; Anderson, 2004; Yu and Kim, 2011a]

°C

(a)

(c) (d)

(b)Group1: Prolonged–decay

Group3: Symmetric–decay D20C anomalies

Group2: Abrupt–decay

J

120°E 180° 180°

0.0

0.5

1.0

0.5

0.0

–0.5

120°W

120°E 180°

0.5

0.0

0.5

0.5

0.0

120°W

120°W

–2.5

–1.5

–0.5

0.5

1.5

120°E

AS

Yea

r (0

)Y

ear

(+1)

ON

DJ

FM

AM

JJ

AS

Yea

r (0

)Y

ear

(+1)

Year (0) Year (+1)

ON

DJ

FM

AM

J

J A S O

Group1

Group3Group2

N D J F M A M J

10

5

0

–5D20

C (

m)

–10

–15

JA

S

Yea

r (0

)Y

ear

(+1)

ON

DJ

0.5

0.0

FM

AM

J

Figure 1.3 Evolution of composite equatorial Pacific (5°S–5°N) SST anomalies for (a) the group of CP El Niño events that has a prolonged decay (i.e., 1968–1969, 1990–1991, and 1991–1992), (b) the group that has an abrupt decay (i.e., 1963–1964, 1977–1978, and 1987–1988), and (c) the group that has a symmetric decay (i.e., 1994–1995, 2002–2003, and 2004–2005). The evolution is shown from July of the El Niño year (0) to June of the following year (+1) and is calculated from the Extended Reconstruction of Historical Sea Surface Temperature dataset (ERSST) [Smith et al., 2008]. Panel (d) shows the evolution of the depth anomalies of the zonal‐mean 20 °C isotherm (D20C) at the equator (5°S–5°N) for the three groups of events. The zonal mean is averaged between 120°E and 80°W and is calculated using the Simple Ocean Data Assimilation Reanalysis data (SODA) [Carton et al., 2000]. Adapted from Yu and Kim [2010b].

0003060439.INDD 12 6/5/2017 7:05:47 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 13

have suggested that the initial warming off Baja California is forced by atmospheric fluctuations, particularly those associated with NPO. When the Pacific SST anomalies are regressed onto the NPO index (Fig. 1.5), it is obvious that the NPO first induces positive SST anomalies off Baja California via surface heat fluxes. The SST anoma­lies then extend southwestward toward the tropical Pacific. After the SST anomalies reach the tropical Pacific, an El Niño event occurs and develops there. The evolution shown in Figure 1.5 is very similar to the evolu­tion of SST anomalies shown in Figure 1.4 preceding the

development of a CP El Niño event. This similarity offers evidence that there exists a close relationship between the subtropical Pacific precursors and CP ENSO.

Ocean‐atmosphere coupling in the subtropical Pacific is crucial for the spread of the NPO‐induced SST anoma­lies southward into the deep tropics. The coupling pro­cesses begin as the SST anomalies off Baja California feed back onto and modify near surface winds via con­vection. The wind anomalies induced by the convection tend to be located to the southwest of initial subtropical SST anomalies [Xie and Philander, 1994], where new positive

20°N

Lag (–9)

Lag (–6)

Lag (–3)

Lag (0)

Lag (–8)

Lag (–5)

Lag (–2)

Lag (+1)

Lag (–7)

Lag (–4)

Lag (–1)

Lag (+2)

10°N

0°

10°S

20°S

20°N

10°N

0°

10°S

20°S

20°N

10°N

0°

10°S

20°S

20°N

10°N

0°

10°S

20°S

20°N

10°N

0°

10°S

20°S120°E 150°E 180° 150°W120°W 90°W 120°E 150°E 180° 150°W120°W 90°W 120°E 150°E 180° 150°W120°W 90°W

Lag (–12) Lag (–11) Lag (–10)

Figure 1.4 Evolution of tropical Pacific SST anomalies of the CP ENSO. Values shown are SST anomalies regressed on the CP ENSO index from a lag of −12 mo to a lag of +2 mo. The contour interval is 0.1 °C. The HadISST dataset for the period 1948–2010 is used.

0003060439.INDD 13 6/5/2017 7:05:48 PM

14 FORCINGS OF CLIMATE EXTREMES

SST anomalies can be formed through a reduction in evaporation. The atmosphere then continues to respond to the new SST anomalies by producing wind anomalies farther southwestward. Through this wind‐evaporation‐SST (WES) feedback [Xie and Philander, 1994], the SST anomalies initially induced by the NPO off Baja California can extend southwestward into the deep tropics. This series of subtropical Pacific coupling processes is referred to as the seasonal footprinting mechanism [Vimont et al., 2001, 2003, and 2009]. This mechanism also offers a way to explain how the subtropical SST anomalies can be sustained from boreal winter, when extratropical atmos­pheric variability (e.g., the NPO) is largest, to the following spring or summer to excite CP ENSO events.

As mentioned, the SST anomaly pattern of the sub­tropical precursor to ENSO (in particular, the CP ENSO) strongly resembles the PMM that is the leading coupled variability mode of the subtropical Pacific. A strong association between the spring PMM index and the fol­lowing winter’s ENSO index was demonstrated in Chang et al. [2007], who found that a majority of El Niño events

over the past four decades were preceded by SST and sur­face wind anomalies similar to those associated with the PMM. Recently, the winter SST anomalies in the western North Pacific have been suggested as a subtropical ENSO precursor, which precede ENSO development in the f ollowing year [S.‐Y. Wang et al., 2012, 2013]. There are several ways to link these ENSO precursor‐associated anomalies to the generation of ENSO events in the equa­torial Pacific. One explanation emphasizes how the surface wind anomalies associated with the subtropical precursor can directly or indirectly (through the reflec­tion of off‐equatorial Rossby wave in the western Pacific) excite downwelling Kelvin waves along the equatorial thermocline that propagate eastward to trigger El Niño events in the eastern Pacific [e.g., Alexander et al., 2010; S.‐Y.Wang et  al., 2012, 2013]. This view appears more related to the onset of EP ENSO events. However, when the subtropical precursor reaches the equatorial Pacific, it can also force a CP ENSO event by interacting with the local ocean mixed layer. Figure 1.6 illustrates this view of how the underlying dynamics of the two types of ENSO

120°E 150°E 180° 150°W120°W 90°W

120°E 150°E 180° 150°W120°W 90°W 120°E 150°E 180° 150°W120°W 90°W

60°N

45°N

30°N

30°S

15°N

15°S

0°

60°N

45°N

30°N

30°S

15°N

15°S

0°

60°N

45°N

30°N

30°S

15°N

15°S

0°

Lag (0) Lag (+2) Lag (+4)

Lag (+6)

Lag (+12)

Lag (+8) Lag (+10)

Figure 1.5 SST anomalies regressed on the NPO index from a lag of 0 mo to a lag of +12 mo. The contour interval is 0.05 °C and only the coefficients exceeding the 99% confidence level are shown. Positive lags denote that the NPO leads the SST anomalies. The National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis [Kalnay et al., 1996] and HadISST datasets for the period of 1948–2010 are used.

0003060439.INDD 14 6/5/2017 7:05:48 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 15

can interact with the subtropical Pacific precursor. In this view, the generation of the CP ENSO is more related to mixed layer ocean dynamics. Dommenget [2010] and Clement et al. [2011] have demonstrated that ENSO‐like events can be produced in coupled models where the ocean component consists of a mixed layer only without any thermocline dynamics. The generation of the EP ENSO is considered more related to the thermocline dynamics depicted by the delayed‐oscillator [Suarez and  Schopf 1988; Battisti and Hirst 1989] and charge‐recharged oscillator theories [e.g., Wyrtki 1975; Zebiak 1989; Jin 1997].

1.3.3. Links Between the Atlantic Multidecadal Oscillation and the Central‐Pacific ENSO

The strength of the subtropical Pacific coupling, there­fore, plays a key role in determining how efficiently the subtropical Pacific precursors can penetrate into the trop­ics to excite ENSO events, particularly CP ENSO events. As mentioned, Yu et al. [2012a] found that the early 1990s is the time when ENSO changed from the EP type to the CP type. Yu et al. [2015a] further confirmed that the sub­tropical Pacific coupling was stronger after the early 1990s. They analyzed the subtropical Pacific coupling strength by examining the correlation coefficient between

the SST and surface wind anomalies associated with the PMM. As shown in Figure 1.7, the coupling strength was at a relatively constant level before and after the 1976–1977 climate shift, which suggests that the phase change in the PDO cannot affect the subtropical Pacific coupling. In contrast, the subtropical Pacific coupling strength increased after the early‐1990s climate shift. The stronger subtropical Pacific coupling has made it easier for the subtropical precursors to penetrate into and to influence the deep tropics, and, as a result, the occurrence of CP ENSO events increases. As mentioned, the early‐1990s climate shift is related to the phase change of the AMO. Yu et al. [2015a] coupled an atmospheric general circula­tion model (AGCM) with a slab ocean model to demon­strate the relationship between the AMO and the CP ENSO. In the coupled model, they prescribed a positive phase of the AMO SST anomalies in one experiment and a negative phase of the AMO in the other experiment. They showed that the Pacific SST variability produced by the coupled model increased most in the tropical central Pacific when a positive AMO was prescribed in the Atlantic. They also laid out the chain of events to explain how the AMO influenced the CP ENSO. First, a switch of the AMO to its positive phase in the early 1990s led  to an intensification of the Pacific subtropical high. The  intensified high resulted in stronger‐than‐average

Extratropicalatmospheric

forcing

Subtropical pacific coupling

Pacific meridional mode

Subtropical pacific precursor

Mixed-layer dynamics Eastward propagating

Kelvin waves

CP ENSO EP ENSO

Thermocline dynamics

Figure 1.6 A schematic diagram illustrating the possible relationships between the subtropical Pacific precursors and the two types of ENSO (CP and EP). (See insert for color representation of the figure.)

0003060439.INDD 15 6/5/2017 7:05:49 PM

16 FORCINGS OF CLIMATE EXTREMES

background trade winds that enhanced the WES feed­back mechanism, strengthening the subtropical Pacific coupling between the atmosphere and ocean, making the subtropical Pacific precursors more capable of pene­trating into the deep tropics, and ultimately leading to increased occurrence of CP ENSO events. The study of Yu et al. [2015a] suggests that the recent shift from more conventional EP to more CP type of ENSO events can thus be at least partly understood as a Pacific Ocean response to the phase change in the AMO.

1.3.4. The Roles of the Hadley and Walker Circulations

The emergence of CP ENSO represents a change in ENSO dynamic from one that emphasizes the tropical Pacific coupling processes to another that emphasizes the subtropical Pacific forcing. This recent change in ENSO dynamic can also be understood via the different roles played by the Walker and Hadley circulations in the ENSO dynamic, as noted by Yu et al. [2010]. Figure 1.8 shows the correlation coefficients between SLP anomalies and the EP and CP ENSO indices. The EP ENSO is asso­ciated with a southern‐oscillation pattern characterized by out‐of‐phase SLP anomalies between the eastern and western tropical Pacific. The SLP variations over the mar­itime continent region are linked to SLP over the eastern equatorial Pacific through the Walker circulation. The SLP anomaly pattern associated with CP ENSO does not really resemble the Southern Oscillation. Instead, the SLP variations over the maritime continent are linked to SLP

variations in the subtropics of both the North and South Pacific, suggesting that a connection through local Hadley circulation may be more important. Therefore, Yu et al. [2010] argued that the tropical Pacific Ocean interacts with the Walker circulation to produce EP ENSO but the local Hadley circulation to produce CP ENSO. The alter­nation between the EP and CP types of ENSO can be viewed as a result of changes in the relative strengths of the Hadley and Walker circulations. A stronger Walker circulation may lead to a dominance of the EP type, while a stronger Hadley circulation may favor the CP type. This view was further examined in Yu et al. [2012a], in which they compared the strength of the time‐mean Walker and Hadley circulations before and after the early‐1990s climate shift. Consistent with the view of Yu et al. [2010], the period of postclimate shift has a stronger mean Hadley circulation compared to the period of preclimate shift. The stronger Hadley circulation could help the NPO‐induced SST variations to spread into the tropical central Pacific. In association with the changes in the strength of  mean atmospheric circulations, the tropical central Pacific has become more responsive to the forcing and influence from the extratropical atmosphere, which has resulted in more frequent occurrence of CP ENSO events. The relative dominance of the two types of ENSO may be related to the relative strength of the Hadley and Walker circulations. As for the mean strength of the Walker circu­lation, it is in debate whether it has weakened [e.g., Vecchi et  al., 2006; Vecchi and Soden, 2007; Xie et  al., 2010; Tokinaga et  al., 2012; Yu et  al., 2012a] or strengthened

1.0

0.8

Cor

rela

tion

coef

ficie

nt0.6

0.4

0.2

1970 1980 1990 2000

AMOphase

Year

PDOphase

Early

–1990s clim

ate shift

1976–

77clim

ate shift

Variation in PMM coupling strength

Figure 1.7 Coefficients of the 10 yr running correlation between the PMM‐SST and PMM‐wind indices in boreal spring (March–May). The ERSST and NCAR/NCEP reanalysis datasets are used. The red line indicates the mean of the correlation coefficients during each period. The shading at the top and bottom show the phases (positive red/negative blue) of the PDO and the AMO, respectively. Modified from Yu et al. [2015a].

0003060439.INDD 16 6/5/2017 7:05:49 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 17

[e.g., Dong and Lu, 2013; Hu et al., 2013; L’Heureux et al., 2013] in recent decades, although it is more generally agreed that the circulation has strengthened in the most recent decade [e.g., England et al., 2014].

1.3.5. Other Possible Mechanisms

In addition to the extratropical atmospheric forcing mechanism, other mechanisms have been proposed to explain the generation of the two types of ENSO. For example, Yu et al. [2009] used ocean‐atmosphere coupled GCM (CGCM) experiments to suggest that the CP ENSO can be excited by forcing associated with the Asian and Australian monsoons. Chung and Li [2013] argued that the La Niña‐like (stronger zonal SST gradient) dec­adal background state since 1999 is responsible for the more frequent occurrence of CP El Niño. Their numeri­cal experiments showed that with the increased back­ground zonal SST gradient, the anomalous wind and precipitation response to EP or CP SST forcing shifts to the west, leading to a westward shift in the SST anomaly tendency that resembles the CP ENSO. Ren and Jin [2013] found that with the decadal signal removed, the ocean heat content shows recharge‐discharge processes through­out the life cycles of both ENSO types. They showed that

the different zonal locations of the SST and easterly wind anomalies and the associated thermocline feedback contribute to the growth and phase transitions of the two ENSO types. This view of varying thermocline feedbacks for the two ENSO types is consistent with the study of Hu et al. [2012b] who suggested that the subsurface ocean temperature anomalies are weaker in the CP ENSO than in the EP ENSO. Hu et al. [2012b] also found differences in the westerly wind bursts (WWBs) between the two types of ENSO. They suggested that the CP (EP) ENSO is followed by weaker and more westward confined (stronger and eastward extended) WWBs along the equatorial Pacific, which are associated with suppressed (enhanced) air‐sea interactions over the cold tongue/Intertropical Convergence Zone and weaker (stronger) thermocline feedbacks. Fedorov et al. [2014] and Hu et al. [2014] also reached a similar conclusion on the differ­ences in the importance of WWB for the two ENSO types. These two modeling studies showed that the EP El Niño develops when WWBs are imposed in an initial recharge ocean state, while the CP El Niño develops in a neutral ocean state in the presence of a WWB or in a recharge state without a WWB. All these WWB studies suggest that properly accounting for the preceding westerly wind signal and air‐sea interaction may be necessary for

60°N

30°N

0°

30°S

60°S

60°N(b)

(a)

0.60.50.40.30.20.10

–0.1–0.2–0.3–0.4–0.5–0.6

0° 45°E 135°E 180° 135°W 90°W 45°W

SLP correlation with CP ENSO index

SLP correlation with EP ENSO index

90°E

0° 45°E 135°E 180° 135°W 90°W 45°W90°E

30°N

0°

30°S

60°S

Figure 1.8 Correlation of SLP anomalies with (a) the EP and (b) CP ENSO indices. Only the coefficients exceeding the 95% confidence level are shown. The NCEP/NCAR reanalysis dataset for the period of 1948–2010 is used.

0003060439.INDD 17 6/5/2017 7:05:53 PM

18 FORCINGS OF CLIMATE EXTREMES

accurate ENSO predictions [Hu et  al., 2012b; Fedorov et al., 2014; Hu et al., 2014; D. Chen et al., 2015].

1.4. CHANGES IN ENSO TELECONNECTIONS AND ASSOCIATED CLIMATE EXTREMES

The recent identification of the two distinct types of ENSO offers a new way to consider the impact of ENSO on global climate. There are indications that the atmos­pheric response to ENSO events can be sensitive to the exact location of their SST anomalies [e.g., Mo and Higgins 1998; Hoerling and Kumar, 2002; Barsugli and Sardeshmukh, 2002; DeWeaver and Nigam, 2004; Larkin and Harrison, 2005b]. An increasing number of recent studies have looked into the different impacts produced by these two types of ENSO on various parts of global climate. Many of these studies have revealed that the CP ENSO can affect global climate differently from the EP ENSO [e.g., Larkin and Harrison, 2005b], indicating that one should not simply use the existing understanding of the conventional EP ENSO to anticipate the climate impacts and teleconnections associated with the CP ENSO. A systematic investigation of the global and regional impacts of the CP ENSO is required. In this s ection, we use the differing impacts of the two types of  ENSO on global climate to illustrate how ENSO’s teleconnection and associated climate extremes have been changing in the past few decades.

1.4.1. North America

It is well recognized that North American climate is significantly impacted by ENSO [e.g., Ropelewski and Halpert, 1986, 1989; Kiladis and Diaz, 1989; Livezey et al., 1997; Cayan et al., 1999; Larkin and Harrison, 2005b; and many others]. However, the existence of two types of ENSO was not taken into account in most of the past studies. The El Niño impact on US winter temperatures is traditionally characterized as a north‐south dipole pat­tern, in which above normal temperatures are found over the northern states and below normal temperatures over the southern states [e.g., Ropelewski and Halpert, 1986]. Larkin and Harrison [2005a] first pointed out that sea­sonal US weather anomalies associated with the two types of El Niño are quite different. Mo [2010] consid­ered a priori the two types of ENSO to examine ENSO’s impacts on North American climate. The author con­trasted the impacts of ENSO on air temperature and pre­cipitation over the United States during the period 1915–1960 when the EP type dominated and the period 1962–2006 when the CP type had increased in impor­tance, and concluded that the EP El Niño produced a north‐south contrast pattern in the air temperature vari­ations, similar to the traditional view. For the CP El Niño,

the author found the temperature variations are charac­terized by more of an east‐west contrast pattern with warming over the Pacific Northwest and cooling in the Southeast. Yu et al. [2012b] also found that the two types of ENSO produced different impacts on US winter tem­perature. Their result is similar to that of Mo [2010] but with significant differences in the details. Yu et al. [2012b] first separately regressed winter (January–March) surface air temperature anomalies to the EP and CP ENSO indi­ces to show that, during EP El Niño events (Fig. 1.9a), positive winter temperature anomalies are concentrated mostly over the northeastern part of the United States and negative anomalies are most obvious over the south­western states. During CP El Niño events (Fig. 1.9b), the warm anomalies are located in the northwestern United States and the cold anomalies are centered in the south­eastern United States. The US temperature anomaly pat­terns are rotated by about 90° between these two types of El Niño. Adding these two impact patterns together can result in a pattern that resembles the traditional El Niño impact on the US winter temperature (i.e., a warm‐north and cold‐south pattern). It indicates that the traditional view of El Niño impact is likely a mixture of the impacts of the two types of El Niño. Yu et al. [2012b] was able to reproduce these two different impact patterns in forced model experiments performed with the Community Atmosphere Model version 4 (CAM4) [Neale et al., 2013] from the National Center for Atmospheric Research (NCAR) (shown in Fig. 1.9c and d], in which SST anom­alies characteristic of the EP El Niño and CP El Niño were separately used to force the model in two different sets of ensemble experiments. The regressed impact pat­terns can also be identified in the four strongest EP El Niño events (in 1997, 1982, 1972, and 1986) and four of the top five strongest CP El Niño events (in 2009, 1957, 2002, and 2004). Both Mo [2010] and Yu et al. [2012b] argued that the cause of the different impacts on US win­ter temperature is the different wave train patterns excited by these two types of El Niño in the extratropical atmos­phere. The winter atmosphere produces a Pacific–North American (PNA) teleconnection pattern [Wallace and Gutzler, 1981] during the CP El Niño, which consists of a positive anomaly center extending from eastern Alaska to the northwestern United States and a negative anom­aly center over the southeastern United States, resulting in a warm‐northwest and cold‐southeast pattern of tem­perature anomalies. During the EP El Niño, the winter atmosphere produces a poleward wave train emanating from the tropical eastern Pacific, across the southwestern United States, and into the northeastern United States, leading to the cold‐southwest and warm‐northeast pat­tern in US winter temperatures. This wave train pattern resembles the Tropical North Hemisphere (TNH) pattern [Mo and Livezey, 1986].

0003060439.INDD 18 6/5/2017 7:05:53 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 19

Concerning the US winter precipitation, the classical view of El Niño’s impact is one of negative precipitation anomalies in the northern United States and positive precipitation anomalies in the southern United States due to a southward displacement of the tropospheric jet  stream and associated winter storm activity [e.g., Ropelewski and Halpert, 1986, 1989; Kiladis and Diaz, 1989; Dettinger et al., 1998; Mo and Higgins, 1998; Cayan et al., 1999]. Mo [2010] suggested that, during the EP El Niño events, precipitation decreases in the Pacific Northwest but increases in northern California. For the CP El Niño, its influence on precipitation increases in the Southwest but decreases in the Ohio valley compared to the EP El Niño. Yu and Zou [2013] regressed US winter precipitation anomalies on the EP and CP ENSO indices to identify the impact patterns (Fig. 1.10a and b). Both types of El Niño produce dry‐north and wet‐south anomaly patterns, similar to the seesaw pattern that has

traditionally been used to describe El Niño impacts on US winter p recipitation. However, the dry anomalies produced by the CP El Niño are of larger magnitude and cover larger areas than those produced by the EP El Niño. For example, the dry anomalies cover only the Great Lakes region during the EP El Niño, but extend southwestward through the Ohio‐Mississippi valley toward the Gulf Coast during the CP El Niño. In con­trast, the wet anomalies tend to have smaller magnitudes during the CP El Niño than during the EP El Niño, a phenomenon that is  most obvious over the southeast United States. In Figure  1.10, a and b indicate that the  CP El Niño tends to intensify the dry anomalies but  weaken the wet anomalies of the impact pattern p roduced by the EP El Niño. This important difference is  clearly revealed in Figure 1.10c, where the precipita­tion anomalies regressed on the EP El Niño were sub­tracted from the anomalies regressed in the CP El Niño.

45°NN

CE

P–

NC

AR

CA

M4

35°N

25°N

45°N

35°N

25°N

45°N

35°N

25°N120°W 100°W 80°W 120°W

(f)(e)

(d)(c)

(b) CP EI Nino impactEP EI Nino impact(a)

100°W 80°W

6543210–1–2

°C

0.50.40.30.20.10–0.1–0.2–0.3–0.4–0.5

Figure 1.9 EP (left panels) and CP (right panels) El Niño impacts on US winter (January–March) surface air tempera-tures. Observed temperature anomalies regressed onto the (a) EP and (b) CP ENSO indices. Coefficients exceeding the 90% confidence level are dotted. The NCEP/NCAR reanalysis dataset for the period 1950–2010 is used. Temperature patterns reproduced by the Community Atmosphere Model version 4 (CAM4) in the (c) EP and (d) CP experiments. Only differences exceeding the 90% confidence level are shaded. Schematic diagrams of the (e) EP and (f) CP El Niño impacts on temperatures. Adapted from Yu et al. [2012b]. (See insert for color representation of the figure.)

0003060439.INDD 19 6/5/2017 7:05:53 PM

20 FORCINGS OF CLIMATE EXTREMES

The negative values in Figure 1.10c indicate that a shift in El Niño from the EP type to the CP type enhances the dry anomalies over the Pacific Northwest and along the Ohio‐Mississippi valley and weakens the wet anomalies over the Southeast. As a result, the emerging CP El Niño produces an overall drying effect during the US winter. The different impacts on winter precipitation can lead to an asymmetric response between the two types of ENSO in the springtime soil water hydrology over the Mississippi River basin as shown by Liang et  al. [2014], who also

indicated that the winter to spring m emory was carried through by subsurface hydrological storage processes.

Winter precipitation over the United States is primarily associated with winter storms, whose tracks are con­trolled by the locations of tropospheric jet streams. Yu and Zou [2013] related the enhanced drying effect of the CP ENSO to a more southward displacement of the trop­ospheric jet streams. Previous studies have suggested that such an equatorward shift of the tropospheric jet streams during El Niño events results from El Niño‐induced Rossby wave trains and strengthening of the Hadley cir­culation [e.g., Wang and Fu, 2000; Seager et al., 2003; Lu et al., 2008]. As the jet streams shift southward, winter storms shift south as well, leading to a dry‐north and wet‐south pattern of precipitation anomalies during the El Niño. Yu and Zou [2013] found that the jet streams were displaced more southward during CP El Niño than dur­ing EP El Niño. The increased southward displacements of the jet streams cause dry anomalies over the northern United States (including the Northwest and Ohio‐Mississippi valley), which expand and intensify more significantly during the CP El Niño than during the EP El Niño. Similarly, the wet anomalies over the southern United States expand over the Southwest and extend into Mexico during the CP El Niño. However, the same south­ward displacements over the East Coast push the core of the jet stream (and therefore the storm tracks) south of the US continent and into the Gulf of Mexico and the Caribbean, which results in only a small area of wet anomalies left in the southeast United States during the CP El Niño. One major implication of the findings of Yu and Zou [2013] is that the droughts occurring in the Ohio‐Mississippi valley and Pacific Northwest during El Niño years may be intensified as El Niño becomes more fre­quently of the CP type and that the Southeast should not expect as much enhancement of winter precipitation d uring El Niño years as in the past. In addition to cold season impacts, Liang et al. [2015] suggested that the two ENSO types have different influences on the warm s eason (May–June) Great Plains low‐level jet (GPLLJ) during their decaying phases. They found that the EP El Niño intensifies the GPLLJ while CP El Niño weakens the GPLLJ, the latter leading to drier conditions in the c entral United States.

Capotondi and Alexander [2010] have shown that CMIP3 models are capable of reproducing the general features of the observed precipitation pattern over North America despite the coarse resolutions of these global models. The six models they analyzed captured the observed precipitation maxima in the American North­west and Southeast and the lower values in the central United States. Their analysis also showed that the sea­sonal cycles of precipitation produced by the CMIP3 models in the Great Plains were realistic, with a rainy

50°N

40°N

30°N

50°N

40°N

30°N

50°N

40°N

30°N

120°W 90°W 60°W

120°W 90°W 60°W

120°W

(b)

(a)

(c)

90°W 60°W

–0.2

–0.4

–0.6

–0.8

1

0.8

EP impact / regression of reanalysis

CP impact / regression of reanalysis

CP- EP / regression of reanalysis

0.6

0.4

0.2

0

Figure 1.10 US winter (January–March) precipitation anoma-lies associated with (a) the EP El Niño and (b) the CP El Niño. The difference between the two types of El Niño (i.e., the CP impact minus EP impact) is shown in (c). The values shown are obtained by regressing US winter precipitation anomalies onto the EP and CP ENSO i ndices. Values shown are in units of mm/day, and the areas passing the 90% confidence level based on the Student’s t‐test are dotted. The NCEP/NCAR reanalysis dataset for the period of 1948–2010 is used. Adapted from Yu and Zou [2013].

0003060439.INDD 20 6/5/2017 7:05:54 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 21

season that starts in early summer and slowly winds down toward winter. Mo [2010] also showed that the CMIP3 simulations could be useful in identifying the CP and EP ENSO’s impacts on US winter climate. A series of studies organized by the National Oceanic and Atmospheric Administration (NOAA) Modeling, Analysis, Predictions, and Projections (MAPP) Program’s CMIP5 Task Force have also demonstrated that CMIP5 models are useful in understanding the climate over North America [Sheffield et al., 2013a, 2013b, 2014; Maloney et al., 2014]. These recent studies demonstrated that CMIP simulations were useful for studying the regional climate impacts produced by global‐scale climate phenomena. The performance of coupled climate models in simulating the EP and CP types of ENSO was also examined for both the CMIP3 models [Yu and Kim, 2010a] and the CMIP5 models [Kim and Yu, 2012]. Compared to the CMIP3 models, the CMIP5 models are found to better simulate the observed spatial patterns of the two types of ENSO and to have a significantly smaller intermodel diversity in ENSO inten­sity. The improved performance of the CMIP5 models is particularly apparent in the simulation of EP ENSO intensity, although it is still more difficult for the models to reproduce the observed EP ENSO intensity than the  observed CP ENSO intensity. The performance of CMIP5 models in simulating the different impacts of the two types of ENSO on US winter climate was also exam­ined in Zou et al. [2014], where 30 CMIP5 preindustrial simulations were examined to conclude that the CMIP5 models can simulate the observed EP El Niño impacts but not the observed CP El Niño impacts. The responses of observed US winter temperature to the EP El Niño are reasonably simulated by most of the CMIP5 models. The northeast‐warm and southwest‐cold pattern can be seen in many of the CMIP5 models. However, the observed northwest‐warm and southeast‐cold response pattern to the CP ENSO can only be found in some of the CMIP5 models. In some of the other CMIP5 models, the US winter temperature anomalies are even opposite from the observed.

Why are CMIP5 models less successful in simulating CP ENSO impact on the US winter climate? Heating associated with deep convection is a key process by which ENSO influences the atmosphere. The intensity and loca­tion of deep convection can be represented by the outgo­ing longwave radiation (OLR). The importance of using OLR anomalies to understand the atmospheric telecon­nections from the tropics has been emphasized in many earlier studies [e.g., Heddinghaus and Krueger, 1981; Lau and Chan, 1983, 1985; Gruber and Krueger, 1984; Ardanuy and Kyle, 1986; Chiodi and Harrison, 2013]. Zou et  al. [2014] looked into the relationships between SST and OLR anomalies for the two types of El Niño and exam­ined how well the observed relationships were simulated

in the CMIP5 models. It was found that during the EP ENSO, the strongest SST anomalies are located in the eastern equatorial Pacific and can easily influence the strength of the Walker circulation, giving rise to basinwide OLR anomalies. The modeled atmospheric responses to the EP El Niño are thus not sensitive to SST anomaly structure and can be well simulated by most of the CMIP5 models. In contrast, the SST anomalies of the CP El Niño are located in the central equatorial Pacific and can induce only local OLR anomalies to the west of the SST anomalies. The modeled atmospheric responses to the CP El Niño are, therefore, different among the mod­els depending on the simulated magnitudes and locations of the CP El Niño SST anomalies. The finding reported by Zou et  al. [2014] implies that it may become more c hallenging for climate models to simulate and predict ENSO’s impacts on US climate as ENSO changes from the EP type to the CP type.

1.4.2. Western Pacific and Indian Oceans

ENSO events also produce significant impacts on the climate of the western Pacific and the Indian Ocean basin. In the conventional point of view, El Niño events shift tropical convection eastward to the equatorial c entral‐to‐eastern Pacific resulting in an anomalous Walker circulation that has an anomalous descending branch over the western Pacific, Maritime Continent, and northern Australia (Fig.  1.8a). This anomalous descending motion then induces two modes of Indian Ocean (IO) variability through surface heat fluxes and oceanic Rossby waves [e.g., Klein et al., 1999; Venzke et al., 2000; Alexander et al., 2002; Xie et al., 2002; Yu and Lau, 2004; Yu et al., 2005; Liu and Alexander, 2007]. One of the modes is the Indian Ocean Dipole [Saji et al., 1999; Webster et  al., 1999] that has one SST anomaly center in the tropical western Indian Ocean and the other center in the southeast Indian Ocean, and the other mode is the Indian Ocean basinwide warming (IOBW). The IOD tends to peak during El Niño’s devel­oping phase during autumn (September–November; SON) and IOBW peaks during El Niño’s decaying phase during spring (March–May) [e.g., Schott et  al., 2009]. These ENSO‐induced Indian Ocean variability modes are known to affect Indian summer monsoon, Australia and East Africa, the western North Pacific, and Southern Hemisphere high latitudes [e.g., Saji et al., 1999; Ashok et al., 2001, 2003; Lau and Nath, 2004; Yang et al., 2007; Xie et al., 2009]. Associated with the differ­ent Walker circulation response to two ENSO types (Fig. 1.8a and b), Wang and Wang [2014] suggested that the EP El Niño and some CP El Niño (El Niño Modoki I) events induce the positive IOD, while the other CP El Niño (El Niño Modoki II) events produce the negative

0003060439.INDD 21 6/5/2017 7:05:54 PM

22 FORCINGS OF CLIMATE EXTREMES

IOD. However, Yu et al. [2015b] argued that the EP El Niño events are more associated with the IOD while the CP El Niño events are not. Paek et  al. [2015] showed that the EP El Niño events tend to be accompanied by basinwide warming in the Indian Ocean but not the CP El Niño events. The different Indian Ocean responses induced by the two types of ENSO bring different ENSO impacts to the neighboring regions. Kumar et al. [2006] used a historical rainfall record and AGCM experiments to suggest that the CP El Niño is more effective in enhancing subsidence over the Indian subcon­tinent during Indian summer monsoon season (June–September) leading to more severe droughts than during EP El Niños. Yadav et al. [2013] found that the EP El Niño leads to more precipitation over the north and central India regions during El Niño peak season (December–February). Ratman et  al. [2014] reported significant drought in austral summer over southern Africa during EP El Niños but not during CP El Niños. Significant decreases in rainfall in Australia have been reported during CP El Niño events compared to the EP El Niño events [Wang and Hendon, 2007; Lim et  al., 2009; Taschetto and England, 2009].

1.4.3. East Asia

ENSO has profound impacts on East Asian climate variability such as the East Asian monsoon [e.g., Zhang et al., 1999; Chang et al., 2000; Zhou and Chan, 2007; Li and Yang, 2010], East Asian rainfall [e.g., McBride et al., 2003; Juneng and Tangang, 2005; Zhou et al., 2009], west­ern Pacific tropical cyclone (TC) activity [e.g., Chan, 2000; Chia and Ropelewski, 2002; Camargo and Sobel, 2005; Ho et al., 2004], and SSTs over the South China Sea [e.g., He and Guan, 1997; Xie et  al., 2003; Wang et  al., 2006; Zhang et al., 2010]. The anomalous western North Pacific subtropical high (WNPSH) or low‐level Philippine Sea anticyclone is a potentially important atmospheric component connecting ENSO and East Asian climate as suggested by several authors [e.g., Zhang et  al., 1999; Wang et al., 2000; Lau and Nath, 2006; Chou et al., 2009]. ENSO induces SST anomalies in the western Pacific to initiate the WNPSH during the ENSO developing sea­sons and the WNPSH variability is sustained until the following summer (June–August) through local air‐sea interactions [Wang et al., 2000], or the IO SST anomalies (i.e., IOBW) to excite a Kelvin wave to the east maintain­ing the WNPSH variability [Yang et al., 2007; Xie et al., 2009], or the tropical central Pacific SST anomalies that produce the WNPSH as a Rossby wave response to the west of the SST anomalies [B. Wang et al., 2013], or the SST anomalies over the maritime continent that give rise to a local Hadley circulation descending into the WNPSH region [Sui et al., 2007].

Recently, Zhang et  al. [2011] contrasted the different impacts of the two ENSO types on the WNPSH. They found that during the El Niño developing phase in autumn, an intensified WNPSH for the EP El Niño but a weakened WNPSH for the CP El Niño. Yuan et al. [2012] further documented that from the El Niño developing summer to decaying summer, the EP El Niño is accompa­nied by a strong WNPSH that extends from the Indian Ocean to the east of the Philippines, while the CP El Niño is accompanied by a weak, short‐lived WNPSH that is confined to the west of the Philippines. They also argued that the IO SST anomalies are responsible for the persis­tence of WNPSH into the ENSO following summer for the EP El Niño, but not for the CP El Niño. Paek et al. [2015] suggested in their observational and AGCM study, that the 3–5‐yr band of summer WNPSH variability is more related to the EP ENSO and its SST anomalies over the Indian Ocean, whereas the 2–3‐yr band of WNPSH variability is related to CP ENSO and its SST anomalies in the central Pacific and maritime continent. These dif­ferent WNPSH responses to the two ENSO types are then shown to modulate East Asian climate. For example, southern China experiences increased rainfall during the EP El Niño but decreased rainfall during the CP El Niño due to the different WNPSH response and associated moisture transport into the region in El Niño developing phase in autumn [Zhang et  al., 2011], mature phase in winter [Weng et al., 2009], and decaying phase in spring [Feng and Li, 2011]. During El Niño decay in summer, Feng et  al. [2011] found increased rainfall south to the Yangtze River and decreased rainfall in the northern Yangtze–Huaihe river region for the EP El Niño, and increased rainfall in the Huaihe River–Yellow River region and decreased rainfall in southern China for the CP El Niño. J.‐S. Kim et al. [2012] showed a significant increase in precipitation over South Korea and southern Japan during the CP El Niño that represents conditions opposite of those found for the EP El Niño. Weng et al. [2007] and Yuan and Yang [2012] also found that the d ifferent WNPSH can exert a different impact on the summer and winter monsoon over East and Southeast Asia, which implies difficulty in predicting East Asian monsoon variability based on the conventional mon­soon‐ENSO relationships [e.g., Yang and Jiang, 2014]. The WNPSH and associated convection also modulate TC activity over the South China Sea and western North Pacific during the El Niño developing phase in summer and autumn. The CP El Niño tends to enhance TC activity in the regions but the EP El Niño suppresses TC activity, which increases the TC landfall in Taiwan, South China, Korea, and Japan particularly in autumn [Chen and Tam, 2010; Chen, 2011; Hong et al., 2011; Kim et al., 2011; Wang et al., 2014]. In the South China Sea, anoma­lous net surface heat flux associated with the different

0003060439.INDD 22 6/5/2017 7:05:54 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 23

atmospheric circulation responses to the two ENSO types leads to a strong warm basin mode during the EP El Niño and a warm semibasin mode during the CP El Niño [Liu et al., 2014].

1.4.4. Europe and the Atlantic Ocean

ENSO can influence the Atlantic/European sector via the tropospheric temperature mechanism [Chiang and Sobel, 2002] and the atmospheric bridge mechanism [Klein et al., 1999]. The tropospheric temperature mecha­nism involves an ENSO‐induced tropical tropospheric warming that spreads to the Atlantic as a Kelvin wave response to ENSO forcing, which then increases SST anomalies there [e.g., Yulaeva and Wallace, 1994; Enfield and Mayer, 1997; Chiang and Sobel, 2002]. In the atmos­pheric bridge mechanism, ENSO weakens the Walker cir­culation and excites a PNA teleconnection pattern that reduces trade wind and evaporation in the tropical North Atlantic, leading to a warming there [e.g., Lau and Nath, 2004; Klein et  al., 1999; Alexander et  al., 2002]. The canonical El Niño is known to suppress Atlantic TC activity through changes in vertical wind shear over the main TC development region [e.g., Gray, 1984; Goldenberg and Shapiro, 1996] and changes in tropospheric and sur­face temperatures [Tang and Neelin, 2004]. Recently, Kim et al. [2009] found that in contrast to the EP El Niño, the CP El Niño tends to enhance Atlantic TC activity and increase landfall potential along the coast of the Gulf of Mexico and Central America, which is associated with the different impacts on the vertical wind shear in the main TC development region by the two ENSO types. In contrast, S.‐K. Lee et al. [2010] and Larson et al. [2012] suggested that the CP El Niño has an insubstantial impact on Atlantic TC activity, which is associated with a weak warming of the tropical North Atlantic during the CP El Niño and a strong warming during the EP El Niño [Rodrigues et al., 2011; Amaya and Foltz, 2014]. Graf and Zanchettin [2012] found different impacts on European winter temperatures. They showed that EP El Niño events  lead to weak warming while CP El Niño events lead to significant cooling. They further attributed the European cooling to the negative phase of North Atlantic Oscillation (NAO) through the CP ENSO‐NAO teleconnection in the troposphere.

1.4.5. Arctic

For the Northern Hemisphere high‐latitude and Arctic climate, the ENSO signal can be transmitted there via a stratospheric pathway. Canonical El Niño events tend to be accompanied by a weakened Northern Hemisphere polar vortex. The positive PNA pattern and a deepened Aleutian low in the troposphere during El Niño may

enhance planetary wave activities and increase the upward propagation of the waves into the stratosphere [e.g., Garfinkel and Hartmann, 2008; Garfinkel et  al., 2010; Nishii et al., 2010], where they dissipate and result in a weakened polar vortex and more frequent polar strato­spheric warmings [e.g., Sassi et al., 2004; García‐Herrera et al., 2006; Manzini et al., 2006]. The warm anomalies then descend to the troposphere in boreal winter through wave–mean flow interactions [e.g., Holton, 1976; McIntyre, 1982; Zhou et al., 2002; Manzini et al., 2006]. This weakened polar vortex and descending warm anom­alies excite the Arctic Oscillation (AO) and the NAO [e.g., Thompson and Wallace, 1998; Baldwin and Dunkerton, 1999; Zhou et  al., 2002; Kunz et  al., 2009]. Recently, d iffering polar teleconnections between the two types of ENSO have been reported. Hegyi and Deng [2011] used post‐1979 observations to suggest that in contrast to the EP El Niño, the CP El Niño can lead to a stronger polar vortex and a positive phase of the AO. They argued that the positive AO tendency associated with the CP El Niño events in recent decades has contributed to decadal changes in the Arctic precipitation. Xie et al. [2012] also reported a weakened midlatitude upward propagation of wave activity during the 1979–2010 CP El Niños, leading to a stronger and colder polar vortex. In contrast, Graf and Zanchettin [2012] showed both types of El Niño since 1948 were accompanied by a weakened polar vortex. Zubiaurre and Calvo [2012] and Sung et al. [2014] found a polar response (that is not statistically significant) to the CP El Niño that is related to a delayed deepening of the Aleutian low and weakened upward wave propagation to stratosphere. Garfinkel et al. [2013] compared the differ­ent composites used in previous studies [Hegyi and Deng, 2011; Graf and Zanchettin, 2012; Xie et  al., 2012; Zubiaurre and Calvo, 2012] to show that the diverse CP ENSO‐Arctic teleconnections are very sensitive to the selection of CP El Niño events from the short observa­tional record. Iza and Calvo [2015] emphasized the role of stratospheric sudden warming (SSW) events for the downward propagation. They found that with the SSWs removed, the different EP and CP El Niño responses are robust regardless of the CP El Niño definition and the composite size. In a modeling study, Xie et  al. [2012] showed that whether the CP El Niño induces a strength­ened or a weakened vortex depends on the phase of the quasi‐biennial oscillation (QBO), while other studies [Garfinkel et al., 2013; Hurwitz et al., 2013; Hurwitz et al., 2014] found a consistent weakening of the polar vortex under CP Niño SST forcing using a single AGCM. In the CMIP5 multimodel simulations, Hurwitz et  al. [2014] found that the multimodel means show a weakening of the polar vortex during the EP El Niño in boreal winter, and a nonsignificant strengthening of the vortex during the CP El Niño.

0003060439.INDD 23 6/5/2017 7:05:54 PM

24 FORCINGS OF CLIMATE EXTREMES

1.4.6. Antarctica

In the Southern Hemisphere, ENSO has been sug­gested to influence mid‐to‐high‐latitude climate [e.g., Karoly, 1989; Mo, 2000; Yuan, 2004], often via the two leading Southern Hemisphere atmospheric variability modes: the southern annular mode (SAM) [Hartmann and Lo, 1998; Thompson and Wallace, 2000] and the Pacific–South American teleconnection (PSA) [Mo and Ghil, 1987] pattern. For example, previous studies found that ENSO, together with the SAM and PSA, influenced South American rainfall [e.g., Ropelewski and Halpert, 1987, 1989; Grimm, 2003], the Southern Ocean SSTs [e.g., Ciasto and Thompson, 2008; Yeo and Kim, 2015], Antarctic sea‐ice concentrations [e.g., Liu et al., 2004; Stammerjohn et al., 2008; Yuan and Li, 2008], and Antarctic surface air temperatures [e.g., Kwok and Comiso, 2002; Ding et al., 2011; Schneider et  al., 2012]. Recently, several studies have begun to examine the different impacts of the two types of ENSO on the Southern Hemisphere climate. The westward shift of tropical Pacific convection during the CP El Niño can excite Rossby wave train (i.e., PSA) toward higher latitudes, which then results in suppressed storm track activity over Australia and enhanced activity over central Argentina [Ashok et al., 2009], a northward displacement of the South American precipitation pat­tern [Hill et  al., 2009], the melting of Antarctic ice by inducing a stationary anticyclone and enhancing the eddy heat flux into the region [T. Lee et al., 2010], and stronger Antarctic sea‐ice variability [Song et al., 2011]. In addi­tion to the PSA pattern, early studies showed that the ENSO signals can be transmitted to high latitudes through a tropospheric pathway mechanism in which an ENSO‐induced equatorward shift in the subtropical jet can change the midlatitude eddy activity and its associ­ated eddy–mean flow interactions to modulate the SAM [e.g., Seager et al., 2003; L’Heureux and Thompson, 2006; Lu et  al., 2008; Chen et  al., 2008]. Recent studies have used observations and GCM experiments to also suggest a stratospheric pathway mechanism in which the CP ENSO can affect the upward propagation of planetary waves into the stratosphere and induce polar temperature and vortex anomalies that subsequently descend into the troposphere to excite the SAM [e.g., Mechoso et al., 1985; Hurwitz et  al., 2011a,b; Lin et  al., 2012; Zubiaurre and Calvo, 2012; Son et al., 2013; Evtushevsky et al., 2014].

Fogt and Bromwich [2006] noticed that the relationships among ENSO, PSA, and SAM can change from decade to decade, which is attributed to the strengthened ENSO–Southern Hemisphere influences from the 1980s to the 1990s. More recently, Yu et al. [2015b] further suggested a shift in the ENSO–Southern Hemisphere teleconnections occurred in the early 1990s. They conducted a 15 yr run­ning correlation analysis among the SAM, PSA, and

ENSO indices during austral spring (SON) for the period 1948–2014. For an ENSO index, they used cold tongue index (CTI; SST anomalies averaged over 6°S–6°N and 180°–90°W). They found that the correlation coefficient between the CTI and SAM index (Fig.  1.11e) dramati­cally increased from insignificantly small to significantly large values after the early 1990s, while the CTI‐PSA cor­relation (Fig. 1.11f) stayed at significant values through­out the analysis period. The SST anomaly pattern correlated with the SAM or PSA index after the early 1990s is shown to be similar to the CP ENSO‐SST anom­aly pattern [Kao and Yu,, 2009], whereas the SST anom­aly pattern correlated with the PSA index before the early 1990s is more similar to the pattern for the EP ENSO. Their composite analyses also showed both types of ENSO (Fig. 11a and b) can excite the PSA (Fig. 1.11d) while only CP type can excite the SAM (Fig.  1.11c) through tropospheric and stratospheric pathway mecha­nisms. The EP ENSO is found to induce the Indian Ocean Dipole but not the CP ENSO (Fig. 1.11a and b). Yu et al. [2015b] showed that the different influences of the two types of ENSO on the Indian Ocean SST anomalies e nable these two ENSO types to produce difference influ­ences on the SAM through eddy–mean flow interactions. They further suggested that the different ENSO–Southern Hemisphere teleconnections and more frequent occur­rence of the CP ENSO in recent decades have led to a more in‐phase SAM‐PSA relationship after the early 1990s, which consequently impacted the post‐1990s Antarctic climate in some aspects: an increase in the geopotential height anomalies over the Amundsen‐Bellingshausen seas, a stronger Antarctic sea‐ice dipole structure [e.g., Yuan and Martinson, 2001] between the Ross Sea and Weddell Sea, and a shift in the phase rela­tionships of surface air temperature anomalies among the Antarctic Peninsula, East and West Antarctica. The findings of Yu et  al. [2015b] imply that the linkages between ENSO and Antarctic climate depend on the dominant ENSO type and have become tighter during the recent two decades and will likely remain so in the coming decades if the dominance of the CP ENSO persists.

1.5. ENSO IN THE FUTURE

The changes in ENSO in the future warmer world are one major concern of the climate research community due to the profound impacts produced by ENSO. Climate models are generally used in ENSO projections; however, the behavior of ENSO differs strongly from model to model and is thus highly uncertain. Despite a number of studies, there is still no consensus on how ENSO SST variability will change in response to global warming [i.e., Meehl et  al., 1993; Knutson et  al., 1997; Collins, 2000b;

0003060439.INDD 24 6/5/2017 7:05:54 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 25

Zelle et al., 2005; Guilyardi, 2006; Merryfield, 2006; An et al., 2008; Yeh et al., 2009; Newman et al., 2011; Kim and Yu, 2012; Hu et al., 2012a; Stevenson, 2012; Kim et al., 2014; Cai et al., 2014; L. Chen et al., 2015]. Model projec­

tions have yet to be conclusive on whether ENSO activity will be enhanced, weakened, or remain the same. The uncertainties come at least from three sources. The first source is related to the diverse abilities of climate models

–2 2–1.6 1.6–1.2 1.2–0.8 –0.4 0.4 0.8°C

COR(CTI, SAM)

–1.0

–0.8

–0.6

–0.4

–0.2

0.0

0.2

1950 1960 1970 1980 1990 2000 2010

Year

Cor

rela

tion

coef

ficie

nt

(e) COR(CTI, PSA)

–1.0

–0.8

–0.6

–0.4

–0.2

0.0

0.2

1950 1960 1970 1980 1990 2000 2010

Year

(f)

30W 30E

0

60W

90W

120W

150W

180

150E

120E

90E

60E

(c) Z500: SAM

30W 30E

0

60W

90W

120W

150W

180

150E

120E

90E

60E

(d) Z500: PSA

30°S

30°N

15°S

15°N

0°

0° 45°E 45°W90°E 90°W135°E 180° 180°135°W

SST: CP EI Niño(a)

30°S

30°N

15°S

15°N

0°

0° 45°E

SST: EP EI Niño(b)

45°W90°E 90°W135°E 135°W

Figure 1.11 Upper panels show the composite SST anomalies for the (a) CP El Niño and (b) EP El Niño events during austral spring (SON). Middle panels show the 500 hPa geopotential height (Z500) anomaly structures of the (c) SAM and (d) PSA. Dots indicate areas where the values exceed the 95% confidence interval using a two‐tailed Student’s t‐test. Contour interval is 8 m. Lower panels show the 15 yr running correlation coefficients (e) between the cold tongue index (CTI) and SAM index, and (f) between the CTI and PSA index. The green lines indicate the 90% confidence interval using a two‐tailed Student’s t‐test. Modified from Yu et al. [2015b].

0003060439.INDD 25 6/5/2017 7:05:55 PM

26 FORCINGS OF CLIMATE EXTREMES

to represent the physical processes responsible for deter­mining the characteristics of ENSO. The second source comes from the different model projections of how the mean state may change under warming scenarios, which can modulate the strengths of the atmosphere‐ocean feedbacks that characterize ENSO activity [e.g., Collins, 2000a, b]. The third source of uncertainty is related to the fact that intrinsic ENSO modulation can occur on inter­decadal or intercentennial timescales even without external forcing [e.g., Wittenberg, 2009], which makes it difficult to determine if the ENSO changes produced in model projections are really global‐warming related.

Early studies analyzing ENSO behavior in response to increasing greenhouse gas (GHG) were mostly based on a single CGCM [Meehl et al., 1993; Tett, 1995; Knutson et al., 1997; Timmermann et  al., 1999; Collins, 2000a, b; Zelle et al., 2005]. For example, Meehl et al. [1993] used a CGCM with a coarse horizontal resolution to indicate little change in ENSO amplitude or period as the global atmosphere warms. A similar conclusion was obtained by Tett [1995] using a different climate model. Knutson et al. [1997] sug­gested that the responses of ENSO to global warming in models can depend on the model resolution. In coarse‐res­olution models, the impact of increased GHG on ENSO is unlikely to be clearly distinguishable from the climate sys­tem “noise” in the near future. Timmermann et al. [1999] used a model with a horizontal resolution high enough to adequately reproduce the n arrow upwelling and low‐fre­quency waves on the equator to suggest that ENSO events will become more frequent, stronger, and skew toward the La Niña phase as warming progresses. By comparing the ENSO behaviors under an identical 4 x CO2 scenario using two versions of the Hadley Centre coupled model, Collins [2000a] found s ignificant changes in ENSO amplitude, fre­quency, and phase locking to the seasonal cycle in one ver­sion but not in the other version. Such discriminations are attributed to the differences in the mean pattern of climate change brought about by subtle changes in the physical parameterization between the two versions of the model. Collins [2000a], therefore, suggested that the uncertainties in the projection of future ENSO behavior also come from the model projections of the future mean climate states. The tropical Pacific mean state was projected to become more El Niño‐like (i.e., with greater warming in the east than in the west) [e.g., Meehl and Washington, 1996; Timmermann et al., 1999], more La Niña‐like [e.g., Noda et al., 1999], or uniformly warm [e.g., Meehl et al., 2000] by various models, each of which may give rise to different changes in ENSO. In addition to the mean background SST state and its zonal gradient, the vertical contrast between the mean surface and subsurface temperatures (i.e., the upper ocean stratification) in the tropical Pacific is another mean state variable of crucial importance in determining future ENSO changes [An et al., 2008].

To reduce the uncertainty related to the diverse model dynamics and physics, multimodel ensembles have been increasingly used for ENSO projections in recent years. This is particularly so after the CMIP3 and CMIP5 m ultimodel outputs were made available to the research community and were analyzed for the Intergovernmental Panel on Climate Change (IPCC) Assessment Reports. Among the studies that analyzed the CMIP3 models, van Oldenborgh et al. [2005] suggested that there are no statistically significant changes in the amplitude of ENSO variability in the projections. Guilyardi [2006] found that models capable of realistically simulating SST‐wind feedback tend to project a significant increase in ENSO amplitude under elevated CO2. In contrast to the change in ENSO amplitude, there were no clear indications of an El Niño frequency change. Merryfield [2006] suggested that changes in ENSO amplitude under CO2 doubling are associated with the meridional width of the zonal wind stress response to ENSO in models. The models were found to have a 5% decrease in ENSO periods, which he argued resulted from an increase in equatorial wave speed through an associated speedup of the delayed‐oscillator feedback. Yeh and Kirtman [2007] suggested that whether the model ENSO is in the linear or nonlinear regime determines the changes in ENSO statistics among various climate change projections.

Although the CMIP5 models represent an improve­ment over the CMIP3 models in many aspects (i.e., reso­lution, physics, etc.), the response of ENSO to global warming as seen in the CMIP5 output has not resolved this issue. The response of ENSO amplitude to global warming in CMIP5 shows great diversity from model to model, which is similar to that found in the CMIP3 m odels [L. Chen et  al., 2015]. Many studies found no c onsistent changes across the CMIP5 projections in the location, magnitude, frequency, or temporal evolution of ENSO events [Stevenson, 2012; Taschetto et al., 2014]. An and Choi [2015] suggested that the feedback between ENSO and tropical climate change induced by global warming is very weak, which may be the reason why the global warming trend cannot significantly modify ENSO amplitude. Other CMIP5 studies have shown that green­house warming can alter both the amplitude of ENSO [Kim et  al., 2014] and the frequency of extreme ENSO events [Cai et al., 2014, 2015]. Kim et al. [2014] analyzed the nine CMIP5 CGCMs that best reproduce the observed Bjerknes coupled stability index to project that, under global warming, ENSO amplitude is likely to increase after 2010, remain essentially unchanged until approxi­mately 2040, and then decrease again afterward. Cai et al. [2014] suggested that global warming will increase the occurrence of extreme El Niño events, due to a faster warming in the eastern equatorial Pacific that enables

0003060439.INDD 26 6/5/2017 7:05:55 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 27

deep convection to occur more easily in the region and provide a coupled feedback to grow the ENSO ampli­tude. Cai et al. [2015] found an increased occurrence of extreme La Niña events under global warming due to an increased warming of the maritime continent compared to the equatorial central Pacific that enables convection to move to the maritime continent more easily to trigger the Bjerknes feedback favoring the extreme events.

After the recent recognition of different types of ENSO, how the ENSO type responds to future warming has also been explored in literature. Modeling studies have suggested that anthropogenic greenhouse gas forc­ing can be one of the causes for the increased occurrence of the CP El Niño [Yeh et al., 2009; Collins et al., 2010; Na et al., 2011; Kim and Yu, 2012]. Yeh et al. [2009] used the CMIP3 models to suggest that the occurrence fre­quency of the CP ENSO will increase in the future, while Kim and Yu, [2012] used CMIP5 models to suggest the amplitude of CP ENSO will increase too. Na et al. [2011] projected that the occurrence frequency of the CP El Niño during 2007–2056 will be 2.5 times that in the 1980–2006 period.

In addition to the response of ENSO properties to global warming, changes in ENSO teleconnection under global warming scenarios have attracted substantial attention [i.e., Meehl and Teng, 2007; Kug et  al., 2010; Zou et  al., 2014]. Based on a six‐member multimodel ensemble, Meehl and Teng [2007] identify some possible changes in the future El Niño teleconnections. Among them, the anomalous low in the North Pacific is sug­gested to weaken and shift eastward and northward in future events. This location shift may lead to a decreased anomalous warming over northern North America, increased cooling over southern North America, and enhanced precipitation in the Pacific Northwest compared to present‐day El Niño event teleconnections. Kug et al. [2010] also found that the atmospheric teleconnection patterns over the Northern Hemisphere may shift eastward due to an eastward migration of the convection centers in  the tropical Pacific associated with the future ENSO events. Zou et  al. [2014] also found the ENSO‐forced Pacific–North American teleconnection pattern to become stronger and more eastward displaced in the warming projections. These are just a few examples of possible ways that the changing ENSO may impact a future warmer world.

1.6. SUMMARY

The changing properties of the ENSO observed dur­ing the past two decades and the recent identification of two distinct types of ENSO have prompted the climate research community to revisit traditional views of ENSO dynamics and impacts and how they may change as

global climate changes. A better understanding of ENSO diversity is important for the further development of models used for the prediction of ENSO and its asso­ciated climate extremes. One source of uncertainty in c limate predictions and projections is associated with whether or not modern climate models can realistically simulate both types of ENSO and the alternation between them, and capture their different impacts. If ENSO dynamics are conclusively shown to be changing, the ENSO prediction models and strategies have to be revised to include this changing dynamic. For instance, successful prediction and modeling of the ENSO in the recent and coming decades may depend more on a better understanding and improved skill in the modeling of the subtropical Pacific precursors and their underlying gen­eration mechanisms. Prediction systems based on this framework would be different from the prediction and modeling s ystems the climate research community has relied on since the 1980s and 1990s for the conventional ENSO that emphasize subsurface ocean dynamics in the equatorial Pacific. Larson and Kirtman [2014] have reported some skill in using the PMM to forecast ENSO events with the North American multimodel ensemble (NMME) experiments. Yang and Jiang [2014] showed that the National Centers for Environmental Prediction (NCEP) Climate Forecast System had substantial skill in predicting both types of ENSO and the relationship between the EP ENSO and the Asian monsoon. However, the skill in predicting the relationship between CP ENSO and the monsoon is low. Nevertheless, in order to utilize the subtropical precursors, particularly the PMM, to forecast ENSO events, coupled atmosphere‐ocean mod­els have to be able to realistically simulate the precursor events. Lin et al. [2014] examined 23 CMIP5 models to conclude that the PMM structure can be reasonably s imulated in most of the coupled models. However, the so‐called seasonal footprinting mechanism that sustains an equatorward extension of the PMM is not well simu­lated in a majority of the CMIP5 models. Therefore, it is necessary to improve the subtropical Pacific coupling in these models in order for them to apply them success­fully in forecasts of ENSO occurrence. Currently, pale­oclimate proxy records have mostly been interpreted from the point of view that there is a single type of ENSO (EP) and one set of conventional global impacts (such as precipitation) to derive the history of ENSO. The view that the two types of ENSO can produce d istinct global teleconnection offers an additional way to interpret the paleoclimate proxies. A deeper understanding of how the changing El Niño affects global precipitation may also offer different ways to interpret paleoclimate proxies, which may lead to a new understanding of the history of Earth’s climate and new implications for future climate changes.

0003060439.INDD 27 6/5/2017 7:05:55 PM

28 FORCINGS OF CLIMATE EXTREMES

ACKNOWLEDGMENTS

This work was supported by the National Science Foundation’s Climate and Large Scale Dynamics Program (Grants AGS‐1233542 and AGS‐1505145), the  National Key Research Program of China (Grant 2014CB953900), the National Oceanic and Atmospheric Administration’s Modeling, Analysis, Predictions, and Projections Program (Grant NA11OAR4310102), and the National Natural Science Foundation of China (Grants 41422601, 41376025, 41690123 and 41690120).

REFERENCES

Alexander, M. A., and C. Deser (1995), A mechanism for the recurrence of wintertime midlatitude SST anomalies, J. Phys. Oceanogr., 25, 122–137.

Alexander, M. A., D. J. Vimont, P. Chang, and J. D. Scott (2010), The impact of extratropical atmospheric variability on ENSO: Testing the seasonal footprinting mechanism using coupled model experiments, J. Climate, 23, 2885–2901; doi:10.1175/2010JCLI3205.1.

Alexander, M. A., I. Blade, M. Newman, J. R. Lanzante, N.‐C. Lau, and J. D. Scott (2002), The atmospheric bridge: The influence of ENSO teleconnections on air‐sea interaction over the global oceans, J. Climate, 15, 2205–2231.

Amaya, D.J., and G.R. Foltz (2014), Impacts of canonical and Modoki El Niño on tropical Atlantic SST, J. Geophys. Res. Oceans, 119 (2), 777–789.

An, S.‐I., and B. Wang (2000), Interdecadal change of the s tructure of ENSO mode and its impact on the ENSO f requency, J. Climate, 13, 2044–2055.

An S.‐I., and F.‐F. Jin (2001), Collective role of zonal advective and thermocline feedbacks in ENSO mode, J. Climate, 14, 3421–3432; doi: http://dx.doi.org/10.1175/1520‐0442[2001] 014<3421:CROTAZ>2.0.CO;2.

An, S.‐I., and F.‐F. Jin (2004), Nonlinearity and Asymmetry of ENSO, J. Climate, 17, 2399–2412; doi: http://dx.doi.org/10.1175/1520‐0442[2004]017<2399:NAAOE>2.0.CO;2.

An, S‐I., and J. Choi (2015), Why the twenty‐first century tropi­cal Pacific trend pattern cannot significantly influence ENSO amplitude? Climate Dyn., 44 (1–2), 133–146.

An, S.‐I., J.‐S. Kug, Y.‐G. Ham, and I.‐S. Kang (2008), Successive modulation of ENSO to the future greenhouse warming, J. Climate, 21, 3–21.

Anderson, B. T. (2004), Investigation of a large‐scale mode of ocean atmosphere variability and its relation to tropical Pacific sea surface temperature anomalies, J. Climate, 17, 1089–4098; doi: 10.1175/1520‐0442[2004]017<4089:IOALMO>2.0.CO;2.

Ardanuy, P. E., and H. L. Kyle (1986), Observed perturbations of the Earth’s radiation budget: A response to the El Chichón stratospheric aerosol layer?, J. Climate Appl. Meteor., 25, 505–516; doi:10.1175/1520‐0450[1986]025<0505:OPOTER> 2.0.CO;2.

Ashok, K., C. Tam, and W. Lee (2009), ENSO Modoki impact on the Southern Hemisphere storm track activity during

extended austral winter, Geophys. Res. Lett., 36, L12705; doi:10.1029/2009GL038847.

Ashok, K., Z. Gaun, and T. Yamagata (2001), Impact of the Indian ocean dipole on the relationship between the Indian monsoon rainfall and ENSO, Geophys. Res. Lett., 28, 4459–4502.

Ashok, K., Z. Gaun, and T. Yamagata (2003), Impact of the Indian Ocean dipole on the Australian winter rainfall. Geophys. Res. Lett., 30, 1821; doi:10.1029/2003GL017926.

Ashok, K., S. K. Behera, S. A. Rao, H. Weng, and T. Yamagata (2007), El Niño Modoki and its possible tele­connection, J.  Geophys. Res., 112, C11007; doi:10.1029/ 2006JC003798.

Baldwin, M. P., and T. J. Dunkerton (1999), Propagation of the Arctic Oscillation from the stratosphere to the troposphere, J. Geophys. Res., 104, 30937–30946; doi:10.1029/ 1999JD900445.

Barnett, T. P. (1991), The interaction of multiple time scales in the tropical climate system, J. Climate, 4, 269–285; doi: http://dx.doi.org/10.1175/1520‐0442[1991]004<0269:TIOMTS>2. 0.CO;2.

Barnett, T. P., D. W. Pierce, M. Latif, D. Dommenget, and R. Saravanan (1999), Interdecadal interactions between the tropics and midlatitudes in the Pacific basin, Geophys. Res. Lett., 26, 615–618.

Barnston, A. G., M. K. Tippett, M. L. L’Heureux, S. Li, and D. G. DeWitt (2012), Skill of real‐time seasonal ENSO model predictions during 2002–11: Is our capability increasing?, Bull. Amer. Meteor. Soc., 93, 631–651.

Barsugli, J. J., and P. D. Sardeshmukh (2002), Global a tmospheric sensitivity to tropical SST anomalies throughout the Indo‐Pacific basin, J. Climate, 15, 3427–3442.

Battisti, D. S., and A. C. Hirst (1989), Interannual variability in the tropical atmosphere‐ocean model: Influence of the basic state, ocean geometry and nonlinearity, J. Atmos. Sci., 45, 1687–1712; doi:10.1175/1520‐0469%281989%29046<1687% 3AIVIATA>2.0.CO%3B2.

Behera, S. K., and T. Yamagata (2003), Influence of the Indian Ocean dipole on the Southern Oscillation, J. Meteor. Soc. Japan, 81, 169–177.

Bjerknes, J. (1969), Atmospheric teleconnections from the equa­torial Pacific, Monthly Weather Rev., 97, 163–172; doi: http://dx.doi.org/10.1175/1520‐0493[1969]097<0163:ATFTEP>2.3. CO;2.

Cai, W., G. Wang, A. Santoso, M. J. McPhaden, L. Wu, F.‐F. Jin, A. Timmermann, M. Collins, G. Vecchi, M. Lengaigne, M. H. England, D. Dommenget, K. Takahashi, and E. Guilyardi (2015), Increased frequency of extreme La Niña events under greenhouse warming, Nature Climate Change, 5, 132–137; doi:10.1038/nclimate2492.

Cai, W., S. Borlace, M. Lengaigne, P. van Rensch, M. Collins, G. Vecchi, A. Timmermann, A. Santoso, M. J. McPhaden, L. Wu, M. H. England, G. Wang, E. Guilyardi, and Fei‐Fei Jin (2014), Increasing frequency of extreme El Niño events due to greenhouse warming. Nature Climate Change, 4, 111–116; doi: 10.1038/nclimate2100.

Camargo, S. J., and A. H. Sobel (2005), Western North Pacific tropical cyclone intensity and ENSO, J. Climate, 18, 2996–3006.

0003060439.INDD 28 6/5/2017 7:05:55 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 29

Capotondi, A., and M. A. Alexander (2010), Relationship between precipitation in the Great Plains of the United States and global SSTs: Insights from the IPCC‐AR4 models, J. Climate, 23, 2941–2958.

Capotondi, A., A. T. Wittenberg, M. Newman, E. Di Lorenzo, J.‐Y. Yu, P. Braconnot, J. Cole, B. Dewitte, B. Giese, E. Guilyardi, F.‐F. Jin, K. Karnauskas, B. Kirtman, T. Lee, N. Schneider, Y. Xue, and S.‐W. Yeh (2015), Understanding ENSO diversity, Bull. Amer. Meteor. Soc.; doi:10.1175/BAMS‐D‐13‐00117.1.

Carton, J. A., G. A. Chepurin, X. Cao, and B. S. Giese (2000), A simple ocean data assimilation analysis of the global upper ocean 1950–1995, Part I: Methodology, J. Phys. Oceanogr., 30, 294–309; doi:10.1175/1520‐0485[2000]030<0294:ASODAA>2.0.CO;2.

Cayan, D. R., K. T. Redmond, and L. G. Riddle, 1999), ENSO and hydrologic extremes in the western United States, J. Climate, 12, 2881–2893.

Chan, J. C. L. (2000), Tropical cyclone activity over the western North Pacific associated with El Niño and La Niña events, J. Climate, 13, 2960–2972.

Chang, C. P., Y. Zhang, and T. Li (2000), Interannual and inter­decadal variations of the East Asian summer monsoon and tropical Pacific SSTs, Part I: Roles of the subtropical ridge, J. Climate, 13, 4310–4325.

Chang, P., B. S. Giese, L. Ji, H. F. Seidel, and F. Wang (2001), Decadal change in the south tropical Pacific in a global assimilation analysis, Geophys. Res. Lett., 28, 3461–3464.

Chang, P., B. Wang, T. Li, and J. Lin (1995), Interactions between the seasonal cycle and El Niño–Southern Oscillation in an intermediate coupled ocean–atmosphere model, J. Atmos. Sci., 52, 2353–2372.

Chang, P., L. Ji, H. Li, and M. Flügel (1996), Chaotic dynamics versus stochastic processes in El Niño–Southern Oscillation in coupled ocean‐atmosphere models, Physica D: Nonlinear Phenomena, 98 (2), 301–320.

Chang, P., L. Zhang, R. Saravanan, D. J. Vimont, J. C. H. Chiang, L. Ji, H. Seidel, and M. K. Tippett (2007), Pacific meridional mode and El Niño–Southern Oscillation, Geophys. Res. Lett., 34, L16608; doi:10.1029/2007GL030302.

Chen, D., T. Lian, C. Fu, M. A. Cane, Y. Tang, R. Murtugudde, X. Song, Q. Wu, and L. Zhou (2015), Strong influence of westerly wind bursts on El Niño diversity, Nat. Geosci., 8 (5), 339–345.

Chen, G. (2011), How does shifting pacific ocean warming modulate on tropical cyclone frequency over the South China Sea?, J. Climate, 24, 4695–4700.

Chen, G., and C.‐Y. Tam (2010), Different impacts of two kinds of Pacific Ocean warming on tropical cyclone frequency over the western North Pacific, Geophys. Res. Lett., 37, L01803; doi:10.1029/2009GL041708.

Chen, G., J. Lu, and D. M. W. Frierson (2008), Phase speed spectra and the latitude of surface westerlies: Interannual variability and global warming trend, J. Climate, 21, 5942–5959.

Chen, L., T. Li, and Y. Yu (2015), Causes of strengthening and weakening of ENSO amplitude under global warming in four CMIP5 models, J. Climate, 28, 3250–3274; doi:10.1175/JCLI‐D‐14‐00439.1.

Chia, H. H., and C. F. Ropelewski (2002), The interannual vari­ability in the genesis location of tropical cyclones in the northwest Pacific, J. Climate, 15, 2934–2944.

Chiang, J. C. H., and D. J. Vimont (2004), Analogous Pacific and Atlantic meridional modes of tropical atmosphere‐ocean variability, J. Climate, 17, 4143–4158; doi:10.1175/JCLI4953.1.

Chiang, J. C., and Sobel, A. H. (2002), Tropical tropospheric temperature variations caused by ENSO and their influence on the remote tropical climate, J. Climate, 15, 2616–2631.

Chiodi, A. M., and D. E. Harrison (2013), El Niño impacts on seasonal U.S. Atmospheric circulation, temperature, and precipitation anomalies: The OLR‐event perspective, J.  Climate, 26, 822–837; doi: http://dx.doi.org/10.1175/JCLI‐D‐12‐00097.1.

Choi, J., S. ‐I. An, and S. W. Yeh (2012), Decadal amplitude modulation of two types of ENSO and its relationship with the mean state, Climate Dyn., 38 (11–12), 2631–2644.

Choi, J., S. ‐I. An, B. Dewitte, and W. W. Hsieh (2009), Interactive feedback between the tropical Pacific Decadal Oscillation and ENSO in a coupled general circulation model, J Climate, 22, 6597–6611.

Chou, C., L.‐F. Huang, J.‐Y. Tu, L. Tseng, and Y.‐C. Hsueh (2009), El Niño impacts on precipitation in the western North Pacific–East Asian sector, J. Climate, 22, 2039–2057.

Chung, P.‐H., and T. Li (2013), Interdecadal relationship between the mean state and El Niño types, J. Climate, 26, 361–379; doi:10.1175/JCLI‐D‐12‐00106.1.

Ciasto, L. M., and D. W. J. Thompson (2008), Observations of large scale ocean‐atmosphere interaction in the Southern Hemisphere, J. Climate, 21, 1244–1259; doi:10.1175/2007JCLI1809.1.

Clement, A., P. DiNezio, and C. Deser (2011), Rethinking the ocean’s role in the Southern Oscillation, J. Climate, 24, 4056–4072; doi:10.1175/2011JCLI3973.1.

Collins, M. (2000a), Understanding uncertainties in the response of ENSO to greenhouse warming, Geophys. Res. Lett., 27, 3509–3513.

Collins, M. (2000b), The El Niño Southern Oscillation in the second Hadley centre coupled model and its response to greenhouse warming, J. Climate, 13, 1299–1312.

Collins, M., and coauthors (2010), The impact of global warm­ing on the tropical Pacific Ocean and El Niño, Nat. Geosci., 3, 391–397; doi:10.1038/NGEO868.

Cummins, P. F., G. S. E. Lagerloef, and G. Mitchum (2005), A  regional index of northeast Pacific variability based on s atellite altimeter data, Geophys. Res. Lett., 32, L17607; doi:10.1029/2005GL023642.

Dettinger, M. D., D. R. Cayan, H. F. Diaz, and D. M. Meko (1998), North‐south precipitation patterns in western North America on interannual‐to‐decadal timescales, J. Climate, 11, 3095–3111.

DeWeaver, E., and S. Nigam (2004), On the forcing of ENSO teleconnections by anomalous heating and cooling, J. Climate, 17, 3225–3235.

Ding, Q., E. J. Steig, D. S. Battisti, and M. Kuettel (2011), Winter warming in West Antarctica caused by central tropical Pacific warming, Nat. Geosci., 4, 398–403; doi:10.1038/ngeo1129.

Dommenget, D. (2010), The slab ocean El Niño, Geophys. Res. Lett., 37, L20701; doi:10.1029/2010/GL044888.

0003060439.INDD 29 6/5/2017 7:05:55 PM

30 FORCINGS OF CLIMATE EXTREMES

Dong, B., R. T. Sutton, and A. A. Scaife (2006), Multidecadal modulation of El Niño–Southern Oscillation (ENSO) vari­ance by Atlantic Ocean sea surface temperatures, Geophys. Res. Lett., 33, L08705; doi:10.1029/2006GL025766.

Dong, B. W., and R. Y. Lu (2013), Interdecadal enhancement of the Walker circulation over the tropical Pacific in the late 1990s, Adv. Atmos. Sci., 30 (2), 247–262; doi: 10.1007/s00376‐012‐2069‐9.

Duan, W., H. Xu, and M. Mu (2008), Decisive role of nonlinear temperature advection in El Niño and La Niña amplitude asymmetry, J. Geophys. Res., 113, C01014; doi:10.1029/2006JC 003974.

Enfield, D. B., and D. A. Mayer (1997), Tropical Atlantic sea surface temperature variability and its relation to El Niño–Southern Oscillation, J. Geophys. Res., 102, 929–945; doi:10.1029/96JC03296.

England, M. H., S. McGregor, P. Spence, G. A. Meehl, A. Timmermann, W. Cai, A. S. Gupta, M. J. McPhaden, A. Purich, and A. Santoso (2014), Recent intensification of wind‐driven circulation in the Pacific and the ongoing warm­ing hiatus, Nature Climate Change, 4, 222–227; doi:10.1038/nclimate2106.

Evtushevsky, O. M., V. O. Kravchenko, L. L. Hood, and G. P. Milinevsky (2014), Teleconnection between the central tropi­cal Pacific and the Antarctic stratosphere: Spatial patterns and time lags, Climate Dyn., 1–15.

Fedorov, A. V., and S. G. Philander (2000), Is El Niño changing?, Science, 288, 1997–2002.

Fedorov, A. V., S. Hu, M. Lengaigne, and E. Guilyardi (2014), The impact of westerly wind bursts and ocean initial state on the development, and diversity of El Niño events, Climate Dyn., 44 (5–6), 1381–1401.

Feng, J., and J. Li (2011), Influence of El Niño Modoki on spring rainfall over South China, J. Geophys. Res., 116, D13102; doi:10.1029/2010JD015160.

Feng, J., W. Chen, C.‐Y. Tam, and W. Zhou (2011), Different impacts of El Niño and El Niño Modoki on China rainfall in  the decaying phases, Int. J. Climatol., 31, 2091–2101; doi:10.1002/joc.2217.

Fogt, R. L., and D. H. Bromwich (2006), Decadal variability of the ENSO teleconnection to the high‐latitude South Pacific governed by coupling with the southern annular mode, J. Climate, 19, 979–997; doi:10.1175/JCLI3671.1.

Frauen, C., and D. Dommenget (2010), El Niño and La Niña amplitude asymmetry caused by atmospheric feedbacks, Geophys. Res. Lett., 37, L18801; doi:10.1029/2010GL044444.

Fu, C., H. F. Diaz, J. O. Fletcher (1986), Characteristics of the response of sea surface temperature in the central Pacific associated with warm episodes of the Southern Oscillation. Monthly Weather Rev., 114, 1716–1738.

García‐Herrera, R., N. Calvo, R. R. Garcia, and M. A. Giorgetta (2006), Propagation of ENSO temperature signals into the middle atmosphere: A comparison of two general circulation models and ERA‐40 reanalysis data, J. Geophys. Res., 111, D06101; doi:10.1029/2005JD006061.

Garfinkel, C. I., and D. L. Hartmann (2008), Different ENSO teleconnections and their effects on the stratospheric polar vor­tex, J. Geophys. Res., 113, D18114; doi:10.1029/2008JD009920.

Garfinkel, C. I., D. L. Hartmann, and F. Sassi (2010), Tropospheric precursors of anomalous Northern Hemisphere

stratospheric polar vortices, J. Climate, 23, 3282–3299; doi:10.1175/2010JCLI3010.1.

Garfinkel, C. I., M. M. Hurwitz, D. W. Waugh, and A. H. Butler (2013), Are the teleconnections of central Pacific and eastern Pacific El Niño distinct in boreal wintertime?, Climate Dyn., 41, 1835–1852; doi:10.1007/s00382‐012‐1570‐2.

Goldenberg, S. B., and L. J. Shapiro (1996), Physical mecha­nisms for the association of El Niño and West African r ainfall with Atlantic major hurricane activity, J. Climate, 9, 1169–1187; doi:10.1175/1520‐0442[1996] 009<1169:PMFTAO >2.0.CO;2.

Graf, H. F., and D. Zanchettin (2012), Central Pacific El Niño, the “subtropical bridge,” and Eurasian climate, J. Geophys. Res., 117, D01102; doi:10.1029/2011JD016493.

Gray, W. M. (1984), Atlantic seasonal hurricane frequency. Part I: El Niño and 30‐mb quasi‐biennial oscillation influences. Monthly Weather Rev., 112, 1649–1668.

Grimm, A. M. (2003), The El Niño impact on the summer m onsoon in Brazil: Regional processes versus remote influences, J. Climate, 16, 263–280.

Gruber, A., and A. F. Krueger (1984), The status of the NOAA outgoing longwave radiation data set, Bull. Amer. Meteor. Soc., 65, 958–962; doi:10.1175/1520‐0477[1984]065<0958:TSOTNO>2.0.CO;2.

Gu, D., and S. G. H. Philander (1995), Secular changes of annual and interannual variability in the tropics during the past century, J. Climate, 8, 864–876; doi: http://dx.doi.org/10.1175/1520‐0442[1995]008<0864:SCOAAI>2.0.CO;2.

Gu, D., and S. G. H. Philander (1997), Interdecadal climate fluctuations that depend on exchanges between the tropics and extratropics, Science, 275, 805–807.

Guilyardi, E. (2006), El Niño‐mean state‐seasonal cycle inter­actions in a multi‐model ensemble, Climate Dyn., 26, 329–348.

Hannachi, A., D. Stephenson, and K. Sperber (2003), Probability‐based methods for quantifying nonlinearity in the ENSO, Climate Dyn., 20, 241–256.

Hartmann, D. L., and F. Lo (1998), Wave‐driven zonal flow vacillation in the Southern Hemisphere, J. Atmos. Sci., 55(8), 1303–1315.

He, H., J. Yang, D. Gong, R. Mao, Y. Wang, and M. Gao (2015), Decadal changes in tropical cyclone activity over the western North Pacific in the late 1990s, Climate Dyn., 1–13.

He, Y., and C. Guan (1997), Interannual and interdecadal v ariability in heat content of the upper ocean of South China Sea, Adv. Atmos. Sci., 16, 23–29.

Heddinghaus, T. R., and A. F. Krueger (1981), Annual and interannual variations in outgoing longwave radiation over the tropics, Monthly Weather Rev., 109, 1208–1218; doi: http://dx.doi.org/10.1175/1520‐0493[1981]109<1208:AAIVIO>2.0.CO;2.

Hegyi, B. M., and Y. Deng (2011), A dynamical fingerprint of tropical Pacific sea surface temperatures on the decadal‐scale variability of cool‐season Arctic precipitation, J. Geophys. Res., 116, D20121; doi:10.1029/2011JD016001.

Hill, K. J., A. S. Taschetto, and M. H. England (2009), South American rainfall impacts associated with inter‐El Niño v ariations, Geophys. Res. Lett., 36, L19702; doi:10.1029/ 2009GL040164.

0003060439.INDD 30 6/5/2017 7:05:56 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 31

Ho, C. H., J. J. Baik, J. H. Kim, D. Y. Gong, and C. H. Sui (2004), Interdecadal changes in summertime typhoon tracks, J. Climate, 17, 1767–1776.

Hoerling, M. P., and A. Kumar (2002), Atmospheric response patterns associated with tropical forcing, J. Climate, 15, 2184–2203.

Holton, J. R. (1976), A semi‐spectral numerical model for wave‐mean flow interactions in the stratosphere: applications to sudden stratospheric warmings, J. Atmos. Sci., 33, 1639–1649.

Hong, C.‐C., Y.‐H. Li, T. Li, and M.‐Y. Lee (2011), Impacts of central Pacific and eastern Pacific El Nino on tropical cyclone tracks over the western North Pacific, Geophys. Res. Lett., 38, L16712; doi:10.1029/2011GL048821.

Horii, T., I. Ueki, and K. Hanawa (2012), Breakdown of ENSO predictors in the 2000s: Decadal changes of recharge/discharge‐SST phase relation and atmospheric intraseasonal forcing, Geophys. Res. Lett., 39, L10707; doi:10.1029/2012GL051740.

Hsieh, W. W. (2004), Nonlinear multivariate and time series analysis by neural network methods, Rev. Geophys., 42, RG1003; doi:10.1029/2002RG000112.

Hu, S., A. V. Fedorov, M. Lengaigne, and E. Guilyardi (2014), The impact of westerly wind bursts on the diversity and pre­dictability of El Niño events: an ocean energetics perspective. Geophys. Res. Lett., 41, 4654–4663; doi:10.1002/2013GL058954.

Hu, Z.‐Z., A. Kumar, B. Jha, and B. Huang (2012a), An analy­sis of forced and internal variability in a warmer climate in CCSM3, J. Climate, 25 (7), 2356–2373; doi: 10.1175/JCLI‐D‐11‐00323.1.

Hu, Z.‐Z., A. Kumar, B. Jha, W. Wang, B. Huang, and B. Huang (2012b), An analysis of warm pool and cold tongue El Niños: Air‐sea coupling processes, global influences, and recent trends, Climate Dyn., 38 (9–10), 2017–2035; doi: 10.1007/s00382‐011‐1224‐9.

Hu, Z.‐Z., A. Kumar, H.‐L. Ren, H. Wang, M. L’Heureux, and F.‐F. Jin (2013), Weakened interannual variability in the trop­ical Pacific Ocean since 2000, J. Climate, 26 (8), 2601–2613; doi: 10.1175/JCLI‐D‐12‐00265.1.

Huang, H. P., R. Seager, and Y. Kushnir (2005), The 1976/77 transition in precipitation over the Americas and the influ­ence of tropical sea surface temperature, Climate Dyn., 24 (7–8), 721–740.

Hurwitz, M. M., C. I. Garfinkel, P. A. Newman, and L. D. Oman (2013), Sensitivity of the atmospheric response to warm pool El Niño events to modeled SSTs and future c limate forcings, J. Geophys. Res. Atmos., 118 (13), 371–13,382; doi:10.1002/2013JD021051.

Hurwitz, M. M., I. S. Song, L. D. Oman, P. A. Newman, A. M. Molod, S. M. Frith, and J. E. Nielsen (2011b), Response of the Antarctic stratosphere to warm pool El Niño events in the  GEOS CCM, Atmos. Chem. Phys., 11, 9659–9669; doi:10.5194/acp‐11‐9659‐2011.

Hurwitz, M. M., N. Calvo, C. I. Garfinkel, A. H. Butler, S. Ineson, C. Cagnazzo, E. Manzini, and C. Peña‐Ortiz (2014), Extra‐tropical atmospheric response to ENSO in the CMIP5 models, Climate Dyn., 43 (12), 3367–3376; doi:10.1007/s00382‐014‐2110‐z.

Hurwitz, M. M., P. A. Newman, L. D. Oman, and A. M. Molod (2011a), Response of the Antarctic stratosphere to two types of El Niño events, J. Atmos. Sci., 68, 812–822.

Iza, M., and N. Calvo (2015), Role of stratospheric sudden warmings on the response to central Pacific El Niño, Geophys. Res. Lett., 42, 2482–2489; doi:10.1002/2014GL062935.

Jiang, N., J. D. Neelin, and M. Ghil (1995), Quasi‐quadrennial and quasi‐biennial variability in the equatorial Pacific, Climate Dyn., 12, 101–112.

Jin, F.‐F. (1996), Tropical ocean‐atmosphere interaction, the pacific cold tongue, and the El Niño Southern Oscillation, Science, 274, 76–78; doi: 10.1126/science.274.5284.76.

Jin, F.‐F. (1997), An equatorial recharge paradigm for ENSO, I. Conceptual model, J. Atmos. Sci., 54, 811–829.

Jin, F.‐F., J. D. Neelin and M. Ghil (1994), El Niño on the devil’s staircase: Annual and subharmonic steps to chaos, Science, 264, 70–72; doi: 10.1126/science.264.5155.70.

Juneng, L., and F. T. Tangang (2005), Evolution of ENSO‐related rainfall anomalies in Southeast Asia region and its relationship with atmosphere ocean variations in Indo‐Pacific sector, Climate Dyn., 25(4), 337–350; doi:10.1007/s00382‐005‐0031‐6.

Kalnay, E., et al. (1996), The NCEP/NCAR 40‐year reanalysis project, Bull. Amer. Meteor. Soc., 77, 437–471.

Kao, H. Y., and J. ‐Y. Yu (2009), Contrasting eastern‐Pacific and central‐Pacific types of ENSO, J. Climate, 22, 615–632; doi:10.1175/2008JCLI2309.1.

Karoly, D. J. (1989), Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events, J. Climate, 2, 1239–1252; doi:10.1175/1520‐0442[1989]002, 1239:SHCFAW.2.0.CO;2.

Kerr, R. A. (2000), A North Atlantic climate pacemaker for the  centuries, Science, 288, 1984–1986; doi:10.1126/science. 288.5473.1984.

Kessler, W. S. (2002), Is ENSO a cycle or a series of events?, Geophys. Res. Lett., 29[23], 2125; doi:10.1029/2002GL015924.

Kiladis, G. N., and H. F. Diaz (1989), Global climatic anoma­lies associated with extremes in the Southern Oscillation, J. Climate, 2, 1069–1090.

Kim, H.‐M., P. J. Webster, and J. A. Curry (2009), Impact of  shifting patterns of Pacific Ocean warming on north Atlantic tropical cyclones, Science, 325, 77–80.

Kim, H. ‐M., P. J. Webster, and J. A. Curry (2011), Modulation of North Pacific tropical cyclone activity by three phases of ENSO, J. Climate, 24, 1839–1849.

Kim, J.‐S., W. Zhou, X. Wang, and S. Jain (2012), El Niño Modoki and the summer precipitation variability over South Korea: A diagnostic study, J. Meteor. Soc. Japan, 90, 673–684; doi: 10.2151/jmsj.2012‐507.

Kim, S. ‐T., and J.‐Y. Yu (2012), The two types of ENSO in CMIP5 models, Geophys. Res. Lett., 39, L11704; doi:10.1029/ 2012GL052006.

Kim, S. ‐T., J.‐Y. Yu, A. Kumar, and H. Wang (2012), Examination of the two types of ENSO in the NCEP CFS model and its extratropical associations, Monthly Weather Rev., 140, 1908–1923; doi:10.1175/MWR‐D‐11‐00300.1.

Kim, S.‐T., W. Cai, F.‐F. Jin, A. Santoso, L. Wu, E. Guilyardi, and S.‐I. An (2014), Response of El Niño sea surface tem­perature variability to greenhouse warming, Nature Climate Change, 4, 786–790; doi: 10.1038/NCLIMATE2326.

Kinter, J. L., K. Miyakoda, and S. Yang (2002), Recent change in the connection from the Asian monsoon to ENSO, J. Climate, 15, 1203–1215.

0003060439.INDD 31 6/5/2017 7:05:56 PM

32 FORCINGS OF CLIMATE EXTREMES

Klein, S. A., B. J. Soden, and N.‐C. Lau (1999), Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge, J. Climate, 12, 917–932.

Knutson, T. R., S. Manabe, and D. Gu (1997), Simulated ENSO in a global coupled ocean‐atmosphere model: Multidecadal amplitude modulation and CO2 sensitivity, J. Climate, 10, 131–161.

Kohl, A., D. Dommenget, K. Ueyoshi, and D. Stammer (2006), The global ECCO 1952 to 2001 ocean synthesis, ECCO Tech. Rep., 40.

Kug, J.‐S., F.‐F. Jin, and S.‐I. An (2009), Two types of El Niño events: Cold tongue El Niño and warm pool El Niño, J. Climate, 22, 1499–1515; doi:10.1175/2008JCLI2624.1.

Kug, J.‐S., J. Choi, S.‐I. An, F.‐F. Jin, and A. T. Wittenberg (2010), Warm pool and cold tongue El Niño events as simu­lated by the GFDL CM2.1 coupled GCM, J. Climate, 23, 1226–1239; doi:10.1175/2009JCLI3293.1.

Kumar, A., and Z.‐Z. Hu (2014), Interannual and interdecadal variability of ocean temperature along the equatorial Pacific in conjunction with ENSO, Climate Dyn., 42 (5–6), 1243–1258; doi: 10.1007/s00382‐013‐1721‐0.

Kumar, K. K., B. Rajagopatan, M. Hoerling, G. Bates, and M. Cane (2006), Unraveling the mystery of Indian monsoon f ailure during El Niño, Science, 314, 115–119.

Kunz, T., K. Fraedrich, and F. Lunkeit (2009), Impact of s ynoptic‐scale wave breaking on the NAO and its connection with the stratosphere in ERA‐40, J. Climate; doi:10.1175/ 2009JCLI2750.1.

Kwok, R., and J. C. Comiso (2002), Spatial patterns of variability in Antarctic surface temperature: Connections to the South Hemisphere annular mode and the Southern Oscillation, Geophys. Res. Lett., 29, 1705; doi:10.1029/2002GL015415.

Larkin, N. K., and D. E. Harrison (2005a), On the definition of  El Niño and associated seasonal average U.S. weather anomalies, Geophys. Res. Lett., 32, L13705; doi:10.1029/ 2005GL022738.

Larkin, N. K., and D. E. Harrison (2005b), Global seasonal temperature and precipitation anomalies during El Niño autumn and winter, Geophys. Res. Lett., 32, L16705; doi:10.1029/2005GL022860.

Larson, S., S.‐K. Lee, C. Wang, E.‐S. Chung, and D. Enfield (2012), Impacts of non‐canonical El Nino patterns on Atlantic hurricane activity, Geophys. Res. Lett., 39, L14706; doi:10.1029/ 2012GL052595.

Larson, S. M., and B. P. Kirtman (2014), The Pacific meridional mode as an ENSO precursor and predictor in the North American multimodel ensemble, J. Climate, 27, 7018–7032; doi: http://dx.doi.org/10.1175/JCLI‐D‐14‐00055.1.

Latif, M., and T. P. Barnett (1994), Causes of decadal climate variability over the North Pacific and North America, Science, 266, 634–637.

Latif, M., D. Anderson, T. Barnett, M. Cane, R. Kleeman, A. Leetmaa, J. O’Brien, A. Rosati, and E. Schneider (1998), A review of the predictability and prediction of ENSO, J. Geophys. Res., 103 (C7), 14375–14393; doi:10.1029/ 97JC03413.

Lau, K.‐M., and P. H. Chan (1983), Short‐term climate varia­bility and atmospheric teleconnections from satellite‐observed outgoing longwave radiation. Part I: Simultaneous

relationships, J. Atmos. Sci., 40, 2735–2750; doi:10.1175/1520‐0469[1983]040<2735:STCVAA>2.0.CO;2.

Lau, K.‐M., and P. H. Chan (1985), Aspects of the 40–50 day oscillation during the northern winter as inferred from o utgoing longwaveradiation, Monthly. Weather Rev., 113, 1889–1909; doi:10.1175/1520‐0493[1985]113<1889:AOTDOD>2.0.CO;2.

Lau, N. C., and M. J. Nath (2004), Coupled GCM simulation of atmosphere‐ocean variability associated with zonally asym­metric SST changes in the tropical Indian Ocean, J. Climate, 17, 245–265.

Lee, E.‐J., K.‐J. Ha, and J.‐G. Jhun (2014), Interdecadal changes in interannual variability of the global monsoon precipi­tation and interrelationships among its subcomponents, Climate Dyn., 42, 2585–2601.

Lee, S.‐K., C. Wang, and D. B. Enfield (2010), On the impact of central Pacific warming events on Atlantic tropical storm activity, Geophys. Res. Lett., 37, L17702; doi:10.1029/2010GL044459.

Lee, T., and coauthors (2010), Record warming in the South Pacific and western Antarctica associated with the strong central Pacific El Niño in 2009–10, Geophys. Res. Lett., 37, L19704; doi:10.1029/2010GL044865.

Lee, T., and M. J. McPhaden (2010), Increasing intensity of El Niño in the central‐equatorial Pacific, Geophys. Res. Lett., 37, L14603; doi:10.1029/2010GL044007.

L’Heureux, M. L., and D. W. J. Thompson (2006), Observed relationships between the El Niño–Southern Oscillation and the extratropical zonal‐mean circulation, J. Climate, 19, 276–287; doi:10.1175/JCLI3617.1.

L’Heureux, M. L., S. Lee, and B. Lyon (2013), Recent multi­decadal strengthening of the Walker circulation across the tropical Pacific, Nature Climate Change, 3, 571–576.

Li, Y., and S. Yang (2010), A dynamical index for the east Asian winter monsoon, J. Climate, 23, 4255–4262.

Liang, Y.‐C., J.‐Y. Yu, M.‐H. Lo, and C. Wang (2015), The changing influence of El Niño on the Great Plains low‐level jet, Atmosph. Sci. Lett.; doi: 10.1002/asl.590.

Liang, Y.‐C., M.‐H. Lo, and J.‐Y. Yu (2014), Asymmetric responses of land hydroclimatology to two types of El Niño in the Mississippi River basin, Geophys. Res. Lett., 41, 582–588; doi:10.1002/2013GL058828.

Lim, E.‐P., and H. H. Hendon, D. Hudson, G. Wang, and O. Alves (2009), Dynamical forecast of inter–El Niño variations of tropical SST and Australian spring rainfall, Monthly Weather Rev., 137, 3796–3810; doi:10.1175/2009MWR2904.1.

Lin, C.‐Y., J.‐Y. Yu, and H. H. Hsu (2014), CMIP5 model s imulations of the Pacific meridional mode and its connection to the two types of ENSO, Int. J. Climatol.; doi:10.1002/joc.4130.

Lin, P., Q. Fu, D. L. Hartmann (2012), Impact of tropical SST on stratospheric planetary waves in the Southern Hemisphere, J. Climate, 25, 5030–5046; doi:10.1175/JCLI‐D‐11‐00378.1.

Linkin, M. E., and S. Nigam (2008), The north Pacific Oscillation‐west Pacific teleconnection pattern: Mature‐phase structure and winter impacts, J. Climate, 21, 1979–1997; doi:10.1175/2007JCLI2048.1.

Liu, J., J. A. Curry, and D. G. Martinson (2004), Interpretation of recent Antarctic sea ice variability, Geophys. Res. Lett., 31, L02205; doi:10.1029/2003GL018732.

0003060439.INDD 32 6/5/2017 7:05:56 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 33

Liu, Q.‐Y., D. Wang, X. Wang, Y. Shu, Q. Xie, J. Chen (2014), Thermal variations in the South China Sea associated with the eastern and central Pacific El Niño events and their mechanisms, J. Geophys. Res. Oceans, 119, 8955–8972; doi:10.1002/2014JC010429.

Liu, Z., and M. Alexander (2007), Atmospheric bridge, oceanic tunnel, and global climatic teleconnections, Rev. Geophys., 45, RG2005; doi:10.1029/2005RG00017.

Livezey, R. E., M. Masutani, A. Leetmaa, A., H. Rui, M. Ji, and A. Kumar (1997), Teleconnective response of the Pacific–North American region atmosphere to large central equatorial Pacific SST anomalies, J. Climate, 10, 1787–1820.

Lu, J., G. Chen, and D. M. W. Frierson (2008), Response of the zonal mean atmospheric circulation to El Niño versus global warming, J. Climate, 21, 5835–5851.

Luo, J. J., S. Masson, S. Behera, S. Shingu, and T. Yamagata (2005), Seasonal climate predictability in a coupled OAGCM using a different approach for ensemble forecasts, J. Climate, 18, 4474–4497.

Maloney, E. D., and coauthors (2014), North American climate in CMIP5 experiments: Part III: Assessment of twenty‐first‐century projections, J. Climate, 27, 2230–2270; doi: http://dx.doi.org/10.1175/JCLI‐D‐13‐00273.1.

Mantua, N. J., and D. S. Battisti (1995), Aperiodic variability in the Zebiak‐cane coupled ocean‐atmosphere model: Air‐sea interactions in the western equatorial Pacific, J. Climate, 8, 2897–2927; doi: http://dx.doi.org/10.1175/1520‐0442[1995] 008<2897:AVITZC>2.0.CO;2.

Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis (1997), A Pacific interdecadal climate oscillation with impacts on salmon production, Bull. Amer. Meteor. Soc., 78, 1069–1079.

Manzini, E., M. A. Giorgetta, M. Esch, L. Kornblueh, and E. Roeckner (2006), The influence of sea surface temperatures on the northern winter stratosphere: Ensemble simulations with the MAECHAM5 model, J. Climate, 19, 3863–3881; doi:10.1175/JCLI3826.1.

McBride, J., M. Haylock, and N. Nicholls (2003), Relationships between the maritime continent heat source and the El Niño–Southern Oscillation phenomenon, J. Climate, 16 (17), 2905–2914; doi:10.1175/1520–0442 [2003]016<2905:RBTMCH> 2.0.CO;2.

McCreary, J. P., and P. Lu (1994), Interaction between the sub­tropical and equatorial ocean circulations: The subtropical cell, J. Phys. Oceanogr., 24, 466–497.

McIntyre, M. E. (1982), How well do we understand the dynamics of stratospheric warming?, J. Meteor. Soc. Japan, 60, 37–64.

McPhaden, M. J. (2008), Evolution of the 2006–2007 El Niño: The role of intraseasonal to interannual time scale dynamics, Adv. Geosci., 14, 219–230; doi:10.5194/adgeo‐14‐219‐2008.

McPhaden, M. J. (2012), A 21st century shift in the relationship between ENSO SST and warm water volume anomalies, Geophys. Res. Lett., 39, L09706; doi:10.1029/2012GL051826.

McPhaden, M. J., and D. Zhang (2002), Slowdown of the meridional overturning circulation in the upper Pacific Ocean, Nature, 415, 603–608.

McPhaden, M. J., T. Lee, and D. McClurg (2011), El Niño and its relationship to changing background conditions in the

tropical Pacific Ocean, Geophys. Res. Lett., 38, L15709; doi:10.1029/2011GL048275.

Mechoso, C. R., D. L. Hartmann, and J. D. Farrara (1985), Climatology and interannual variability of wave, mean‐flow interaction in the Southern Hemisphere, J. Atmos. Sci., 42, 2189–2206.

Meehl, G. A., and H. Teng (2007), Multi‐model changes in El  Niño teleconnections over North America in a future warmer climate, Climate Dyn., 29, 779–790; doi:10.1007/s00382‐007‐0268‐3.

Meehl, G. A., and W. M. Washington (1996), El Niño‐like c limate change in a model with increased atmospheric CO2 concentrations, Nature, 382 (6586), 56–60.

Meehl, G. A., et  al. (2007), The WCRP CMIP3 multimodel dataset: A new era in climate change research, Bull. Amer. Meteorol. Soc., 88, 1383–1394; doi:10.1175/BAMS‐88‐9‐1383.

Meehl, G. A., G. W. Branstator, and W. M. Washington (1993), Tropical Pacific interannual variability and CO2 climate change, J. Climate, 6, 42–63.

Meehl, G. A., W. D. Collins, B. A. Boville, J. T. Kiehl, T. M. L.Wigley, and J. M. Arblaster (2000), Response of the NCAR climate system model to increased CO2 and the role of p hysical processes, J. Climate, 13 (11), 1879–1898.

Meinen, C. S., and M. J. McPhaden (2000), Observations of  warm water volume changes in the equatorial Pacific and their relationship to El Niño and La Niña, J. Climate, 13, 3551–3559; doi:10.1175/1520‐0442[2000] 013<3551: OOWWVC>2.0.CO;2.

Merrifield, M. A. (2011), A shift in western tropical Pacific sea level trends during the 1990s, J. Climate, 24, 4126–4138; doi: http://dx.doi.org/10.1175/2011JCLI3932.1.

Merryfield, W. J. (2006), Changes to ENSO under CO2 d oubling in a multimodel ensemble, J. Climate, 19, 4009–4027.

Minobe, S. (2000), Spatio‐temporal structure of the pentadecadal variability over the North Pacific, Prog. Oceanogr., 47, 381–408; doi:10.1016/S0079‐6611[00]00042‐2.

Minobe, S. (2002), Interannual to interdecadal changes in the Bering Sea and concurrent 1998/99 changes over the North Pacific, Prog. Oceanogr., 55, 45–64; doi:10.1016/S0079‐6611[02]00069‐1.

Mo, K. C. (2000), Relationships between interdecadal variability in the Southern Hemisphere and sea surface temperature anomalies, J. Climate, 13, 3599–3610.

Mo, K. C. (2010), Interdecadal modulation of the impact of ENSO on precipitation and temperature over the United States, J. Climate, 23, 3639–3656; doi: 10.1175/2010JCLI3553.1.

Mo, K. C., and M. Ghil (1987), Statistics and dynamics of persistent anomalies, J. Atmos. Sci., 44, 877–902; doi:10.1175/ 1520‐0469[1987]044,0877:SADOPA.2.0.CO;2.

Mo, K. C., and R. E. Livezey (1986), Tropical‐extratropical geopotential height teleconnections during the Northern Hemisphere winter, Monthly Weather Rev., 114, 2488–2515.

Mo, K. C., and R. W. Higgins (1998), Tropical convection and precipitation regimes in the western United States, J. Climate, 11, 2404–2423.

Monahan, A. H. (2001), Nonlinear principal component analysis: Tropical Indo‐Pacific sea surface temperature and sea level pressure, J. Climate, 14, 219–233.

Na, H., B.‐G. Jang, W.‐M. Choi, and K.‐Y. Kim (2011), Statistical simulations of the future 50‐year statistics of cold‐tongue

0003060439.INDD 33 6/5/2017 7:05:56 PM

34 FORCINGS OF CLIMATE EXTREMES

El Niño and warm‐pool El Niño, Asia‐Pacific J. Atmos. Sci., 47, 223–233; doi:10.1007/s13143‐011‐0011‐1.

Neale, R. B., J. Richter, S. Park, P. H. Lauritzen, S. J. Vavrus, P. J. Rasch, and M. Zhang (2013), The mean climate of the community atmosphere model (CAM4) in forced SST and fully coupled experiments, J. Climate, 26, 5150–5168; doi:10.1175/JCLI‐D‐12‐00236.1.

Neelin, J. D. (1991), The slow sea surface temperature mode and the fast‐wave limit: Analytic theory for tropical interannual oscillations and experiments in a hybrid coupled model, J. Atmos. Sci., 48, 584–606.

Neelin, J. D., and F.‐F. Jin (1993), Modes of interannual tropical ocean‐atmosphere interaction‐a unified view, II, Analytical results in the weak coupling limit, J. Atmos. Sci., 50, 3504–3522.

Neelin, J. D., D. S. Battisti, A. C. Hirst, F.‐F. Jin, Y. Wakata, T.  Yamagata, and S. E. Zebiak (1998), ENSO theory, J.  Geophys. Res., 103 (C7), 14261–14290; doi:10.1029/ 97JC03424.

Newman, M., et  al. (2015), The Pacific decadal oscillation, revisited, Bull. Amer. Meteor. Soc. (submitted).

Newman, M., S.‐I. Shin, and M. A. Alexander (2011), Natural variation in ENSO flavors. Geophys. Res. Lett., L14705; doi:10.1029/2011GL047658.

Nishii, K., H. Nakamura, and Y. J. Orsolini (2010), Cooling of the wintertime Arctic stratosphere induced by the western Pacific teleconnection pattern, Geophys. Res. Lett., 371, L13805; doi:10.1029/2010GL043551.

Noda, A., K. Yoshimatsu, S. Yukimoto, K. Yamahuchi, and S. Yamaki (1999), Relationship between natural variability and CO2‐induced warming pattern: MRI AOGCM experiment, in 10th Symposium on Global Change Studies, American Meteorological Society Publication.

Paek, H., J.‐Y., Yu, J.‐W. Hwu, M.‐M. Lu, and T. Gao (2015), A  source of AGCM bias in simulating the western Pacific subtropical high: Different sensitivities to the two types of ENSO, Monthly Weather Rev., 143, 2348–2362.

Peterson, W. T., and F. B. Schwing (2003), A new climate regime in northeast pacific ecosystems, Geophys. Res. Lett., 30, 1896; doi:10.1029/2003GL017528.

Philander, S. G. (1999), A review of tropical ocean–atmosphere interactions, Tellus A, 51 (1), 71–90.

Pierce, D. W., T. P. Barnett, and M. Latif (2000), Connections between the Pacific Ocean tropics and midlatitudes on d ecadal timescales, J. Climate, 13, 1173–1194.

Qian, C., J. ‐Y. Yu, and G. Chen (2014), Decadal summer drought frequency in China: The increasing influence of the Atlantic multi‐decadal oscillation, Environ. Res. Lett., 9 (12), 124004; doi:10.1088/1748‐9326/9/12/124004.

Qiu, B., and S. Chen (2012), Multidecadal sea level and gyre circulation variability in the northwestern tropical Pacific Ocean, J. Phys. Oceanogr., 42, 193–206; doi: http://dx.doi.org/10.1175/JPO‐D‐11‐061.1.

Rajagopolan, B., U. Lall, and M. Cane (1997), Anomalous ENSO occurrences: An alternate view, J. Climate, 10, 2351–2357.

Rasmusson, E. M., and T. H. Carpenter (1982), Variations in tropical sea surface temperature and surface wind fields a ssociated with the Southern Oscillation/El Niño, Monthly

Weather Rev., 110, 354–384; doi: http://dx.doi.org/10.1175/ 1520‐0493[1982]110<0354:VITSST>2.0.CO;2.

Rasmusson, E. M., X. Wang, and C. F. Ropelewski (1990), The biennial component of ENSO variability, J. Mar. Syst., 1, 71–90; doi:10.1016/0924‐7963[90]90153‐2.

Ratnam, J. V., S. K. Behera, Y. Masumoto, and T. Yamagata (2014), Remote effects of El Niño on Modoki events on the austral summer precipitation of southern Africa, J. Climate, 27, 3802–3815; doi:10.1175/JCLI‐D‐13‐00431.1.

Rayner, N. A., et al. (2003), Global analyses of sea surface t emperature, sea ice, and night marine air temperature since the late nineteenth century, J. Geophys. Res., 108 (D14), 4407.

Ren, H.‐L., and F.‐F. Jin (2011), Niño indices for two types of  ENSO, Geophys. Res. Lett., 38, L04704; doi: 10.1029/ 2010GL046031.

Ren, H.‐L., and F.‐F. Jin (2013), Recharge oscillator mecha­nisms in two types of ENSO, J. Climate, 26, 6506–6523. doi:10.1175/JCLI‐D‐12‐00601.1.

Rodgers, K. B., P. Friederichs, and M. Latif (2004), Tropical Pacific decadal variability and its relation to decadal m odulations of ENSO, J. Climate, 17, 3761–3774.

Rodrigues, R. R., R. J. Haarsma, E. J. D. Campos, T. Ambrizzi (2011), The impacts of inter‐El Niño variability on the t ropical Atlantic and northeast Brazil climate, J. Climate, 24, 3402–3422.

Rogers, J. C. (1981), The North Pacific Oscillation, Int. J. Climatol., 1, 39–57; doi:10.1002/joc.3370010106.

Ropelewski, C. F., and M. S. Halpert (1986), North American precipitation and temperature patterns associated with the El Niño/Southern Oscillation, Monthly Weather Rev., 114, 2352–2362.

Ropelewski, C. F., and M. S. Halpert (1987), Global and regional scale precipitation patterns associated with the El Nino Southern Oscillation, Monthly Weather Rev., 115, 1606–1626.

Ropelewski, C. F., and M. S. Halpert (1989), Precipitation p atterns associated with the high index phase of the Southern Oscillation, J. Climate, 2, 268–284.

Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T.  Yamagata (1999), A dipole mode in the tropical Indian Ocean, Nature, 401, 360–363.

Sassi, F., D. Kinnison, B. A. Boville, R. R. Garcia, and R. Roble (2004), Effect of El Niño–Southern Oscillation on the dynami­cal, thermal, and chemical structure of the middle atmosphere, J. Geophys. Res., 109, D17108; doi:10.1029/2003JD004434.

Schlesinger, M. E., and N. Ramankutty (1994), An oscillation in the global climate system of period 65–70 years, Nature, 367, 723–726; doi:10.1038/367723a0.

Schneider, D. P., Y. Okumura, and C. Deser (2012), Observed Antarctic interannual climate variability and tropical l inkages, J. Climate, 25, 4048–4066; doi:10.1175/JCLI‐D‐ 11‐00273.1.

Schopf, P. S., and R. J. Burgman (2006), A simple mechanism for ENSO residuals and asymmetry, J. Climate, 19, 3167–3179.

Schott. F. A., S. P. Xie, and J. McCreary (2009), Indian Ocean circulation and climate variability, Rev. Geophys., 47, RG1002; doi:10.1029/2007RG000245.

0003060439.INDD 34 6/5/2017 7:05:56 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 35

Seager, R., N. Harnik, Y. Kushnir, W. Robinson, and J. Miller (2003), Mechanisms of hemispherically symmetric climate variability, J. Climate, 16, 2960–2978.

Sheffield, J., and coauthors (2013a), North American climate in  CMIP5 experiments, Part I: Evaluation of historical s imulations of continental and regional climatology, J. Climate, 26, 9209–9245; doi: http://dx.doi.org/10.1175/JCLI‐D‐12‐00592.1.

Sheffield, J., and coauthors (2013b), North American climate in  CMIP5 experiments, Part II: Evaluation of 20th century intraseasonal to decadal variability, J. Climate, 23, 9247–9290.

Sheffield, J., and coauthors (2014), Regional climate processes and projections for North America: CMIP3/CMIP5 differ­ences, attribution and outstanding issues, NOAA Tech. Rep., OAR CPO‐2.

Smith, T. M., R. W. Reynolds, T. C. Peterson, and J. Lawrimore (2008), Improvements to NOAA’s historical merged land‐ocean surface temperature analysis (1880–2006), J. Climate, 21, 2283–2296.

Son, S. ‐W., A. Purich, H. H. Hendon, B. M. Kim, and L. M. Polvani (2013), Improved seasonal forecast using ozone hole variability? Geophys. Res. Lett., 40, 6231–6235; doi:10.1002/ 2013GL057731.

Song, H.‐J., E. Choi, G.‐H. Lim, Y. H. Kim, J.‐S. Kug, and S.‐W. Yeh (2011), The central Pacific as the export region of the El Niño–Southern Oscillation sea surface temperature anomaly to Antarctic sea ice, J. Geophys. Res., 116, D21113; doi:10.1029/2011JD015645.

Stammerjohn, S. E., D. G. Martinson, R. C. Smith, X. Yuan, and D. Rind (2008), Trends in Antarctic annual sea ice retreat and advance and their relation to El Niño–Southern Oscillation and southern annular mode variability, J. Geophys. Res., 113, C03S90; doi:10.1029/2007JC004269.

Stevenson, S., B. Fox‐Kemper, M. Jochum, R. Neale, C. Deser, and G. Meehl (2012), Will there be a significant change to El Niño in the twenty‐first century?, J. Climate, 25, 2129–2145; doi:10.1175/JCLI‐D‐11‐00252.1.

Su, J., R. Zhang, T. Li, X. Rong, J.‐S. Kug, and C.‐C. Hong (2010), Causes of the El Niño and La Niña amplitude asymmetry in the equatorial eastern Pacific, J. Climate, 23, 605–617.

Suarez, M. J., and P. S. Schopf (1988), A delayed action oscillator for ENSO, J. Atmos. Sci., 45, 3283–3287; doi:10.1175/1520‐0469 %281988%29045<3283%3AADAOFE>2.0.CO%3B2.

Sui, C. H., P. H. Chung, and T. Li (2007), Interannual and i nterdecadal variability of the summertime western North Pacific subtropical high, Geophys. Res. Lett., 34, L11701; doi:10.1029/2006GL029204.

Sun, D.‐Z., and T. Zhang (2006), A regulatory effect of ENSO on the timemean thermal stratification of the equatorial upper ocean, Geophys. Res. Lett., 33, L07710; doi:10.1029/ 2005GL025296.

Sun, F., and J. ‐Y. Yu (2009), A 10–15‐yr modulation cycle of ENSO intensity, J. Climate, 22, 1718–1735; doi: http://dx.doi.org/10.1175/2008JCLI2285.1.

Sung, M.‐K., B.‐M. Kim, and S.‐I. An (2014), Altered atmos­pheric responses to eastern Pacific and central Pacific El Niños over the North Atlantic region due to stratospheric interference, Climate Dyn., 42, 159–170; doi:10.1007/s00382‐012‐1661‐0.

Takahashi, K., A. Montecinos, K. Goubanova, and B. Dewitte (2011), ENSO regimes: Reinterpreting the canonical and Modoki El Niño, Geophys. Res. Lett., 38, L10704; doi: 10.1029/2011GL047364.

Tang, B. H., and J. D. Neelin (2004), ENSO influence on Atlantic hurricanes via tropospheric warming, Geophys. Res. Lett., 31, L24204; doi:10.1029/2004GL021072.

Taschetto, A. S., and M. H. England (2009), El Niño Modoki impacts on Australian rainfall, J. Climate, 22, 3167–3174.

Taschetto, A. S., A. Sen Gupta, N. C. Jourdain, A. Santoso, C. C. Ummenhofer, and M. H. England (2014), Cold tongue and warm pool ENSO events in CMIP5: Mean state and future projections, J. Climate, 27, 2861–2885; doi:10.1175/JCLI‐D‐ 13‐00437.1.

Taylor, K. E., R. J. Stouffer, and G. A. Meehl (2012), An over­view of CMIP5 and the experiment design, Bull. Amer. Meteor. Soc., 93, 485–498.

Tett, S. (1995), Simulation of El Niño–Southern Oscillation‐like variability in a global AOGCM and its response to CO2 increase, J. Climate, 8 (6), 1473–1502.

Thompson, D. W. J., and J. M. Wallace (1998), The Arctic Oscillation signature in the wintertime geopotential height and temperature fields, Geophys. Res. Lett., 25, 1297–1300; doi:10.1029/98GL00950.

Thompson, D. W. J., and J. M. Wallace (2000), Annular modes in the extratropical circulation, Part I: Month‐to‐month v ariability, J. Climate, 13, 1000–1016.

Timmermann, A. (2003), Decadal ENSO amplitude modulations: A nonlinear mechanism, Global Planet. Change, 37, 135–156.

Timmermann, A., and F.‐F. Jin (2002), A nonlinear mechanism for decadal El Niño amplitude changes, Geophys. Res. Lett., 29, 1003; doi:10.1029/2001GL013369.

Timmermann, A., J. Oberhuber, A. Bacher, M. Esch, M. Latif, and E. Roeckner (1999), Increased El Niño frequency in a climate model forced by future greenhouse warming, Nature, 398, 694–697.

Tokinaga, H., S.‐P. Xie, C. Deser, Y. Kosaka, and Y. M. Okumura (2012), Slowdown of the Walker circulation driven by tropical Indo‐Pacific warming, Nature, 491, 439–443; doi:10.1038/nature11576.

Torrence, C., and P. J. Webster (1999), Interdecadal changes in the ENSO‐monsoon system, J. Climate, 12, 2679–2690.

Trenberth, K. E., and D. P. Stepaniak (2001), Indices of El Niño evolution, J. Climate, 14, 1697–1701. doi: http://dx.doi.org /10 .1175 /1520‐0442 [2001 ]014<1697 :LIOENO >2.0.CO;2.

Trenberth, K. E., and T. J. Hoar (1996), The 1990–1995 E1 Niño–Southern Oscillation event: Longest on record, Geophys. Res. Lett., 23, 57–60.

Tu, J., C. Chou, and P. Chu (2009), The abrupt shift of typhoon activity in the vicinity of Taiwan and its association with western North Pacific–East Asian climate change, J. Climate, 22, 3617–3628.

Tziperman, E., L. Stone, M. Cane, and H. Jarosh (1994), El Niño chaos: Overlapping of resonances between the seasonal cycle and the Pacific Ocean‐atmosphere oscillator, Science, 264, 72.

Tziperman, E., M. Cane, and S. E. Zebiak (1995), Irregularity and locking to the seasonal cycle in an ENSO prediction

0003060439.INDD 35 6/5/2017 7:05:56 PM

36 FORCINGS OF CLIMATE EXTREMES

model as explained by the quasi‐periodicity route to chaos, J. Atmos. Sci., 52, 293–306.

van Oldenborgh, G. J., S. Y. Philip, and M. Collins (2005), El Niño in a changing climate: A multi‐model study, Ocean Sci., 1, 81–95; doi:10.5194/os‐1‐81‐2005.

Vecchi, G. A., and B. J. Soden (2007), Global warming and the  weakening of the tropical circulation, J. Climate, 20, 4316–40.

Vecchi, G. A., et  al. (2006), Weakening of tropical Pacific atmospheric circulation due to anthropogenic forcing, Nature, 441, 73–76.

Venzke, S., M. Latif, and A. Villwock (2000), The coupled GCM ECHO‐2, Part II: Indian Ocean response to ENSO, J. Climate, 13, 1317–1383.

Verdon, D. C., and S. W. Franks (2006), Long‐term behaviour of ENSO: Interactions with the PDO over the past 400 years inferred from paleoclimate records, Geophys. Res. Lett., 33, L06712; doi:10.1029/2005GL025052.

Vimont, D. J., D. S. Battisti, and A. C. Hirst (2001), Footprinting: A seasonal connection between the tropics and mid‐latitudes, Geophys. Res. Lett., 28, 3923–3926; doi:10.1029/2001GL013435.

Vimont, D. J., J. M. Wallace, and D. S. Battisti (2003), The sea­sonal footprinting mechanism in the Pacific: Implications for ENSO, J. Climate, 16, 2668–2675; doi:10.1175/1520‐0442%282003%29016<2668%3ATSFMIT>2.0.CO%3B2.

Vimont, D. J., M. Alexander, and A. Fontaine (2009), Midlatitude excitation of tropical variability in the Pacific: The role of thermodynamic coupling and seasonality, J. Climate, 22 (3), 518–534; doi:10.1175/2008JCLI2220.1.

Walker, G. T., and E. W. Bliss (1932), World weather V Mem, R. Meteor. Soc., 4, 53–84.

Wallace, J. M., and D. S. Gutzler (1981), Teleconnections in the  potential height field during the Northern Hemisphere winter, Monthly Weather Rev., 109, 784–812.

Wang, B., and S. ‐I. An (2001), Why the properties of El Niño  change in the late 1970s?, Geophys. Res. Lett., 28, 3709–3712.

Wang, B., and Y. Wang (1996), Temporal structure of the Southern Oscillation as revealed by a waveform and a wavelet transform, J. Climate, 9, 1586–1598. doi: http://dx.doi.org/10.1175/1520‐0442[1996]009<1586:TSOTSO>2.0.CO;2.

Wang, B., B. Xiang, and J.‐Y. Lee (2013), Subtropical high p redictability establishes a promising way for monsoon and tropical storm predictions, PNAS, 110, 2718–2722.

Wang, B., R. Wu, and X. Fu (2000), Pacific–East Asian tele­connection: How does ENSO affect East Asian climate?, J. Climate, 13, 1517–1536.

Wang, C., and R. H. Weisberg (1998), Climate variability of the coupled tropical‐extratropical ocean‐atmosphere system, Geophys. Res. Lett., 25, 3979–3982.

Wang, C., and X. Wang (2013), Classifying El Niño Modoki I and II by different impacts on rainfall in southern China and typhoon tracks, J. Climate, 26, 1322–1338; doi: 10.1175/JCLI‐D‐12‐00107.1.

Wang, C., C. Deser, J.‐Y. Yu, P. DiNezio, and A. Clement (2015), El Nino and Southern Oscillation (ENSO): A review, A c hapter for Springer book: Coral Reefs of the Eastern Pacific, in press.

Wang, C., R. H. Weisberg, and J. I. Virmani (1999), Western Pacific interannual variability associated with the El Niño–Southern Oscillation, J. Geophys. Res., 104, 5131–5149. DOI: 10.1029/1998JC900090.

Wang, D., Q. Liu, R. X. Huang, Y. Du, and T. Qu (2006), Interannual variability of the South China Sea throughflow inferred from wind data and an ocean data assimilation product, Geophys. Res. Lett., 33, L14605; doi:10.1029/ 2006GL026316.

Wang, G., and H. H. Hendon (2007), Sensitivity of Australian rainfall to inter–El Niño variations, J. Climate, 20, 4211–4226.

Wang, H., and R. Fu (2000), Influences of ENSO SST anoma­lies and winter storm tracks on the interannual variability of  upper‐troposphere water vapor over the Northern Hemisphere extratropics, J. Climate, 13, 59–73.

Wang, S.‐Y., M. L’Heureux, and H. ‐H. Chia (2012), ENSO prediction one year in advance using western North Pacific sea surface temperatures, Geophys. Res. Lett., 39, L05702; doi:10.1029/2012GL050909.

Wang, S.‐Y., M. L’Heureux, and J. H. Yoon (2013), Are green­house gases changing ENSO precursors in the western North Pacific?, J. Climate, 26 (17), 6309–6322.

Wang, X., and C. Wang (2014), Different impacts of various El Niño events on the Indian Ocean dipole, Climate Dyn., 42, 991–1005; doi: 10.1007/s00382‐013‐1711‐2.

Wang, X., W. Zhou, C. Li, and D. Wang (2014), Comparison of the impact of two types of El Niño on tropical cyclone g enesis over the South China Sea, Int. J. Climatol., 34, 2651–2660; doi: 10.1002/joc.3865.

Weare, B. C., A. R. Navato, and R. E. Newell (1976), Empirical orthogonal analysis of Pacific sea surface temperatures, J. Phys. Ocean, 6, 671–678.

Webster, P. J., A. M. Moore, J. P. Loschnigg, and R. R. Leben (1999), Coupled ocean‐atmosphere dynamics in the Indian Ocean during 1997–98, Nature, 401, 356–360.

Weisberg, R. H., and C. Wang (1997), A western Pacific oscillator paradigm for the El Niño–Southern Oscillation, Geophy. Res. Lett., 24, 779–782; doi: 10.1029/97GL00689.

Weng, H., K. Ashok, S. K. Behera, S. A. Rao, and T. Yamagada (2007), Impacts of recent El Nino Modoki on droughts/floods in the Pacific rim during boreal summer, Climate Dyn., 29, 113–129; doi:10.1007/s00382‐007‐0234‐0.

Weng, H., S. K. Behera, and T. Yamagada (2009), Anomalous winter climate conditions in the Pacific rim during recent El  Nino Modoki and El Nino events, Climate Dyn., 32, 663–674; doi:10.1007/s00382‐008‐0394‐6.

Wittenberg, A. T. (2009), Are historical records sufficient to constrain ENSO simulations?, Geophys. Res. Lett., 36, L12702; doi:10.1029/2009GL038710.

Wyrtki, K. (1975), El Niño: The dynamic response of the e quatorial Pacific Ocean to atmospheric forcing, J. Phys. Oceanogr., 5, 572–584; doi:10.1175/1520‐0485%281975 %29005<0572%3AENTDRO>2.0.CO%3B2.

Xiang, B., B. Wang, and T. Li (2013), A new paradigm for the predominance of standing Central Pacific warming after the  late 1990s, Climate Dyn., 41 (2), 327–340; doi: 10.1007/s00382‐012‐1427‐8.

Xie, F., J. Li, W. Tian, J. Feng, and Y. Huo (2012), Signals  of  El  Niño Modoki in the tropical tropopause

0003060439.INDD 36 6/5/2017 7:05:56 PM

ThE ChANGING EL NIñO–SOuThERN OSCILLATION ANd ASSOCIATEd CLIMATE EXTREMES 37

layer and stratosphere, Atmos. Chem. Phys., 12, 5259–5273; doi:10.5194/acp‐12‐5259‐2012.

Xie, S.‐P., and S. G. H. Philander (1994), A coupled ocean‐atmosphere model of relevance to the ITCZ in the eastern Pacific, Tellus, 46A, 340–350; doi:10.1034/j.1600‐0870.1994.t01‐1‐00001.x.

Xie, S.‐P., C. Deser, G. A. Vecchi, J. Ma, H. Teng, and A. T. Wittenberg (2010), Global warming pattern formation: Sea surface temperature and rainfall, J. Climate, 23, 966–986; doi:10.1175/2009JCLI3329.1.

Xie, S.‐P., H. Annamalai, F. Schott, and J. P. McCreary Jr. (2002), Origin and predictability of South Indian Ocean cli­mate variability, J. Climate, 15 (8), 864–874.

Xie, S. ‐P., K. Hu, J. Hafner, H. Tokinaga, Y. Du, G. Huang, and T. Sampe (2009), Indian Ocean capacitor effect on Indo–Western Pacific climate during the summer following El Niño, J. Climate, 22 (3).

Xie, S.‐P., Q. Xie, D. Wang, and W. T. Liu (2003), Summer upwelling in the South China Sea and its role in regional c limate variations, J. Geophys. Res., 108 (C8), 3261; doi:10.1029/2003JC001867.

Xu, B., J. Cao, J. Hansen, T. Yao, D. Joswia, N. Wang, G. Wu, M. Wang, H. Zhao, W. Yang, X. Liu, and J. He (2009), Black soot and the survival of Tibetan glaciers, Proc. Nat. Acad. Sci., 106, 22114.

Xu, Z., K. Fan, and H. Wang (2015), Decadal variation of s ummer precipitation over China and associated atmospheric circulation after the late 1990s, J. Climate, 28 (10), 4086–4106.

Xue, Y., M. Chen, A. Kumar, Z.‐Z. Hu, and W. Wang (2013), Prediction skill and bias of tropical Pacific sea surface tem­peratures in the NCEP climate forecast system version 2, J. Climate, 26, 5358–5378; doi:10.1175/JCLI‐D‐12‐00600.1.

Yadav, R. K., D. A. Ramu, and A. P. Dimri (2013), On the r elationship between ENSO patterns and winter precipitation over north and central India, Global Planet. Change, 107, 50–58; doi:10.1016/j.gloplacha.2013.04.006.

Yang, J., Q. Liu, S.‐P. Xie, Z. Liu, and L. Wu (2007), Impact of the Indian Ocean SST basin mode on the Asian summer monsoon, Geophys. Res. Lett., 34, L02708; doi:10.1029/ 2006GL028571.

Yang, S., and X. Jiang (2014), Prediction of eastern and central Pacific ENSO events and their impacts on East Asian c limate by the NCEP climate forecast system, J. Climate, 27, 4451–4472.

Yeh, S.‐W., and B. P. Kirtman (2004), Tropical Pacific decadal variability and ENSO amplitude modulation in a CGCM, J. Geophys. Res., 109, C11009; doi:10.1029/2004JC002442.

Yeh, S. ‐W., and B. P. Kirtman (2007), ENSO amplitude changes due to climate change projections in different coupled m odels, J. Climate, 20, 203–217.

Yeh, S.‐W., B. P. Kirtman, J.‐S. Kug, W. Park, and M. Latif (2011), Natural variability of the central Pacific El Niño event on multi‐centennial timescales, Geophys. Res. Lett., 38, L02704; doi:10.1029/2010GL045886.

Yeh, S. W., J. S. Kug, B. Dewitte, M. H. Kwon, B. P. Kirtman, and F.‐F. Jin (2009), El Niño in a changing climate, Nature, 461, 511–514; doi:10.1038/nature08316.

Yeh, S.‐W., X. Wang, C. Wang, and B. Dewitte (2015), On the relationship between the North Pacific climate variability

and the central Pacific El Niño, J. Climate, 28, 663–677; doi: http://dx.doi.org/10.1175/JCLI‐D‐14‐00137.1.

Yeo, S. ‐R., and K. ‐Y. Kim (2015), Decadal changes in the Southern Hemisphere sea surface temperature in association with El Niño–Southern Oscillation and southern annular mode, Climate Dyn., 1–16, http://dx.doi.org/10.1007/s00382‐015‐2535‐z.

Yeo, S. ‐R., K. Y. Kim, S. W. Yeh, B. M. Kim, T. Shim, and J. G. Jhun (2014), Recent climate variation in the Bering and Chukchi seas and its linkages to large‐scale circulation in the Pacific, Climate Dyn., 42 (9–10), 2423–2437.

Yu, J.‐Y., and H.‐Y. Kao (2007), Decadal changes of ENSO p ersistence barrier in SST and ocean heat content indices: 1958–2001, J. Geophys. Res., 112, D13106; doi:10.1029/ 2006JD007654.

Yu, J.‐Y., and K. M. Lau (2004), Contrasting Indian Ocean SST variability with and without ENSO influence: A coupled atmosphere‐ocean GCM study, Meteor. Atmos. Phys., 90, 179–191; doi:10.1007/s00703‐004‐0094‐7.

Yu, J.‐Y., and S. T. Kim (2010a), Identification of central‐Pacific and eastern‐Pacific types of ENSO in CMIP3 models, Geophys. Res. Lett., 37, L15705; doi:10.1029/ 2010GL044082.

Yu, J.‐Y., and S. T. Kim (2010b), Three evolution patterns of central‐Pacific El Niño, Geophys. Res. Lett., 37, L08706; doi:10.1029/2010GL042810.

Yu, J.‐Y., and S. T. Kim (2011a), Relationships between extrat­ropical sea level pressure variations and the central‐Pacific and eastern‐Pacific types of ENSO, J. Climate, 24, 708–720; doi:10.1175/2010JCLI3688.1.

Yu, J.‐Y., and S.‐T. Kim (2011b), Reversed spatial asymmetries between El Niño and La Niña and their linkage to decadal ENSO modulation in CMIP3 models, J. Climate, 24, 5423–5434.

Yu, J.‐Y., and Y. Zou (2013), The enhanced drying effect of Central‐Pacific El Niño on US winter, Environ. Res. Lett., 8; doi:10.1088/1748‐9326/8/1/014019.

Yu, J.‐Y., H.‐Y. Kao, and T. Lee (2010), Subtropics‐related interannual sea surface temperature variability in the equato­rial central Pacific, J. Climate, 23, 2869–2884; doi: http://dx.doi.org/10.1175/2010JCLI3171.1.

Yu, J.‐Y., H.‐Y. Kao, T. Lee, and S. T. Kim (2011), Subsurface ocean temperature indices for central‐Pacific and eastern‐Pacific types of El Niño and La Niña events, Theor. App. Climatol.; doi: 10.1007/s00704‐010‐0307‐6.

Yu, J.‐Y., F. Sun, and H.‐Y. Kao (2009), Contributions of Indian Ocean and monsoon biases to the excessive biennial ENSO in CCSM3, J. Climate, 22, 1850–1858.

Yu, J.‐Y., H. Paek, E. S. Saltzman, and T. Lee (2015b), The early‐1990s change in ENSO‐PSA‐SAM relationships and its impact on Southern Hemisphere climate, J. Climate; doi: 10.1175/JCLI‐D‐15‐0335.1, in press.

Yu, J.‐Y., M.‐M. Lu, and S. T. Kim (2012a), A change in the relationship between tropical central Pacific SST variability and the extratropical atmosphere around 1990, Environ. Res. Lett., 7, 034025; doi:10.1088/1748‐9326/7/3/034025.

Yu, J.‐Y., Y. Zou, S. T. Kim, and T. Lee (2012b), The changing impact of El Niño on US winter temperatures, Geophys. Res. Lett., 39, L15702; doi:10.1029/2012GL052483.

0003060439.INDD 37 6/5/2017 7:05:56 PM

38 FORCINGS OF CLIMATE EXTREMES

Yu, J.‐Y., P. ‐K. Kao, H. Paek, H. ‐H. Hsu, C. ‐W. Hung, M. ‐M. Lu, and S. ‐I. An (2015a), Linking emergence of the Central‐Pacific El Niño to the Atlantic multi‐decadal oscillation, J. Climate, 28, 651–662; doi:10.1175/JCLI‐D‐14‐00347.1.

Yu, W., B. Xiang, L. Liu, and N. Liu (2005), Understanding the  origins of interannual thermocline variations in the t ropical Indian Ocean, Geophys. Res. Lett., 32, L24706; doi:10.1029/2005GL024327.

Yuan, X. (2004), ENSO‐related impacts on Antarctic sea ice: A synthesis of phenomenon and mechanisms, Antarctic Sci., 16 (04), 415–425.

Yuan, X., and C. Li (2008), Climate modes in southern high latitudes and their impacts on Antarctic sea ice, J. Geophys. Res., 113, C06S91; doi:10.1029/2006JC004067.

Yuan, X., and D. G. Martinson (2001), The Antarctic dipole and its predictability, Geophys. Res. Lett., 28, 3609–3612.

Yuan, Y., and S. Yang (2012), Impacts of different types of El  Niño on East Asian climate: Focus on ENSO cycles, J. Climate, 25, 7702–7722.

Yuan, Y., S. Yang, and Z. Zhang (2012), Different evolutions of the Philippine Sea anticyclone between eastern and central Pacific El Niño: Possible effect of Indian Ocean SST, J. Climate, 25, 7867–7883.

Yulaeva, E., and J. M. Wallace (1994), The signature of ENSO in global temperature and precipitation fields derived from the microwave sounding unit, J. Climate, 7, 1719–1736; doi:10.1175/1520‐0442[1994] 007<1719:TSOEIG>2.0.CO;2.

Zebiak, S. E. (1989), Ocean heat content variability and El Niño cycles, J. Phys. Oceanogr., 19, 475–486; doi:10.1175/1520‐0485%281989%29019<0475%3AOHCVAE>2.0.CO%3B2.

Zelle, H., G. J. van Oldenborgh, G. Burgers, and H. Dijkstra (2005), El Niño and greenhouse warming: Results from ensemble simulations with the NCAR CCSM, J. Climate, 18, 4669–4683.

Zhang, L., L. Wu, X. Lin, and D. Wu (2010), Modes and mech­anisms of sea surface temperature low‐frequency variations over the coastal China Seas, J. Geophys. Res., 115, C08031; doi:10.1029/2009JC006025.

Zhang, R., and T. L. Delworth (2007), Impact of the Atlantic  multidecadal oscillation on North Pacific climate

variability, Geophys. Res. Lett., 34, L23708; doi:10.1029/ 2007GL031601.

Zhang, R., A. Sumi, and M. Kimoto (1999), A diagnostic study of the impact of El Nino on the precipitation in China, Adv. Atmos. Sci., 16, 229–241.

Zhang, W., F. Jin, J. Li, and H. Ren (2011), Contrasting impacts of two‐type El Niño over the western north Pacific during boreal autumn, J. Meteor. Soc. Japan., 89, 563–569.

Zhang, X. B., J. Sheng, and A. Shabbar (1998), Modes of inter­annual and interdecadal variability of Pacific SST, J. Climate, 11, 2556–2569.

Zhang, Y., J. M. Wallace, and D. S. Battisti (1997), ENSO‐like  interdecadal variability: 1900–93, J. Climate, 10, 1004–1020.

Zhou, L.‐T., C.‐Y. Tam,W. Zhou, and J. C. L. Chan (2009), Influence of South China Sea SST and the ENSO on winter rainfall over South China, Adv. Atmos. Sci., 27, 832–844.

Zhou, S., A. J. Miller, J. Wang, J. K. Angell (2002), Downward‐propagating temperature anomalies in the preconditioned polar stratosphere, J. Climate, 15, 781–792.

Zhou, W., and J. C. L. Chan (2007), ENSO and the South China Sea summer monsoon onset, Int. J. Climatol., 27, 157–167.

Zhou W., X. Wang, T. J. Zhou, C. Y. Li, and J. C. L. Chan (2007), Interdecadal variability of the relationship between the East Asian winter monsoon and ENSO, Meteor. Atmos. Phys., 98, 283–293; doi: 10.1007/s00703‐007‐0263‐6.

Zhou, Z. Q., S. P. Xie, X. T. Zheng, Q. Y. Liu, and H. Wang (2014), Global warming–induced changes in El Niño tele­connections over the North Pacific and North America, J. Climate, 27, 9050–9064.

Zhu, Y., W. Wang, W. Zhou, and J. Ma (2011), Recent changes in the summer precipitation patterns in East China and the background circulation, Climate Dyn., 36, 1463–1473.

Zou, Y., J.‐Y. Yu, T. Lee, M.‐M. Lu, and S. T. Kim (2014), CMIP5 model simulations of the impacts of the two types of El Niño on US winter temperature, J. Geophys. Res. Atmos., 119, 3076–3092; doi:10.1002/2013JD021064.

Zubiaurre, I., and N. Calvo (2012), The El Niño–Southern Oscillation [ENSO] Modoki signal in the stratosphere, J. Geophys. Res., 117, D04104; doi:10.1029/2011JD016690.

0003060439.INDD 38 6/5/2017 7:05:56 PM

3

1

Geological Carbon Storage: Subsurface Seals and Caprock Integrity, Geophysical Monograph 238, First Edition. Edited by Stephanie Vialle, Jonathan Ajo-Franklin, and J. William Carey. © 2019 American Geophysical Union. Published 2019 by John Wiley & Sons, Inc.

1.1.  INTRODUCTION

Geological storage of carbon dioxide (CO2) has been mooted as a greenhouse gas mitigation strategy for over 20 years. The practical mechanics of such a strategy have been tested out at small scale at sites such as the Otway Basin in Australia [Sharma et al., 2009] and Frio in Texas [Doughty et  al., 2008] and during industrial‐scale pro‑jects, for example, at Sleipner [Arts et  al., 2008] and In Salah [Ringrose et al., 2013]. Many years of effort have been put into defining the critical parameters for poten‑tial CO2 storage sites [e.g., IPCC, 2005], and these include depth, storage capacity of the site, injectivity of the reser‑voir, and the containment integrity of the structure into which the CO2 is injected. Containment integrity is usu‑ally thought of in similar terms as traps and seals in

petroleum systems, and similar technologies can be used to evaluate the properties of the fault and/or top seals that provide the trapping mechanisms for keeping injected CO2 in the deep subsurface. Fault seals usually result from the incorporation of material into the fault zone during fault movement, and this can comprise smearing out of ductile clay‐rich units, abrasion of harder shales, cataclasis of rigid grains, and syn‐/post‐kinematic cemen‑tation of the fault rock products [e.g., Lindsay et  al., 1993; Yielding et  al., 1997; Fisher and Knipe, 1998; Dewhurst et al., 2005]. Top seals are usually characterized in terms of their thickness (especially in relation to fault throw), areal extent, seal capacity (pore‐scale capillary properties), and seal integrity (mechanical properties). There are multiple techniques for assessing the potential sealing capacity of faults [e.g., Watts, 1987; Lindsay et al., 1993; Yielding et al., 1997], and these will not be discussed further here. This paper will concentrate on methods that

Microstructural, Geomechanical, and Petrophysical Characterization of Shale Caprocks

David N. Dewhurst, Claudio Delle Piane, Lionel Esteban, Joel Sarout, Matthew Josh, Marina Pervukhina, and M. Ben Clennell

ABSTRACT

Geological storage of carbon dioxide requires extensive characterization of potential selected sites in terms of injectivity, storage capacity, and containment integrity. The latter item on that list requires a multi‐scale evalua‑tion of all aspects of the subsurface geology that can trap CO2 underground and keep it there long term. One part of a containment integrity strategy includes characterization of the caprock at a given site. Many selected sites have clay‐rich shales as caprocks, and this contribution will concentrate on workflows and methods for characterizing such rocks in the laboratory. Shale preservation is the most critical step in the process as dehydra‑tion from the native in situ water content significantly affects shale properties. Various mineralogical, micro‑scopical, petrophysical, and geomechanical properties and associated testing methods are discussed, and where possible, examples are shown of the impact of lack of preservation. Results are discussed in the context of inter‑action of CO2 with caprocks and trapping mechanisms. Finally, the discussion looks at a number of the uncer‑tainties associated with laboratory testing of shales in terms of both results obtained to date and our limited understanding as yet of the behavior and interaction of supercritical CO2 with clay‐rich caprocks.

CSIRO Energy, Perth, Australia

0004182022.INDD 3 10/16/2018 7:14:56 AM

COPY

RIGH

TED

MATER

IAL

4 GEOLOGICAL CARBON STORAGE

can be used to characterize caprocks in the laboratory and the relationship between these measurement tech‑niques and the properties noted above. In this contribu‑tion, we will concentrate on shale‐rich caprocks but acknowledge that other rocks such as anhydrites [e.g., Hangx et al., 2010] are being evaluated as caprocks for CO2 storage sites. However, it should be emphasized that any seal evaluation for a storage site or petroleum prospect should be fully integrated across both fault and top seals and for a wide range of scales.

Shale caprock properties are dependent on a number of factors, including depositional environment and resultant lithology, electrochemical conditions at deposition, min‑eralogy, the presence of organic matter, compaction, and diagenetic alteration. All of these processes have a significant impact on porosity and permeability, as well as the mechanical, capillary, and petrophysical properties of shales [e.g., Bennett et  al., 1991a,b; Vernik and Liu, 1997; Dewhurst et al., 1998, 1999a, 1999b; Clennell et al., 2006]. A number of these properties are also controlled by human intervention during and after the coring pro‑cess, such as stress relief microfracture development and drying out and desiccation of recovered core, and care must be taken for certain properties that adequate sample preservation is undertaken [e.g., Schmitt et  al., 1994; Dewhurst et al., 2012; Ewy, 2015]. This study will there‑fore review possible preservation methods and discuss multiple mechanical and petrophysical characterization techniques that can be used to either directly measure or estimate relevant properties required for shale caprocks.

1.2.  SHALE PRESERVATION

The most critical stage for deriving high‐quality labora‑tory results from shales is their immediate preservation on recovery. Loss of pore water from the in situ state can result in changing mechanical, physical, and petrophysi‑cal properties [Schmitt et  al., 1994] no matter whether the shale is soft, weak, and ductile or hard, strong, and brittle. Some pore fluid will always be lost from shales on recovery due to outgassing as cores are depressurized from the in situ conditions to the Earth’s surface [Schmitt et al., 1994]. However, most techniques that look to mea‑sure mechanical and rock physics properties, for example, would look to test the shale using a chemically compat‑ible pore fluid under pressure, and this would generally drive any air into solution at fluid pressures >0.5 MPa. Running such tests under undrained conditions at low strain rates (< 10−7  s−1) allows monitoring of the pore pressure response either through Skempton B tests [Skempton, 1954] or during axial loading. Pore pressure increase under such conditions is indicative of full satura‑tion. Hence, the slight loss of pore fluid during recovery can be alleviated for such tests. Equilibrating in relative

humidity (RH) environments equivalent to shale native water activity can also mitigate this effect [e.g., Steiger and Leung, 1991; Ewy, 2014]. Other tests such as compo‑sition via X‐ray diffraction (XRD), cation exchange capacity (CEC), or specific surface area (SSA) measure‑ments generally would not be significantly affected by core preservation, although one should be careful to verify whether the presence of salts (e.g., halite, sylvite) or gypsum is real or artifacts of core storage [e.g., Milliken and Land, 1994].

Ideally, fully saturated core samples should be pre‑served under a nonpolar mineral spirit (e.g., Ondina 15 or Ondina 68) such that the fluid does not interact with the clays present and prevents native pore fluids escaping from the sample. Oil cannot intrude fully water saturated nanopores in shale at ambient pressures due to immisci‑bility and wettability issues. Other potential fluids that can be used for shale preservation include decane [e.g., Ewy et  al., 2008; Ewy, 2014]. Core plugs subsampled from recovered core should also be sealed in glass vials immersed in an appropriate preservative solution. Should such materials not be available, a short‐term solution would be to coat cores or plugs in cling film, tin foil, and wax as a short‐period (weeks to months) stopgap and kept cool in a fridge (but not frozen). However, it would be preferable to immerse in the fluids suggested above as soon as possible as wax is slightly permeable to air and samples will eventually begin to desiccate.

In order to avoid sample desiccation and concomitant alteration of rock properties (see examples below), a workflow has been developed to maximize high‐quality results from preserved shale cores (Fig. 1.1). Initially, a whole core is X‐ray CT scanned in order to look for fractures, limestone stringers, nodules, and the like. This allows the development of a coring plan (Fig. 1.2) directly linked to the workflow which avoids such features and means that when core plugs are taken, exposure to air is minimized. While conventional rotary coring is some‑times used for harder and more isotropic shales, in gen‑eral a Murg diamond wire bandsaw is used to take core plugs, and these plugs are finished off on a cylindrical grinder. This allows significantly increased core plug recovery and better quality of plugs taken in these noto‑riously difficult‐to‐prepare rocks. Shales are at their weakest in tension parallel to bedding [e.g., Fjær et  al., 2008], and rotary core plugs often lead to biscuiting due to closely spaced fracture development parallel to the fabric anisotropy. The lack of stress induced by torque in the case of the diamond wire bandsaw means that plugs can be taken in more difficult rocks without imposing stresses on the intrinsic planes of weakness in the shale and better plug condition and recovery is the result.

This paper will discuss the methods and application of a number of the tests shown in the workflow with a view

0004182022.INDD 4 10/16/2018 7:14:56 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 5

to characterizing shale top seals for CO2 storage and will also discuss the impact of poor preservation on results. Where possible, examples from CO2 storage sites will be used; otherwise examples from top seals in petroleum systems will be shown.

1.3.  MINERALOGY AND MICROSTRUCTURE

For a carbon capture and storage (CCS) project to be successful, it needs to guarantee the long‐term trapping of the sequestered CO2 underground. In this context, an important aspect to be considered is fluid‐rock interac‑tion which results from the contrast in physical properties between the injected CO2 and the in situ pore fluids permeating the storage/sealing rocks and the potential reactivity between CO2 as a free phase and as dissolved carbonic acid in the aqueous phase and the rock it comes in contact with. Through these physical and chemical interactions, the transport, elastic, and mechanical prop‑erties as well as the sealing effectiveness of shale caprocks could be changed, affecting the success of CO2 geose‑questration. A key factor to the understanding and

prediction of shale behavior is the proper characteriza‑tion of the various elements that collectively constitute their microstructure. Here we regard the microstructure of shales in the context of CO2 geosequestration as sub‑divided into two components: (i) the minerals and (ii) the pore space; specific techniques to characterize these two components are discussed below.

A detailed characterization of the mineral component of shales is desirable as a number of common minerals have been identified as being reactive in the presence of sCO2 and water at temperature and pressure conditions relevant for subsurface carbon dioxide storage. Numerous studies indicate that CO2 trapping via mineral carbon‑ation can occur by dissolution of albite (NaAlSi3O8) and the precipitation of dawsonite [NaAlCO3(OH)2; e.g., Romanov et al., 2015] as well as the dissolution of Mg, Fe‐silicates and sulfides, with precipitation of Mg, Fe‑carbonates such as siderite and magnesite [e.g., Kaszuba et  al., 2005]. Moreover, recent studies showed that electrical and capillary interactions can occur bet‑ween clay minerals and CO2 in the liquid and supercrit‑ical state. CO2 diffusion into the clay layered structure

8. Mercury intrusion porosimetry7. Electronmicroscopyand FIB-SEM

Resination, polishing

Offcuts andend pieces

divided

10a. CEC, XRD, XRF

10b. Ethylene glycolsurface area

9. Gas adsorption isotherms

6. X-ray micro-tomography

11. Dielectricanalysis of paste

Rock Mechanics

1a. Whole core X-ray CT Plugs cut,preservedunder oil

2a. Low-fieldNMR:T1, T2

diffusion

5a. Ultra-sonicVp,Vs

5b.Acousticemission

2b. High-field NMR:

solid-state andsurface interactions

Powdersand pastes

Intactpieces

3. Electricalimpedancespectroscopy

4. Dielectricspectroscopy

1b. X-ray CTof plugs

+Selectedcuttings

Figure 1.1 Schematic workflow for shale characterization. CT scanning of whole core and core plugs is followed by nondestructive petrophysical testing (NMR, electrical/dielectric properties) with plugs ending up in the rock mechanics laboratory for static and dynamic mechanical testing. Offcuts are sent for mineralogical, geochemical, and physicochemical properties testing, along with characterization of the shale microstructure and pore systems. Modified from Clennell et al. [2006]. Reprinted with permission of SPWLA.

0004182022.INDD 5 10/16/2018 7:14:56 AM

6 GEOLOGICAL CARBON STORAGE

results in changes in the molecular clay‐water chemistry leading to polarity changes in the internal electrical forces, eventually resulting in intraparticle/interparticle repulsion [e.g., Espinoza and Santamarina, 2012; Berrezueta et al., 2013]. Electrical and capillary effects are modulated by the surface area of the particular clay min‑erals which in turn is known to be related to the type of clay [e.g., Josh, 2014]. Therefore, identification of the clay minerals in a shale caprock may help predict the potential fluid‐rock interactions which could adversely affect the mechanical strength and seal capacity of the formation.

The importance of clay types has been shown, for example, by Busch et  al. [2008] who indicated that the CO2 sorption capacity of Muderong Shale (a regional seal in offshore Western Australia) could be attributed entirely to its clay mineral constituents.

Identification of clay minerals is traditionally achieved on powdered samples by XRD [Brindley, 1952; Moore and Reynolds, 1997], a well‐established technique that only requires a few grams of material for quantification. This, however, is a destructive procedure, meaning that all spatial information on the arrangement of the identified

1.5 inches (3.81 cm)

1.5 inches

1.5 inches

1 inch1 inch

7 cm

9.2 cm

max. 3.5 cm

1 cm

1 cm

3 inches(7.6 cm)

4 cm

Facing view

4 in

ches

B12

B11 B11 B12

B11

B12

B13

B113

B112

B111

B13

B13

To rock mechanics tests

To electrical tests and NMR

To dielectric tests

To rock mechanics tests10cm

10cm

10cm

10cm

10cm

Vertical drill

4 inches

45° drill

2.4 cm

= 7.6 cm

3 in

ches

Offcuts (i.e.,leftovers aftercore drillings)go to dielectric

To permeability tests

To permeability tests

To NMR, AMS, and electricaltests

To NMR, AMS, and electrictests

A11

A14 A13 A11

A12

A13

A14

A141

A142

A112

A111

A12

1.5 inches

1 inch

1 inch

1 inch

To dielectrictests

To rockmechanics tests

Figure 1.2 Example coring plan for subsampling whole cores normal to bedding (top) and parallel to bedding (bottom). Planning from medical CT scan images reduces exposure time to air, preventing further loss of shale pore fluids.

0004182022.INDD 6 10/16/2018 7:14:57 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 7

minerals is lost. Also, the technique is only sensitive to crystalline material such that any organic or amorphous matter will not be identified nor quantified; this could prove detrimental as organic matter is often found in shales and it is known to play a role in CO2 sorption [Krooss et al., 2002; Busch et al., 2008]. Accurate determi‑nations of grain size fractions (< 2 μm and <0.2 μm) are critical for determination of important clay parameters and composition. The clay fraction is defined as the wt% of material <2 μm in grain size and has been shown to be related to shale compaction and permeability [e.g., Aplin et al., 1995; Yang and Aplin, 1998]. Accurate determina‑tion of wt% of clay minerals is also important as this defines the clay content parameter, which is directly linked to shale geomechanical and petrophysical prop‑erties [e.g., Clennell et al., 2006; Josh et al., 2012; Dewhurst et al., 2015]. Finally, accurate determination of the com‑position and ordering of mixed layer clays, especially illite‐smectites, requires analyses to be performed on the <0.2 μm fraction to allow clear distinction of peaks, espe‑cially in the low reflection angle region [Moore and Reynolds, 1997].

Complementary methods that can preserve the spatial information regarding the arrangement of different min‑eral phases at the microscale and allow for their identification and quantification are normally based on petrographic analysis of polished rock samples. As a result of the fine‐grained nature of shales, the use of scanning electron microscopy (SEM) is necessary to resolve many of the single components of the rock matrix. SEM allows the visualization of microstructural features down to a resolution of few tens of nanometers and, at its best, can be used to visualize areas on the order of cm2. Additionally, the interaction between the incident electron beam and the analyzed material gives rise to the emission of X‐rays which with the use of appropriate detectors can be analyzed in terms of energy or wave‑length to infer chemical information of the emitting sub‑stance. Examples of the use of SEM‐based automated mineral analysis for the quantification of shale miner‑alogy include Golab et al. [2012] and Timms et al. [2015]. In geological materials, electron bombardment also stim‑ulates the emission of light at relatively low energy by cathodoluminescence (CL); spectral analysis of the CL emission in the UV to IR region offers the ability to mea‑sure trace ionic species with relatively short acquisition times enabling large areas to be mapped with detection limits orders of magnitude below elemental detections levels acquired using X‐rays. This technique can be use‑ful, for example, in mapping the distribution of authi‑genic quartz cement in shale formations to understand their diagenetic history and help evaluate their mechanical and elastic properties [Peltonen et al., 2009; Thyberg et al., 2010; MacRae et al., 2014; Delle Piane et al., 2015].

All of the above techniques are applicable to study small samples in great detail and provide high‐quality ref‑erence points along a depth profile. There may be cases though where continuous depth profiling on recovered cores is desirable, and this can be achieved, for example, via hyperspectral logging [e.g., Haest et  al., 2012]. The technique has been recently used to assess the variation in clay mineral content in cores extracted from the Harvey 1 well in the clay‐rich, Late Triassic unit known as the Yalgorup Member of the Lesueur Sandstone [Olierook et al., 2014], in order to evaluate potential containment stratigraphic horizons for CO2 injection in the Perth Basin, Western Australia, as part of the South West Hub Australian Flagship Carbon Capture and Storage Project [Stalker et al., 2013].

The second component of shale microstructure to be characterized in the context of CO2 geosequestration is pore space. Research on this topic has increased dramati‑cally in recent years, particularly since shale gas systems have become commercial hydrocarbon production tar‑gets [e.g., Loucks et al., 2009; Chalmers et al., 2012]. SEM imaging is certainly the most direct approach to investi‑gate porosity, but it is generally limited by (i) the poor quality of the mechanically prepared surfaces and (ii) the relatively low resolution achievable in a traditional fila‑ment instrument with respect to shale pore sizes. The introduction of field emission gun (FEG) SEMs has allowed the attainment of much higher resolution with images being acquired with pixel sizes down to a few nanometers. Also, traditional sample preparation for SEM imaging include several steps of mechanical grind‑ing and polishing of the surface under investigation via the use of fine grits and diamond suspensions. While effective for most geological materials, these methods tend to produce topographic irregularities due to the differential hardness of the fine‐grained components of shales. Loucks et al. [2009] showed that these irregular‑ities greatly exceed the size of nanopores typical of shales. To eliminate these conventional preparation limitations, argon‐ion milling has been introduced as a technique to produce a much flatter surface where the minor topo‑graphic variations are unrelated to differences in hardness of the sample forming minerals but, rather, to slight vari‑ations in the path of the Ar‐ion beam. Both focused ion beam (FIB) and broad ion beam (BIB) [e.g., Holzer et al., 2004; Desbois et al., 2009, respectively] preparations have been used to establish a pore classification scheme based on the morphology of pores from representative sample areas. Desbois et  al. [2009] defined three main types of pore morphology occurring at the interfaces between mineral particles: (i) type I (elongated pores between sim‑ilarly oriented clay sheets), (ii) type II (crescent‐shaped pores in saddle reefs of folded sheet of clay, and (iii) type III (large jagged pores surrounding clast grains). Type III

0004182022.INDD 7 10/16/2018 7:14:57 AM

8 GEOLOGICAL CARBON STORAGE

pores are typically larger than 1 μm, type II pores are bet‑ween 1 μm and 100 nm, and type I pores are <100 nm. It should be noted that both FIB‐SEM and BIB‐SEM have relatively small areas/volumes of investigation, with the former on the scale of a few square microns and the latter around 1 mm2.

Loucks et al. [2009], on the other hand, focused on the porosity observed within the organic matter defined as intraparticle organic nanopores and later introduced a simple classification scheme for the spectrum of pore types normally found in mudrocks, dividing them into mineral matrix pores (intraparticles and interparticles) and organic matter pores [Loucks et  al., 2012]. The classification follows the schemes normally adopted to describe pore space in sandstones and carbonates with the additional variable of pores contained within the organic matter. The classification is based on a ternary diagram populated via point counting of pores from SEM images. The authors concluded that interparticle and organic matter pores have generally better connec‑tivity than intraparticle pores. The latter pores will pro‑vide storage and contribute to some permeability but will not have the same level of connectivity as the other pore types [Loucks et al., 2012].

The 2D assessment of pore morphology has been com‑plemented by the recent development of field emission microscopes coupled with ion milling (FIB) tools. This advance has allowed the production of in situ high‐quality polished cross sections suitable for high‐resolu‑tion SEM imaging of pores down to the nanoscale and the backstripping of the sample surface, in at best 5 nm steps, to visualize three‐dimensional (3D) volumes of the specimen. Previous studies based on mercury injection highlighted that pore sizes of preserved shales are often well below the micron scale [e.g., Dewhurst et al., 1998, 1999a, 1999b, 2002; Yang and Aplin, 1998]. FIB nanoto‑mography is therefore ideal to describe porous networks in detail as it enables the 3D reconstruction of micro‑structural features on the 5–100 nm scale and can serve as a basis for quantitative microstructural analysis. Keller et al. [2011] used this approach on samples of Opalinus Clay considered as a candidate host rock formation for the disposal of radioactive waste. Their image acquisition and analysis workflow was adapted to the study of pores as small as about 10 nm, revealing preferential alignment of the pore paths along the bedding planes of the rock. Heath et  al. [2011] investigated the properties of pore types and networks from a variety of geologic environ‑ments using core samples from continental and marine mudstones. They recognized seven dominant pore types distinguished by geometry and connectivity based on the  quantitative and qualitative 3D observations. In particular, a dominant planar pore type was recognized in all investigated mudstones generally characterized by a

number of neighboring connected pores. The authors argued that connected networks of pores of this type likely control most matrix transport in these mudstones. Each sample was defined using cubes of 3D pore network reconstructions of 101.5 μm3 obtained through the inter‑polation of 2D slices collected at a spacing of 25 nm. Thyberg et al. [2010] also noted planar geometrical bodies of diagenetic quartz cement in mudstones which may have precipitated in such pores and would serve to further impede fluid transport in materials already at nano‐Darcy level permeability.

FIB nanotomography is a very powerful technique that allows imaging of porosity in situ; that is, the spatial relationships between pores and surrounding mineral matrix are preserved and constitute part of the visualiza‑tion results. However, the technique has strong limita‑tions in terms of sample volume and requires some image manipulation procedures in order to obtain quantitative information on the pore space. A novel approach is to complement the information provided by local high‐resolution imaging with bulk measurements of pore size distribution (size ranging from nanometers to microme‑ters) obtained from small and ultra small angle neutron scattering (SANS/USANS) on cm‐scale shale specimens [Gu et al., 2015; King et al., 2015] or small angle X‐ray scattering (SAXS) [e.g., Mildner et al., 1986; North et al., 1990] to gain a more complete representation of the pore architecture and connectivity in shales (see below).

In terms of material preservation, mineralogical com‑position, CEC, and SSA are not significantly affected by desiccation, and indeed, generally the samples are dried out and powdered before such tests are made. However, microstructural imaging and pore size distribution mea‑surements can be affected by lack of preservation. In soft and stiff clays, particle orientation can change on desicca‑tion as can clay mineral shape, with resultant impacts on pore size distribution [Diamond, 1970; Delage et al., 1982; Griffiths and Joshi, 1989, 1990]. Capillary threshold pressure and resultant seal capacity calculations for shales using mercury porosimetry can also be affected by drying method. Dewhurst et al. [2002] found that freeze‐drying produced the most consistent seal capacity calculations, whereas air‐drying and vacuum‐drying results were more scattered.

1.4.  GEOMECHANICAL PROPERTIES

Seal integrity for both hydrocarbons and CCS is usu‑ally assessed from geomechanical properties of the rocks in question. Laboratory properties of shales can be used to inform geomechanical models used to investigate the impact of injection on a prospective CCS site [e.g., Rinaldi et al., 2015; Zhang et al., 2015a] or as inputs for geome‑chanical techniques such as slip tendency [Morris et al.,

0004182022.INDD 8 10/16/2018 7:14:57 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 9

1996], critically stressed fractures [Barton et al., 1995], or fracture stability [Mildren et  al., 2005]. These latter approaches have been used to assess potential CCS sites [e.g., Vidal‐Gilbert et  al., 2010; Rasouli et  al., 2013; Tenthorey et  al., 2013, 2014; Zhang et  al., 2015b] in combination with magnitude and orientation of the situ stress field and pore pressure measurements. However, these field‐scale approaches are beyond the scope of this paper which is reviewing laboratory measurements, and as such, we will concentrate on the latter in this section.

Standard triaxial testing and multistage triaxial testing are the best methods to obtain shale geomechanical prop‑erties [e.g., Marsden et  al., 1992; Horsrud et  al., 1998; Dewhurst and Hennig, 2003; Nygard et al., 2004; Dewhurst et al., 2015; Skurtveit et al., 2015]. While unconfined com‑pressive strength (UCS) testing and Brazilian testing are also standard tests, they are not recommended for shales as samples can dry out during testing and thus properties can be altered (see below). UCS can be calculated from standard triaxial tests anyway [e.g., Fjær et  al., 2008]. Standard triaxial testing involves placing samples in a cell surrounded by hydraulic oil at different confining pres‑sures and loading axially until failure occurs and residual strength is reached. Best practice multistage triaxial test‑ing involves axially loading samples to 80–90% of their failure strength at a given confining pressure, unloading to zero axial load, increasing confining pressure, allowing consolidation, and repeating axial loading [e.g., Fjær et al., 2008; Dewhurst et  al., 2015]. To judge where individual axial loading stages should be terminated, assessment of proximity to failure can be determined through deviations

from linearity in the load‐displacement curve (or stress‐strain curve) or by monitoring volumetric strain and the onset of dilatancy [e.g., Fjær et al., 2008; Youn and Tonon, 2010; Dewhurst et al., 2015]. This train of events is then repeated multiple times, and the final stage is taken through to failure and residual strength. See Dewhurst et al. [2015] for a full description of the multistage meth‑odology and associated limitations, plus Dewhurst and Siggins [2006] for definitions of geomechanical termi‑nology used in this paper.

Geomechanical properties were required to evaluate the fault and top seal for the CO2CRC Otway Project [Vidal‐Gilbert et al., 2010], a pilot‐scale CO2 storage site in Victoria, Australia. Standard triaxial testing was per‑formed at varying confining pressures on seven 50 mm long × 25 mm diameter core plugs of preserved Belfast Mudstone plugged parallel to bedding. This orientation was necessary due to closely spaced stress relief fractures parallel to bedding which did not allow the recovery of long enough plugs normal to bedding. Two plugs failed early, splitting axially along the bedding, and were excluded from the analysis of rock strength, so results from five plugs are shown here (Fig.  1.3). The Belfast Mudstone is shown to be a weak rock with a cohesive strength of ~2 MPa and a friction coefficient <0.5 (Fig. 1.4). It should be remembered that these properties are mea‑sured parallel to bedding and that shales are significantly anisotropic in regard to strength and elastic properties [e.g., Niandou et al., 1997; Ajalloeian and Lashkaripour, 2000]. In general, shales are stronger and stiffer normal and parallel to bedding and weaker at 30–60° to bedding,

0

10

20

30

40

50

0 0.5 1 1.5

4.9 MPa

10.7 MPa

27.0 MPa

5.8 MPa

10.7 MPa

Diff

eren

tial s

tres

s (M

Pa)

Axial strain (%)

Figure 1.3 Stress‐strain curves for Belfast Mudstone under increasing confining pressures with deformation parallel to bedding. The samples seem to become more brittle with increasing confining pressure, with the largest stress drop observed at the highest confining pressure used.

0004182022.INDD 9 10/16/2018 7:14:57 AM

10 GEOLOGICAL CARBON STORAGE

usually weakest at ~45° [e.g., Valès et al., 2004; Skurtveit et al., 2015]. However, it is not clear whether shales are stronger normal to bedding or parallel to bedding or if the strengths in these orientations are similar as various past research shows contrasting results [e.g., Niandou et  al., 1997; Ajalloeian and Lashkaripour, 2000; Valès et  al., 2004; Skurtveit et  al., 2015]. The anisotropy of shale properties will be returned to in the discussion sec‑tion below. These results were used in field‐scale geome‑chanical assessments of the Otway site and the nearby Iona gas storage site by Tenthorey et al. [2013] and Vidal‐Gilbert et al. [2010].

As noted above, preservation of shales for geomechan‑ical testing is critical in terms of measurement of rock properties that would be applicable to in situ scenarios. Loss of in situ pore fluids from shales can completely change the rock properties being measured, and this is particularly the case in terms of parameters derived from triaxial testing such as strength and static/dynamic elastic properties such as Young’s modulus or Poisson’s ratio [e.g., Valès et  al., 2004; Dewhurst et  al., 2012; Pervukhina et  al., 2015]. An example of the change of properties on loss of water is shown in Figure 1.5. Here a preserved set of samples of Pierre Shale have been deformed at room temperature under different confining pressures to determine a failure envelope, which shows a cohesive strength of ~3 MPa and a friction coefficient of 0.39. A second set of Pierre Shale samples were equil‑ibrated in a desiccator in a lower RH environment, resulting in a water saturation of 59%. Hence, these samples are no longer fully preserved and have lost some of their in situ water content. It should be noted that the 59% water saturation quoted is the value measured before shearing and that saturation will change during deformation due to porosity loss. These samples were deformed under the same effective stress and temperature

conditions as the preserved samples. However, these partially saturated samples are significantly stronger than the preserved samples, with a cohesive strength of >12 MPa. The friction coefficient remains similar though for both sets of samples. Other examples of the effects of changing water saturation in shales can be found in Pervukhina et al. [2015]. Again, this shows the significant impact the lack of preservation can have on the results of rock properties tests on shales and why it is critical to preserve samples as soon as a shale core is recovered to the surface.

1.5.  POROSITY AND PERMEABILITY

Transport properties and sealing capacity are two fundamental aspects for shale barriers when evaluating a site for CO2 geosequestration. These parameters will con‑trol the volume and displacement velocity of fluid(s) able to move through the sealing layer. Other factors also affect flow properties, namely, (i) pore network tortu‑osity, (ii) pore network geometry (e.g., equant/tubular pores, planar microcracks) and grain surface, (iii) pore network connectivity, and (iv) wettability. Most of the characteristics of the pore network are affected by stress (nature and magnitude) to variable extents; for instance, microcracks are more prone to deformation and closure under compressive stress than equant pores [e.g., Paterson, 1977]. In addition, these characteristics of the pore net‑work are not independent from each other, and in fact, a complex interaction between them is usually observed; as such, microcrack closure under increasing compressive stress affects pore network connectivity and hence fluid transport properties.

00

10

20

She

ar s

tres

s (M

Pa)

t f=12.74 + 0.40σ′n

t f=3.21+0.39σ′n

30

40

10 20 30 40

Effective normal stress (MPa)

50 60

Sw= 100%

Sw= 59%

70 80

Figure 1.5 Failure envelopes for Pierre Shale deformed when fully water saturated (grey line) and partially water saturated (black line). Sw is the percentage of water‐saturated porosity before samples were confined and axially loaded. The Mohr circles for the upper failure envelope have been omitted for clarity. Pierre Shale becomes substantially stronger as water saturation decreases although the friction coefficient does not change much.

0

10

20

30

40

0 10 20 30 40 50 60 70 80

She

ar s

tres

s (M

Pa)

Effective normal stress (MPa)

tf= 2.23 + 0.48σ’n

Figure 1.4 The Mohr circles derived from the peak strength takes from the stress‐strain curves in Figure  1.3. The failure envelope appears to be linear with low cohesive strength and a friction coefficient lower than the standardly used values for non‐shaly rocks.

0004182022.INDD 10 10/16/2018 7:14:58 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 11

In clay‐rich shales, porosity and permeability prop‑erties are linked by scale from nanometers to kilometers as illustrated by Neuzil [1994] and Sondergeld et al. [2010]. Over this wide range of scale, the notions of free fluid and bound fluid are fundamental as they affect the storage/sealing capacity and flow properties in different manner and intensity. At the reservoir scale, fractures can play a major role in the leakage risk and can be defined as the free fluid component of the shale formation. The fractures can have various origins such as overpressure, tectonic or induced during drilling and/or injection. In most cases, fractures, if preexisting, can be identified with logging tools and seismic exploration, and their scale and connec‑tivity will govern the overall permeability (integrity) of the sealing barrier. The large‐scale effects are beyond the scope of the paper however and will not be discussed further here [see, e.g., Neuzil, 1994; Ingram and Urai, 1999; Dewhurst and Hennig, 2003; Lash, 2006].

At the pore scale, the pore sizes generally range from sub‐nanometer to micron scale in clay‐bearing shale for‑mations, making the notion of free fluid and effective porosity less applicable in the very small pores. Bound water is related to the interactions of the fluid(s) with min‑erals, mostly the clay minerals due to their strong affinity for water. As a result of the extreme surface area of clays, the water binds to clays over a large volume that is in most cases the dominant contribution to the total water content. Some clays, for example, smectites, can swell significantly by incorporating water (and ions) into interlayers in the mineral structure, which can complicate porosity determi‑nation [Brown and Ransom, 1996]. The various different types of water (free, clay‐bound, capillary, and interlayer) make porosity determination in these materials rather more complex than it first appears [e.g., Pearson, 1999].

1.5.1. Porosity and Seal Capacity Determination

The recent interests in shales by the petroleum industry pushed development of new and existing technologies to investigate microscale and even nanoscale porosity in these rocks. Beyond the knowledge of porosity to estimate volumes and some fluid transport aspects, the small porosity scales occurring in shales have been found to have an impact on seismic‐scale phenomena and rock physics modeling, for instance. Thorough porosity anal‑ysis in shales requires combined techniques to overlap all scales of investigations. The first rule to apply when investigating porosity in shales is to have preserved sam‑ples (i.e., original saturated water content) and handle samples with an extreme care to avoid dehydration and crack generation. Porosity can be assessed by various lab‑oratory approaches: (i) 2D and 3D images, (ii) gas methods, and (iii) liquid methods. All of these approaches and methods have pros and cons.

1.5.1.1. Image ApproachesWhile assessment of nanopore scales and above have

become easier and more accurate with the recent technol‑ogies such as SEM or X‐ray micro‐CT images, porosity determination remains challenging at the nanometer scale. Sample preparation can dry the samples that will change texture and potentially form some microcracks, although sometimes this can be mitigated by freeze‐drying or critical point drying [Chiou et al., 1991; Dewhurst et al., 1998; Holzer et al., 2010]. However, such images are still useful to detect clay locations that could affect the flow properties (e.g., pore throat clogging, grain coating, pore/crack bridging) and will give essential insights about diagenetic processes. In all cases, image processing is not trivial, particularly the segmentation to separate pores from the various minerals composing the rock. Establishing representative elementary volumes for material property determination from images is also a non‑trivial task [e.g., Keller et al., 2013].

1.5.1.2. Gas ApproachesThese methods are usually used on unpreserved samples

with all the inherent risks on result quality (e.g., clay shrinkage, cracks, residual pore fluid), although again, such methods could use freeze‐dried or critical point dried samples. Nitrogen or helium can be injected into solid plugs. The principle of gas methods is to invade the pore network with a high‐diffusivity inert gas. Gas expan‑sion is monitored, and the pore volume (i.e., porosity) of the rock sample is calculated from Boyle’s law:

V

P VP1

2 2

1

where V1 is the volume of gas permeating the rock sample, P2 and V2 are the pressure and the calibrated volume of gas before being released into the sample, and P1 is the pressure of gas after sample infiltration. Overburden stress can be applied to close any potential microcracks. The low permeability of shales makes such measurements very time consuming to allow gas to permeate through the entire volume of the sample. Variations in tempera‑ture during the experiment will also affect the results. As a result of these limitations, such a method is rarely accu‑rate and repeatable [e.g., Sinha et al., 2013; Wang et al., 2013; Heller et al., 2014].

Helium pycnometry on granulated samples and pow‑ders is a good method to measure bulk and grain density in order to compute the total porosity. The best proce‑dures involve granulation of preserved samples into 0.5–1 mm diameter size and immediately perform accu‑rate mass and volume measurement by gas pycnometry to compute bulk density. The measurement is then repeated after drying the granulated samples in oven at 105°C for up to 2 weeks until mass stabilization occurs.

0004182022.INDD 11 10/16/2018 7:14:58 AM

12 GEOLOGICAL CARBON STORAGE

Figure 1.6 Opalinus Clay preserved (left) and the same sample after a few seconds in a non‐appropriate brine solution (right): note the massive structural deterioration.

The bulk density and grain density of the sample are therefore accurately measured under the same conditions, and total porosity can be calculated (assuming negligible amounts of residual salts from pore fluid evaporation) from

grain bulk

grain fluid

100

where ϕ is the porosity in percent and densities (ρ) are in g/cm3. Gas pycnometry uses the same gas expansion prin‑ciple described above but works at very low gas pressure (usually around 20 psi) to measure the sample volume with an accuracy usually around 0.0001 cm3. Most of the issues with grain density computations in laboratory come from the drying process. Many experiments have demonstrated that solid chunks/plugs cannot be fully dehydrated at 105°C due to isolated water‐filled pores and clay‐bound water that require time to dehydrate by diffusion transport, and this can result in lower than expected grain density measurements.

The molecular size of the gas used for the porosity measurement in the laboratory (N2, 0.421 nm equivalent spherical van der Waals radius; He, 0.356 nm) compared to the actual size of carbon dioxide (CO2, 0.454 nm) can affect the gas measurement results [Sondergeld et  al., 2010]. Helium is often used for porosity measurements, although its size is significantly smaller than carbon dioxide. As pore throats approach the molecular diam‑eter of the gas used for testing, helium can pass into adja‑cent pores where CO2 will not be able to pass. Layers of bound ions or water make the constriction filtration effect even more prominent. Therefore, helium deter‑mined porosity is potentially greater than the effective porosity to CO2. The magnitude of these porosity dis‑crepancies is controlled by pore size distributions, which are poorly constrained and can reach up to a factor of 2.

1.5.1.3. Liquid ApproachesSeveral methods exist to access the porosity using injec‑

tion of liquid (brine or mercury). Some of the techniques try to use the original water content without affecting the preservation state of the samples, and others push liquid through the pore network using dried samples (freeze‐dried, critical point dried, or unpreserved).

Water immersion porosity (WIP) consists of saturating a solid plug or solid chunk of shale with water to invade all the pores. The wettability, fluid type, fluid composi‑tion, and pore pressure will affect the water mass intake of the sample during saturation, leading to considerable inconsistency. The saturation process can take a very long time as all fluids move at a diffusive speed and will only access the connected pores with a capillary force less or equal to the water injection pressure. In other words, it is virtually impossible to reach nanopores that require a large capillary force to invade, especially as the nanopores are water wet. However, it is fundamental to understand in the case of injecting water that any variation in the salinity and/or salt composition of the injected water from that in equilibrium with the shale will generate osmotic forces. Clay‐water reactions and capillary forces in the case of partial saturation can quickly destroy the shale structure during water injection and re‐saturation (Fig. 1.6) and is particularly significant in smectite‐rich shales that can swell or shrink. However, the presence of organic matter and organic matter coatings can result in different wettability, and this may affect how clay min‑erals and shales as a whole respond to water injection.

If the sample is well preserved (i.e., approximately fully saturated), the sample can be dried under 105°C until mass stabilization and the difference in saturated and dry mass corresponds to the total water content. Assuming or knowing a pore fluid density allows the equivalent porosity to be computed from the sample volume [e.g.,

0004182022.INDD 12 10/16/2018 7:14:58 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 13

Head, 1980]. This is a relatively standard method and easy to use for well‐preserved samples, and porosity cal‑culated this way is directly related to the strength of over‑burden shales [Dewhurst et al., 2015].

Mercury intrusion porosimetry (MIP) consists of recording the volume of injected non‐wettable mercury at increasing injection pressures through a dry solid resin‐coated shale cylindrical plug or, in a worst case scenario, cuttings [e.g., Vavra et  al., 1992; Dewhurst et  al., 2002; Esteban et  al., 2006]. Assuming cylindrical pores, the Laplace‐Washburn equation [Washburn, 1921] can relate the pressure of mercury injection to an equivalent pore throat size. The cumulative pore throat volumes or total volume of injected mercury corresponds to the connected porosity. The typical maximum pressure of injection is ~400 MPa which is equivalent to a cylindrical pore throat diameter of ~4 nm. However, MIP is always lower than water or gas‐derived porosity as mercury cannot access all pores. Some authors also demonstrated that pore ori‑entations in shales can be anisotropic [e.g., Keller et al., 2013], such that there is a strong orientation of pores in the foliation plane of shales resulting from grain shape and compaction, although MIP measurements per‑formed parallel and normal to bedding in shales often show little difference in intrusion spectra [e.g., Dewhurst et al., 2002; Sarout et al., 2014].

MIP has also be used to estimate the capillary seal capacity of shales, that is, the height of a hydrocarbon or CO2 column that can be retained by a given seal under in situ conditions before the non‐wetting phase begins to move into and potentially breach the seal, assuming an idealized cylindrical tube pore morphology. Essentially, a capillary threshold pressure (Pth) for the air‐Hg system can be calculated from a plot of pressure against cumulative intruded volume of mercury (or mercury sat‑uration). This is considered to be the point where a con‑tinuous filament of non‐wetting phase traverses the seal. There are a number of ways to pick Pth, including (i) the inflection point on the pressure‐cumulative volume curve [e.g., Dewhurst et al., 2002]; (ii) a percentage of mercury saturation, usually between 5 and 10% [e.g., Schowalter, 1979]; or (iii) by using the entry (or displacement) pressure where mercury first intrudes the seal sample [e.g., Vavra et  al., 1992]. None of these techniques has much rigor around them, and in the main, they are based on experi‑mental results from Schowalter [1979] or from Katz and Thompson [1986] who evaluated mercury penetration through sandstones using MIP and electrical properties. Once an air‐Hg threshold pressure has been calculated, this can be converted to an in situ capillary pressure for a given CO2 column height through the use of air‐Hg and brine‐CO2 contact angles and interfacial tensions; finally, a column height can be calculated from the seal and res‑ervoir threshold pressures and the difference in brine and

CO2 fluid densities. There are multiple publications that cover these methods in great detail [e.g., Schowalter, 1979; Watts, 1987; Vavra et al., 1992]. It should be noted that estimated MIP threshold pressures for CO2 do not often match CO2 breakthrough pressures from laboratory experiments on shales [Hildenbrand et  al., 2002]. In addition, while it has long been known that surface con‑ditions of samples need to be corrected for [so‐called con‑formance; e.g., Vavra et al., 1992], more recently, it has been suggested that for tight samples such as shales, parts of the MIP curve represent sample damage or compress‑ibility and not intrusion of mercury [e.g., Clarkson et al., 2012], and there is some concern that in tight shales, mer‑cury might not even be entering the pores at all [Hildenbrand and Urai, 2003; Q. Fisher, pers. comm., 2015]. As a final word of warning, MIP assumes that the shale system is water wet and there is some uncertainty around wettability in CO2‐mineral systems, although there is minimal research as yet on clay minerals which usually form the bulk (the load‐ and pore‐bearing frame‑work) of shaly caprocks [Iglauer et al., 2015]. Capillary pressure techniques for measuring mudrock properties have been extensively examined by Busch and Amann‐Hildenbrand [2013] and further discussed by Amann‐Hildenbrand et al. [2013] in the context of CO2 storage.

Ferrofluid injection combined with magnetic suscepti‑bility consists of injection of a non‐wettable ferrofluid into the pore network under high pressure. Ferrofluid is a liquid made of low‐viscosity isoparaffin saturated with nanoparticles of magnetite. Such nanoparticles will be able to invade the small pores and will return a strong magnetic signal correlated to their amount [Pfleiderer and Halls, 1994]. Magnetic susceptometers can measure the intensity of the magnetic signal of the ferrofluid‐sat‑urated shales. The amount of ferrofluid in the shales (i.e., porosity, assuming full saturation) can be calibrated against the magnetic signal from a specific amount of fer‑rofluid. The use of an MIP apparatus with mercury replaced by ferrofluid was successfully applied on shales and provided information about porosity and pore shape/pore connectivity [Esteban et al., 2006, 2007].

Low‐field nuclear magnetic resonance (NMR) is prob‑ably the most promising tool for porosity and pore size distribution in shales. NMR is a nondestructive method that records the processing of protons occurring in liquid water for shales [Martinez and Davis, 2000; Dunn et al., 2002]. The magnitude and time relationship of the mag‑netization decay signal after applying an external magnetic field (the so‐called relaxation time) is directly proportional to the amount of protons and their interac‑tions with the pore volumes and mineral surfaces. The water content measured by NMR can then be converted into equivalent porosity (assuming that water fully satu‑rates the pore network) knowing the sample volume.

0004182022.INDD 13 10/16/2018 7:14:58 AM

14 GEOLOGICAL CARBON STORAGE

Beyond porosity measurements, NMR can provide details of the pore fluid distribution such as the clay‐bound water, capillary water, and potential movable water in cracks. Models can be used to tentatively com‑pute permeability [Coates et  al., 1991; Martinez and Davis, 2000; Hidajat et al., 2003; Arns et al., 2004; Minh and Sundararaman, 2006; Josh et al., 2012; Rezaee et al., 2012]. However, extreme care must be taken due to NMR machine resolution as a 2 MHz spectrometer (equivalent to NMR logging tool frequency) will not be able to resolve pores smaller than 10 nm [Nicot et  al., 2006]. Therefore, the total NMR porosity usually underesti‑mates the real porosity by missing the part of the water signal from clay‐bound and interlayer water. To over‑come this limitation, a 20 MHz NMR spectrometer is more adapted to resolve nanopores, but the macroporos‑ity will not be detected. A combination of 2 and 20 MHz NMR instruments is a good compromise to fully assess the pore size distribution and porosity of shales. In regard to the NMR technique and sample preservation, two spectra are shown in Figure 1.7, illustrating the impact of drying on NMR response in Opalinus Clay. The solid line shows the NMR spectrum for a preserved sample and clearly has a large peak at relaxation times between 0.4 and ~3 μs, which corresponds to capillary and clay‐bound water. On dehydration, this water is mostly lost, and all that remains is water with relaxation times below 0.4 μs which is likely tightly bound and interlayer water in mixed layer illite‐smectite.

Taken individually, each method will generate different results for porosity measurements [Howard, 1991]. Method combinations are necessary to overlap the differ‑ent scales accessible by each method [Al Hinai et al., 2014] and for quality control in the overlap regions (Fig. 1.8). Repeatability of the measurements is the major labora‑tory issue due to difficult pore fluid accessibility related to the different sample treatment (grinding, sieving, drying, and partial re‐saturation, same environmental conditions). Therefore, the best measurements are always achieved on preserved shales, particularly given the issues with crushing samples, limited pore access using MIP, or the clay‐water reactions that can result from liquid re‐sat‑uration [Kuila, 2013]. Consistent sample treatment and environmental conditions (temperature, pressure, and RH) for each method, and ideally a combination of methods, can help overcome this repeatability limitation.

1.5.1.4. Radiation‐based ApproachesIn recent years, with the advent of gas shales, radiation‐

based approaches such as SAXS and SANS/USANS have come back into fashion after initial application to shales back in the 1980s [e.g., Hall et  al., 1986], and neutron‐based methods have recently been applied to CO2 storage research by Busch et  al. [2014]. These methods use an incident beam of thermal neutrons or X‐rays of known intensity and fixed wavelength impact‑ing on the sample and then recording of the angle and intensity of scattering [e.g., Radlinski et al., 2004]. These

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0.01 0.1 1 10 100 1000

Incr

emen

tal e

quiv

alen

t wat

er-f

illed

por

osity

(%

)

Relaxation time constant T2 (ms)

Water distribution in shale (NMR 2 MHz)

Preserved

Dehydrated (65°C) and rehydrated under room condition relative humidity

Figure 1.7 Nuclear magnetic resonance spectra for preserved and dehydrated Opalinus Clay showing significant differences in response due to dehydration. A clay‐bound water peak is visible in the preserved sample but is almost gone in the dried‐out sample.

0004182022.INDD 14 10/16/2018 7:14:59 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 15

parameters are dependent on the geometry of the pore‐matrix interface where scattering occurs [e.g., Hall et al., 1986; North and Dore, 1993; Radlinski et al., 2004]. Data obtained from such experiments can give information on total porosity, SSA, and pore size distribution (over a range from ~0.5 nm to ~10 μm) as long as certain reason‑able assumptions are made (e.g., spherical pore models). However, results can show significant anisotropy on sec‑tions normal to bedding due to preferred orientation of particles (especially clays), although isotropy is usually noted in the bedding plane [e.g., Hall et  al., 1986; Gu et al., 2015]. Essentially, preferred orientation of particles resulting from compaction and diagenesis leads to preferred orientation and shape of pores, with strongly oriented slit and tube‐shaped pores causing the observed anisotropy. This is observed in overburden shales as well as gas shales [e.g., Keller et  al., 2011]. This anisotropy gives different porosity and SSA in different orientations although this can be corrected for [e.g., Gu et al., 2015].

Various studies have also compared the results of SAXS and SANS/USANS to pore size distribution and porosity determined through MICP and N2 adsorption, for example. Differences are often noted between the results of these tests as they measure different properties, for example, total porosity and pore size distribution for the radiation‐based methods and connected porosity and pore size distribution from MICP and N2 adsorption [Hall et al., 1986; Clarkson et al., 2013; Gu et al., 2015]. The pore size distributions calculated from these methods also dif‑fer in that some measure pore bodies (N2) and others pore throats (MICP), while the radiation‐based methods do not distinguish between throats and bodies [e.g., Clarkson et al., 2013]. SAXS and SANS/USANS usually give higher

porosities than N2 adsorption which in turn is higher than MICP due to pore accessibility issues for the larger mole‑cules, as well as different assumptions between scattering models, adsorption models, and MICP models used to estimate porosity and SSA [Gu et al., 2015].

SAXS and SANS/USANS have different sensitivities to water‐filled pores, especially when pore sizes are <2 nm and such pores are expected to hold water under ambient conditions [Hall et al., 1986]. For SANS/USANS, the con‑trast between hydrocarbon‐ or water‐filled pores is strong compared to empty pores, but for SAXS the contrast is much less [e.g., Hall et al., 1986]. While SANS/USANS, for example, nominally evaluates total porosity, Clarkson et  al. [2013] used deuterated methane (CD4) to look at connected porosity in gas shales with SANS/USANS and noted that not all micropores (< 2 nm) and surprisingly not all large pores (> 0.3 μm) were accessible to CD4.

Overall, radiation‐based methods provide a further string to the bow in analyzing and evaluating shale properties, especially those that are porosity and pore size distribution based. In combination with other techniques which analyze different parts of the porosity spectrum, they can provide additional information useful in fully characterizing shale behavior at the nanometer to micron scale.

1.5.2. Permeability Techniques

Clay‐rich caprocks are common in sites evaluated for geological storage of CO2 due to high capillary seal capacity and often nano‐Darcy level permeability. Additionally, it is known that CO2 can be adsorbed onto clay minerals; as such, if capillary leakage occurred, this could result in extra storage capacity [e.g., Busch et al.,

< 0.1 nm 1 nm 10 nm 100 nm 1 µm 10 µm 100 µm 1 mm

Pore size range detections per method for shales

Methods of porosimetry

He/N2 pycnometry on plugs

GIR

N2 adsorption

MIP

WIP

NMR 2 MHz

NMR 20 MHz

Ferrofluid porosimetry (magnetics)

Density (bulk and grain)

Figure 1.8 Overview of the pore size ranges detected by different methods of porosimetry for clay‐bearing shales.

0004182022.INDD 15 10/16/2018 7:14:59 AM

16 GEOLOGICAL CARBON STORAGE

2008]. Once capillary seal leakage occurs, then fluid transport properties such as water and effective CO2 per‑meability become important [Hildenbrand et  al., 2002]. There are a number of methods for measuring water per‑meability of materials such as constant head tests, falling head tests, steady state flow, and transient pulse decay methods. The first two are generally unsuitable for clay‐bearing materials, especially softer materials due to high hydraulic gradients which are imposed that can lead to sample deformation [Olsen et  al., 1985]. Pulse decay methods work well and have been used by Brace et  al. [1968] and Kwon et al. [2004], for example, for testing of low‐permeability materials as the technique is often held to be the most rapid of all the low‐permeability testing techniques. However, in general, steady state methods have been used for the assessment of CO2‐related cap‑rocks in that this method can assure full saturation of a given caprock sample with brine under pressure as well as measuring permeability to said brine before CO2 flooding through a sample is attempted [e.g., Hildenbrand et  al., 2002; Wollenweber et al., 2010].

An example instrument for steady state flow testing by water injection and measurement of permeability is shown in Figure  1.9, along with some results on a cemented shale. Nominally, a disc of 25.4 mm in diameter and 10–15 mm in length is confined between two platens inside a Viton jacket and surrounded by pressurized hydraulic oil in a small steel cell. During a test, tempera‑ture and confining pressure are maintained at constant

values, typically 22°C and 20 MPa, respectively. Saturation with an appropriate brine composition (determined from petrophysical measurements) occurs through injection at the upstream side of the sample (bottom end) for several days while monitoring the downstream fluid pressure (top end) with the valve closed (Fig.  1.9). The down‑stream fluid circuit is initially flooded with brine at atmo‑spheric pressure and then closed. During the saturation process, it shows an increase in pressure up to the desired value, equaling the pressure imposed at the upstream end (typically 10 MPa). The injection procedure helps us ensure that the specimen is saturated prior to the perme‑ability test, although 100% water saturation is still an assumption, given there are almost no methods that can assess this properly in tight shales. The relatively high pore fluid pressure imposed (10 MPa) allows any residual air within the pore network to be dissolved in the brine. Following saturation, the downstream pressure is regulated at a specific pressure, typically 10 MPa. The upstream pressure is suddenly increased from 10 MPa to a higher pressure (15 MPa, for instance) and regulated at that value using a high‐accuracy volumetric pump. The resolution of the upstream pump volume monitoring is 10−6 ml/min. The fluid pressure difference between the upstream and downstream circuits is then held constant at ∆P = 5 MPa throughout the test. Once a steady flow is achieved (straight line in Fig. 1.9), the constant fluid flow imposed by the pump is used to compute permeability using Darcy’s law. Once this steady state flow regime is

Time [days]

Vol

ume

inje

cted

at c

onst

ant D

PP (

ml)

Linear permanent regimeInitial transitory regimeLinear regression

Constant DPP = PPU - PP

DDownstream fluid reservoir:constant rservoir pressure PP

D

Topplaten

Bottomplaten

Confiningpressure

cell

High-accuracy volumetric pump:regulated injection pressure PP

U

Cylindricalshale

specimen

PPD

PPU

ControlledDPP

Vitonsleeve

ResultingQ

L

k: permeability in flow direction

Darcy’s law:Q = A

MESE permeability method Permeability of Mancos Shale

0 1 2 3 40

0.02

0.04

0.06

0.08

0.10

0.12

0.14

Slope: Fluid flux Q

k ΔPpμ L

(a) (b)

Figure 1.9 (a) Schematics of an experimental setup for permeability measurement of a cylindrical shale specimen. (b) Example of a permeability measurement for Mancos Shale, parallel to bedding, showing the fit of displaced water volume versus time during the permeability test.

0004182022.INDD 16 10/16/2018 7:15:00 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 17

established, the average pore pressure within the shale sample is expected to be 12.5 MPa so that the sample experiences a mean effective stress of 7.5 MPa.

Figure 1.9 shows the evolution of the volume of water displaced by the injection pump in order to maintain the constant pressure difference across the specimen. Steady state is achieved after about a day from the beginning of the permeability test for this Mancos Shale sample cut parallel to bedding. The interval between 1 and 4 days is chosen for the determination of the permeability of this shale as a linear fit with a good correlation factor can be obtained. From Darcy’s law,

Q A

k PL

where Q is the fluid flow rate in (m3/s), k is the perme‑ability in (m2), μ is the fluid dynamic viscosity (equal to 10−3 Pa.s for water at room temperature), ∆P is the difference in fluid pressure across the specimen in (Pa), A is the cross‐sectional area of the cylindrical specimen in (m2), and L is its length in (m). In this case, the perme‑ability of the Mancos Shale parallel to bedding was 1.6 nD (L  =  12.9 mm and Q  =  1.82 10−5 ml/min for ΔP = 4.98 MPa), and that of the Mancos Shale normal to bedding was 1.1 nD (L = 15.1 mm and Q = 1.08 10−5 ml/min for ΔP = 4.77 MPa). Anisotropy is a topic that will be returned to in the discussion. During these tests, the flow through the specimens reaches a steady value that is consistent with time if left over longer time periods. The small length of the specimens used for the permeability tests helps to reduce the duration of the test to few days. Note that very small flow rates are achieved during the test; however, thanks to the high accuracy of the pump volume monitoring, this flow rate is resolvable.

Following full brine saturation of a shale sample and measurement of water permeability, CO2 breakthrough experiments are then performed [e.g., Hildenbrand et al., 2002; Angeli et  al., 2009; Wollenweber et  al., 2010]. Essentially, the pressure of CO2 is raised significantly on one side of the specimen and breakthrough monitored on the opposite side through pressure transducers, geochem‑ical sampling, and/or measurements of P‐wave velocity. Both drainage and imbibition‐type experiments have been used to attempt to define a seal capacity to CO2 for such experiments, although there does not seem to be good agreement between these methods nor with estima‑tions of seal capacity from mercury porosimetry tests [e.g., Hildenbrand et al., 2002; Wollenweber et al., 2010], with potentially all methods having issues that are not yet  fully resolved in terms of getting consistent results. These inconsistencies could be caused by the various dif‑ferent gas injection methods that are used, methods of interpreting seal capacity, and issues mentioned with

mercury porosimetry earlier. However, what appears certain is that effective permeability to CO2 is at least an order of magnitude lower than the brine permeability in the measured rocks and, in most cases, the brine perme‑ability was already at the nano‐Darcy level. Such tests also noted increases in both diffusivity and permeability with repeated cycling of pressure, some of which was ascribed to mineral reactions or adsorption of CO2 on mineral surfaces [e.g., Chiquet et al., 2007; Busch et al., 2008; Shah et  al., 2008]. However, Wollenweber et  al. [2010] noted no significant alteration of samples during such tests although they raised the possibility that redis‑tribution of phases could be occurring, that is, dissolu‑tion and re‐precipitation on the experimental timescale, which would not be detectable by traditional mineralog‑ical analysis methods.

It is not clear as yet as to the specific mechanisms behind CO2 breakthrough in such rocks. Traditional seal  capacity analysis would mean that when capillary breakthrough occurs, the non‐wetting fluid flows through the pore structure, which remains unaltered by either the fluid pressure (or effective stress change) or chemical reactions, that is, essentially a passive process. However, experiments where sample dimensions and P‐wave velocity have been monitored during CO2 injection [e.g., Angeli et  al., 2009; Skurtveit et  al., 2012, 2015] seem to  suggest that the process is far from passive and that either pre‑existing microcracks are dilated or new ones are forcibly formed. This is also consistent with suggested methods of gas flow through argillaceous rocks in the nuclear waste disposal industry [e.g., Harrington and Horseman, 1999; Harrington et al., 2009]. This topic will be expanded upon in the discussion.

1.6.  DIELECTRIC METHODS

Most mudrocks contain abundant charge carriers in the pore fluids and on surfaces that are mobilized and/or become reoriented in an electrical field, in a process termed polarization. The degree of polarization can be measured in terms of the dielectric permittivity of a given material (also known as the dielectric constant, ε′). The anomalously high conductivity (low resistivity) of clay‐bearing rocks containing pore fluid of a given salinity is well known in petrophysics and can be related to electrical conduction on and close to clay mineral surfaces. The charges on and around clay particles become increasingly polarized by space‐charge mechanisms as frequency decreases so that a clay or mudrock can have a dielectric permittivity that easily exceeds 100 at a frequency of 10 MHz. The electrical properties of clays and shales exhibit a complex frequency dependence controlled by (i) water content and its salinity, (ii) surface area and surface charge density that are essentially controlled by clay

0004182022.INDD 17 10/16/2018 7:15:01 AM

18 GEOLOGICAL CARBON STORAGE

content and clay type, and (iii) other microstructural characteristics including pore and grain shape and pore space tortuosity and interconnectedness. In general, dielectric properties of clay‐rich shales are dominated by water content at high frequency (≳ 300 MHz) and by CEC at frequencies ≲50 MHz [e.g., Myers, 1991; Josh, 2014]. These aspects make dielectric properties appealing in clay‐bearing shales as CEC and water content are directly related to strength [Dewhurst et  al., 2015] and brittleness/ductility [Marsden et  al., 1992; Petley, 1999; Dewhurst and Hennig, 2003] in overburden shales.

The dielectric spectrum of preserved shales can be measured with appropriate instruments in the laboratory, including (i) parallel‐plate‐based methods with an imped‑ance analyzer and (ii) coaxial or end‐terminated probe measurements using a vector network analyzer [see Josh et  al., 2012; Josh, 2014 for details]. Cuttings measure‑ments can also be useful to characterize mudrock sequences [e.g., Leung and Steiger, 1992]; much useful information about mineralogy and surface activity of shales can be derived from studies of pulverized cuttings using a simple end‐terminated dielectric probe without using complex preparation steps found in previous litera‑ture [Josh, 2014]. An example of dielectric and conduc‑tivity response with frequency using the end‐loaded probe method [Josh et  al., 2012; Josh, 2014] on standard clay mineral dry powders and pastes (Fig. 1.10) shows larger differences at low frequency (kHz) than at higher fre‑quency (GHz). In the case of dry powders, small amounts of moisture are adsorbed from air, and this occurs to a larger degree for the minerals with a higher SSA. In the case of Figure 1.10 and SSA, smectite SSA > illite > kao‑linite > quartz, and as such, at high frequency, the smec‑tite contains more water than illite, kaolinite, and quartz and has a higher dielectric constant. The same result is visible for the mineral pastes, except that the illite and kaolinite seem to have similar responses. However, in this particular case, the illite used is Ward Illite, a clay mineral standard, which is around 50% illite and 50% quartz. Given illite usually has about twice the SSA of kaolinite [e.g., van Olphen, 1977] and quartz has negligible SSA, it is consistent that the results are similar for these two min‑erals in the hydrated state. Such results are interesting from a point of view of mineral typing in shales, with larger dispersion (frequency dependence) of the dielectric constant likely to relate to increased smectite content, which would also likely equate to increasing ductility of a given caprock and less likelihood of significant frac‑turing. In terms of pure minerals, the order of smectite‐illite‐kaolinite‐quartz is the order of increasing strength of those minerals as well as decreasing CEC/SSA. Hence, in addition to the trends linking shale strength and duc‑tility to water content and CEC [Dewhurst et al., 2015], there are sound mineralogical and theoretical reasons to

tie dielectric response to these parameters [see also Steiger and Leung, 1988; Leung and Steiger, 1992]. Strength and ductility are important parameters to understand in any evaluation of caprock integrity for CO2 storage.

As the dielectric response of shales is strongly corre‑lated to water content, it should come as no surprise that poor sample preservation will significantly affect dielectric results. Figure 1.11 shows the dielectric results obtained on multiple samples of the same shale kept in different RH atmospheres using a parallel plate capaci‑tance cell [Josh et  al., 2012; Josh, 2014]. This method covers a wider range of frequencies than the end‐loaded probe used in generating the results in Figure 1.10. The sample at 97% RH would have the highest water satura‑tion, and as the RH decreases, so does water saturation. The figure shows that the dielectric constant at high fre‑quency increases slightly with increasing RH as would be expected as there is more brine in the samples. However, as the frequency decreases, the difference between the degrees of dispersion is seen to significantly increase bet‑ween samples at different RH and is highest in the 97% RH sample. This is likely to do with clay‐fluid charged surface interactions increasing with increasing fluid satu‑ration allowing increasing ion mobility.

Overall, the techniques outlined above provide a brief outline of various potential shale characterization tech‑niques and show some example workflows, procedures, and results. There are a number of aspects of shale behavior that have not been covered or only lightly touched on in the above descriptions, and we will use the discussion to expand on some of these issues further.

1.7.  DISCUSSION

A number of experimental techniques for character‑izing shale caprocks for CO2 storage and containment have been discussed here. The review is intended to be in the area of the authors’ expertise and as such is not fully comprehensive. For example, considerable effort has been invested in CO2 breakthrough, permeability, and diffusion experiments on mudrocks [e.g., Hildenbrand et al., 2002, 2004; Amann‐Hildenbrand et al., 2013; Busch et al., 2008, 2009; Wollenweber et  al., 2010; Skurtveit et  al., 2012, 2015], and the reader is referred to these papers and refer‑ences therein for full details. Most of these papers show that tight mudrocks have high CO2 breakthrough pres‑sures or do not break through at all under simulated in situ pressure and temperature conditions. In addition, in general, little mineral alteration is noted in these intact siliciclastic mudrocks as any CO2 penetration is along dis‑crete flow paths which only exposes a small amount of mineral surface to flowing sCO2 [Amann‐Hildenbrand et al., 2015]. Harrington et al. [2009] measured compressibility, permeability, and its anisotropy, plus specific storage on

0004182022.INDD 18 10/16/2018 7:15:01 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 19

water‐saturated Nordland Shale from the North Sea. Permeability was at the micro‐Darcy level with an anisot‑ropy of ~2.5 to water. Nitrogen gas breakthrough pres‑sures were ~3 MPa. Modeling indicated that gas would have a greater anisotropy than water although the experi‑mental data could not be matched. However, it appeared that standard concepts of two‐phase flow through an unchanged pore system were not adequate to describe gas flow through this shale. It appeared that pressure‐induced

discrete pathways were developing rather than passive Darcy flow [e.g., Harrington and Horseman, 1999]. Skurtveit et  al. [2012, 2015] also looked at CO2 break‑through during triaxial testing of shales and noted both physical sample dilation on radial gauges during CO2 injection associated and a drop in P‐wave and S‐wave velocity in Draupne Shale from the North Sea. The velocity drops could not be accounted for by saturation changes as these were very low due to flow along discrete

0

10

20

30

40

50

60

6 7 8 9Log10 frequency (Hz)

Log10 frequency (Hz)

1

10

100

1000

10,000

6 7 8 9

0

10

20

30

40

50

60

6 7 8 9Log10 frequency (Hz)

Log10 frequency (Hz)

Illite εr′

Illite εr′

Quartz εr′

Kaolinite εr′

Rea

l rel

ativ

e di

elec

tric

per

mitt

ivity

ε′ r

Smectite εr′

Smectite εr′

1

10

100

1000

10,000

6 7 8 9

Rea

l rel

ativ

e di

elec

tric

per

mitt

ivity

ε′ r

Quartz εr′Kaolinite εr′

Figure 1.10 Dielectric response of dry powders of pure clay minerals and quartz (top) and pastes made from the same materials mixed with water (bottom). The dry powders are room dry and so contain moisture adsorbed from the air. The dielectric constant increases with decreasing frequency in both cases, the degree of dispersion related directly to the CEC of the minerals (smectite > illite > kaolinite > quartz). At high frequency, the dielectric constant is generally related to the water content alone and is much higher in the pastes than in the room dry materials. From Josh et al. [2012] and Josh [2014].

0004182022.INDD 19 10/16/2018 7:15:01 AM

20 GEOLOGICAL CARBON STORAGE

flow paths as noted above. Once CO2 flow was established, there was no change in velocity with changes in differential pressure across the sample. They noted that the CO2 breakthrough pressure depended on both confining pressure and effective stress but not the gas pressure difference induced across the sample. At the highest con‑fining pressures (~20 MPa), no CO2 breakthrough was observed. This suggested that in addition to capillary dis‑placement, there was a pressure‐induced microfracture opening to accommodate the CO2 transport through the shale and noted that the flow through the sample was con‑sistent with fracture flow models, similar to the pressure‐induced gas flow pathways noted by Harrington and Horseman [1999] and Harrington et al. [2009]. The effec‑tive CO2 permeability was more sensitive to volumetric deformation than the water permeability and fitted a power law dependency which is typical of flow in a frac‑tured medium. Skurtveit et al. [2012, 2015] also noted sim‑ilarities between the CO2 breakthrough pressure in the triaxial rig and that estimated from MIP tests, although this is not always the case [e.g., Hildenbrand et al., 2002].

One aspect only briefly discussed so far is that of anisotropy. In general, shales are thought of as trans‑versely isotropic materials, with similar rock properties within the bedding or clay foliation plane and different properties orthogonal or at angles to that plane. Strength anisotropy is a critical parameter both for drilling wells

and for fracture/fault reactivation [e.g., Fjær et al., 2008]. In general, shales are weaker at 30–60° to bedding and often weakest at 45° as compared to deformation normal to or parallel to bedding. However, although elastic deformation shows that static and dynamic properties in shales are almost always stiffer parallel to bedding [e.g., Dewhurst et  al., 2011], it is unclear in terms of failure properties which orientation is weaker. Some studies sug‑gest both orientations have similar strength [e.g., Niandou et  al., 1997; Valès et  al., 2004], while others note that parallel to bedding is in fact stronger than normal to bed‑ding in shales [Ajalloeian and Lashkaripour, 2000; Fjær and Nes, 2014]. Clennell et  al. [1999] investigated the development of anisotropic permeability in clays under‑going consolidation and found that the intrinsic anisot‑ropy is only modest when there is no compositional layering, that is, where fabric anisotropy comes from clay particle alignment alone. The most extreme cases were found in pure clays and may reach a factor of 3–4. However, in natural shales, anisotropy of permeability, and other physical properties, can be enhanced by orders of magnitude where shales contain silty laminations [e.g., Armitage et al., 2012]. Note that many studies claiming high values of permeability anisotropy in mudrocks are either looking at laminated rocks, containing silt layers, or else are obviously affected by microcracks, which close at higher confining stress. There should be some general

1

10

100

1000

1000 psi saturated

3 4 5 6 7 8

Frequency log10 (Hz)

Rel

ativ

e di

elec

tric

per

mitt

ivity

97%

52%43%

29%

11%

Dry (baked)

Figure 1.11 Dielectric response of multiple samples of the same shale re‐saturated under different relative humidity (RH) environments. The 97% RH sample has the highest water saturation, with lower water saturations commensurating with relative humidity levels. Decreased saturation as a result of poor preservation would there-fore completely change dielectric response.

0004182022.INDD 20 10/16/2018 7:15:01 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 21

correspondence between electrical/dielectric anisotropy and permeability anisotropy, as both depend on particle shape and pore space tortuosity [e.g., Clennell, 1997]. Anisotropy of P‐wave and S‐wave velocity (and other rock physics properties such as attenuation) are well doc‑umented [e.g., Banik, 1984; Johnston and Christensen, 1994, 1995; Vernik and Liu, 1997; Dewhurst and Siggins, 2006; Sarout and Guéguen, 2008; Delle Piane et al., 2011, 2014] and in experiments on shales can range up to 30% for P waves and 70% for S waves in extreme cases. However, of the petrophysical properties, electrical prop‑erties (resistivity and conductivity, including dielectric methods) usually show the most extreme anisotropy, in some cases up to a factor of 6–8 [Clavaud, 2008; Josh and Clennell, 2015]. As such, dependent on the properties required and local conditions, anisotropy should not be neglected when evaluating caprock shales for CO2 storage purposes.

A significant amount of research has looked at the chemical interaction between sCO2 and both the reservoir and caprocks. It seems that local conditions and miner‑alogy strongly govern whether significant alteration to mineralogy and pore structure does occur. Studies in siliciclastic reservoir rocks which contain natural CO2 accumulations report minor to significant mineralogical reactions in such systems. For example, Major et al. [2014] examined a natural CO2 reservoir‐seal system in sand‑stones and cemented siltstones. They noted that the reser‑voir sands that had hosted CO2 were weaker than those that had not, but the sealing siltstones showed little change in properties measured. The weakness was sug‑gested as the result of the dissolution of iron‐rich cements in the framework and their re‐precipitation as secondary carbonates and clays. Hangx et  al. [2015] noted that in general, unreacted rock strength was similar to sCO2‐sat‑urated rock strength in reservoir sandstones, except for where anhydrite dissolution had occurred. Watson et al. [2004] compared two adjacent gas fields with low and high concentrations of CO2 to study potential mineral reactions in siliciclastic reservoirs. In the high CO2 case, most of the reactive minerals have altered or dissolved, with kaolinite, quartz, and less soluble carbonates being precipitated. Haese and Watson [2014] noted that the CO2 mineral trapping potential in two siliciclastic reservoirs was low compared to volcaniclastic formations with abundant reactive minerals. Higgs et al. [2015] note alter‑ation of chlorite to kaolinite and iron‐bearing carbonates at moderate to high CO2 concentrations, but no effect where CO2 concentrations were low. While these studies were performed in reservoir rocks, it seems that in clay‐rich caprocks at least, chemical reactions tend to have much lesser effects, although they can be observed in microfabric analysis or measured through changes in pore structure and water chemistry. Liu et al. [2012] noted

minor dissolution of feldspar and anhydrite in Eau Claire shale on exposure to supercritical CO2, associated with precipitation of pore‐bridging illite and smectite, as well as minor precipitation of siderite adjacent to pyrite. Their review of the literature around caprock mineral interac‑tion with sCO2 showed that in silicate‐rich systems, reac‑tivity is low and generally confined to the reservoir‐caprock interface. Reaction rates were temperature controlled unsurprisingly, with little occurring at temperatures of 50–80°C. sCO2 can react with organic matter, potentially resulting in alteration of pore structure [Busch et  al., 2008]. Lu et  al. [2009] investigated a natural CO2 field sealed by Kimmeridge Clay and noted through geochem‑ical examination that there was ~12 m zone of interaction in the caprock, involving dissolution of carbonates, decarboxylation of organic matter, and infiltration of reservoir CO2. Garrido et al. [2013] noted that exposure to sCO2 resulted in increases in pore volume in pore sizes <10 nm in a Tournemire shale sample that contained a significant amount of calcite. This change in pore struc‑ture was attributed to calcite dissolution, although minor aluminosilicate dissolution was also indicated by pore water chemistry. Wollenweber et al. [2010] evaluated CO2 transport in a clay‐rich shale and a marl under single‐phase and multiphase flow conditions. Cyclic applica‑tions of gas pressure led to a reduction in capillary threshold pressure and decreased effective diffusion coef‑ficients. Water permeability also increased after each cycle, although there was little change in composition (by XRD), SSA (BET), or pore size distribution (MIP). Overall, it appears that mineral reactions and changes in properties due to sCO2 exposure are limited in larger intact volumes of clay‐rich shales. Extensive reviews of caprock‐CO2 interaction have been performed by Liu et al. [2012], Griffith et al. [2011], and Shukla et al. [2010].

One area of research which is not yet well understood is the adsorption of sCO2 onto clay surfaces and into clay interlayers. It has been previously noted [e.g., Busch et al., 2008] that shales of all mineralogies (i.e., with any or a mixture of the common clays such as kaolinite, illite, illite‐smectite, and smectite) can strongly adsorb sCO2, although it should be noted that many of such tests have been done on powdered samples which would significantly increase surface area able to be contacted by CO2 compared to the case should sCO2 enter a caprock along discrete pressure‐driven microfracture pathways as discussed above. While there is a vast literature on the incorporation of water into smectite interlayers, there is little published about the incorporation of sCO2 into such interlayers, either in pure dry form or containing dissolved water. Loring et al. [2014] worked on smectitic clay standards at realistic in situ pres‑sures and temperatures. They noted that sCO2 can interca‑late into the smectite interlayers but generally only if levels of water intercalation were low. Michels et al. [2015] noted

0004182022.INDD 21 10/16/2018 7:15:01 AM

22 GEOLOGICAL CARBON STORAGE

that smectitic clays can capture CO2 in the interlayer posi‑tions and that the degree of capture was dependent upon the interlayer cation type (e.g., Na+, Li+, etc.). Although the conditions they used to test the samples were very dif‑ferent to those encountered in situ. De Jong et al. [2014] noted swelling strains of up to ~2.5% in sCO2 atmospheres at 45°C and CO2 pressures up to 15 MPa in smectitic clay standards. Such swelling if occurring in situ would likely close pores and microfractures. Mechanisms of CO2 adsorption however are not well understood. Little swell‑ing is noted in dry clays, for example [e.g., Giesting et al., 2012], and dehydration can occur when there are more than two water interlayers in smectite, dependent on the interlayer cation [Schaef et al., 2012]. Jin and Firoozabadi [2014] use molecular dynamics models of montmoril‑lonite to show that the presence of water strongly reduces sCO2 adsorption. Under water‐saturated conditions and in pore sizes >2 nm, water forms the first layer on the clay surface, and CO2 forms the second through interaction between the CO2 and water molecules. When smectite has an RH representative of discrete states, so 0 (dry), 1, or 2 water layers, no or little swelling is observed. When the states are in between these discrete states, then significant swelling is measured [Busch et al., 2016].

However, it is not just smectitic clays which can adsorb CO2. Illite and kaolinite have also been observed to adsorb CO2, which therefore must be a mineral surface phenomenon rather than an interlayer one [Busch et al., 2008]. Most work on smectite swelling looks at a smear of clay paste and examines swelling through the change in d‐spacings. The de Jong et al. [2014] study used larger samples (1 mm cubes) in order to investigate a more bulk rock scale sample size to try to draw conclusions regarding bulk scale parameters such as geomechanical properties. In their tests, they noted that the amount of swelling strain was highest in the highest hydration state, although these hydration states lay between fully dehydrated and one interlayer of water. CO2 contact with a smectitic cap‑rock is likely to be limited to the reservoir‐caprock inter‑face and small volumes around preexisting fractures [de Jong et al., 2014]. Any swelling that occurred in smectites with lower amounts of interlayer water (< 1) would likely lead to the closure of pores and small fractures, resulting in lower permeability, and would thus be a self‐limiting process. However, where smectites are highly hydrated, with two or more water interlayers, dehydration may occur in contact with sCO2, which may result in shrink‑age, and this would be an issue in regard to the potential for fracture development [de Jong et al., 2014]. There is currently little data under realistic in situ conditions for such smectites. It is also unclear as yet as to how CO2 swelling would affect clay‐rich fault gouges and whether this would impact on fault stability [de Jong et al., 2014], although natural leakage of CO2 along faults in uplifted

terrains is observed [e.g., Shipton et al., 2004]. However, all these results above looking at d‐spacings or small cubes were all undertaken under unconfined conditions. When pressure‐driven sCO2 injection is performed on smectitic shales under a simulated in situ confining pressure, it appears that such shales dilate and small microfractures are generated that allow sCO2 flow through the rock [e.g., Harrington et al., 2009; Skurtveit et  al., 2012, 2015]. It may be possible that confining pressure suppresses any swelling that might occur and that the injection of sCO2 results in a pressure‐driven flow that opens dynamic and localized microscale fluid pathways [Harrington and Horseman, 1999; Hildenbrand and Krooss, 2003]. After CO2 breakthrough and establish‑ment of a steady flow, reduction in fluid pressure resulted in decreasing CO2 flow and eventually complete cessation at ~1 MPa due to capillary re‐imbibition of water.

There is a large uncertainty around wettability in clay‐bearing caprocks as few measurements have been made on shales [Iglauer et  al., 2015]. No measurements of contact angles for clay‐water‐sCO2 systems have been performed on pure clays due to their fine grain size, although Borysenko et al. [2009] made measurements on shales in brine‐hydrocarbon systems. Micas such as mus‑covite and biotite are usually used as substitutes for illite, assuming that, for example, the chemical similarity of muscovite and illite would automatically mean a simi‑larity in wettability. This has not been verified, and it is likely that illite, for example, has much higher surface charge density than muscovite and certainly the SSA and CEC of illite is much higher than that of muscovite. Amann‐Hildenbrand et  al. [2013] note that the assump‑tion of a water‐wet caprock may not always be correct and there have been studies in quartz, calcite, and micas which suggest water wetness may not be endemic [see review by Heath et al., 2012]. However, once again, this latter review found no papers investigating wettability of clay minerals in a CO2 storage context. The vastly differ‑ent surface characteristics of clays compared to the min‑erals above and their well‐known strongly hydrophilic behavior [e.g., van Olphen, 1977] should be noted, in combination with experimental breakthrough pressures and transport rates of CO2 through shales under pressure which approach molecular diffusion slowness [e.g., Amann‐Hildenbrand et al., 2013]. However, natural CO2 gas accumulations are trapped in nature over long time periods [Watson et  al., 2004; Haszeldine et  al., 2005; Major et al., 2014], suggesting that shale caprock wetta‑bility or its alteration may not be too significant on the timescale required for CO2 injection. However, long‐term exposure to organic molecules may potentially change caprock wettability [e.g., Heath et al., 2012; Iglauer et al., 2015], although bitumen coatings have been noted to preserve otherwise reactive minerals [Hangx et al., 2015].

0004182022.INDD 22 10/16/2018 7:15:01 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 23

It is fair to say that the wettability of clays is currently a known unknown in terms of shale seals for CO2 storage.

The discussion above highlights a number of uncer‑tainties around testing of shales for CO2 storage, and it is difficult to determine the relative importance and prior‑ities of such research. However, as noted in the main sec‑tion of the manuscript, there are many techniques available for shale characterization for CO2 storage, many of which have long‐term provenance in the evaluation of  hydrocarbon trapping in the oil and gas industry. It should be remembered however that although these mul‑tiple techniques are useful for application to CO2 contain‑ment integrity, experimental procedures such as these should not be used as stand‐alone techniques but should form part of a microscale to macroscale evaluation of seal integrity and containment running from seismic scale through a full structural fault and top seal analysis down to pore‐scale evaluation of capillary and flow properties. Full‐scale integration of all available measurements and data, modeling, and calibration to larger scales are critical in order to fully evaluate containment integrity for CO2 storage.

1.8.  SUMMARY

In this contribution, we have described a workflow and a number of methods for characterizing caprocks for geological CO2 storage sites. These methods are predi‑cated on high‐quality samples being obtained and pre‑served at the point of recovery, so they can be tested with minimized loss of in situ pore fluids. Dehydration of clay‐rich shales causes significant changes to their mechanical, physical, and petrophysical properties. Ideally, properties that are critical to assess include min‑eralogy, microstructure, porosity, seal capacity, perme‑ability (both water and CO2), diffusivity, electrical response, and geomechanics. Such properties can then feed into larger‐scale integrated site analysis and models of subsurface behavior. There are still significant uncer‑tainties around the interaction of supercritical CO2 and clay‐rich caprocks. These uncertainties include chemical reactivity of supercritical CO2 with caprock minerals, although in general it appears that such reactions would only occur close to the reservoir‐caprock interface and reactive surface area would be small due to the discrete flow paths taken by sCO2 should it penetrate into the cap‑rock. Swelling of smectitic clays through incorporation of sCO2 into interlayers can be significant, but in general such tests have been run under unconfined conditions, and it is not clear whether such a response would occur in situ. Uncertainty also surrounds the wettability in the shale‐brine‐sCO2 system and whether the assumption of water wetness is a good one. There are no data at all on wettability of individual clay minerals, so this remains a

fertile area for research. However, even with the uncer‑tainties around reactivity, clay swelling, and wettability, what can be stated is that there are numerous natural CO2 accumulations that have survived for hundreds of thou‑sands to millions of years, long in excess of the timescales currently being investigated for geological storage of CO2, and some of those have seals which seem consider‑ably more permeable than typical shale caprocks dis‑cussed here. As a final point, it should be noted that experimental characterization of shales such as that dis‑cussed here would be one part of a much larger strategy to evaluate containment integrity which should cover an  integrated fault and top seal methodology across a range of scales from kilometer to nanometer, integrating regional seismic, wireline log and well data with labora‑tory approaches and modeling.

ACKNOWLEDGMENTS

Much of the work carried out for this paper would not have been possible without the expertise and efforts of Bruce Maney, Stephen Firns, Shane Kager, Leigh Kiewiet, David Nguyen, Ian Penny, and Neil Sturrock. We thank Andreas Busch and an anonymous reviewer for their sterling efforts to improve the original manuscript.

REFERENCES

Ajalloeian, R. and G. R. Lashkaripour, (2000), Strength anisot‑ropies in mudrocks, Bulletin of Engineering Geology and the Environment, 59, 195–199.

Al Hinai, A., R. Rezaee, L. Esteban, and M. Labani, (2014), Comparisons of pore size distribution: a case from the west‑ern Australian gas shale formations, J. Unconv. Oil Gas Res., 8, 1–13. DOI:10.1016/j.juogr.2014.06.002.

Amann‐Hildenbrand, A., P. Bertier, A. Busch, and B. M. Krooss, (2013), Experimental investigation of the sealing capacity of generic clay‐rich caprocks, International Journal of Greenhouse Gas Control, 19, 620–641. DOI:10.1016/j.ijggc.2012.12.040.

Amann‐Hildenbrand, A., B. M. Krooss, P. Bertier, and A.  Busch, (2015), Laboratory testing procedure for CO2 capillary entry pressures on caprocks. In Carbon Dioxide Capture for Storage in Deep Geological Formations, edited by Gerdes, K.F., vol. 4, pp. 355–384, CPL Press and BPCNAI.

Angeli, M., M. Soldal, E. Skurtveit, and E. Aker, (2009), Experimental percolation of supercritical CO2 through a cap‑rock, Energy Procedia, 1, 3351–3358. DOI:10.1016/j.egypro. 2009.02.123.

Aplin, A.C., Y. Yang, and S. Hansen, (1995), Assessment of β, the compression coefficient of mudstones and its relationship with detailed lithology, Marine and Petroleum Geology, 12, 955–963. DOI:10.1016/0264‐8172(95)98858‐3.

Armitage, P.J. D.R. Faulkner, R.H. Worden, A.C. Aplin, A.R. Butcher, and J. Iliffe, (2012), Experimental measurement of, and controls on, permeability and permeability anisotropy of

0004182022.INDD 23 10/16/2018 7:15:02 AM

24 GEOLOGICAL CARBON STORAGE

caprocks from the CO2 storage project at the Krechba Field, Algeria, Journal of Geophysical Research, 116, B12208. DOI:10.1029/2011JB008385.

Arns, H., H. Averdunk, F. Bauget, A. Sakellariou, T. J. Senden, A. P. Sheppard, R. M. Sok, W. V. Pinczewski, and M. A. Knackstedt, (2004), Digital core laboratory: reservoir core analysis from 3D images, In: Gulf Rocks 2004, the 6th North America Rock Mechanics Symposium (NARMS), ARMA Paper 04–498, Houston, Texas.

Arts, R., A. Chadwick, O. Eiken, S. Thibeau, and S. L. Nooner, (2008), Ten Years’ experience of monitoring CO2 injection in the Utsira sand at Sleipner, Offshore Norway, First Break, 26, 65–72.

Banik, N.C., (1984), Velocity anisotropy of shales and depth estimation in the North Sea basin, Geophysics, 49, 1411–19. DOI:10.1190/1.1441770.

Barton, C. A., M. D. Zoback, and D. Moos, (1995), Fluid flow along potentially active faults in crystalline rock, Geology, 23, 683–686. DOI:10.1130/0091‐7613.

Bennett, R. H., W. R. Bryant, and M. H. Hulbert, (1991a), Microstructures of Fine Grained Sediments, From Mud to Shale, Springer, New York, 582 pp. DOI:10.1007/978‐1‐4612‐4428‐8.

Bennett, R. H., N. R. O’Brien, and M. H. Hulbert, (1991b), Determinants of clay and shale microfabric signatures: processes and mechanisms. In Microstructures of Fine Grained Sediments, edited by Bennett, R.H., W. R. Bryant, and M. H. Hulbert, pp. 5–32, Springer‐Verlag, New York.

Berrezueta, E., L. González‐Menéndez, D. Breitner, and L. Luquot, (2013), Pore system changes during experimental CO2 injection into Detritic rocks: studies of potential storage rocks from some sedimentary basins of Spain, International Journal of Greenhouse Gas Control, 17, 411–22. DOI:10.1016/j.ijggc.2013.05.023.

Borysenko, A., B. Clennell, R. Sedev, I. Burgar, J. Ralston, M. Raven, D. Dewhurst, and K. Liu, (2009), Experimental inves‑tigations of the wettability of clays and shales, J. Geophys. Res., 114, B07202. DOI:10.1029/2008JB005928.

Brace, W. F., J. B. Walsh, and W. T. Frangos, (1968), Permeability of granite under high pressure, J. Geophys. Res., 73, 2225–2236. DOI:10.1029/JB073i006p02225.

Brindley, G.W., (1952), Identification of clay minerals by X‐ray diffraction analysis, Clays and Clay Minerals, 1(1), 119–129. DOI:10.1346/CCMN.1952.0010116.

Brown, K. M., and B. Ransom, (1996), Porosity corrections for smectite‐rich sediments: impact on studies of compaction, fluid generation, and tectonic history, Geology, 24, 843–846. DOI:10.1130/0091‐7613(1996)024.

Busch, A. and A. Amann‐Hildenbrand, (2013), Predicting cap‑illarity of mudrocks, Marine and Petroleum Geology, 45, 208–223. DOI:10.1016/j.marpetgeo.2013.05.005.

Busch, A., P. Bertier, Y. Gensterblum, G. Rother, C. J. Spiers, M.  Zhang, and H. M. Wentinck, (2016), On sorption and swelling of CO

2 in clays, Geomechanics and Geophysics for Geo‐Energy and Geo‐Resources, 2(2), 111–130. DOI:10.1007/s40948‐016‐0024‐4.

Busch, A., N. Kampman, S. J. Hangx, J. Snippe, M. Bickle, P.  Bertier, H. Chapman, C. J. Spiers, R. Pijnenburg, J. Samuelson, J. P. Evans, A. Maskell, J. Nicholl, V. Pipich, Z. Di, G. Rother, and M. Schaller, (2014), The Green River

natural analogue as a field laboratory to study the Long‐term fate of CO2 in the subsurface, Energy Procedia, 63, 2821–2830. DOI:10.1016/j.egypro.2014.11.304.

Busch, A., S. Alles, B. M. Krooss, H. Stanjek, and D. Dewhurst, (2009), Effects of physical sorption and chemical reactions of  CO2 in shaly caprocks, Energy Procedia, 1, 3229–3235. DOI:10.1016/j.egypro.2009.02.107.

Busch, A., S. Alles, Y. Gensterblum, D. Prinz, D. N. Dewhurst, M. D. Raven, and B. M. Krooss, (2008), Carbon dioxide storage potential of shales, International Journal of Greenhouse Gas Control, 2, 297–308. DOI:10.1016/j.ijggc.2008.03.003.

Chalmers, G. R., R. M. Bustin, and I. M. Power, (2012), Characterization of gas shale pore systems by porosimetry, pycnometry, surface area, and field emission scanning elec‑tron microscopy/ transmission electron microscopy image analyses: examples from the Barnett, Woodford, Haynesville, Marcellus, and Doig units, AAPG Bulletin, 96, 1099–1119. DOI:10.1306/10171111052

Chiou, W. A., W. R. Bryant, and R. H. Bennett, (1991), Clay fabric of gassy submarine sediments. In Microstructures of Fine Grained Sediments, edited by Bennett, R. H., W. R. Bryant, and M. H. Hulbert, pp. 333–52, Springer‐Verlag, New York.

Chiquet, P., D. Broseta, and S. Thibeau, (2007), Wettability alteration of caprock minerals by carbon dioxide, Geofluids, 7, 1–11. DOI:10.1111/j.1468‐8123.2007.00168.x.

Clarkson, C. R., J. L. Jensen, and S. Chipperfield, (2012), Unconventional gas reservoir evaluation: what do we have to consider, Journal of Natural Gas Science and Engineering, 8, 9–33. DOI:10.1016/j.jngse.2012.01.001.

Clarkson, C. R., N. Solano, R. M. Bustin, A. M. M. Bustin, G. R. L. Chalmers, L. He, Y. B. Melnichenko, A. P. Radlinski, and T. P. Blach, (2013), Pore structure characterization of north American shale gas reservoirs using USANS/SANS, gas adsorption, and mercury intrusion, Fuel, 103, 606–616. DOI:10.1016/j.fuel.2012.06.119.

Clavaud, J.‐B., (2008), Intrinsic electrical anisotropy of shales: the effect of compaction, Petrophysics, 49, 243–260.

Clennell, M. B., (1997), Tortuosity: a guide through the maze. In Development in Petrophysics, edited by Lovell, M. A. and P. K. Harvey, Geological Society London Special Publications, 122, 299–344. 10.1144/GSL.SP.1997.122.01.18.

Clennell, M. B., D. N. Dewhurst, and M. Raven, (2006), Shale petrophysics: electrical, dielectric and nuclear magnetic reso‑nance studies of shales and clays. Proceedings of the 47th SPWLA Annual Logging Symposium, Veracruz, Mexico, 4–7 June, 13 pp.

Clennell, M. B., D. N. Dewhurst, K. M. Brown, and G. K. Westbrook, (1999), Permeability anisotropy of consolidated clays. In Muds and Mudstones: Physical and Fluid Flow Properties, edited by Aplin, A. C., A. J. Fleet, and J. H. S. Macquaker, Geological Society of London Special Publication, 158, pp. 79–96. 10.1144/GSL.SP.1999.158.01.07.

Coates, G. R., R. C. A. Peveraro, and D. Roberts, (1991), The magnetic resonance imaging log characterized by comparison with petrophysical properties and laboratory core data. SPE Paper 22723, 66th Annual Technical Conference and Exhibition of the SPE, Dallas, Texas, 6–9 October.

0004182022.INDD 24 10/16/2018 7:15:02 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 25

de Jong, S. M., C. J. Spiers, and A. Busch, (2014), Development of swelling strain in smectite clays through exposure to carbon dioxide, International Journal of Greenhouse Gas Control, 24, 149–161. DOI:10.1016/j.ijggc.2014.03.010.

Delage, P., D. Tessier, and M. Marcel‐Audiguier, (1982), Use  of  the Cryoscan apparatus for observation of freeze‐fractured planes of a sensitive Quebec clay in scanning electron microscopy, Canadian Geotechnical Journal, 19, 111–114.

Delle Piane, C., B. S. G. Almqvist, C. M. MacRae, A. Torpy, A.  J. Mory, and D. N. Dewhurst, (2015), Texture and diagenesis of ordovician shale from the canning basin, western Australia: implications for elastic anisotropy and geomechanical properties, Marine and Petroleum Geology, 59, 56–71. DOI:10.1016/j.marpetgeo.2014.07.017.

Delle Piane, C., C. Madonna, J. Sarout, E.H. Saenger, D.N. Dewhurst, and M.D. Raven, (2014), Frequency dependent seismic attenuation in shales: experimental results and theo‑retical analysis, Geophysical Journal International, 1–12. DOI:10.1093/gji/ggu148.

Delle Piane, C., D. N. Dewhurst, A. F. Siggins, and M. D. Raven, (2011), Stress‐induced anisotropy in saturated shale, Geophysical Journal International, 184, 897–906. DOI:10.1111/ j.1365‐246X.2010.04885.x, 10 pp.

Desbois, G., J. L. Urai, and P. A. Kukla, (2009), Morphology of the pore space in claystones  –  evidence from BIB/FIB ion beam sectioning and cryo‐SEM observations, Earth Discussions, 4, 1–19. DOI:10.5194/eed‐4‐1‐2009.

Dewhurst, D. N. and A. Hennig, (2003), Geomechanical prop‑erties related to top seal leakage in the Carnarvon Basin, Northwest Shelf, Australia, Petroleum Geoscience, 9, 255–263. DOI:10.1144/1354‐079302‐557.

Dewhurst, D. N. and A. F. Siggins, (2006), Impact of fabric, microcracks and stress field on shale anisotropy, Geophysical Journal International, 165, 135–148. DOI:10.1111/j.1365‐ 246X.2006.02834.x

Dewhurst, D. N., A. C. Aplin, J.‐P. Sarda and Y. Yang, (1998), Compaction‐driven evolution of porosity and permeability in natural mudstones: an experimental study, Journal of Geophysical Research, Solid Earth, 103(B1), 651–661. DOI:10.1029/97JB02540.

Dewhurst, D. N., Y. Yang, and A. C. Aplin, (1999a), Fluid flow through natural mudstones. In Muds and Mudstones: Physical and Fluid Flow Properties, edited by Aplin, A. C., A. J. Fleet, and J. H. S. Macquaker, Geological Society of London Special Publication, 158, pp. 23–43. DOI:10.1144/GSL.SP.1999.158.01.03.

Dewhurst, D. N., A. C. Aplin, and J.‐P. Sarda, (1999b), Influence of clay fraction on pore‐scale properties and hydraulic con‑ductivity of experimentally compacted mudstones, Journal of Geophysical Research, Solid Earth, 104(B12), 29261–29274. DOI:10.1029/1999JB900276.

Dewhurst, D. N., P. J. Boult, R. M. Jones, and S. A. Barclay, (2005), Fault healing and fault sealing in impure sandstones. In Evaluating Fault and Cap Rock Seals, edited by Boult, P. J. and J. Kaldi, AAPG Hedberg Series, no. 2, pp. 37–56. DOI:10.1306/1060755H23161.

Dewhurst, D. N., R. M. Jones, and M. D. Raven, (2002), Microstructural and petrophysical characterisation of

Muderong shale: application to top seal risking, Petroleum Geoscience, 8, 371–383. DOI:10.1144/petgeo.8.4.371.

Dewhurst, D. N., B. Maney, B. Clennell, C. Delle Piane, C. Madonna, E. Saenger, and N. Tisato, (2012), Impact of satu‑ration change on shale properties. Extended abstract for 3rd EAGE Shale Conference, Barcelona, 4 pp. DOI:10.3997/2214‐4609.20143919.

Dewhurst, D. N., J. Sarout, C. Delle Piane, A. F. Siggins, and M. D. Raven, (2015), Empirical strength prediction for preserved shales, Marine and Petroleum Geology, 67, 512–525. DOI:10.1016/j.marpetgeo.2015.06.004.

Dewhurst, D. N., A.F. Siggins, J. Sarout, M. Raven, and H. M. Nordgard‐Bolas, (2011), Geomechanical and ultrasonic characterisation of a Norwegian Sea shale, Geophysics, 76(3), WA101–WA111, DOI:10.1190/1.3569599.

Diamond, S., (1970), Pore size distribution in clays, Clays and Clay Minerals, 18, 7–23. DOI:10.1346/CCMN.1970.0180103.

Doughty, C., B. Freifeld, and R. Trautz, (2008), Site character‑ization for CO

2 geologic storage and vice versa: the Frio brine pilot, Texas USA as a case study, Environ. Geol., 54, 1635–1656. DOI:10.1007/s00254‐007‐0942‐0.

Dunn, K. J., D. J. Bergmann, and G. A. Latorraca, (2002), Nuclear Magnetic Resonance; Petro‐Physical and Logging Applications. Handbook of Geophysical Exploration, Amsterdam, Pergamon.

Espinoza, D. N., and J. C. Santamarina, (2012), Clay interac‑tion with liquid and supercritical CO2: the relevance of electrical and capillary forces, International Journal of Greenhouse Gas Control, 10, 351–362. DOI:10.1016/j.ijggc.2012.06.020.

Esteban, L., J. L. Bouchez, and A. Trouiller, (2006), The Callovo‐Oxfordian argillites from eastern Paris basin: magnetic data and petrofabrics, Comptes Rendus Geosciences, 338, 867–881. 10.1016/j.crte.2006.03.011.

Esteban, L., Y. Géraud, and J. L. Bouchez, (2007), Pore network connectivity anisotropy in Jurassic argillite specimens from eastern Paris Basin (France), Physics and Chemistry of the Earth, 32, 161–169. DOI:10.1016/j.pce.2005.11.001.

Ewy, R. T., (2015), Shale/claystone response to air and liquid exposure, and implications for handling, sampling and test‑ing, International Journal of Rock Mechanics and Mining Sciences, 80, 388–401. DOI:10.1016/j.ijrmms.2015.10.009.

Ewy, R. T., (2014), Shale swelling/shrinkage and water content change due to imposed suction and due to direct brine contact, Acta Geotechnica, 9, 869–886. DOI:10.1007/s11440‐013‐0297‐5.

Ewy, R. T., C. A. Bovberg, and R. J. Stankovich, (2008), Shale Triaxial strength alteration due to brine exposure. Proceedings of the 42nd US Rock Mechanics Symposium, ARMA 08‐304, San Francisco, CA, 29 June–2 July, 9 pp.

Fisher, Q. J. and R. J. Knipe, (1998), Fault sealing processes in siliciclastic sediments. In Faulting, Fault Sealing and Fluid Flow in Hydrocarbon Reservoirs, edited by Jones, G., Q. J. Fisher, and R. J. Knipe, Geological Society of London, Special Publication, 147, pp. 117–134. DOI:10.1144/GSL.SP.1998.147.01.08.

Fjær, E. and O.‐M. Nes, (2014), The impact of heterogeneity on the anisotropic strength of an outcrop shale, Rock Mechanics and Rock Engineering, 47, 1603–1611. DOI:10.1007/ s00603‐ 014‐0598‐5.

0004182022.INDD 25 10/16/2018 7:15:02 AM

26 GEOLOGICAL CARBON STORAGE

Fjær, E., R. M. Holt, P. Horsrud, A. M. Raaen, and R. Risnes, (2008), Petroleum Related Rock Mechanics. Developments in Petroleum Science, 33, Elsevier, Amsterdam, 338 pp.

Garrido, D. R. R., S. Lafortunea, H. Soulib, and P. Dubujet, (2013), Impact of supercritical CO2/water interaction on the caprock nanoporous structure, Procedia Earth and Planetary Science, 7, 738–741. DOI:10.1016/j.proeps.2013.03.123.

Giesting, P., S. Guggenheim, A. F. K. van Groos, A. Busch, (2012), Interaction of carbon dioxide with Na‐exchanged montmorillonite at pressures to 640 bars: implications for CO2 sequestration, International Journal of Greenhouse Gas Control, 8, 73–81. DOI:10.1016/ j.ijggc.2012.01.011.

Golab, A. R. Romeyn, H. Averdunk, M. Knackstedt, and T. J. Senden, (2012), 3D characterisation of potential CO2 reser‑voir and seal rocks, Australian Journal of Earth Sciences, 59, 1–13. DOI:10.1080/08120099.2012.675889.

Griffith, C., D. Dzombak, and G. Lowry, (2011), Physical and chemical characteristics of potential seal strata in regions con‑sidered for demonstrating geological saline CO2 sequestration, Environmental Earth Sciences, 64, 925–948. DOI:10.1007/s12665‐011‐0911‐5.

Griffiths, F. J. and R. C. Joshi, (1989), Change in pore size dis‑tribution due to consolidation of clays, Geotechnique, 39, 159–167.

Griffiths, F. J. and R. C. Joshi, (1990), Clay fabric response to consolidation, Applied Clay Science, 5, 37–66. DOI:10.1016/ 0169‐1317(90)90005‐A.

Gu, X., D.R. Cole, G. Rother, D. F. R. Mildner, and S. L. Brantley, (2015), Pores in Marcellus shale: A neutron scattering and FIB‐SEM study, Energy and Fuels, 29, 1295–1308. DOI:10.1021/acs.energyfuels.5b00033.

Haese, R. R. and M. Watson, (2014), Comparison of the min‑eral trapping capacity in three reservoirs with variable min‑eral compositions under CO2 saturated conditions, Energy Procedia, 63, 5479–5482. DOI:10.1016/j.egypro.2014.11.579

Haest, M., T. Cudahy, C. Laukamp, and S. Gregory, (2012), Quantitative mineralogy from infrared spectroscopic data. I. Validation of mineral abundance and composition scripts at the Rocklea channel iron deposit in Western Australia, Economic Geology, 107, 209–228. DOI:10.2113/econgeo. 107.2.209.

Hall, P. L., D. F. R. Mildner, and R. L. Borst, (1986), Small‐angle scattering studies of the pore spaces of Shaly rocks,  Journal of Geophysical Research, 91(B2), 2183–2192. DOI:10.1029/ JB091iB02p02183.

Hangx, S. J. T., C. J. Spiers, and C. J. Peach, (2010), Mechanical behaviour of anhydrite caprock and implications for CO2  sealing capacity, J Geophys Res., 115, B07402. DOI:10.1029/2009JB006954

Hangx, S., E. Bakker, P. Bertier, G. Nover, and A. Busch, (2015), Chemical‐mechanical coupling observed for depleted oil res‑ervoirs subjected to long‐term CO2 exposure – a case study of the Werkendam natural CO2 analogue field, Earth and Planetary Science Letters, 428, 230–242. DOI:10.1016/j.epsl.2015.07.044.

Harrington, J. F., D. J. Noy, S. T. Horseman, D. J. Birchall, and R. A. Chadwick, (2009), Laboratory study of gas and water flow in the Nordland Shale, Sleipner, North Sea. In Carbon Dioxide Sequestration in Geological Media  –  State of the

Science, edited by Grobe, M, J. C. Pashin, and R. L. Dodge, AAPG Studies in Geology, 59, pp. 521–543. DOI:10. 1306/13171259St593394.

Harrington, J. F. and S. T. Horseman, (1999), Gas transport properties of clays and mudrocks. In Muds and Mudrocks: Physical and Fluid‐Flow Properties, edited by Aplin, A. C., A. J. Fleet, and J. H. S. Macquaker, Geological Society of London Special Publications, 158, pp. 107–124. DOI:10.1144/GSL.SP.1999.158.01.09.

Haszeldine, R. S., O. Quinn, G. England, M. Wilkinson, Z. K. Shipton, J. P. Evans, J. Heath, L. Crossey, C. J. Ballentine, and C. M. Graham, (2005), Natural geochemical analogues for carbon dioxide storage in deep geological porous reservoirs, a  United Kingdom perspective, Oil & Gas Science and Technology, 60, 33–49. DOI:10.2516/ogst:2005004.

Head, K. H., (1980), Manual of Soil Laboratory Testing. Volume 1: Soil Classification and Compaction Tests. Pentech Press Ltd, Plymouth, 339 pp.

Heath, J. E., T. A. Dewers, B. J. O. L. McPherson, M. B. Nemer, and P. G. Kotula, (2012), Pore‐lining phases and capillary breakthrough pressure of mudstone caprocks: sealing efficiency of geologic CO

2 storage sites, International Journal of Greenhouse Gas Control, 11, 204–220. DOI:10.1016/j.ijggc.2012.08.001.

Heath, J. E., T. A. Dewers, B. J. O. L. McPherson, R. Petrusak, T. C. Chidsey, A. J. Rinehart, and P. S. Mozley, (2011), Pore networks in continental and marine mudstones: characteris‑tics and controls on sealing behavior, Geosphere, 7, 429–454. DOI:10.1130/GES00619.1

Heller, R., J. Vernylen, and M. Zoback, (2014), Experimental investigation of matrix permeability of gas shales, AAPG Bull., 98(5), 975–995. DOI:10.1306/09231313023.

Hidajat, I., M. Singh, and K. K. Mohanty, (2003), NMR response of porous media by random walk algorithm: a parallel implementation, Chem. Eng. Commun. 190, 1661–1680. DOI:10.1080/00986440390225233.

Higgs, K. E., R. Haese, S. D. Golding, U. Schacht, and M. N. Watson, (2015), The Pretty Hill Formation as a natural ana‑logue for CO2 storage; An investigation of mineralogical and isotopic changes associated with sandstones exposed to low, intermediate and high CO2 concentrations over geological time, Chemical Geology, 399, 36–64. DOI:10.1016/ j.chem geo.2014.10.019.

Hildenbrand, A. and B. Krooss, (2003), CO2 migration processes in argillaceous rocks: pressure‐driven volume flow and diffu‑sion, Journal of Geochemical Exploration, 78, 169–172. DOI:10.1016/S0375‐6742(03)00077‐3.

Hildenbrand, A., and J. L. Urai, (2003), Investigation of the morphology of pore space in mudstones  –  first results, Marine and Petroleum Geology, 20, 1185–1200. DOI:10.1016/ j.marpetgeo.2003.07.001.

Hildenbrand, A., S. Schlömer, B. Krooss, and R. Littke, (2004), Gas breakthrough experiments on pelitic rocks: comparative study with N2, CO2 and CH4, Geofluids, 4, 61–80. DOI:10.1111/ j.1468‐8123.2004.00073.x.

Hildenbrand, A., S. Schlömer, and B. M. Krooss, (2002), Gas breakthrough experiments on fine‐grained sedimentary rocks, Geofluids, 2, 3–23. DOI:10.1046/j.1468‐8123.2002. 00031.x.

0004182022.INDD 26 10/16/2018 7:15:02 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 27

Holzer, L., F. Indutnyi, P. Gasser, B. Münch, and M. Wegmann, (2004), Three‐dimensional analysis of porous BaTiO3 ceramics using FIB nanotomography, Journal of Microscopy, 216, 84–95. DOI:10.1111/j.0022‐2720.2004.01397.x.

Holzer, L., B. Münch, M. Rizzi, R. Wepf, P. Marschall, and T.  Graule, (2010), 3D microstructure analysis of hydrated bentonite with cryo‐stabilized pore water, Applied Clay Science, 47, 330–342. DOI:10.1016/j.clay.2009.11.045.

Horsrud, P., E. F. Sønstebø, and R. Bøe, (1998), Mechanical and petrophysical properties of North Sea shales, Int. J. Rock Mech. Min. Sci., 35, 1009–1020. DOI:10.1016/S0148‐9062 (98)00162‐4.

Howard, J. J., (1991), Porosimetry measurement of shale fabric  and its relationship to illite/smectite diagenesis, Clays and Clay Minerals, 39, 355–361. DOI:10.1346/CCMN. 1991.0390403

Iglauer, S., C. H. Pentland, and A. Busch, (2015), CO2 wetta‑bility of seal and reservoir rocks and the implications for carbon geo‐sequestration, Water Resources Research, 50, 1–46. DOI:10.1002/2014 WR015553.

Ingram, G. M. and J. L. Urai, (1999), Top seal leakage through faults and fractures: the role of mudstone properties. In Muds and Mudstones: Physical and Fluid Flow Properties, edited by Aplin, A. C., A. J. Fleet, and J. H. S. Macquaker, Geological Society of London Special Publication, 158, pp. 125–135. DOI:10.1144/GSL.SP.1999.158.01.10

IPCC, (2005), Special Report on Carbon Dioxide Capture and Storage: Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge and New York, 442 pp.

Jin, Z. and A. Firoozabadi, (2014), Effect of water on methane and carbon dioxide sorption in clay minerals by Monte Carlo simulations, Fluid Phase Equilibria, 382, 10–20. DOI:10. 1016/j.fluid.2014.07.035.

Johnston, J. E. and N. I. Christensen, (1994), Elastic constants and velocity surfaces of indurated anisotropic shales, Surveys in Geophysics, 15, 481–494. DOI:10.1007/BF00 690171.

Johnston, J. E. and N. I. Christensen, (1995), Seismic anisot‑ropy of shales, Journal of Geophysical Research, 100, 5991–6003. DOI:10.1029/95JB00031.

Josh, M., (2014), Dielectric permittivity: A petrophysical parameter for shales, Petrophysics, 55, 319–332.

Josh, M. and B. Clennell, (2015), Broadband electrical prop‑erties of clays and shales: comparative investigations of remolded and preserved samples, Geophysics, 80, D129–D143. DOI:10.1190/GEO2013‐0458.1.

Josh, M., L. Esteban, C. Delle Piane, J. Sarout, D.N. Dewhurst, and M.B. Clennell, (2012), Laboratory characterisation of shale properties, Journal of Petroleum Science and Engineering, 88–89, 107–124. DOI:10.1016/ j.petrol.2012.01.023.

Kaszuba, J. P., D. R. Janecky, and M. G. Snow, (2005), Experimental evaluation of mixed fluid reactions between supercritical carbon dioxide and NaCl brine: relevance to  the integrity of a geologic carbon repository, Chemical  Geology, 217, 277–293. DOI:10.1016/j.chemgeo. 2004.12.014.

Katz, A. J. and Thompson, A. H., (1986), Quantitative predic‑tion of permeability in porous rock, Physical Reviews, B34, 8179–8181. DOI:10.1103/PhysRevB.34.8179.

Keller, L. M., L. Holzer, P. Schuetz, and P. Gasser, (2013), Pore space relevant for gas permeability in Opalinus clay: statistical analysis of homogeneity, percolation, and representative volume element, Journal of Geophysical Research Solid Earth, 118, 2799–2812. DOI:10.1002/jgrb.50228.

Keller, L. M., L. Holzer, R. Wepf, and P. Gasser, (2011), 3D geometry and topology of pore pathways in Opalinus clay: implications for mass transport, Applied Clay Science, 52, 85–95. DOI:10.1016/j.clay.2011.02.003.

King, H. E., A. P. R. Eberle, C. C. Walters, C. E. Kliewer, D. Ertas, and C. Huynh, (2015), Pore architecture and connec‑tivity in gas shale, Energy and Fuels, 29, 1375–1390. DOI:10.1021/ef502402e.

Krooss, B. M, F. van Bergen, Y. Gensterblum, N. Siemons, H. J. M. Pagnier, and P. David, (2002), High‐pressure methane and carbon dioxide adsorption on dry and moisture‐equilibrated Pennsylvanian coals, International Journal of Coal Geology, 51, 69–92. DOI:10.1016/ S0166‐5162(02)00078‐2.

Kuila, U., (2013), Measurement and interpretation of porosity and pore‐size distribution in mudrocks, PhD Thesis, Colorado School of Mines, 269 pp.

Kwon, O., A. K. Kronenberg, A. F. Gangi, B. Johnson, and B.  E.  Herbert, (2004), Permeability of illite‐bearing shale: 1.  Anisotropy and effects of clay content and loading, Journal of Geophysical Research, 109, B10205. DOI:10.1029/ 2004JB003052.

Lash, G., (2006), Top seal development in the shale‐dominated upper Devonian Catskill Delta complex, western New York State, Marine and Petroleum Geology, 23(3), 317–335. DOI:10.1016/j.marpetgeo.2006.02.001.

Leung, P. K. and R.P. Steiger, (1992), Dielectric constant mea‑surements: a new, rapid method to characterize shale at the wellsite. IADC/SPE Drilling Conference, Paper IADC/SPE 23887, pp. 401–408. DOI:10.2118/23887‐MS.

Lindsay, N. G., F. C. Murphy, J. J. Walsh, and J. Watterson, (1993), Outcrop studies of shale smears on fault surfaces. In The Geological Modelling of Hydrocarbon Reservoirs and Outcrop Analogues, edited by Flint, S. S. and I. D. Bryant, International Association of Sedimentologists Special Publication, 15, 113–123. DOI:10.1002/9781444303957.ch6.

Liu, F., P. Lua, C. Griffith, S. W. Hedges, Y. Soong, H. Hellevang, and C. Zhu, (2012), CO

2–brine–caprock interaction: reactivity experiments on Eau Claire shale and a review of relevant literature, International Journal of Greenhouse Gas Control, 7, 153–167. DOI:10.1016/ j.ijggc.2012.01.012.

Loring, J. S., E.S. Ilton, J. Chen, C. J. Thompson, P. F. Martin, P. Bénézeth, K. M. Rosso, A. R. Felmy, and H. T. Schaef, (2014), In situ study of CO2 and H2O partitioning between Na−montmorillonite and variably wet supercritical carbon dioxide, Langmuir, 30, 6120–6128, DOI:10.1021/la500682t.

Loucks, R. G., R. M. Reed, S. C. Ruppel, and U. Hammes, (2012), Spectrum of pore types and networks in mudrocks and a descriptive classification for matrix‐related mudrock pores, AAPG Bulletin, 96, 1071–1098. DOI:10.1306/08171111061.

Loucks, R. G., R. M. Reed, S. C. Ruppel, and D. M. Jarvie, (2009), Morphology, genesis, and distribution of nanometer‐scale pores in siliceous mudstones of the Mississippian Barnett shale, Journal of Sedimentary Research, 79, 848–861. DOI:10.2110/jsr.2009.092.

0004182022.INDD 27 10/16/2018 7:15:02 AM

28 GEOLOGICAL CARBON STORAGE

Lu, J., M. Wilkinson, R. S. Haszeldine, and A. E. Fallick, (2009), Long‐term performance of a mudrock seal in natural CO2 storage, Geology, 37, 35–38. DOI:10.1130/G25412A.1.

MacRae, C. M., N. C. Wilson, A. Torpy, S. Kidder, Z. Li, C. Delle Piane, and D. N. Dewhurst, (2014), Quartz overgrowths in shales and sandstones studied by EPMA and SIMS. Microscopy and Microanalysis, 20, 914–915. DOI:10.1017/S1431927614006291.

Major, J. R., P. Eichhubl, T. A. Dewers, A. S. Urquhart, J. E. Olson, and J. Holder, (2014), The effect of CO2‐related dia‑genesis on geomechanical failure parameters: fracture testing of CO2‐altered reservoir and seal rocks from a natural analog at crystal geyser, Utah. Proceedings of the 48th US Rock Mechanics/Geomechanics Symposium, Minneapolis, MN, 1–4 June, ARMA 14‐7463, 5 pp.

Marsden, J. R., R. M. Holt, S. J. Nakken, and A. M. Raaen, (1992), Mechanical and petrophysical characterisation of highly stressed mudstone. In Rock Characterisation, edited by Hudson, J. R.. Proceedings of the ISRM Conference, Eurock, 92, pp. 51–56. DOI:10.1680/rc.35621.0010.

Martinez, G. A. and L. A. Davis, (2000), Petrophysical mea‑surements on shales using NMR. Paper presented at SPE/AAPG Western Regional Meeting, SPE Paper number 62851‐MS, Long Beach, CA. DOI:10.2118/62851‐MS.

Michels, L., J. O. Fossum, Z. Rozynek, H. Hemmen, K. Rustenberg, P. A. Sobas, G. N. Kalantzopoulos, K. D. Knudsen, M. Janek, T. S. Plivelic, and G. J. da Silva, (2015), Intercalation and retention of carbon dioxide in a Smectite clay promoted by interlayer cations. Scientific Reports, 5, 8775. DOI:10.1038/srep08775.

Mildner, D. F. R., R. Rezvani, P. L. Hall, and R. L. Borst, (1986), Small‐angle scattering of shaley rocks with fractal pore interfaces, Applied Physics Letters, 48, 1314–1316. DOI:10.1063/1.96964

Mildren, S. D., R. R. Hillis, D. N. Dewhurst, P. J. Lyon, J. J. Meyer, and P. J. Boult, (2005), FAST: a new technique for Geomechanical assessment of the risk of reactivation‐related breach of fault seals. In Evaluating Fault and Cap Rock Seals, edited by Boult, P. and J. Kaldi, AAPG Hedberg Series, 2, pp. 73–85. DOI:10.1306/1060757H23163.

Milliken, K. and L. S. Land, (1994), Evidence of fluid flow in microfractures in geopressured shales: discussion, AAPG Bulletin, 78, 1637–1640.

Minh, C. C. and P. Sundararaman, (2006), NMR petrophysics in thin sand/shale laminations. SPE Paper 102435, 81st SPE Annual Technical Conference and Exhibition, San Antonio, TX. DOI:10.2118/102435‐MS.

Moore, D. M. and R. C. Reynolds, (1997), X‐Ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford: Oxford University Press, 378 pp.

Morris, A., D. A. Ferrill, and D. B. Henderson, (1996), Slip ten‑dency analysis and fault reactivation, Geology, 24, 275–278. DOI:10.1130/0091‐7613(1996)024.

Myers, M. T., (1991), A Saturation Model for the Dielectric Constant of Shaly Sands. Paper 9118, Preprints, SCA Annual Conference, Dallas, TX, 21–22 August.

Neuzil, C. E., (1994), How permeable are clays and shales? Water Resources Research, 30, 145–150. DOI:10.1029/93WR02930.

Niandou, H., J. F. Shao, J. P. Henry, and D. Fourmaintreaux, (1997), Laboratory investigation of the mechanical behav‑iour of Tournemire shale, International Journal of Rock Mechanics and Mining Sciences, 34, 3–16. DOI:10.1016/S0148‐9062(96)00053‐8.

Nicot, B., S. Gautier, M. Fleury, and S. Durucan, (2006), Pore Structure Analysis of Coals Using Low Field NMR Measurements and Thermogravimetric Analysis. SCA paper 2006‐27, International Symposium of the Society of Core Analysts, Trondheim, Norway, 12 pp.

North, A. N. and J. C. Dore, (1993), Characterisation of porous solids using small‐angle x‐ray and neutron scattering. In Multifunctional Mesoporous Inorganic Solids, edited by Sequeira, C. A. C. and M. J. Hudson, NATO Advanced Science Institutes Series, Series C, Mathematical and Physical Sciences, 400, Kluwer Academic Publishers, Dordrecht/Boston, 379–400.

North, A. N., J. C. Dore, A. R. Mackie, A. M. Howe, and J. Harries, (1990), Ultrasmall‐angle X‐ray scattering studies of heterogeneous systems using synchrotron radiation tech‑niques, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 47, 283–290. DOI:10.1016/0168‐583X(90)90758‐M.

Nygard, R., M. Gutierrez, R. Gautam, and K. Høeg, (2004), Compaction behaviour of argillaceous sediments as function of diagenesis, Marine and Petroleum Geology, 21, 349–362. DOI:10.1016/j.marpetgeo.2004.01.002.

Olierook, H. K. H., C. Delle Piane, N. E. Timms, L. Esteban, R. Rezaee, A.J. Mory, and L. Hancock, (2014), Facies‐based rock properties characterization for CO

2 sequestration: GSWA Harvey 1 well, Western Australia, Marine and Petroleum Geology, 50, 83–102. DOI:10.1016/j.marpetgeo.2013.11.002.

Olsen, H. W., R. W. Nichols, and T. L. Rice, (1985), Low gra‑dient permeability measurements in a triaxial system, Geotechnique, 35, 145–157.

Paterson, M. S., (1977), Experimental Rock Deformation – The Brittle Field. Springer‐Verlag, New York, 254 pp.

Pearson, F. J., (1999), What is the porosity of a mudrock? In Muds and Mudstones: Physical and Fluid Flow Properties, edited by Aplin, A. C., A. J. Fleet, and J. H. S. MacQuaker, Geological Society of London, Special Publication, 158, 9–21. DOI:10.1144/GSL.SP. 1999.158.01.02

Peltonen, C., Ø. Marcussen, K. Bjørlykke, and J. Jahren, (2009), Clay mineral diagenesis and quartz cementation in mud‑stones: the effects of smectite to illite reaction on rock prop‑erties, Marine and Petroleum Geology, 26, 887–898. DOI:10.1016/j.marpetgeo.2008.01.021.

Pervukhina, M., B. Gurevich, D. N. Dewhurst, P. Golodoniuc, and M. Lebedev, (2015), Rock physics analysis of shale reser‑voirs. In Fundamentals of Gas Shale Reservoirs, edited by Rezaee, R., John Wiley & Sons, Inc., Hoboken, NJ, pp. 191–205.

Petley, D. N., (1999), Failure envelopes of mudrocks at high confining pressures. In Muds and Mudstones: Physical and Fluid Flow Properties, edited by Aplin, A. C., A. J. Fleet, and J. H. S. Macquaker, Geological Society of London Special Publication, 158, 61–71. DOI:10.1144/GSL.SP.1999.158.01.05

0004182022.INDD 28 10/16/2018 7:15:02 AM

MICROSTRuCTuRAL, GEOMEChANICAL, ANd PETROPhySICAL ChARACTERIzATION Of ShALE CAPROCkS 29

Pfleiderer, S. and H. C. Halls, (1994), Magnetic pore fabric anal‑ysis: a rapid method for estimating permeability anisotropy, Geophys. J. Int., 116, 39–45. DOI:10.1111/j.1365‐246X.1994.tb02125.x

Radlinski, A. P., M. Mastalerz, A. L. Hinde, M. Hainbuchner, H. Rauch, M. Baron, J. S. Lin, L. Fan, and P. Thiyagarajan, (2004), Application of SAXS and SANS in evaluation of porosity, pore size distribution and surface area of coal, International Journal of Coal Geology, 59, 245–271. DOI:10.1016/j.coal.2004.03.002.

Rasouli, V., M. Pervukhina, T. M. Müller, and R. Pevzner, (2013), In‐situ stresses in the southern Perth Basin at the GSWA Harvey‐1 well site, Exploration Geophysics, 44, 289–298. DOI:10.1071/EG13046.

Rezaee, R., A. Saeedi, and B. Clennell, (2012), Tight gas sands permeability estimation from mercury injection capillary pressure and nuclear magnetic resonance data, J. Pet. Sci. Eng., 88–89, 92–99. DOI:10.1016/j.petrol.2011.12.014.

Rinaldi, A. P., V. Vilarrasa, J. Rutqvist, and F. Cappa, (2015), Fault reactivation during CO2 sequestration: effects of well orientation on seismicity and leakage, Greenhouse Gases‐Science and Technology, 5, 645–656. DOI:10.1002/ghg.1511.

Ringrose, P. S., A. S. Mathieson, I. W. Wright, F. Selama, O. Hansen, R. Bissell, N. Saoula, J. Midgley, (2013), The In Salah CO2 storage project: lessons learned and knowledge transfer, Energy Procedia, 37, 6226–6236. DOI:10.1016/j.egypro.2013.06.551.

Romanov, V., Y. Soong, C. Carney, G. E. Rush, B. Nielsen, and W. O’Connor, (2015), Mineralization of carbon dioxide: a lit‑erature review, ChemBioEng Reviews. DOI:10. 1002/cben.201500002.

Sarout, J. and Y. Guéguen, (2008), Anisotropy of elastic wave velocities in deformed shales: part 1 – experimental results, Geophysics, 73(5), D75–D89. DOI:10.1190/1.2952744.

Sarout, J., L. Esteban, C. Delle Piane, B. Maney, and D. N. Dewhurst, (2014), Elastic anisotropy of Opalinus clay under variable saturation and triaxial stress, Geophysical Journal International, 198, 1662–1682. DOI:10.1093/gji/ggu231.

Schaef, H. T., E. S. Ilton, O. Qafoku, P. F. Martin, A. R. Felmy, K. M. Rosso, (2012), In situ XRD study of Ca2+ saturated montmorillonite (STX‐1) exposed to anhydrous and wet supercritical carbon dioxide, International Journal of Greenhouse Gas Control, 6, 220–229. DOI:10.1016/j.ijggc.2011.11.001.

Schmitt, L., T. Forsans, and F. J. Santarelli, (1994), Shale testing and capillary phenomena, Int. J. Rock Mech. M. Sci. and Geomech. Abstr., 31, 411–427. DOI:10.1016/0148‐9062 (94)90145‐7

Schowalter, T. T., (1979), Mechanisms of secondary hydro‑carbon migration and entrapment, Bulletin of the American Association of Petroleum Geologists, 63, 723–760.

Shah, V., D. Broseta, and G. Mouronval, (2008), Capillary alteration of caprocks by acid gases. SPE Paper 113353. SPE/DOE Improved Oil Recovery Symposium, Tulsa, OK. DOI:10.2118/113353‐MS

Sharma, S., P. Cook, T. Berly, and M. Lees, (2009), The CO2CRC Otway project: overcoming challenges from planning to execution of Australia’s first CCS project, Energy Procedia, 1, 1965–1972. DOI:10.1016/j.egypro.2009.01.256.

Shipton, Z. K., J. P. Evans, D. Kirschner, P. T. Kolesar, A. P. Williams, and J. Heath, (2004), Analysis of CO

2 leakage through ‘low‐permeability’ faults from natural reservoirs in the Colorado plateau, east‐central Utah. In Geological Storage of Carbon Dioxide, edited by Baines, S. J. and R. H. Worden, Geological Society of London Special Publications, 223, 43–58. DOI:10.1144/GSL.SP.2004.233.01.05

Shukla, R., P. Ranjith, A. Haque, and X. Choi, (2010), A review of studies on CO2 sequestration and caprock integrity, Fuel, 89, 2651–2664. DOI:10.1016/j.fuel.2010.05.012.

Sinha, S., E. M. Braun, M. D. Determan, Q. R. Passey, S. A. Leonardi, J. A. Boros, A. C. Wood, T. Zirkle, and R. A. Kudva, (2013), Steady‐state permeability measurements on intact shale samples at reservoir conditions – effect of stress, temperature, pressure, and type of gas. SPE 164263, 15 pp. DOI:10.2118/164263‐MS

Skempton, A. W., (1954), The pore pressure coefficients A and B, Geotechnique, 4, 143–147.

Skurtveit, E., E. Aker, M. Soldal, M. Angeli, and Z. Wang, (2012), Experimental investigation of CO2 breakthrough and flow mechanisms in shale, Petroleum Geoscience, 18, 3–15. DOI:10.1144/1354‐079311‐016

Skurtveit, E., L. Grande, O. Y. Ogebule, R. H. Gabrielsen, J. I. Faleide, N. H. Mondol, R. Maurer, and P. Horsrud, (2015), Mechanical testing and sealing capacity of the Upper Jurassic Draupne Formation, North Sea. Proceedings of the 49th US Rock Mechanics/Geomechanics Symposium, San Francisco, CA, 28 June–1 July, ARMA 15‐331, 8 pp.

Sondergeld, C. H., K. E. Newsham, J. T. Comisky, M. C. Rice, and C. S. Rai, (2010), Petrophysical considerations in evalu‑ating and producing shale gas resources. SPE Unconventional Conference, Pittsburg, Pennsylvania. SPE Paper 131768. DOI:10.2118/131768‐MS.

Stalker, L., S. Varma, D. van Gent, J. Haworth, and S. Sharma, (2013), South west hub: a carbon capture and storage project, Australian Journal of Earth Sciences, 60, 45–58. DOI:10.1080/08120099.2013.756830.

Steiger, R. P. and P. K. Leung, (1988), Quantitative determina‑tion of the mechanical properties of shales. SPE Paper 18024, 8 pp. DOI:10.2118/18024‐PA

Steiger, R. P. and P. K. Leung, (1991), Consolidated undrained triaxial test procedure for shales. Proceedings 32nd U.S. Symposium on Rock Mech., ARMA Paper 91‐637, Norman, OK, Balkema, Rotterdam, pp. 637–646.

Tenthorey, E., T. Dance, Y. Cinar, J. Ennis‐King, and J. Strand, (2014), Fault modelling and geomechanical integrity associ‑ated with the CO2CRC Otway 2C injection experiment, Int. J. Greenhouse Gas Control, 30, 72–85. DOI:10.1016/j.ijggc.2014.08.021.

Tenthorey, E., S. Vidal‐Gilbert, G. Backé, R. Puspitasari, Z. J. Pallikathekathil, B. Maney, and D. N. Dewhurst, (2013). Modelling the geomechanics of gas storage: a case study from the Iona gas field, Australia, Int. J. Greenhouse Gas Control, 13, 138–148. DOI:10.1016/ j.ijggc.2012.12.009

Thyberg, B., J. Jahren, T. Winje, K. Bjørlykke, J. I. Faleide, and Ø. Marcussen, (2010), Quartz cementation in late cretaceous mudstones, northern North Sea: changes in rock properties due to dissolution of smectite and precipitation of

0004182022.INDD 29 10/16/2018 7:15:02 AM

30 GEOLOGICAL CARBON STORAGE

micro‐quartz crystals, Marine and Petroleum Geology, 27, 1752–1764. DOI:10.1016/j.marpetgeo.2009.07.005.

Timms, N. E, H. K. H. Olierook, M. E. J. Wilson, C. Delle Piane, P. J. Hamilton, P. Cope, and L. Stütenbecker, (2015), Sedimentary facies analysis, mineralogy and diagenesis of the mesozoic aquifers of the Central Perth Basin, Western Australia, Marine and Petroleum Geology, 60, 54–78. DOI:10.1016/j.marpetgeo.2014.08.024.

Valès, F., D. Nguyen Minh, H. Gharbi, and A. Rejeb, (2004), Experimental study of the influence of the degree of satura‑tion on physical and mechanical properties in Tournemire shale (France), Applied Clay Science, 26, 197–207. DOI:10.1016/j.clay.2003.12.032.

Van Olphen, H., (1977), An Introduction to Clay Colloid Chemistry, 2nd edition, John Wiley, New York.

Vavra, C. L., J. G. Kaldi, and R. M. Sneider, (1992), Geological applications of capillary‐pressure – a review, Bulletin of the American Association of Petroleum Geologists, 76, 840–850.

Vernik, L. and X. Liu, (1997), Velocity anisotropy in shales; a petrophysical study, Geophysics, 62, 521–532. DOI:10.1190/ 1.1444162.

Vidal‐Gilbert, S., E. Tenthorey, D. Dewhurst, J. Ennis‐King, P. van Ruth, and R. Hillis, (2010), Geomechanical analysis of the Naylor field, Otway Basin, Australia: implications for CO

2 injection and storage, International Journal of Greenhouse Gas Control, 4, 827–839. DOI:10.1016/j.ijggc.2010.06.001.

Wang, R., N.‐S. Zhang, X.‐J. Liu, X.‐M. Wu, and J. Yan, (2013), Apparent permeability of shale considering the adsorption and diffusion of gas and the effects of temperature and pressure on it. Research progress of mechanism of adsorp‑tion and desorption of gas in shale, Science Technology and Engineering, 2013‐19.

Washburn, E. W., (1921), A note on the method of determining the distribution of pore sizes in a porous material, Proceedings of the National Academy, 7, 115–116.

Watson, M. N., N. Zwingmann, and N. M. Lemon, (2004), The Ladbroke Grove–Katnook carbon dioxide natural laboratory: a recent CO2 accumulation in a lithic sandstone reservoir, Energy, 29, 1457–1466. DOI:10.1016/j.energy.2004.03.079.

Watts, N., (1987), Theoretical aspects of cap‐rock and fault seals for single‐ and two‐phase hydrocarbon columns, Marine and Petroleum Geology, 4, 274–307. DOI:10.1016/0264‐8172(87)90008‐0.

Wollenweber, J., S. Alles, A. Busch, B. M. Krooss, H. Stanjek, and R. Littke, (2010), Experimental investigation of the CO2 sealing efficiency of caprocks, International Journal of Greenhouse Gas Control, 4, 231–241. DOI:10.1016/j.ijggc.2010.01.003.

Yang, Y. and A. C. Aplin, (1998), Influence of lithology and compaction on the pore size distribution and modelled per‑meability of some mudstones from the Norwegian Margin, Marine and Petroleum Geology, 15, 163–175. DOI:10.1016/S0264‐8172(98)00008‐7.

Yielding, G., B. Freeman, and T. D. Needham, (1997), Quantitative fault seal prediction, AAPG Bulletin, 81, 897–917.

Youn, H. and F. Tonon, (2010), Multi‐stage triaxial test on brittle rock, International Journal of Rock Mechanics and Mining Sciences, 47, 678–684. DOI:10.1016/j.ijrmms.2009.12.017.

Zhang, Y., L. Langhi, C. Delle Piane, P. Schaubs, D. Dewhurst, L. Stalker, and K. Michael, (2015a), Geomechanical stability of CO2 containment at the southwest hub, Western Australia: a coupled geomechanical‐fluid flow modelling approach, International Journal of Greenhouse Gas Control, 37, 12–23. DOI:10.1016/j.ijggc.2015.03.003.

Zhang, Y., L. Langhi, C. Delle Piane, P. Schaubs, D. Dewhurst, L. Stalker, and K. Michael, (2015b), Geomechanical evaluation of fault reactivation potential and uplift at the South West Hub geological CO2 storage site, Western Australia. Paper 15‐127, 49th ARMA Conference, San Francisco, CA, 8 pp.

0004182022.INDD 30 10/16/2018 7:15:02 AM

3

Water‐Energy‐Food Nexus: Principles and Practices, Geophysical Monograph 229, First Edition. Edited by P. Abdul Salam, Sangam Shrestha, Vishnu Prasad Pandey, and Anil Kumar Anal. © 2017 American Geophysical Union. Published 2017 by John Wiley & Sons, Inc.

The Need for the Nexus Approach

P. Abdul Salam1, Vishnu Prasad Pandey2, Sangam Shrestha3, and Anil Kumar Anal4

1

1.1. INTRODUCTION

The water, energy, and food resources share a lot in common, including inaccessibility to billions of people, rapidly growing demand, strong interdependencies with climate change, different regional availability, and variations in supply and demand [Bazilian et al., 2011; Walsh et al., 2015]. Apart from the similarities, there is a growing sense of awareness of the linkages among water, energy, and food sectors (Figure 1.1) and that the actions in one sector would inadvertently affect one or both of the other sectors. The growing population, rapid economic growth, and changing consumption trends has increased the urgency to act through the utilization of integrated approaches that

encompasses all three sectors. This ensures that there is a proper balance among the different user goals and inter­ests while at the same time protecting the ecosystem.

It was acknowledged at the Bonn 2011 Nexus Con­ference that integrated policies related to water, energy, and food are required in the face of growing concerns over the future availability and sustainability of these resources. The continuation of isolated policies which are predominant in developing countries will unavoidably affect other sectors and eventually lead to the acceleration of ecosystem degradation. Hence, a better understanding of the strong linkages and trade‐offs with respect to the water‐energy‐food (WEF) nexus is important for sustain­able long‐term development growth as well as for human well‐being. A nexus approach is based on three guiding principles [Bonn 2011 Conference, 2011]:

1. Placing people and their basic human rights as the basis of the nexus

2. Creating public awareness and the political will for successful implementation

3. Involving local communities in the planning and implementation processes in order to create a sense of participation and ownership

ABSTRACT

The water, energy, and food resources share a lot in common; they have strong interdependencies and are inad­vertently affected by action in any one of them. Therefore, the nexus approach (integrated policies related to water, energy, and food) is required in the face of growing concerns over the future availability and sustainability of these resources. The nexus approach can help achieve at least some of the “Sustainable Development Goals (SDGs)” (e.g., SDG 2, 6, 7, 12, 13, 15). This chapter discusses trends in availability and consumption of the three key resources (i.e., water, energy, and food) and interactions between them, and finally provides some reasons why the nexus approach can help achieve social and economic development goals.

1 Energy Field of Study, Asian Institute of Technology, Klong Luang, Thailand

2 International Water Management Institute (IWMI), Nepal Office, Lalitpur, Nepal

3 Water Engineering and Management, Asian Institute of Technology, Klong Luang, Thailand

4 Food Engineering and Bioprocess Technology, Asian Institute of Technology, Klong Luang, Thailand

0003116743.INDD 3 6/29/2017 8:58:33 AM

COPY

RIGH

TED

MATER

IAL

4 WATER‐ENERGY‐FOOD NEXUS

The practical implementation is proven as difficult mainly due to the vastness of the individual sectors, the multidimensional interlinkages among the sectors, and the fact that stakeholders in all disciplines and at all levels need to be involved. In addition, significant financial investment would also be required in the restructuring of existing infrastructure to suit the nexus approach. The develop­ment of robust analytical tools, conceptual models, and robust data sets which can be used to supply information on the future use of energy, water, and food is vital toward making the WEF nexus a reality [Bazilian et al., 2011].

The Sustainable Development Goals (SDGs) have set targets for each of the nexus sectors explicitly under SDG 2 (zero hunger), SDG 6 (clean water and sanitation), and SDG 7 (affordable and clean energy). In order to satisfy the stipulated goals, a shift to more sustainable produc­tion and consumption patterns (SDG 12) will be required while tackling climate change (SDG 13) and ensuring a balance in ecosystem both on land and water (SDG 14 and SDG 15). The interconnection between the SDGs emphasizes the need for a nexus approach in achieving the individual goals.

1.2. AVAILABILITY AND CONSUMPTION TRENDS OF THE NEXUS COMPONENTS

The growing demand for water, energy, and food are driven by common factors: population growth and mobility, sustainable development, international trade, urbanization,

changing lifestyles, cultural and technological changes, and climate change [FAO, 2014]. The exploitation of more resources will definitely be required to meet the growing demand. However, it is possible to slow down this growing demand by reducing wastage and loss incurred in the water, energy, and food stream, which would also help in  saving embedded resources during production and reducing environmental impacts. Reduction of water and energy through conservation and efficient use will be crucial in the coming decade.

1.2.1. Water

The world has enough freshwater to supply the global demand but nonuniform distribution of these reserves and other reasons have led to shortages in certain loca­tions. The United Nations (UN) estimates indicate that there are 1.2 billion people living in areas of physical water scarcity and another 1.6 billion people facing economic water shortage [Walsh et al., 2015]. In terms of water quality, there are 748 million people who lack access to an improved drinking water source [UNESCO, 2015]. The shortage in both quantity and quality may likely spread and become more acute due to growing demands, unsus­tainable withdrawal rates, degradation of source water quality, and changing climate patterns. Understandably, the main impact of water shortage is on direct human consumption but other indirect impacts include those on energy supply, food production, and ecosystem.

IrrigationFertilizersHarvestingProcessingStorage

IrrigationProcessing

PumpingWater treatmentDrainageWater distribution

Energy generationCoolingExtractionTransport

Water quality

Bioenergy production

FoodWaterEnergy

Figure 1.1 Interactions of the water‐energy‐food nexus. Source: IRENA [2015], “Renewable Energy in the Water, Energy & Food Nexus.”

0003116743.INDD 4 6/29/2017 8:58:33 AM

ThE NEED FOR ThE NEXUS AppROAch 5

Traditionally, the expansion of water resources mainly depended on the need of the expanding population for food, clothing, and modern energy. More recently, the rising standards of the middle‐income group has led to sudden and sharp increases in the water consumption in both production and use. Economic growth coupled with higher living standards could be the reason that the growth of water demand is double that of population growth in the twentieth century.

The global water withdrawals in 2009 stood at 4500 billion m3 (BCM) of which 70% was used for agriculture, 17% for industry, and 13% for municipal and domestic purposes [2030 Water Resources Group, 2009]. According to the 2030 Water Resources Group [2009], the projected demand of 6900 BCM in 2030 under the business‐as‐usual scenario is 40% more than the currently assessed water supplies (ground and surface) that are accessible, reliable, and sustainable. In another report by UN Water, the water demand is projected to increase by around 55% in 2050, which will mainly be attributed to growing demands in the manufacturing sector, thermal power plants, and domestic use [UNESCO, 2015].

The gap between future availability and demand can be closed not through the discovery of more water supplies but through effective demand‐side management, which will definitely need effective policy interventions.

1.2.2. Energy

Energy demand is increasing primarily due to drivers like growth in population and gross domestic product (GDP). Though there are diverse sources of energy, fossil fuels are expected to continue as the dominant fuel source and would account for almost 80% of the total energy supplies in 2035 [BP, 2016]. Gas is expected to

gain popularity along with renewable energy though the share of the latter would still be below 10% in 2035. On the other hand, coal will exhibit a decreasing trend while oil remains steady. The additional energy demand will come from growing and emerging economies while Organization for Economic Cooperation and Development (OECD) countries will hardly show any growth. Apart from the need of energy to support the increased GDP in the developing countries, the push for global electrifica­tion will drive the steady growth for power generation. China will be a key player in the future energy demand as Figure 1.2 indicates that they will move toward a more sustainable rate compared to the past.

Most energy projections by various organizations follow the trend as depicted in Figure  1.2. There are international initiatives which look at reducing the demand and dependency on fossil fuels. One such initi­ative is the Sustainable Energy for All (SE4ALL), which was launched by the UN Secretary‐General in 2011. The SE4ALL has set three main objectives to be achieved by 2030: ensure universal access to modern energy ser­vices, double the global rate of improvement in energy efficiency, and double the share of renewable energy in the global context.

1.2.3. Food

There are concerns on whether the world would be able to produce enough food for the growing population. The amount of arable land and water required for agriculture is declining and at the same time has to compete with urbanization and industrialization for the same resources. The most popularly used indicator for measuring and monitoring the world food status is food consumption in kcal/person/day. The world average per capita availability

18

15

12

9

6

3

01965 2000 2035 1965 2000 2035

Renewables*

Billion toeConsumption by region

50%

Shares of primary energy

Oil

Coal

Gas

Hydro

Nuclear

40%

30%

20%

10%

0%

Other

OtherAsia

China

OECD

Figure  1.2 Projected growth in energy consumption. Toe is ton equivalent. * includes biofuels. Source: Reproduced with permission of BP [2016]. (See insert for color representation of the figure.)

0003116743.INDD 5 6/29/2017 8:58:33 AM

6 WATER‐ENERGY‐FOOD NEXUS

of food for direct human consumption was 2770 kcal/person/day in 2005/2007 (Figure 1.3). This world average is, however, misleading as there are areas where the value falls below 2500 kcal and other areas where it is way above 3000 kcal.

By 2050, food production in the global context and for developing countries will need to be increased by 60 and 100% respectively from 2005/2007 figures [UNESCO, 2015]. This translates into a 1.1% annual growth rate increment of total world consumption [Alexandratos and Bruinsma, 2012]. The projected values in million tons for some of the major food groups with respect to 2005/2007 figures are illustrated in Figure  1.4. The drivers for increase will mainly result from increasing population and income as well as structural changes in diet (i.e., shifting to a meat‐based diet) and overnutrition.

1.3. SECTORAL INTERACTIONS

Water, energy, and food are interlinked in many ways. Water is required to produce energy and food. Energy is required to produce water and food. Food can be a source of energy (e.g., biofuel). Therefore, action in one sector will have implications on the others.

1.3.1. Water–Energy Interactions

The water intensity in the energy sector varies depend­ing on the choice of technology, source of water, and type

of fuel. Water is used for the production of fuels origi­nating from fossils, growing of biomass‐related fuel stocks, and generation of energy (e.g., electricity from fossil fuels). Thermal power plants utilize large amounts of water for cooling, of which a fraction is lost to evapo­ration depending on the type of cooling system employed. On the other hand, hydropower plants utilize a large area, which in turn increases the surface area of the water body, further facilitating evaporation. In 2010, energy produc­tion accounted for 15% (580 BCM of water annually) of global freshwater withdrawals, of which 66 BCM was consumed [Walsh et al., 2015]. In the United States, power plants account for the largest share (41%) of fresh­water withdrawal [Union of Concerned Scientists, 2010]. The global energy demand is projected to increase by 35% in 2035, which would increase water withdrawal in the energy sector by 20% and water consumption by 85% [IRENA, 2015]. The life cycle water consumption (gallons/MWh) for some selected electricity generation technolo­gies is illustrated in Figure 1.5.

Renewable energy is very slowly replacing fossil fuels, especially in the power sector, but it is still projected that 75% of the expected energy increase by 2030 will be from fossil fuels [ADB, 2013]. Renewable energy alternatives may be climate‐change‐friendly but may not be favorable when considering water and land requirements. As illus­trated in Figure  1.6, the water requirement for biofuel production is much higher than that required for fossil‐fuel‐based products. The promotion of biofuels in the

4000

World Developed Developing Sub-SaharanAfrica

Near East& NorthAfrica

LatinAmerica & Caribbean

South Asia East Asia

3500

3000

2500

2000

1500

1000

500

0

2005/2007 2050

Figure  1.3 Per capita food consumption (kcal/person/day). Source: Alexandratos and Bruinsma [2012]. Reproduced with permission of FAO.

0003116743.INDD 6 6/29/2017 8:58:33 AM

ThE NEED FOR ThE NEXUS AppROAch 7

1400

Food Nonfood (incl. waste) Total production

1200

1000

455

258 195

341

149282

2068

3009

800

600

400

200

0

2005/2007

Meat Sugarcane/beet inraw sugar equiv.

Oil crops & productsin oil equiv.

Cereals (wheat,coarse grains,

milledrice) - right axis

2005/2007

2050 2050 2005/2007

2050 2005/2007

2050

3500

3000

2500

2000

1500

1000

500

0

Figure 1.4 World production and use of major agricultural products (million tons). Source: Alexandratos and Bruinsma [2012]. Reproduced with permission of FAO.

Wind:on-shore and

off-shore

PVOther C-5

Flat panel Concentrated PV

Flash BinaryDry cooling

Dry cooling

Centrifugal enrichment

Conventional gas

Conventional gas

Open loop cooling Pond cooling Coolingtower

Shale gas

No cooling

Shale gas

Hybrid cooling

Hybrid cooling

BinaryHybrid cooling

Cooling tower

Cooling tower

Geothermal

CSP: power tower

CSP: trough

Nuclear

Natural gas: CT

Natural gas: CC

Coal: IGCC

Dry Open loop

Open loop cooling Cooling tower Pond cooling

Coolingtower

Cooling tower

Pond coolingCooling Cooling

Surface mining

Surface mining

Underground mining

Underground miningCoal: PC

0

Power plant Fuel cycle Operations

Fuel cycle or power plantsubcategories (top)(1- base case:1-variant)Operations subcategories(bottom)(1- base case:|-variant)

200 400 600 800 1000l

l

1200

Gas dry cooling

Diffusion enrichment

Dry cooling

Figure  1.5 Life cycle water consumption for selected electricity generation technologies (gal/MWh). Source: IRENA [2015], “Renewable Energy in the Water, Energy & Food Nexus.”

0003116743.INDD 7 6/29/2017 8:58:34 AM

8 WATER‐ENERGY‐FOOD NEXUS

transport sector through subsidies has led to greater competition for land and water use [ADB, 2013].

Energy is required for the extraction, transportation, and treatment of water. The energy intensity for water will vary depending mainly on the source of water, quality of water, and efficiency of the water system. For example, desalination of seawater would be more energy‐intensive than utilizing surface or groundwater. Surface water was traditionally used for agricultural irrigation but with advances in technologies and inaccessibility to surface water, the use of groundwater has increased stead­ily. This shift to groundwater use comes with increased energy demand and lowered groundwater levels.

Energy is a dominant cost factor in the provision of water and wastewater facilities with estimates of 55% of water utilities operating budget being attributed to the energy cost [IRENA, 2015]. Water purification for indus­trial processes and human consumption requires energy and the amount of energy required depends on the source of water. For example, the purification of lake, river, or groundwater consumes less than 1 kWh/m3 of potable water while purification of seawater can be as high as 8 kWh/m3 [IRENA, 2015].

1.3.2. Water–Food Interactions

The accessibility and availability of water determine the agricultural characteristics of a given locality and the world as a whole. Water is necessary for food production, preparation, and consumption while changes in food consumption patterns or agricultural practices can create a strain on water security. Agriculture can be considered as the largest consumer of freshwater supplies, account­ing for approximately 70% of consumption [Ooska News, 2011]. Water is not only used for growing food crops (i.e., irrigation) but also for processing, distribution, retailing, and consumption [IRENA, 2015]. Agricultural practices also affect water resources via water pollution through fertilizers and pesticides, which in turn affects agriculture itself, thus forming a vicious cycle. Though agriculture accounts for a large share of freshwater withdrawal, most of the water is returned to the surface or groundwater along with pollutants [IRENA, 2015].

The generation of waste or polluted water is unavoidable whenever food is handled, processed, packed, distributed, or stored. It was estimated that the consumption of water in the food industry in England is around 250 million m3

Conventional gas

CoalWithdrawal

ConsumptionShale gas

<1 101 102 103 104 105 106

Liters per toe

107

Refined oil (conventional)

Refined oil (oil sands)

Gas to liquids

Coal to liquids

Refined oil (EOR)

Lignocellulosic ethanol

Palm oil biodiesel

Rapeseed biodiesel

Soybean biodiesel

Corn ethanol

Sugarcane ethanol

Figure  1.6 Water withdrawal and consumption for primary fuel extraction, processing, and transportation. Source: IRENA [2015], “Renewable Energy in the Water, Energy & Food Nexus.”

0003116743.INDD 8 6/29/2017 8:58:34 AM

ThE NEED FOR ThE NEXUS AppROAch 9

(MCM) for 2006 [Klemes et al., 2008]. The cost incurred during supply and disposal could be minimized by reducing the amount of wastewater, which can also lead to saving the loss of potential revenue.

1.3.3. Energy–Food Interactions

The energy–food interaction is more visible and easily felt in the modern context as the variations in food prices are strongly linked to oil price variations [Bazilian et al., 2011]. This is not surprising as the agri‐food supply chain accounts for 30% of the world’s energy consumption [IRENA, 2015]. The main share of the energy consumed in the food sector is required for activities related to processing, distribution, preparation, and cooking. Energy is also accounted for in energy‐intensive products such as pesticides and fertilizers. High‐yield agriculture is heavily dependent on synthetic nitrogen‐based fertilizers, which are almost entirely produced using natural gas [ADB, 2013].

The growing demand for food will be due to the growing population, improved lifestyle, and further mechanization of the food supply chain. The main challenge in the food sector with meeting the growing demand is not actually an increase in food production but rather a reduction in food wastages. The Food and Agricultural Organization (FAO) reported that approximately one‐third of edible food pro­duced for human consumption is lost or wasted [IRENA, 2015]. This accounts for a loss in not only embedded energy but also embedded water and contributes to green­house gas (GHG) emissions.

Food‐processing industries also consume a significant amount of energy for heating and cooling during pro­cessing and storage of food products. For example, 20% of energy in the dairy industry is used for cooling and 80% for heating purposes. Energy consumption of the food industry in the United Kingdom is estimated at 126 TWh/year, which is equivalent to 14% of energy con­sumption in the country [Klemes et al., 2008]. Similarly, the premium energy in the form of biogas can be produced from the effluent of food‐processing plants by running anaerobic digestors. The quality and quantity of gas production depends on the balance of organic materials and process management ranges from 150 to 600 L/kg of volatile solids [Burton and Turner, 2003]. Pure methane has a thermal energy of 53 MJ/kg. Studies show that fuel can be generated from the utilization of organic waste [Bianchi et al., 2006].

1.4. THE NEED FOR THE WATER‐ENERGY‐FOOD (WEF) NEXUS

There are still 1.2 billion people who lack access to elec­tricity, 783 million people without access to potable water, and 842 million people who suffer from chronic hunger

[IRENA, 2015]. Developing countries are expected to see a rise in population and consumption in both developing and developed countries is becoming more resource‐intensive. By 2050, it is expected that global energy demand will double, with water and food demand increas­ing by over 50% [IRENA, 2015]. Climate change impacts such as global temperature increase and extreme weather conditions further compound the challenge of meeting the growing demand.

As the planet approaches the sustainable limit of its resources, competition and scarcity of the resources will become more predominant. There is a likely possibility that economic growth will soon be constrained by short­ages of one or more of these resources. Therefore, water security, energy security, and food security have already been on national and international agendas for quite some time.

The amalgamation of water, energy, and food in a “nexus” framework in order to increase resource effi­ciency can be considered as a necessary way forward in achieving the SDGs. It enables us to take into considera­tion the impacts of a decision for one sector on itself as well as on the other sectors.

The best case example of a complex interaction of the nexus is the emerging trend of biofuels as an energy source in the transport sector. Biofuel production raises the conflict of the use of limited water and land against growing food for human consumption.

1.5. STRUCTURE OF THIS BOOK

In the context of the need to better understand, operationalize, and practice the nexus approach for resource use efficiency vis‐à‐vis the lack of adequate knowledgebase and publications in the arena, this book aims to contribute to the global debate on WEF nexus through knowledgebase generation. A single volume of the book covers theoretical and/or conceptual aspects of the WEF nexus, ways to overcome operational chal­lenges of the nexus approach to resources management, cases of the nexus approach in practice from different regions of the world, and opinions on the future of the nexus agenda. The book is divided into four sections and 19 chapters. The first section on “understanding the nexus” contains five chapters focusing on the need of a nexus approach; its evolution as a policy and develop­ment discourse; its contribution to better water manage­ment and limitation; the emergence of a new paradigm in the nexus approach; and the urban nexus. The second section “operationalizing the nexus” contains six chap­ters focusing on modeling techniques; available tools/models in practice; governing the nexus; the role of international cooperation in operationalizing the nexus; framing nexus cooperation issues in the transboundary

0003116743.INDD 9 6/29/2017 8:58:34 AM

10 WATER‐ENERGY‐FOOD NEXUS

context; and cases of energy‐centric operationalization of the nexus. The third section on the theme of “nexus in practice” covers seven chapters and focuses on vari­ous types of case studies of WEF nexus in various geographical regions of the world. Finally, the fourth section called “future of the nexus agenda” contains only one chapter focusing on how the nexus approach can help achieve the SDGs or the 2030 Agenda.

REFERENCES

2030 Water Resources Group (2009), Charting our water future: Economic frameworks to inform decision‐making.

ADB (Asian Development Bank) (2013), Thinking about Water Differently: Managing the Water‐Energy‐Food Nexus, ADB, Bangkok.

Alexandratos, N., and J. Bruinsma (2012), World Agriculture Towards 2030/2035: The 2012 Revision, FAO, Rome.

Bazilian, M., H. Rogner, M. Howells, S. Hermann, D. Arent, D. Gielen, P. Steduto, A. Mueller, P. Komor, R. S. J. Tol, and K. K. Yumkella (2011), Considering the energy, water and food nexus: Towards an integrated modelling approach, Energy Policy, 39(12), 7896–7906, doi:10.1016/j.enpol.2011.09.039.

Bianchi, M., F. Cherubini, A. De Pascale, A. Peretto, and B. Elmegaard (2006), Cogeneration from poultry industry wastes: Indirectly fired gas turbine application, Energy, 31(10), 1417–1436.

Bonn2011 Conference (2011), Messages from the Bonn2011 Conference: The water, energy and food security nexus—solutions for a green economu.

BP (2016), BP energy outlook 2016: Outlook to 2035.Burton, C. H., and C. Turner (2003), Manure Management:

Treatment Strategies for Sustainable Agriculture, Editions Quae ed., Silsoe Research Institute, Bedford.

FAO (Food and Agricultural Organization) (2014), The Water‐Energy‐Food Nexus: A New Approach in Support of Food Security and Sustainable Agriculture, FAO, Rome.

IRENA (The International Renewable Energy Agency) (2015), Renewable energy in the water, energy and food nexus, IRENA, Abu Dhabi.

Klemes, J., R. Smith, and J.‐K. Kim (2008), Handbook of Water and Energy Management in Food Processing, Elsevier, Burlington, MA.

Ooska News (2011), Conference synopsis, paper presented at the Bonn2011 conference: The water, energy and food security nexus—solutions for the green economy, Bonn, Germany.

UNESCO (United Nations Educational, Scientific and Cultural Organization) (2015), The United Nations World Water Development Report 2015: Water for a Sustainable World, UNESCO, Paris.

Union of Concerned Scientists (2010), The energy‐water collision: 10 things you should know.

Walsh, B. P., S. N. Murray, and D. T. J. O’Sullivan (2015), The water energy nexus, an ISO50001 water case study and the need for a water value system, Water Resour. Ind., 10, 15–28, doi:10.1016/j.wri.2015.02.001.

0003116743.INDD 10 6/29/2017 8:58:34 AM

5

Lithospheric Discontinuities, Geophysical Monograph 239, First Edition. Edited by Huaiyu Yuan and Barbara Romanowicz. © 2019 American Geophysical Union. Published 2019 by John Wiley & Sons, Inc.

1

1.1. INTRODUCTION

Major seismic discontinuities both bound the transition zone (~410 to ~660 km) and occur within it. In this depth range, the seismic discontinuities are attributed mostly to phase transformations in the major mantle minerals [e.g., Ringwood, 1991]. In contrast, the volumetrically domi-nant upper‐mantle mineral olivine suffers no phase trans-formations shallower than ~410 km. (There is a phase transformation in pyroxene ((Mg,Fe)SiO3), the second most abundant mineral in the upper mantle, but the influence of this phase transformation on seismic‐wave velocity is relatively minor [e.g., Matsukage et al., 2005].) Several seismic features interpreted as discontinuities, how-ever, have been reported in the upper mantle [e.g., Thybo and Perchuć, 1997; Kawakatsu et  al., 2009; Rychert and

Shearer, 2009; Tauzin et  al., 2010; Beghein et  al., 2014; Hopper and Fischer, 2015; Wei and Shearer, 2017], among which we focus on two well‐documented discontinuities in the upper mantle: the lithosphere–asthenosphere boundary (LAB) in both oceanic and continental upper mantle and the mid-lithosphere discontinuity (MLD) in the continental upper mantle.

At both the LAB and the MLD seismic velocities decrease with depth. Both the LAB and the MLD have been attributed to the onset of partial melting [e.g., Anderson and Spetzler, 1970; Thybo, 2006], layering in anisotropy [e.g., Yuan and Romanowicz, 2010b], layering in chemical/mineralogical composition [e.g., Selway et  al., 2015], the smooth temperature dependence of seismic velocities including the influence of anelastic relaxation [e.g., Gueguen and Mercier, 1973; Faul and Jackson, 2005], and physical dispersion associated with temperature and water‐dependent anelasticity facilitated by elastically accommodated grain‐boundary sliding

On the Origin of the Upper Mantle Seismic Discontinuities

Shun‐ichiro Karato and Jeffrey Park

ABSTRACT

Recent high‐frequency seismological studies revealed enigmatic structures in the upper mantle: the mid-lithosphere discontinuity (MLD) in the old continents and the oceanic lithosphere–asthenosphere boundary (LAB) at relatively shallow depth. Both discontinuities occur at relatively low temperatures (~1000 °C) making it difficult to attribute them to partial melting. Also the inferred relatively sharp and large velocity drop at these boundaries is difficult to explain with conventional smooth temperature‐dependent elastic and anelastic properties from an absorption‐band Q model. Models that go beyond these two mechanisms are needed to explain these characteristics of the MLD and the LAB. Three models are reviewed: (a) layered anisotropy, (b) layered composition, and (c) elastically accommodated grain‐boundary sliding (EAGBS). A layered anisotropy model is difficult to reconcile with the presence of these discontinuities for diverse ray geometries, and a layered composition model is not supported by the petrological/geochemical observations. EAGBS can explain globally observed substan-tial velocity drop at a modest temperature. This model implies that the link between seismic‐wave velocity and the long‐term creep strength is indirect, and predicts a peak in attenuation near these boundaries. However, exper-imental studies on EAGBS and geophysical tests are incomplete. Directions of future studies to test these models are suggested.

Department of Geology and Geophysics, Yale University, New Haven, CT, USA

0004167983.INDD 5 10/11/2018 8:51:02 PM

COPY

RIGH

TED

MATER

IAL

6 LITHOSPHERIC DISCONTINUITIES

[Karato, 2012; Karato et al., 2015]. These models differ in their implications for the geological significance of the LAB and the MLD. Therefore it is important to under-stand which mechanisms may cause these discontinuities.

In this chapter, we review geophysical observations on upper mantle discontinuities and their interpretation in terms of elastic and nonelastic properties of rocks. In the next section (section 1.2), we review seismological obser-vations on these discontinuities and emphasize the impor-tance of using a broad range of mutually complementary seismological approaches (e.g., high‐frequency body waves, low‐frequency surface waves). Also, we will briefly summarize magnetotelluric (MT) observations on electrical conductivity as far as it is related to the MLD and the LAB (see also Chapters 2 and 5). In section 1.3 we review geological (and petrological) observations related to the MLD and the LAB. We review various models of the LAB and the MLD in section 1.4, including chemical layering, partial melting, and solid‐state relaxa-tion. A detailed account of the EAGBS (elastically accommodated grain‐boundary sliding) model is provided in section 1.5. In section 1.6 we evaluate each model by comparing seismological (and other) observa-tions with the predictions for each model. Finally in sec-tion 1.7 we discuss possible directions of future studies.

1.2. SEISMOLOGICAL OBSERVATIONS RELEVANT TO THE LAB AND THE MLD

1.2.1. General Introduction: Long Wavelength Versus Short Wavelength Seismology

In order to understand the origin of various disconti-nuities, several seismological observations can be used to complement geological observations of rocks in Earth’s interior and the MT observations on electrical conduc-tivity. These seismological observations include (a) the depth, (b) the sharpness, and (c) the magnitude of velocity change. In addition, (d) distribution and nature (geom-etry) of seismic anisotropy and (e) distribution of seismic attenuation will also provide data to evaluate models. In this section, we provide a review of various seismological approaches to address these issues, with specific examples from both oceanic and continental upper mantle.

The sharpness of and the velocity contrast across the mantle discontinuities place important constraints on the  models to explain them. Their resolution is highly sensitive to the method used, particularly to the frequency (wavelength) of the seismic waves. In body‐wave seis-mology, relatively high frequencies (short wavelengths) are used, whereas in surface‐wave seismology to reveal global structure, low‐frequency (long‐wavelength) waves are used. For example, with a seismic wave of 1 s period, the wavelength for shear waves is ~4–5 km in the upper

mantle, whereas for a seismic wave of 100 s period, the wavelength will be ~400–500 km. The spatial resolution of Earth’s structures that can be revealed by each type of wave is different, as shown in Figure 1.1.

Different types of seismic waves are used to investigate different regions, and consequently it is essential to understand the resolution of each study to develop a comprehensive model for the mantle. For example, as we discuss in the next section, early studies on the oceanic upper mantle used long‐wavelength surface waves that revealed global structures with poor depth resolution, whereas recent studies using ocean‐bottom seismometers utilize short wavelength waves that reveal fine‐scale structures.

Scattered waves offer a direct inference of sharp lithospheric interfaces. Revenaugh and Jordan [1991a–c] inferred scattered waves within near‐vertical ScS rever-berations to infer mantle transition‐zone interfaces at 410 km and 660 km and, in oceanic regions, a negative

Velocity Velocity

Dep

th

Dep

th

(a) (b)

High-frequency model

Low-frequency model

Figure 1.1 Schematic diagrams of velocity–depth profiles. (a) A case where an anomalous layer is present with a sharp top boundary but with gradual change in velocity at a deeper region. (b) A case where there is a thin low‐velocity region with sharp boundaries at both top and bottom. Blue lines are velocity–depth profiles caused by relatively sharp changes in velocity with depth. These features can be resolved using high‐frequency (short wave-length) waves, but using low‐frequency (long wavelength) waves these details cannot be resolved and models with smooth depth variation (shown by red curves) would explain seismological observations. Note also that if a velocity–depth variation contains two sharp boundaries (as in (b)), high‐frequency seismic records would show both boundaries but in low‐frequency records only one boundary (or no sharp boundary) would be seen (as in (a)). The estimate of the magnitude of velocity change also depends on the frequency of seismic waves. With a long‐wavelength record, it is not possible to distinguish a large and a sharp velocity change from small and distributed (smooth) velocity change. (See electronic version for color representation of the figure.)

0004167983.INDD 6 10/11/2018 8:51:02 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 7

velocity step at ~60 km depth that they named the Gutenberg discontinuity. ScS reverberations in continental regions showed no evidence of the 60 km fea-ture. Although inferred as a sharp interface, the seismic deconvolution to detect scattered waves used a low‐pass filter with a 40 mHz corner frequency (25 s period), so that a 30 km velocity gradient would appear “sharp” [Bagley and Revenaugh, 2008]. The low‐passed ScS decon-volution would plausibly miss the thin low‐velocity zone features within continental lithosphere inferred from body‐waveform modeling [e.g., Given and Helmberger, 1980]. A subsequent ScS study by [Gaherty et al., 1999] argued that the Gutenberg discontinuity is prevalent only beneath oceans. Interpreting a suite of tomographic models from surface waves and long‐period shear waves, Gung et al. [2003] proposed that the LAB globally repre-sents a transition from stiff lithosphere into deforming anisotropic asthenosphere, at the Gutenberg discontinuity (60–80 km depth) beneath oceans, and at the Lehmann discontinuity (200–250 km depth) beneath stable conti-nents, see also [Eaton et  al., 2009]. Along this line of investigation, no MLD was detected.

Receiver functions have good potential for quantifying the sharpness of lithospheric interfaces, particularly using the higher frequency signals available in P waves (see also Chapter 8). Crustal reverberations, however, can obscure upper‐mantle interfaces in P receiver functions [Zhu and Kanamori, 2000; Park and Levin, 2016], so attention to epicentral moveout effects (effects on the traveltime differences among different stations in an array measurement) is critical. S‐to‐P scattered phases in Sp receiver functions precede the main S wave, avoiding reverberations.

The most clear evidence for an abrupt change in elastic properties at the base of the lithosphere, or within it, comes from scattered teleseismic body waves, e.g., precur-sors to PP and SS waves from reflections at interfaces prior to a free‐surface bounce‐point, and S‐to‐P and P‐to‐S converted phases at interfaces, arriving respectively as a precursor to the direct S phase or within the coda of the P wave. The frequency content of scattered waves is key to resolving the sharpness of boundaries [Olugboji et  al., 2013]. If scattering is distributed evenly within a transition layer corresponding to the seismic wavelength at oscillation period T, the observed wave amplitude will be dampened by phase cancellation. Scattering within a transition layer corresponding to half a wavelength will  suffer no cancellation, and the observed signal will not easily be distinguishable from a zero‐thickness discontinuity. For an SS precursor of 10 s period scat-tered from the lithospheric mantle (VS ~4.7 km s−1), accounting for two‐way traveltime through a ~24 km gra-dient layer will predict substantial cancellation, and a ~12 km gradient zone will resemble a “sharp” interface.

For receiver functions, the accumulation of Ps delay time corresponds roughly to 10 km s−1 in the upper mantle, leading to substantial cancellation at 1 s period for a 10 km transition layer, and 5 km thickness as “sharp.” For S receiver functions the Sp precursor time accumulates at the same rate, so that a 5 s period Sp phase cancels over a 50 km transition, and a 25 km transition is “sharp.”

1.2.2. Some Examples: Isotropic Velocity–Depth Models

1.2.2.1. Oceanic Upper MantleUsing the traveltime records of refracted and reflected

body waves is a classic technique to determine Earth’s structure [e.g., Lay and Wallace, 1995]. Where there is a low‐velocity layer, however, body‐wave raypaths refract downward through the layer without bottoming, leaving the details of the velocity inversion invisible to the travel-time curve. Indeed, the presence of a low‐velocity layer in the upper mantle (now referred to as the asthenosphere) was inferred indirectly based on the absence of seismic signals at a certain range of distance between the source and the receiver by Gutenberg [1926].

Consequently, surface waves are typically used to infer the velocity structure of the low‐velocity layer such as the asthenosphere. In these studies, records of surface waves for a broad range of frequency are used, and the fre-quency‐dependence of phase velocity is inverted for the depth‐dependent velocity [e.g., Kanamori and Press, 1970; Forsyth, 1975, 1977; Yoshii et  al., 1976]. Study of the structures of deep regions requires the use of long‐ wavelength (low‐frequency) waves. This poses a serious limit for the depth resolution of the velocity structure, as discussed previously (see Figure 1.1).

The studies cited above show a smooth transition from the lithosphere to the asthenosphere, including the age dependence of the transition depth (Figure 1.2). Such a feature, including the age dependence of the structure of the oceanic lithosphere, can be explained by various models that include a smooth reduction of velocity due to temperature increase [e.g., Gueguen and Mercier, 1973; Schubert et al., 1976; Yoshii et al., 1976; Parsons and McKenzie, 1978; Faul and Jackson, 2005; Priestley and McKenzie, 2013; see also Chapter  6] (Figure 1.3).

High‐frequency waves are now used more commonly in investigations of the oceanic upper mantle, partly due to the development of ocean‐bottom seismometers (OBS) [e.g., Shimamura and Asada, 1976] that have the ability to detect subtle but sharp changes in seismic‐wave velocities. Such studies have revealed more details of the velocity–depth profile of the oceanic upper mantle, including the sharp and large velocity drop at the oceanic LAB.

0004167983.INDD 7 10/11/2018 8:51:02 PM

8 LITHOSPHERIC DISCONTINUITIES

Kawakatsu et al. [2009] imaged an interface that dipped with the subducting Pacific Plate beneath northern Honshu, Japan based on the data collected by the land‐based Hi‐Net seismographic network and seafloor OBS stations using both Ps and Sp receiver functions (RF). Expressed as a distinct pulse by using data with periods as short as 3 s, the sharpness of the interface implied a transition no thicker than 10–15 km, suggesting a plate thickness that ranged from 55 km to 80 km, respectively, for plate ages from 20 Ma to 130 Ma. Using long‐running permanent stations Kumar and Kawakatsu [2011] detected the LAB beneath subducting slabs with RFs from the Ryukyu Islands to California around the coastline of the northern Pacific, confirming the depth range, sharpness, and magnitude of velocity drop. Olugboji et  al. [2016] identified the LAB from data collected by ocean‐bottom seismometers, mostly on the Philippine Sea plate, with a multiple‐frequency RF estimator for Ps converted waves that detected variations in interface sharpness from 2 to 10 km thickness.

An example of the velocity profile of the old oceanic upper mantle is shown in Figure 1.4a. In contrast to the results from long‐wavelength (long‐period) seismic waves, where velocity changes smoothly with depth (Figure 1.2), a sharp and large velocity drop is seen at ~80 km depth in the old oceanic upper mantle (~130 Ma) when short‐wavelength (short‐period) seismic waves are used. In the old oceans, the depth at which such a sharp and large velocity drop occurs is nearly independent of the age of

the ocean floor, whereas the depth increases with age in young oceans [e.g., Kumar and Kawakatsu, 2011; Rychert et al., 2012; Olugboji et al., 2016] (Figure 1.4b). A similar structure was also inferred using waveform analysis of differences in multiply reflected waves from the Pacific to infer a sharp and large velocity drop at ~60 km depth [Tan and Helmberger, 2007]. The temperature corresponding to these depths can be estimated by a thermal model of the oceanic lithosphere to be ~800–1000 °C [e.g., McKenzie et al., 2005].

Some more detailed results were reported for the oce-anic upper mantle. Mehouachi and Singh [2018] infer from an equatorial Atlantic marine‐reflection study a low‐velocity channel with 12–18 km thickness and 8.5% drop in Vp, the upper surface of which deepens from 72 km in 40 Ma oceanic lithosphere to 88 km in 70 Ma oceanic lithosphere. In the open Pacific Ocean near the Shatsky Rise, Ohira et  al. [2017] interpret reflectors at ~50 km depth from marine reflection and refraction data as packages of low‐velocity rock associated with what they term “oceanic MLDs.” These features also involve thin layers, likely to be expressed weakly in receiver functions at periods T > 2 s. but the global significance of these detailed structures is unknown.

Global‐scale studies by Tharimena et al. [2016, 2017a,b] report the effect of lithospheric interfaces in precursors to SS waves. With dominant wave periods ~20 s, Tharimena et al. [2017a] reported 3–15% drops in VS over potentially a 21 km transition at depths that subside from 30 to 60–80 km depth from mid‐ocean ridges to ~36 Ma seafloor, and flattens beneath older seafloor. The feature is characterized as “sharp” because thermal variations with depth are too smooth to explain it. Tharimena et al. [2017b] detected similar features in low‐passed stacked SS precursors beneath old continents. These SS precur-sors imply a large velocity drop at the relatively shallow (and cold) regions, and together with the results shown in Figure 1.4 these observations provide key constraints to understanding the origin of the LAB.

1.2.2.2. Continental Upper Mantle By modeling the traveltimes and waveforms of 1–10 s P

waves in the tectonically active southwestern United States, Burdick and Helmberger [1978] favored a VP model with a lithospheric “lid” that overlies a low‐velocity asthenosphere at ~60 km depth. Although some trade‐offs in model parameters were acknowledged, Burdick and Helmberger [1978] argued that their P waveforms had resolved a sharp interface, a feature that we would now call the continental LAB. Benz and McCarthy [1994] con-firmed the LAB sharpness in this region with data from the Polar Anglo‐American Conjugate Experiment (PACE) active‐source field experiment. The velocity drops inferred for the asthenosphere in these and other studies of the 1970s

03.5 4.0

VSv (km s–1)

Age (Ma)0–44–2020–5252–110110+

4.5 5.0

100

200

Dep

th (

km)

300

400

Figure 1.2 A seismic structure of the oceanic upper mantle based on long‐period (20–125 s) surface waves: VSv is the velocity of vertically polarized S‐wave. (After Nishimura and Forsyth [1989].)

0004167983.INDD 8 10/11/2018 8:51:03 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 9

were substantial, often 5–10%. Parsons and McKenzie [1978] argued that these velocity drops were too large to be explained by thermal effects alone, and suggested that partial melt pervades the asthenosphere.

Later studies have provided a rich data set on the struc-tures of the continental upper mantle. These studies have

revealed a large regional variation in seismic structure of the continents. In active regions with extension tectonics (e.g., the Basin and Range region in the United States) the high‐velocity lid (the lithosphere) is thin, whereas in the stable continents the lithosphere is thick [e.g., Grand and Helmberger, 1984; Artemieva, 2011].

Seismic wave velocity (kms–1)

4.5(a)

(b)

5.46 5 8 85 8 850

100

200

1.02

1.01

1

0.99

0.98

0.97

0.96

0.95

0.940 50 100 150 200

Dep

th (

km)

Depth (km)

Vs(

z)/V

s (0)

Vs Vp

150 Ma 150 Ma

AHAH + AN* (no water effect)

AH + AN* (water effect)(* absorption band model)

10 Ma 10 Ma

Figure 1.3 Seismic‐wave velocity versus depth relationship. (a) A model of seismic‐wave velocity in the oceanic upper mantle based on models of the temperature and pressure dependence of seismic‐wave velocity where only the anharmonic effect (effect of thermal expansion) is included. VP is P‐wave velocity, VS is S‐wave velocity. (After Schubert et al. [1976]. Reproduced with the permission of Wiley.) (b) A S‐wave velocity versus depth profile of the oceanic upper mantle showing the influence of various processes: AH, anharmonicity (effect of thermal expansion); AN, anelasticity (effect of viscous dissipation). The influence of AN is included only through the absorption band model. In this case the influence of AN is limited to less than V

VQ 1 and is small (Faul and

Jackson [2005] provided essentially the same results as a model of AH + AN, although we argue in the text that their model of grain‐size sensitivity of seismic‐wave velocity is questionable). (After Karato and Jung [1998]. Reproduced with permission of Elsevier.)

0004167983.INDD 9 10/11/2018 8:51:03 PM

10 LITHOSPHERIC DISCONTINUITIES

Figure 1.5 shows a reference velocity–depth model of stable (SNA) and tectonically (TNA) active regions of the continental lithosphere, as determined by S and SS waves with 5–10 s period [Grand and Helmberger, 1984]. A  small velocity reduction is noted in TNA regions at ~100 km depth and SNA regions at ~150–200 km, how-

ever, the velocity–depth relation is nearly flat in the depth range of 50–150 km in the stable continents in this model.

Thybo and Perchuć [1997] reported a substantially different velocity–depth profile. They performed long‐baseline refraction surveys in stable continental regions to document a loss of sharp‐onset P waves at downrange distances between 8° and 11°, consistent with depressed VP between ~100 km depth and the Lehmann discontinuity in all old continents studied, including Europe, North America, and Eurasia, and suggested that this is a global structure.  Thybo [2006] summarized attempts to model refraction data with complex finite‐difference synthetic seismograms, and argued that small‐scale scatterers per-vade the low‐velocity layer at around ~100 km depth in the stable continents (Figure 1.6a).

Thybo’s [2006] inference has been confirmed by a number of later studies using converted waves (RFs; Figure 1.6b) [e.g., Rychert and Shearer, 2009; Abt et al., 2010]. These studies identified interfaces within the depth range 60–130 km at which velocity drops 6–9%. Ford et al. [2010] and Abt et al. [2010] confirmed the observations of such interfaces across Australia and North America with RFs for Sp and Ps and concluded that they were expres-sions of a MLD distinct from the LAB, which would occur at a greater depth in these regions. Further observa-tions established the global character of the MLD, with RF studies in East Africa [Wölbern et al., 2012], China

3 4VS (km s–1)

50

0

0

20

20

40

40

60

60

Age [Ma]

Dep

th [k

m]

Dep

th (

km)

80

80

100

100

120

120140

140

900°CG ?

1100°CPlate model

Mid-oceanridge

(b)(a)Anomalous crust

Normal crustKawakatsu et al., 2009Kumar and Kawakatsu, 2011Schmerr, 2012Gaherty et al., 1999

Bagley and Revenaugh, 2008

50Mantle lid

LVL 100

Tan and Helmberger, 2007

Rychert andShearer, 2011

Figure 1.4 Seismological models of the lithosphere–asthenosphere boundary. (a) A velocity–depth model for ~130 Ma oceanic upper mantle in the Pacific based on the receiver function (RF) observations (T ~3 s for SP RFs;). The shear‐wave velocity drop at the LAB is 7–8% that occurs sharply (<15 km thickness). LVL, low‐velocity layer. (After Kawakatsu et al. [2009]. Reproduced with permission of American Association for the Advancement of Science.) (b) The LAB depth–age relationship. The LAB is located at a relatively cold isotherm, T = 900 ± 200°C. The LAB depth is dependent on age in the young ocean, but it is only weakly dependent on age in the old ocean. G, Gutenberg discontinuity. (After Rychert et al. [2012]. Reproduced with the permission of Wiley.)

VS (km s–1)

4.0 4.5 5.0

Dep

th (

km)

100

200

300

TNA

SNA

Figure  1.5 Velocity–depth models of the continent (North America) by [Grand and Helmberger, 1984] based on long‐period S and SS waves (5–10 s period): TNA, tectonically active regions (Basin and Range region); SNA, stable cratonic regions.

0004167983.INDD 10 10/11/2018 8:51:04 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 11

[Chen et  al., 2014], Scandinavia [Frassetto and Thybo, 2013], Tibet [Shen et al., 2017], southern Europe [Knapmeyer‐Endrun et al., 2017], Namibia [Yuan et al., 2017], India [Mandal, 2017], and denser coverage in North America [Chen et al., 2018; Hopper and Fischer, 2015].

Calò et  al. [2016] combined RFs and surface‐wave dispersion measurements into a Monte‐Carlo inversion scheme that they applied to data from 30 long‐running seismographic stations in North America—seeking evidence for the LAB and the MLD with RFs and con-straining average seismic velocities with surface‐wave dis-persion. Similar to Gaherty et  al. [1999], this study formulates a general taxonomy of upper‐mantle seismic discontinuities, reporting a number of velocity inversions in the depth range of 60–150 km. The inferred velocity profiles of stations in tectonically stable continent have more and deeper discontinuities than those in actively deforming terranes, in addition to a generally higher average VS.

The joint inversion of RFs and surface‐wave dispersion will likely aid greatly the interpretation of seismic features associated with the LAB and various MLDs. Although upper‐mantle velocity inversions based primarily on sur-face‐wave dispersion [Priestley and McKenzie, 2006; Darbyshire et  al., 2007; Pedersen et  al., 2009; Lin et  al., 2016; see also Chapter  6] typically interpret a shallow LAB as gradational or miss the MLD entirely, dispersion observations can also be fit with low‐velocity depth ranges that are bounded by sharp interfaces [Dalton et al., 2017]. The take‐away message is that surface‐wave dispersion can be fit with either sharp or gradational interfaces, and therefore cannot discriminate them well.

Tharimena et  al. [2016] used SS precursors to infer velocity inversions below the Ontong Java Plateau (OJP) at roughly 80 km and 280 km depths, matching the pattern of the MLD and LAB under continents. The SS stacks for the OJP are modeled with sharp interfaces but could be fit with broader transitions.

1.2.3. Anisotropy

The one‐dimensional Preliminary Reference Earth Model (PREM) [Dziewonski and Anderson, 1981] con-tains an anisotropic layer in the top 220 km where SV wave (vertically polarized S‐wave) velocity is slower than that of SH wave (horizontally polarized S‐wave) velocity. In their model SH/SV radial anisotropy does not change at the depths corresponding to the LAB or the MLD.

In one region of the Pacific upper mantle, Tan and Helmberger [2007] infer SH/SV anisotropy with a sharp upper boundary at the base of their lithospheric lid, but most one‐dimensional seismic models lack azimuthal anisotropy. Surface waves, shear‐wave birefringence and RFs have detected shifts in the orientation of azimuthal anisotropy with depth in the shallow mantle, under both continents [Levin and Park, 2000; Yuan and Romanowicz, 2010a,b; Yuan and Levin, 2014] and oceans [Beghein et al., 2014; Lin et al., 2016]. We note that Babuška et al. [1998], Babuška and Plomerová [1989], Plomerová and Babuška [2010], and Plomerová et al. [2002] also reported depth‐dependent, tilted anisotropic structure in the con-tinent using body‐wave traveltime analyses. Although there is reason to associate the LAB with a transition from a relict anisotropy from lithosphere formation to

1000

MOHO

MLD

LAB

250020001500

A B

1000

Distance (km)

5000

ULM

EY

MN

JFW

S

SLM

SIU

CU

TM

TP

LAL

LRA

L

50

100

150

200

250

6 8

Stable (cold)A

ctive (hot)

VP (km s–1)

10

(a) (b)

200

300

Dep

th (

km)

Dep

th (

km)

Figure 1.6 High‐resolution seismic models of the stable continents. (a) A schematic model to explain traveltime anomalies. (From Thybo and Perchuć [1997]. Reproduced with permission of American Association for the Advancement of Science.). (b) The amplitude of conversion of waves in the cratonic region of the United States. The blue region indicates positive velocity–depth gradient, a red region indicates negative. There is a marked presence of a low‐velocity anomalies at ~100 km depth, whereas in the depth regions where the LAB is expected (~200 km), there is no evidence for a velocity drop. (From Abt et al. [2010]. Reproduced with the permission of Wiley.) (See electronic version for color representation of the figure.)

0004167983.INDD 11 10/11/2018 8:51:04 PM

12 LITHOSPHERIC DISCONTINUITIES

anisotropy associated with present‐day asthenospheric shear, the sharpness of the anisotropic transition is not well constrained. Wirth and Long [2014] and Park and Levin [2016] report cases where RFs detect both a MLD and a change in anisotropy, but at distinct depths.

1.2.4. Attenuation

Different models have different links between velocity reduction and attenuation, and therefore observations on seismic wave attenuation combined with velocity models help identify the mechanisms for the LAB and the MLD. In order to understand how to use the observations on attenuation to evaluate models, let us review the basics of attenuation [e.g., Karato, 2008a; Jackson, 2015].

Attenuation is often characterized by a quantity called Q: Q−1 provides a measure of attenuation (Q = 100 means 1% of energy is lost during one cycle of wave propagation). At a very general level, there is a relation between elastic and anelastic parts of wave propagation, i.e., the Kramers–Kronig relation, but the actual relationship between anelasticity (attenua-tion) and elasticity depends on the particular physical mechanism. To see this, let us consider two cases (Figure 1.7). Figure 1.7a is a case of a standard linear solid corresponding to a single peak in anelastic relax-ation, which corresponds to a single mechanism with a single characteristic time. In this case, low velocity could correspond either to low or high attenuation depending on the frequency.

Figure 1.7b is a case where there is a broad distribution of relaxation peaks leading to a weak frequency

dependence of attenuation and velocity (Q , 0 1), the “absorption band model”. In this case, the velocity anomaly is related to anelasticity as [e.g., Minster and Anderson, 1981]

VV

Q T P C d Q T P C d12 2

1 1cot ,, , , , ,W W

(1.1)

where ΔV is the magnitude of velocity reduction, V is the velocity at infinite frequency, T is temperature, P is pressure, Cw is water content, and d is grain‐size. Most previous analyses of velocity and attenuation for the lithosphere–asthenosphere system assumed the absorption band behavior [e.g., Gueguen and Mercier, 1973; Karato and Jung, 1998; Faul and Jackson, 2005; Priestley and McKenzie, 2013].

In this scenario, low velocity always corresponds to high attenuation. Most laboratory studies show this behavior [Jackson, 2015], and many seismological obser-vations also follow this relation [e.g., Anderson and Minster, 1979; Shito et  al., 2004]. One important consequence of this relation is that it (with 0.3) predicts a small influence of anelasticity on velocity anomalies. Therefore an absorption‐band model cannot explain a large velocity drop at the LAB (or MLD).

There are two cases, however, where power‐law dependence in Q (i.e., the absorption band model) is not valid: anelasticity associated with partial melting [Jackson et al., 2004] and EAGBS [Jackson and Faul, 2010; Sundberg and Cooper, 2010; Karato, 2012; Karato et  al., 2015].

(a) (b)

V(ω)

Q–1(ω)

V(ω)

Q–1(ω)

Frequency (ω)Frequency (ω)

An absorption band modelA standard linear solid

Figure 1.7 Two models of anelastic behavior: (a) a standard linear solid that corresponds to the case of a single peak; (b) absorption band model that corresponds to a case of distributed peaks. The majority of seismological and laboratory observations are consistent with this model, but some exceptions are also observed.

0004167983.INDD 12 10/11/2018 8:51:05 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 13

In  both cases, attenuation has a peak at a certain fre-quency for a fixed temperature (or at a certain tempera-ture for a fixed frequency). Therefore detection of such a peak will help us identify the mechanisms of formation of the LAB or the MLD.

Let us now summarize some key seismological observations on attenuation (Figure  1.8). We note first that measurements of attenuation are far more challenging than those of seismic velocities because distinguishing true (intrinsic) attenuation from scattering (extrinsic) attenuation or other geometrical effects is not straightforward [e.g., Fehler et al., 1992; Jin et al., 1994; Romanowicz and Mitchell, 2007]. but we consider that the following observations are robust and important.

1. Generally, attenuation is stronger in the deeper upper mantle (~100–300 km). In the asthenosphere QS (Q for S‐wave) ~50–100 [e.g., Dziewonski and Anderson, 1981].

2. Attenuation in the oceanic lithosphere is very small: QS = 600 [Dziewonski and Anderson, 1981]; QS = 3200

[Takeuchi et  al., 2017], but substantial attenuation (QS = 100–300) is reported in the continental lithosphere [e.g, Mitchell, 1995; Dalton et al., 2009].

3. In the asthenosphere, attenuation follows the power law relations, which is consistent with absorption band behavior [Shito et al., 2004].

4. There is evidence for an attenuation peak in the asthenosphere when compared with attenuation in the lithosphere [e.g., Takeuchi et al., 2017].

Observation 4 suggests either EAGBS or a localized melt‐rich region both of which would show a peak in attenuation within a thin layer near the LAB. Also, observations of modestly high attenuation (QS = 100–300) in the continental lithosphere are surprising when considering that temperatures in the continental litho-sphere are generally lower than those of the oceanic lithosphere [Sclater et al., 1981]. One possible explana-tion is the presence of an attenuation peak in the continental lithosphere, as we shall discuss later in section 1.6.

0.025

0.02

0.015

0.01

Atte

nuat

ion

(1/Q

)

0.005

0.1

(a) (b)

0.01

PREM

PREMYang

Asthenosphere

Booth

QRFSI12

QRFSI12

LithosphereThis Study

This Study

1010.10.01

Frequency (Hz)

0.001

Qs–1

f–0.2

f–0.4

0.0001

04.84.64.44.2

100 km

DataYoung oceansMid-age oceansOld oceansOrogenic/magmaticOld continents

Velocity (km s–1)

Figure 1.8 Anelasticity (intrinsic QS) of the upper mantle inferred from seismological models. (a) A global model based on 50–250 s Rayleigh waves. The data for old continents (~100 km) are for the lithosphere showing QS = 100–300 (similar QS values are reported by Dalton et al. [2017]). The data for oceans are for the asthenosphere and QS = 50–100. (After Dalton et al. [2009]. Reproduced with permission of Elsevier.) (b) A model of the oceanic upper mantle based on both 100 s surface waves and P and S waves near 3 Hz: QS for the oceanic lithosphere is ~1000 depending on the frequency f; QS for the oceanic asthenosphere is less frequency dependent, with QS ~50–100. QRFSI12 is the attenuation model proposed by Dalton et al [2009]. The dashed lines marked “Yang” and “Booth” are attenuation models proposed by Yang et al. [2007] and Booth et al. [2014], respectively, plotted over the frequency ranges of their analyzed data. There is a hint of a peak in attenuation at a few Hz in the asthenosphere. QRFSI12 is the attenuation model proposed by Dalton et al [2009]. The dashed lines marked “Yang” and “Booth” are attenuation models proposed by Yang et al. [2007] and Booth et al. [2014], respectively, plotted over the frequency ranges of their analyzed data. (After Takeuchi et al. [2017]: referred to as “This study” in the figure. Reproduced with permission of American Association for the Advancement of Science.)

0004167983.INDD 13 10/11/2018 8:51:05 PM

14 LITHOSPHERIC DISCONTINUITIES

1.3. GEOLOGICAL/PETROLOGICAL OBSERVATIONS RELEVANT TO THE LAB AND MLD

1.3.1. The LAB in the Oceanic Upper Mantle

The plate‐tectonics model for the formation of oceanic lithosphere provides an important framework within which any model for a seismic discontinuity must be com-patible. Oceanic lithosphere is formed at a mid‐ocean ridge where hot materials rise and undergo partial melting. After reaching the thermal boundary layer, hot mantle rock diverts to near‐horizontal flow and gradually cools. This cooled and hence mechanically strong region is the oceanic lithosphere.

In addition to cooling, oceanic lithosphere develops a distinct chemical composition from the partial melting under the mid‐ocean ridge. The nature of this partial melting is well understood based on experimental studies as well as comparing the chemical compositions of mid‐ocean ridge basalt and xenoliths from the oceanic upper mantle [e.g., Ringwood, 1975]. The major component of oceanic lithosphere is harzburgite, a depleted peridotite that contains less FeO, Al2O3, and CaO than primitive peridotite, as well as less water.

The changes in the major‐element composition listed above do not result in substantial velocity changes (Figure  1.9) [e.g., Matsukage et  al., 2005; Schutt and Lesher, 2006]. In contrast, a change in the water content

4.531 GPa (MM3)

3 GPa (KR4003)

3.5 GPa (MM3)

4 GPa (KR4003)

4.5 GPa (KR4003)

6 GPa (KR4003)

7 GPa (KR4003)

4.494.444.404.35

4.554.514.464.424.37

4.584.534.494.444.40

4.554.514.464.424.37

4.53

VS (

km s

–1)

VS change (%

)

4.494.444.404.35

4.584.534.494.444.40

4.594.544.504.454.41

0 5 10 15 20Melt removed (%)

25 30 35

210–1–2

210–1–2

210–1–2

210–1–2

210–1–2

210–1–2

210–1–2

Figure 1.9 The influence of compositional change due to melt extraction on shear‐wave velocity. After partial melting, major element composition (mineralogy) of the residual rocks changes (Mg# increases, garnet content decreases), but these changes largely cancel out and the net effect of compositional change is small (< ~1%). More drastic changes in composition are required to change the seismic velocity more than a 1–2%. (After Schutt and Lesher [2006]. Reproduced with the permission of Wiley.)

0004167983.INDD 14 10/11/2018 8:51:05 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 15

could result in a large change in seismic velocity and attenuation [e.g., Karato, 1995, 2003, 2012; Karato et al., 2015], but this has not been confirmed by laboratory studies and will be discussed in detail later in section 1.6.

1.3.2. Composition and Evolution of the Continental Lithosphere

The composition of the continental lithosphere can  be inferred from the rocks carried from its deep interior by ascending melts, called “xenoliths” [e.g., Pearson et al., 2003]. In the oceanic regions, xenoliths are available only from shallow depths < 60 km, [e.g., Nicolas, 1989; Peslier et  al., 2010]. From the continents,  there are many xenoliths that originate from depths to ~200 km. In isolated cases, xenoliths from the transition zone or the lower mantle have been identified [Collerson et al., 2000; Pearson et al., 2014; Kaminsky et al., 2015; Nestola, 2017]. We shall focus on xenoliths coming from the lithosphere, down to ~200 km.

Processes that control the composition and mineralogy of continental mantle rocks are more complex than those of the oceanic upper mantle [e.g., Lee, 2003; Carlson et al., 2005; Schutt and Lesher, 2006]. Consequently, there

is a larger degree of chemical and mineralogical variety in rocks in the continental upper mantle. As far as the compositions of dominant peridotites are concerned, however, the influence of composition on seismic velocity is minor (Figure 1.9).

Exceptions are the rocks formed from the freezing of a melt. If a small degree of partial melting occurs, then a large fraction of incompatible elements goes to the melt, including Fe and H (FeO and H2O) [e.g., Ringwood, 1975]. If these melts solidify to form rocks, then seismic velocity could be substantially lower than the original rocks because of high FeO content and a large amount of hydrous minerals; a point discussed later in section 1.6.

Another important constraint on the causes of MLDs is the distribution of mantle‐rock ages [Carlson et  al., 2005], which elucidate the evolution of conti-nents in general. Figure 1.10 displays examples of Re–Os dating studies showing that the ages of mantle rocks drawn from the cratonic upper mantle do not vary drastically with depth and that the ages do not differ greatly from those of the crust. This implies that after its formation, most of the continental lithosphere has remained intact. Therefore, if extensive metasoma-tism were to be invoked, it must have occurred before the continental crust was formed in the Archean Eon (before 2.7 Ga).

2.0–0.9 Ga

Archean kaapvaal craton

Proterozoic Proterozoic

West

SpinelGarnet

Diamond

EastFL

J F K

0.07 0.12 0.151.180.09 0.09

0.09

P NLR A

Crust 2.1–0.9 Ga 3.6–2.7 Ga

2.02.2

2.7

2.7

2.72.7 2.2

2.3 2.0 3.0

1.93.3 2.2

1.61.91.22.21.8

2.61.9

2.7

2.52.2

2.32.3

2.5

1.2 2.8

250km

3.10.9

1.50.5

2.4

2.12.10.61.8

2.0

1.73.5

2.32.6

2.72.72.8

1.6

1.6

1.21.70.81.81.9

Low-T

High-T

Graphite

SpinelGarnet2.0

2.2

2.72.72 72.7 2.2

9.91 91 91.93.3 2.2

1.61.91 2

2.61.9

2.7

2.52 2 2 3

1.21.70 8 Low-T

MLD

Figure 1.10 Distribution of ages of mantle rocks in South Africa determined by Re–Os isotope measurements. The numbers are the ages (in Ga) determined by Re–Os isotope systematics. In most regions, Re–Os ages in the mantle rocks agree roughly with the ages of crust and show small depth variation. Although there are some depth variations in the age, there is no clear trend to show younger ages at or near the MLD. (After Pearson [1999]. Reproduced with permission of Elsevier.)

0004167983.INDD 15 10/11/2018 8:51:06 PM

16 LITHOSPHERIC DISCONTINUITIES

1.4. MODELS FOR THE LAB AND THE MLD

1.4.1. Partial Melting

Partial melting is an obvious mechanism to reduce seismic wave velocities [e.g., Mizutani and Kanamori, 1964; Spetzler and Anderson, 1968; Anderson and Spetzler, 1970; Lambert and Wyllie, 1970], but the following condi-tions need to be met for this model to explain the low velocities observed: (a) the physical conditions must be consistent with the presence of melt; (b) the melt fraction must be sufficient. The critical melt fraction to explain the observed velocity reduction depends greatly on the geometry of melt in a rock [Stocker and Gordon, 1975; Takei, 2002]. For a typical melt geometry in the shallow upper mantle corresponding to the dihedral angle of 20–50° [e.g., Toramaru and Fujii, 1986; Yoshino et  al., 2009], 2–3% (3–5%) of melt is needed to explain ~2–5% (5–10%) of velocity reduction (Figure 1.11). (Holtzman [2016] recently proposed that even a very small amount of melt (say ~0.1%) can substantially reduce the seismic wave velocity—we will evaluate this claim in section 6.)

Maintaining a few percent of melt is not easy in most of the asthenosphere [e.g., Hirschmann, 2010; Karato, 2014], and there are a couple of reasons for this.

1. In most cases, partial melting in the asthenosphere depends on volatile components (water, carbon dioxide); exceptions are the shallow upwelling regions (<50 km) beneath mid‐ocean ridges and hotspots, where decom-pression melting occurs. The amount of volatiles deter-mines the amount of melt produced in a closed system (batch melting), and typically is less than ~0.3%.

2. When melt migrates, then the amount of melt at any point is controlled by the dynamic balance of melt gener-ation and migration, and as melt migration is easy, the amount of melt is small; in the upper mantle beneath a mid‐ocean ridge it is estimated to be ~0.1% [Spiegelman and Elliott, 1993; Karato, 2014].

For a partial‐melt model to explain a large velocity drop, then some mechanisms are needed to accumulate melt. One possibility is melt accumulation at the LAB [e.g., Hirschmann, 2010] (Figure  1.12a), and another is layering of melt‐rich regions by deformation [e.g., Kawakatsu et al., 2009] (Figure 1.12b); the plausibility of these mechanisms will be discussed in section 1.6.

1.4.2. Chemical/Mineralogical Layering

Layering in the major element composition and/or mineralogy is an obvious mechanism to explain a sharp velocity change. For this mechanism to be a good model for the LAB or the MLD, there must be a large composi-tional contrast at that boundary. Existing experimental studies, however, show that seismic‐wave velocities are only weakly dependent on major‐element chemistry and mineralogy as far as typical ranges following partial melting are assumed (Figure  1.9) [e.g., Lee, 2003; Matsukage et al., 2005; Schutt and Lesher, 2006].

There is scant major‐element change in composition expected for the LAB, and therefore this mechanism is unlikely to be a good model. For the MLD, a large con-trast in chemistry and/or mineralogy generated by the complex processes of continent formation could be invoked. Behn and Kelemen [2006] discussed amphibole enrichment within igneous rocks of deeply exposed arc terranes, which are one of the building blocks of conti-nents. Rader et  al. [2015] proposed that metasomatism has occurred globally in the deep upper mantle (~200 km or so), producing volatile‐rich melt. The melt is assumed to be buoyant and to migrate upward until it reaches its solidus. The melt would solidify as mantle cools to pro-duce volatile and FeO‐rich rocks at the observed depth of the MLD (Figure 1.13).

For this model to explain the MLD, three conditions must be met: (a) the velocity drop observed at the MLD (2–6%) must be consistent with the compositional varia-tion; (b) the depth of compositional variation must agree with the depth of the MLD; (c) the compositional

1

0.9

0.8

0.7

0.6

0.50 2 4 6 8 10

Melt fraction (%)

0.01 0.02

0.0050.1

μ/μ 0

0.15

0.20.3

α = 0.5

θ = 20°40°

Dihedral angle

80°60°

Slope

Oblate spheroidmodel

Asthenosphere

AN

0.05

Figure  1.11 Relation between melt fraction and shear mod-ulus (μ) of a partially molten rock: μo, shear modulus of a rock without melt; θ, dihedral angle between the melt and minerals (=20–40o for asthenosphere materials). In order to explain 5–10% velocity reduction (10–20% shear modulus reduction),

a melt fraction of 3–5% is needed (VS , ρ is density).

(Modified from Takei [2002]. Reproduced with the permission of Wiley.)

0004167983.INDD 16 10/11/2018 8:51:06 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 17

variation occurs globally at similar depths. Corollary predictions of this model must be consistent with the rel-evant observations, which include an increase in electrical conductivity across the MLD and geochemical traces of this wet mineralogy in mantle xenoliths. We will review these issues in section 1.6 in detail.

1.4.3. Layering in Anisotropy

If there were a sharp change in anisotropy within the upper mantle with scarce compositional changes, it would resemble a sharp discontinuity in isotropic velocity with respect to wave refraction and conversion, but properties would vary with the direction of wave propagation (Figure  1.14). Evidence for a change in anisotropy has been reported by, for example, Yuan and Romanowicz [2010b] and Sodoudi et al. [2013].

When anisotropy is caused by lattice‐preferred orienta-tion (LPO) of minerals, layering in anisotropy may occur either by a change in the flow geometry or in the defor-mation fabrics, or both. Anisotropy may also be due to

shape‐preferred orientation (SPO) (of a melt‐rich materials), as suggested by Holtzman et  al. [2003] and Kawakatsu et al. [2009].

Yuan and Romanowicz [2010b] suggested a change in azimuthal anisotropy with depth corresponding to the change in flow geometry, and proposed that this is a mechanism for a MLD. Kawakatsu et al. [2009] suggested that layering in SPO is a cause of the LAB. The validity of such models for the LAB and the MLD is discussed in section 1.6.

1.4.4. Temperature Effects

Low seismic velocities in the asthenosphere have been explained classically by the high homologous tempera-ture [Birch, 1952; Schubert et al., 1976]. Physical disper-sion from anelastic effects amplifies this [Gueguen and Mercier, 1973; Faul and Jackson, 2005]. These models, however, fail to explain the sharp velocity drop at the LAB (and the MLD) demonstrated by studies using short wavelength body waves. Temperature increases with

(a)

(b)

Lithosphere

Lithosphere

Figure 1.12 Possible mechanisms of melt accumulation that might explain the large velocity reduction in the asthenosphere: (a) melt accumulation at the LAB; (b) melt‐rich layers in the asthenosphere. (After Hirschmann [2010]. Reproduced with permission of Elsevier.)

0004167983.INDD 17 10/11/2018 8:51:06 PM

18 LITHOSPHERIC DISCONTINUITIES

depth smoothly, and therefore velocity drop caused by temperature is gradual and could not by itself explain the sharp velocity drop shown by RF studies (Figure 1.3).

1.4.5. Temperature and Water Effects

In order to explain a sharp velocity drop, a sharp change in some factor that has a large effect on seismic‐wave velocity would need to be invoked. Knowing that the major element chemistry of the upper mantle does not change seismic velocities much [e.g., Lee, 2003; Matsukage et al., 2005], a possibility is layering in water content. This layering likely arises from the large contrast in water (hydrogen) solubility between minerals and melt, and to the slow diffusion of hydrogen. A large contrast in water solubility causes depletion of water in the solid upon partial melting [e.g., Karato, 1986; Hirth and

Kohlstedt, 1996]. At mid‐ocean ridges the rate of melting increases rapidly at ~65 km depth, so there will be a sharp contrast in water content in a rock at around this depth (water‐poor above this depth, and water‐rich below). Hydrogen diffusion is slow (diffusion distance is ~1–10 km for ~100 Myr) and therefore this layered structure will be maintained for most of the lifetime of an oceanic plate.

Using a close link between plastic deformation and anelastic relaxation [e.g., Karato and Spetzler, 1990], Karato and Jung [1998] suggested that a sharp change in the water content might explain a sharp velocity drop at the LAB. If restricted to the absorption band model, however, the magnitude of a velocity drop by this mecha-nism is limited, as seen from equation (1.1) [e.g., Minster and Anderson, 1981; Karato, 1993]. If Q is 80 [Dziewonski and Anderson, 1981] in the asthenosphere, the velocity drop is ~1% or less (Figure 1.3), which is far smaller than the observed velocity drop (5–10%).

Active margin (young)

Craton (ancient)

Shear velocity

Dep

th

Figure 1.13 A frozen‐melt model of the MLD. In this model, it is assumed that metasomatism occurred in the deep continental upper mantle at the global scale. The hydrous melts thus formed migrated upward until they reached the place where temperature is below the solidus (~1000°C). Minerals with low seismic‐wave velocities (e.g., amphibole, phlogopite, FeO‐rich pyroxenes) are formed to cause the velocity drop at the MLD. If this model is correct, then the age of rocks (frozen melt) near the MLD should be younger than the ages of rocks below and above, and rocks from/near the MLD should contain a large amount of low‐velocity minerals such as amphibole, phlogopite, or FeO‐rich pyroxenes at the global scale. (After Rader et al. [2015]. Reproduced with permission of Elsevier.)

0004167983.INDD 18 10/11/2018 8:51:07 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 19

1.5. ELASTICALLY ACCOMMODATED GRAIN‐BOUNDARY SLIDING MODEL

The above difficulties led Karato [2010] to invoke “elastically accommodated grain‐boundary sliding” (EAGBS) as an alternative model to explain a sharp and large velocity drop at the oceanic LAB. When EAGBS operates, then the close link between attenuation and velocity reduction predicted by the absorption band model (equation 1.1) is no longer valid, and the magni-

tude of velocity reduction can exceed VV

Q 1.

EAGBS is a well‐established physical process that occurs during the deformation of a polycrystalline material. The basic physics were established in the 1940s, and some modifications have been made on theory in the subsequent years. Experimental studies have also been

made in metals, oxides, and silicates. In this section, we will provide a brief review of this model. One major advantage of this model is that it predicts a relatively large velocity drop (a few to 10%) at a modest temperature. Therefore if this process occurs sharply enough, it could explain the key features of LAB (and MLD), namely a large and sharp velocity reduction at a modest tempera-ture on the global scale. In the following, we review the basics of EAGBS.

1.5.1. Deformation of a Polycrystalline Material: the Role of Grain‐Boundary Sliding

When a deviatoric stress is applied to a polycrystalline material, each grain will be deformed corresponding to the local stress. When chemical bonding among atoms is strong everywhere, then the magnitude of all atomic dis-placement is small, and deformation is instantaneous and recoverable, i.e., elastic. In such a case, the elastic constant of a polycrystalline material is some average of the elastic constants of individual mineral grains.

At high temperatures and low frequencies, however, grain boundaries behave like a viscous fluid and fail to sustain shear stresses. Consequently, large stress concentration will occur at the vertices of crystalline grains, leading to excess strain and, effectively, the reduction of elastic constants. Since grain‐boundary sliding dissipates energy, seismic waves attenuate when this process occurs.

The processes of grain‐boundary sliding are schemati-cally illustrated in Figure  1.15. The key point is that grain‐boundary sliding is accommodated in different manners at different stages of deformation. In the initial small strain deformation, accommodation is elastic, leading to high local stress. This EAGBS leads to a peak in attenuation and a reduction in elastic constant (seismic‐wave velocity). The local high stress caused by elastic accommodation is gradually relaxed by diffusional mass transport, leading to a distributed relation time and weakly frequency‐dependent attenuation (the absorption band behavior). Finally, stress distribution and diffu-sional accommodation will produce a balance, leading to steady‐state diffusion creep. Changes in elastic‐wave velocity (elastic constant) and attenuation associated with these processes are shown schematically in Figure 1.16 as a function of frequency of elastic waves.

As illustrated in Figure  1.15, the degree to which an elastic constant is reduced by grain‐boundary sliding is determined by the degree of stress concentration at the mineral scale, which in turn is determined by the grain‐boundary shape, that is, the orientation of each grain‐boundary with respect to the applied stress. Consequently, the magnitude of reduction in elastic constant (seismic‐wave velocity) due to EAGSB is independent of grain‐size.

VSv

VSv

VSv

VSh> VSv VSh< VSv

VSh> VSv

VSh = VSv

VSh = VSv

VSh> VSv or VSh> VSv

VSh> VSv or VSh> VSv

VSh< VSv

or

or

or

(a)

(b)

(c)

Figure  1.14 Schematic diagram showing the role of layered anisotropy in causing the velocity reduction. (a) Layering in the azimuthal anisotropy. The velocity change at the interface can be either positive or negative depending on the nature of change in anisotropy but also on the orientation of seismic waves. (b) Layering in the radial anisotropy (isotropic layer in the deeper part). (c) Layering in the radial anisotropy (isotropic layer in the shallow part). (After Karato et  al. [2015]. Reproduced with the permission of Springer.)

0004167983.INDD 19 10/11/2018 8:51:07 PM

20 LITHOSPHERIC DISCONTINUITIES

hy

τaτa

σnσn

y

y y

y

y

420

–2–4

τa

σn

4 Boundary diffusion Boundary diffusion

20

–2–4

τa

σn

420

–2–4

τa

σn

420

–2–4

Elastic accommodation

Diffusional accommodation

Grain-boundary shape

Volume diffusion Volume diffusion

τa

σn

(a)

(b)

(c)

Figure  1.15 Schematic diagrams showing the processes of accommodation associated with grain‐boundary sliding: σn, normal stress (positive = tension); τa, applied shear stress. (a) Grain‐boundary shape. (b) Stress distri-bution for elastic accommodation. When grain‐boundary sliding is accommodated by elastic deformation, stress concentration occurs at grain‐corners. Sliding eventually stops when stress concentration balances applied stress. (c) Stress distribution for diffusional accommodation. High stress is relaxed by diffusional mass transport. Steady‐state stress distribution is shown that corresponds to steady‐state diffusional creep. Gradual transition from elastic to diffusional accommodation leads to transient diffusional creep associated with weakly frequency‐dependent anelasticity absorption‐band behavior. (After [Raj and Ashby, 1971]. Reproduced with the permission of Springer.)

Frequency

Atte

nuat

ion,

Q–1

Sei

smic

wav

e ve

loci

ty, V

Q–1(ω)

V(ω)

∆V

High water contenthigh temperature

Maxwell body

Absorptionband behavior

Elasticallyaccommodatedgrain-boundary sliding

ωEAGBS

Figure 1.16 A schematic diagram showing the frequency dependence of seismic‐wave velocity (V) and attenuation (Q 1). (From Karato [2012]. Reproduced with permission of Elsevier.). (See insert for color representation of the figure.)

0004167983.INDD 20 10/11/2018 8:51:08 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 21

This can be seen in the following equations for the change in shear modulus by grain‐boundary sliding [Zener, 1941; Raj and Ashby, 1971; Ghahremani, 1980],

relaxed relaxedVV

Zener2

25

7 57 4

1941,

(1.2a)

relaxed relaxedVV

Raj a

21

0 57 1 1.

nnd Ashby,1971

(1.2b)

relaxed relaxedVV

20 86 0 83

11

1 14. .

. 00 83

1980

.

,Ghahremani

(1.2c)

where μrelaxed is the shear modulus of an aggregate relaxed by grain‐boundary sliding, is the shear modulus corresponding to pure elastic deformation, and ν is Poisson’s ratio. Note that these equations do not con-tain grain‐size.

The transition from unrelaxed (high frequency) to relaxed (low frequency) states is associated with a peak in attenuation of seismic waves (Figure 1.16). The frequency at which this peak occurs is dependent on grain‐size as [e.g., Nowick and Berry, 1972],

EAGBS

GB W, ;d T P C (1.3)

where δ is grain‐boundary thickness, d is grain‐size, and ηGB is grain‐boundary viscosity. The presence of a peak frequency for attenuation is a unique feature of velocity reduction by EAGBS and therefore a potential fingerprint of the process.

1.5.2. Experimental Observations on Anelasticity Including EAGBS

1.5.2.1. EAGBSEAGBS was first observed for Al (metal) [Kê, 1947],

where both energy dissipation and modulus reduction were measured. Pezzotti [1999] measured energy dissipa-tion in Al2O3 and MgO but not the modulus reduction. (Modulus reduction can be calculated from energy dissi-pation using the Kramers–Kronig relationship to show that the modulus reduction in these materials is on the order of a few percent.) Two papers were published on

upper mantle materials: Jackson and Faul [2010] (see also [Jackson et al., 2014]) on olivine and Sundberg and Cooper [2010] on an olivine + orthopyroxene mixture. Both energy dissipation and modulus reduction were measured in these studies. For olivine‐rich samples with grain‐size 5–15 µm, a peak in attenuation was observed at around 900–1000 °C, associated with a substantial modulus reduction (Figure 1.17). Jackson and Faul [2010] reported ~7% reduction in shear modulus, while Sundberg and Cooper [2010] reported ~30% reduction.

EAGBS peaks are observed as a small addition to “high‐temperature background” (absorption‐band behavior), so the detailed nature of the anelasticity is dif-ficult to parameterize simply. The reasons for differing modulus reduction between different laboratory studies are unknown. In addition, although theoretical models suggest that the magnitude of velocity reduction is independent of grain size, some dependence on this parameter was seen in experiments.

1.5.2.2. Influence of Water A strong influence of water content on the characteristic

frequency is suggested by experimental observations of a close link between anelasticity and creep, although direct tests of the influence of water on anelasticity have not been conclusive. A study on a natural sample by Aizawa et  al. [2008] observed water to strongly enhance anelas-ticity, while a recent study on synthetic olivine aggre-gates in the Ti‐dominated regime showed small effects [Cline et al., 2018].

Both theoretical consideration and experimental results suggest a close link between microcreep (transient creep) and anelastic relaxation [Karato and Spetzler, 1990; Webb et  al., 1999; Tan et  al., 2001; Webb and Jackson, 2003; Faul and Jackson, 2015]. Also, theoret-ical and experimental studies showed a close connec-tion between transient and steady state creep [e.g., Amin et  al., 1970]. Indeed McCarthy et  al. [2011] suggested a direct link between steady state creep and  anelasticity for an organic material (borneol). Experimental studies showed a substantial effect of water (hydrogen) to enhance steady state creep in olivine [e.g., Karato et  al., 1986; Mei and Kohlstedt, 2000a,b; Karato and Jung, 2003]. Consequently, it is natural to expect that water enhances anelastic relaxa-tion in olivine [Karato, 1995, 2003].

There are a few studies on a natural dunite (olivine‐rich rock) that support this notion [e.g., Jackson et al., 1992; Aizawa et al., 2008]. Interpretation of these earlier studies is complicated, however, because a natural rock was used and so the chemical environment was not fully controlled. In contrast, Cline et  al. [2018] reported results of an experimental study on the influence of water on anelas-ticity where synthetic olivine polycrystals were used. In

0004167983.INDD 21 10/11/2018 8:51:08 PM

22 LITHOSPHERIC DISCONTINUITIES

this study they dissolved hydrogen in their samples in combination with Ti, and explored a range of hydrogen (and Ti) content as well as oxygen fugacity. They found that the influence of hydrogen content is relatively small compared to the influence of oxygen fugacity. This is a surprising result because the influence of water (hydrogen) content (water fugacity) on steady state creep of olivine is

known to be substantially larger than that of oxygen fugacity [e.g., Bai et al., 1991; Karato and Jung, 2003].

The reason for this surprising result is not well under-stood. One possibility is that the hydrogen‐related defect structure and mobility in the Ti‐dominated regime are different from the intrinsic regime where the concen-tration of hydrogen‐related defects is independent of the

0 1 2 3 4–2 –1 0 1 2 3 4–2 –1 0 1 2 3 4–2 –1

Log 1

0 Q

–180

(a)

(b)

(a) (b)

70

60

50

40

30

20

10

0

0

–0.5

–1

–1.5

–2

–2.5

0 1 2 3 4–2 –1 0 1 2 3 4–2 –1 0 1 2 3 4–2 –1

σ = 1.5,Δρ= 7.4%,

log τρr (S) = –3.12,

Eρ= 320 kJ mol–1

σ = 2.5,

Δρ= 7.4%,log τρr (S) = –3.12,

Eρ= 320 kJ mol–1

σ = 2.5,

Δρ= 12.0%,log τρr (S) = –3.7,

Eρ= 320 kJ mol–1

G (

GP

a)

Log10 To (S)

1073 K1123 K1173 K1223 K

0.4

0.2

0.0

–0.2

–0.4

–0.6

–0.8

–1.0–2.5 –2.0 –1.5 –1.0 –0.5

Log10 (frequency (Hz)) Log10 (frequency (Hz))0.0 –2.5 –2.0 –1.5 –1.0 –0.5 0.0

Ol-39 vol % opx Ol-39 vol % opx

Log 1

0 Q

G–1

She

ar m

odul

us (

GP

a)

40

35

30

25

20

15

10

5

0

1200°C1250°C1300°C

1200°C1250°C1300°C

Figure 1.17 Experimental results for EAGBS. (a) On olivine. (From Olugboji et al. [2013] based on the results by Jackson and Faul [2010]. Reproduced with permission of Elsevier.) (b) On olivine + orthopyroxene. (From Sundberg and Cooper [2010]). (Reproduced with permission from Taylor&Francis.)

0004167983.INDD 22 10/11/2018 8:51:09 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 23

concentration of impurities (such as Ti). The Ti content in the mantle olivine is small (~20–30 ppm wt (TiO2) = ~30–50 ppm Ti/Si [Pearson et  al., 2003]), much less than the inferred water content in the asthenosphere (~100 ppm wt (H2O) = ~1500 ppm H/Si [Hirschmann, 2006]). Consequently, hydrogen‐related defects in Earth likely occur in a regime that differs from that studied by Cline et  al. [2018]. We expect that the influence of hydrogen on physical properties such as anelasticity is different among different regimes. A study to assess the differing roles of hydrogen is needed to evaluate the rele-vance of the results by Cline et al. [2018] to Earth’s upper mantle. In addition, the study by Cline et al. [2018] is on the absorption band regime and the influence of water on EAGBS has not been studied.

1.6. DISCUSSION

1.6.1. Partial Melt Model Versus Subsolidus Models for the LAB

Difficulties of the partial melt model for the MLD are obvious because of the low temperature involved, and we will not add any discussions here. The possible role of partial melting as the cause for the LAB, however, does need further scrutiny. We shall discuss three issues: (a) velocity reduction and the melt distribution in the asthenosphere, (b) electrical conductivity, and (c) seismic‐wave attenuation in the asthenosphere. The two prop-erties (b) and (c) are sensitive to partial melting, temperature, and water content.

1.6.1.1. Seismic Wave Velocity and Melt Distribution In section 1.4.1 (Figure 1.11) we discussed that in order

to reduce seismic‐wave velocity by a few percent, a sub-stantial amount of melt (2–5%) is needed. Holtzman [2016], however, discussed that even a small melt fraction (~0.1 %) could reduce seismic wave velocity by a few per-cent, an idea based on a theoretical model by Takei and Holtzman [2009] which suggests that even a small amount of melt might enhance diffusion creep modestly. Holtzman [2016] used this model and the link between long‐term creep and anelasticity to claim that a substantial velocity reduction will occur even with a small amount of melt.

The stress distribution calculated by Takei and Holtzman [2009] has singularities suggesting that the stress distribution is not steady state and enhanced diffusion creep likely corresponds to transient diffusion creep (see Figure  1.15) [Raj and Ashby, 1971]. Such a stress singularity has small effects on dislocation creep, but the presence of substantial seismic anisotropy in the asthenosphere suggests that the dominant mechanism of deformation is dislocation creep. Consequently, the appli-cability of the results by Takei and Holtzman [2009] to

anelasticity and velocity reduction in the asthenosphere is questionable.

How about melt accumulation near the LAB (Figure 1.12a)? Theory of melt migration under gravity suggests that compaction is efficient, being character-ized by a compaction length of ~0.1 km to ~1 km for typical asthenosphere conditions [e.g., Ribe, 1985], too thin to have any effects on geophysically observable properties. Indeed magnetotelluric observations in the NoMelt region (~70 Ma Pacific Ocean) show no evi-dence of a peak in electrical conductivity at the LAB [Sarafian et al., 2015] (see Chapter 2). An exception is possible near trenches. Due to the stress field associ-ated with plate bending, regional tension could be gen-erated near the bottom of the oceanic plate. This would suck melt to cause locally melt‐rich regions that Yamamoto et  al. [2014] proposed might explain the presence of “petit spots” near ocean trenches. The global presence of accumulated melt at the LAB, how-ever, is unlikely.

How about a layered structure (Figure  1.12b)? Kawakatsu et al. [2009] proposed that a sharp and large velocity drop at the LAB is caused by the presence of near horizontal melt‐rich layers in the asthenosphere. They argued that even with a net melt fraction of ~0.1%, as inferred from petrology [e.g., Hirschmann, 2010], if there were melt‐rich bands as suggested by the experi-mental results by Holtzman et al. [2003], then there would be large radial seismic anisotropy (VSh > VSv) in the asthenosphere but not in the lithosphere, leading to a drop in VSv at the LAB.

The layered structure model, however, has some notable difficulties. Horizontal shear deformation is expected in most regions of the asthenosphere, but a melt‐rich layer will not remain horizontal according to laboratory studies [Holtzman et  al., 2003]. In such a case, a large amount of melt cannot be kept in these layers because gravity will separate melt from the solid rock. In addition, if this model were to explain the observed velocity drop of ~5%, the melt fraction in the melt‐rich bands must have a very specific value, 14.3% (Figure 1.18). Deviation of the melt fraction in the melt‐rich bands from this value of more than 1% will lead to velocity perturbations that do not agree with seismolog-ical observations [Karato, 2014, 2018].

The Kawakatsu et al. [2009] model also implies that the magnitude of the velocity drop at the LAB is the same as the magnitude of radial anisotropy. The reported value range of velocity drop at the LAB is 5–10%, whereas the magnitude of radial anisotropy in the oceanic astheno-sphere is usually ~2–5% [e.g., Montagner and Tanimoto, 1990, 1991]. For these reasons, we consider that the observed properties of the LAB are difficult to explain using a layered structure model.

0004167983.INDD 23 10/11/2018 8:51:09 PM

24 LITHOSPHERIC DISCONTINUITIES

1.6.1.2. Electrical Conductivity Some MT studies on the oceanic regions report a peak

in electrical conductivity in the shallow asthenosphere [Evans et al., 2005; Naif et al., 2013]. These studies were close to ocean ridges but Naif et  al.’s [2013] study was also close to a trench. A melt‐rich layer at the top of the asthenosphere could explain the observed peak in con-ductivity. Another study away from the ridge, however, shows no peak in conductivity [Sarafian et  al., 2015], which discourages a melt‐accumulation model for the LAB at least for regions far from ocean ridges (see also Chapter 2). This is a notable point since a very large and sharp LAB is observed in old oceanic upper mantle [Rychert et al., 2005; Kawakatsu et al., 2009]. These two observations suggest that a large and sharp velocity drop at the LAB should be attributed to a process other than partial melting. An obvious alternative explanation is to invoke hydrogen‐enhanced conductivity [Karato, 1990; Wang et al., 2006; Dai and Karato, 2009]. We conclude that accumulated melt could occur in limited regions near ridges but not globally.

1.6.1.3. Seismic Wave Attenuation Both the partial‐melt model and the EAGBS model

have clear predictions for seismic‐wave attenuation: there would be a peak in seismic wave attenuation if there is

sufficient melt [Jackson et al., 2004]; if the EAGBS mech-anism is the cause of the LAB, then similarly an attenua-tion peak near the LAB would be expected (see Figure 1.16). There have been a few observations on the oceanic LAB to suggest the presence of a peak in attenu-ation [e.g., Takeuchi et al., 2017]. This behavior is differ-ent from the known frequency dependence of attenuation in most minerals [e.g., Jackson, 2009] as well as the results of previous seismological studies [e.g., Anderson and Minster, 1979; Shito et al., 2004].

Takeuchi et al. [2017] proposed two explanations of this anomalous frequency dependence. One is the presence of partial melt, and another is the consequence of EAGBS. Jackson et  al. [2004] showed that the presence of melt leads to a peak in attenuation. EAGBS also predicts a peak in attenuation (Figure  1.16). How can we distin-guish these two?

One possible way to distinguish partial melt model from the EAGBS model is to look at the bulk attenua-tion, namely attenuation associated with volumetric strain. Grain‐boundary sliding involves viscous motion within thin grain boundaries, mostly in response to shear stress. In contrast, partially molten materials cause sub-stantial bulk attenuation as viscous melt migrates within the solid matrix according to some models [e.g., Budianski and O’Connell, 1980; Mavko, 1980] (see also a recent experimental study by Cline and Jackson [2016], who reported some difficulties in the experimental study). In this connection, bulk attenuation measurements would provide useful constraints.

1.6.2. The Frozen‐Melt Model for the MLD

Seismic wave velocities and attenuation are little affected by compositional changes at the LAB, except for changes in water (hydrogen) content. A larger degree of composi-tional change, however, may occur in the continents that could be a cause for the MLD. Among various models for the MLD, a frozen‐melt model proposed by Rader et al. [2015] could be an important possibility (Figure 1.13).

Rader et al. [2015] proposed that, after the formation of continents, metasomatism in the deep continental upper mantle occurred globally. The volatile‐rich melts formed by such metasomatism migrated upward to the depth where the temperature lies below the solidus (~1000 °C). At that depth, the melt froze to form minerals enriched with volatiles and FeO that have low seismic‐wave velocities.

This model predicts the following:1. rocks that form near the MLD typically contain a

large amount of low‐velocity minerals such as amphibole, phlogopite, or FeO‐rich pyroxenes at the global scale;

2. the ages of rocks near the MLD are younger than the ages of rocks below and above the MLD;

�0= 1%

20

15

10

5

00 2 4 6 8 10

(Melt fraction in a melt-rich layer)

12

�0= 0.1%

ΔV/V

(%)

14�

Figure 1.18 Relation between the melt fraction in a melt‐rich layer and the velocity reduction of the asthenosphere contain-ing a melt‐rich layer corresponding to the Kawakatsu model: ϕ0 is the melt fraction if melt were homogeneously distributed—in the old oceanic asthenosphere ϕ0 ≈ 0.1% [Hirschmann, 2010]. In order to explain the observed velocity reduction of 5–10% (in the old oceanic asthenosphere [e.g., Kawakatsu et al., 2009; Rychert et al., 2005]) with ϕ0 ≈ 0.1%, the melt fraction in melt‐rich layers (ϕ) should have a very specific value (14.3 +0.5, −1%) that is highly unlikely. (From Karato [2018].)

0004167983.INDD 24 10/11/2018 8:51:09 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 25

3. electrical conductivity increases substantially at the MLD (from above to below), owing to increased water (+FeO) content.

The amount of low‐velocity materials needed to explain the MLD is shown in Figure  1.19. Among various hydrated minerals, the most commonly observed in the continental mantle is amphibole. If 20–30% of amphi-bole is present, this could explain the observed velocity reduction at the MLD. Xenoliths show no strong indica-tion for the global presence of amphibole‐rich peridotites at the MLD depths [e.g., Pearson et al., 2003], but geolog-ical observations are not perfect and it is difficult to rule this model out for this reason alone.

Regarding the global occurrence of metasomatism, clear support would be the depth variation of the age of mantle rocks: if there were metasomatism then the age should have been reset and the regions that have under-gone metasomatism should have younger ages than the surrounding regions. According to the available data, however, there is no indication of global resetting of the age at the MLD depth (Figure 1.10) [e.g., Pearson, 1999; Carlson et al., 2005].

Another consequence of this model is that there will be an increase in electrical conductivity with depth at the MLD. Electrical conductivity of most materials from the volatile‐rich melt is substantially higher than that of

olivine. For example, the electrical conductivity of amphibole exceeds that of olivine by a factor of 102 to 104 [Wang et al., 2012]. In this scenario, melt had migrated through the mantle, and so melt must have been connected (dihedral angle less than 60°). Consequently, these high‐conductivity minerals are likely connected and enhance the bulk electrical conductivity [e.g., McLachlin, 1987]. Currently available MT models do not show a clear increase in electrical conductivity near the MLD depth (e.g., Meqbel et al. [2014]; see also Chapter 5).

None of these points is definitive, but taken together the above factors do not favor the frozen‐melt model. In the model’s defense, observations on mantle samples (mineralogy, major element chemistry, Re–Os dating) are limited. Similarly, details on the depth variation in electrical conductivity have not been resolved in all available MT studies. Therefore it is important to make advances in these two areas (studies on mantle samples and MT) to test the frozen‐melt model for the MLD.

1.6.3. Layered Anisotropy Model for the MLD and the LAB

The model for the LAB proposed by Kawakatsu et al. [2009] invokes a layered (radial) anisotropy caused by partial melting. Difficulties with this model were discussed

10

5

0

She

ar v

eloc

ity r

educ

tion

(%)

[ref

eren

ce =

4.7

5 km

s–1

]

–5

Annite

Phlogopite

Muscovit

e

Fayali

te

Celado

nite

Calcite

Aragonite

Ferro

actin

olite

Ferro

silite

High a

lbite

Low a

lbite

Tsche

rmak

ite

Ferro

glauc

opha

ne

Heden

berg

ite

Trem

olite

Hornblen

de

Pargas

ite

Diopsid

e

Alman

dine

Coesit

e

Ensta

tite

Pumpe

llyite

Jade

ite

Forste

rite

Pyrop

e

SpinelFast

Slow Micas Fe-rich Carbonates Amphiboles % mineralin rock

1

5

10

15

20

MLDvelocity

reduction

Minerals (from Hacker et al., 2003 database)

Gross

ular

Figure 1.19 Influence of secondary phases on the reduction of shear‐wave velocities. (From Rader et al. [2015]. Reproduced with the permission of Wiley.). (See insert for color representation of the figure.)

0004167983.INDD 25 10/11/2018 8:51:10 PM

26 LITHOSPHERIC DISCONTINUITIES

in section 1.6.1. A fabric transition between the oceanic lithosphere (A‐type fabric) and the asthenosphere (E‐type fabric) has been suggested [Karato, 2008b; Karato et al., 2008], but a change in anisotropy between A‐ and E‐type fabrics is subtle and not enough to explain a large velocity drop at the LAB. Furthermore, there is no evi-dence for a fabric transition in the continental lithosphere down to ~200 km from the study of mantle xenoliths [Ben Ismail and Mainprice, 1998].

Layered anisotropy was also invoked to explain the MLD in the continents [e.g., Yuan and Romanowicz, 2010b]. In case of the MLD, layered anisotropy is most likely due to the layering in the geometry of LPO of min-erals in rocks [e.g., Karato et  al., 2008]. Upper mantle xenoliths, however, show no evidence of systematic change in LPO with depth [e.g., Ben Ismail and Mainprice, 1998], and therefore if there is a layering in anisotropy it would be caused by a change in flow geometry. Since the MLD is nearly horizontal, a change in flow geometry would result from a change in azimuthal anisotropy not in radial anisotropy across the MLD. As pointed out by Karato et al. [2015], however, azimuthal anisotropy can simulate a sharp isotropic velocity inversion only if most seismic observations illuminate a small portion of avail-able back azimuths. Layered anisotropy could explain the upper mantle discontinuity in regions with such data restrictions, but it cannot be the cause of these disconti-nuities at a global scale (see Figure 1.14).

1.6.4. EAGBS Model for the MLD and the LAB

The following features are predicted by the EAGBS model:

1. velocity reduction in olivine (or olivine‐rich rocks) occurs where temperature reaches a critical value, ~1000 °C, that depends on the water content (and pressure) and therefore the velocity reduction by this mechanism is global;

2. velocity reduction is large (several percent);3. velocity reduction occurs without compositional lay-

ering, and the ages of rocks do not change across the MLD in the continents if EAGBS is the cause of the MLD;

4. velocity reduction is not associated with a sudden increase in electrical conductivity if EAGBS occurs solely due to elevated temperature (there will be a conductivity jump if EAGBS is associated with an increase in water content);

5. there should be a peak in attenuation in the depth range where velocity reduction occurs.

All these features are consistent with the known geolog-ical and geophysical observations summarized previously. In contrast, alternative models, such as the frozen‐melt model, are not consistent with many of the geological or geophysical observations previously summarized.

There is another observation from the cratonic mantle that provides a useful constraint on models for the velocity reduction at the MLD and the LAB. This is the much larger velocity drop at the MLD (~100 km) than at the LAB (~200 km) (Figure 1.6b) [Abt et al., 2010]. This is a notable observation because for some scenarios, such as partial melt or monotonic temperature increase models, a larger velocity drop would be expected in the deep regions where temperature is high.

In contrast, the EAGBS model provides a natural explanation for this observation for the following rea-sons. Figure 1.16 depicts, from right to left (for a given seismic wave frequency), transition from a shallow region to a deeper region where temperature is higher. In the shallow part, seismic frequency is higher than ωEAGBS, and rocks are in the unrelaxed (high velocity) state; in the deeper part where ωEAGBS increases the seismic frequency becomes lower than ωEAGBS. At that point, a substantial (a few percent) velocity drop occurs. With increasing depth in the absorption band regime, where velocity change and Q are directly linked by equation (1.1) and therefore the magnitude of velocity change is ~ Q−1, and when Q ~100 near the LAB, the velocity drop is ~1% not a few percent.

Regarding attenuation, Takeuchi et al. [2017] provided evidence for the attenuation peak in the asthenosphere that is consistent with the EAGBS model (and also con-sistent with a partial melt model) for the LAB. As to the MLD, there is no direct evidence for the attenuation peak associated with the MLD. The reported rather low intrinsic QS in the continental lithosphere (QS = 100–300, Figure 1.8) could, however, well be caused by an attenua-tion peak near the MLD for the following reasons. If a conventional absorption‐band model (without an attenu-ation peak) is assumed, QS within the continental lithosphere can be calculated from the known QS in the asthenosphere (QS ~80) and the temperatures in the lithosphere. Because temperatures in the continental lithosphere are low, substantially higher QS (~1000 or higher) values would be expected than those reported (QS = 100–300; Figure 1.8). Invoking an attenuation peak (QS = 20) around the MLD (Qs = 20 corresponds to a velocity reduction of 5%) with 20 km width may solve this apparent paradox. Such an attenuation peak will have the same degree of attenuation as a constant QS = 200 distributed in a layer of 200 km thickness.

Another explanation is to attribute the apparent low Qs to scattering [e.g., Fehler et al., 1992; Jin et al., 1994]). Dalton et al.’s [2009] attenuation model is for intrinsic Qs, however, after corrections for scattering and focusing. In addition, the rough agreement between a long‐wavelength study [Dalton et al., 2009] and a short‐wavelength study [Mitchell, 1995] suggests that scattering is not a dominant cause for low QS.

0004167983.INDD 26 10/11/2018 8:51:10 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 27

Regarding the MT observations, there is no clear evidence for an increase in conductivity at ~100 km depth in the cratonic upper mantle, but conductivity becomes high below ~200 km [e.g., Meqbel et al., 2014]. Similarly, an increase in conductivity is inferred across the oceanic LAB [e.g., Evans et  al., 2005; Naif et  al., 2013; Meqbel et al., 2014]. A simple explanation is that the large increase in conductivity across the LAB is due to a change in water content (higher below the LAB [Karato, 1990; Wang et  al., 2006; Dai and Karato, 2009]). The large and sharp velocity reduction for most of the oceanic LAB may also be due to the enhanced EAGBS caused by an increase in water content [Karato, 2012; Olugboji et al., 2013].

In summary, the EAGBS model provides a unified model for the MLD and the LAB in the sense that (a) it explains the global occurrence of these boundaries, (b) the temperature requirement for the MLD and the oceanic LAB is consistent with EAGBS, and (c) it does not require chemical or anisotropy layering. A major limitation of this model, however, is that experimental studies on EAGBS are still preliminary; key features such as the grain‐size sensitivity of velocity reduction and the influence of water on EAGBS are poorly constrained.

1.7. SUMMARY AND FUTURE DIRECTIONS

The causes of both the LAB and the MLD are reviewed in the light of geological, geophysical, and mineral (and rock) physics observations. Among several models for these discontinuities, three are potentially important: compositional layering (in major element and/or miner-alogy), partial melting, and EAGBS.

Geological observations relevant to these models were reviewed, and we conclude that compositional layering model is difficult to reconcile with the observations of mantle rocks including mineralogical layering and rock‐age layering.

Our preferred model for the oceanic LAB is that the large and sharp velocity drop is caused mainly by ther-mally induced EAGBS helped by the jump in water content (Figure 1.20). There should be a peak in attenua-tion at the LAB associated with EAGBS. For the continental upper mantle, the MLD is likely due to ther-mally induced EAGBS, but it may also be influenced by chemical layering (layering in water content). The MLD should be associated with a peak in seismic attenuation if it is mainly caused by EAGBS. If EAGBS occurs only by temperature increase, there should be no large increase in electrical conductivity. In the continental (cratonic) LAB,

VShVSv

VSv

Lithosphere

Asthenosphere

(a) Oceanic upper mantle

(b) Continental upper mantle

Lithosphere

Asthenosphere

~100 km

VSh

VShVSv

VShVSv

~200 km

Thermally induced EAGBS+ compoistional layering(large velocity drop,attenuation peak,small or large conductivity change)

Thermally andwater-induced EAGBS(large velocity drop,attenuation peak,large conductivity increase)

Thermally inducedvelocity reduction withthe absorption band behavior(small velocity drop, large conductivity increase due to water content change)

LAB

LAB

MLD

Figure 1.20 Preferred models for the MLD and the LAB consistent with available geological and geophysical observations. (Modified from Karato [2012]. Reproduced with the permission of Elsevier.)

0004167983.INDD 27 10/11/2018 8:51:10 PM

28 LITHOSPHERIC DISCONTINUITIES

rocks are already in the relaxed state with respect to EAGBS, so that further increase in temperature and/or water content leads only to a minor decrease in velocity; an increase in water content at the LAB, however, will increase electrical conductivity substantially.

Our model for the MLD and the LAB has some implications for change in rheological properties across these boundaries. Inasmuch as the MLD is mainly caused by temperature increase, change in long‐term rheological properties across the MLD will be negligible because EAGBS changes only the short‐term, small

strain rheological properties but not the long‐term steady‐state rheological properties (in this case, a large seismic wave velocity drop is not associated with a large drop in rock strength). In contrast, the LAB in the oce-anic regions and presumably that in the continent is likely associated with a change in water content. Consequently, the LAB, including the LAB in the cra-tonic mantle where velocity reduction is small, is associ-ated with a major reduction in the long‐term rheological properties (creep strength) with depth. The link between seismic‐wave velocities and long‐term rheological properties

(i) Compositional layering

Velocity

Electricalconductivity

Distinct chemistry(secondary phases

(frozen melt))

(ii) EAGBS

Unrelaxed

Relaxed

Velocity

Attenuation (Q–1)Attenuation (Q–1)Attenuation (Q–1)

Electricalconductivity

(ii) Partial melt

Velocity

Electricalconductivity

Partial melt

Hydrogen

Therm

al

Dep

thD

epth

Dep

thD

epth

Figure 1.21 Schematic diagrams showing the expected geophysical observations of three models for the upper mantle discontinuity. For the partial‐melt model, bulk attenuation will be significant, but for other models, shear attenuation will be dominant. Electrical conductivity–depth relation for the EAGBS model differs between ther-mally induced and water‐induced EAGBS; conductivity will jump at the boundary for the latter, but it should be continuous for the former. There should be a large conductivity jump at the MLD for a compositional layering model invoking frozen melts.

0004167983.INDD 28 10/11/2018 8:51:10 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 29

is not direct, particularly when EAGBS plays an impor-tant role in controling the seismic‐wave velocities.

Our preferred model is developed based on experi-mental and theoretical studies on elastic and anelastic properties of rocks in combination with constraints from geological (+ geochemical) and geophysical observations. Geological observations are based on a limited number of rock samples, however, so that our understanding is far from complete. Further studies are warranted, including the depth variation in mineralogy in mantle xenoliths and more precise determination of the depth dependence of the age of mantle xenoliths. Similarly, geophysical observations are also limited.

In the following, we summarize some of the possible future directions on the studies on the origin of the LAB and the MLD.

• Geophysics (see Figure 1.21) ∘ Conduct MT studies on the continental upper mantle. If the frozen melt is the cause of the MLD, there should be a substantial increase in conduc-tivity at the MLD. Thermally induced EAGBS will cause no increase in conductivity at the MLD. ∘ Determine the sharpness of the velocity drop at the MLD. For the thermally induced EAGBS model, the MLD is diffuse. For a compositional layering model, it is sharp. ∘ Determine the seismic‐wave attenuation in the continental lithosphere. For the EAGBS model there is a peak in attenuation. ∘ Determine the magnitude of bulk attenuation. If partial melting were a cause of attenuation, bulk attenuation would be substantial. For other mecha-nisms, shear attenuation dominates.

• Geology (geochemistry) ∘ Conduct detailed studies on the mineralogy, isotopic compositions (ages) of mantle xenoliths as a function of depth. For a frozen‐melt model, there should be marked compositional anomalies at around the MLD at a global scale. For a frozen‐melt model, rocks around the MLD should show younger ages than rocks above and below it.

• Mineral physics ∘ Conduct laboratory studies to test the grain‐size sensitivity of velocity reduction by EAGBS. If such sensitivity is confirmed, develop a model to provide a physical explanation for it. ∘ Conduct laboratory studies to understand the influence of water content on anelasticity, particu-larly EAGBS in the regime where hydrogen is dis-solved in a similar way as in Earth’s upper mantle.

A study of each of these topics involves technical chal-lenges. In addition to making advances in these specific issues, it is also important to integrate all available obser-vations to devise acceptable geological and geophysical

models. Those studies will help us understand the signifi-cance of the MLD and the LAB in the upper mantle in connection to the geological evolution of the upper mantle.

ACKNOWLEDGMENTS

We thank Huaiyu Yuan for inviting us to make this contribution and for his patience in handling our paper. This study is partly supported by the grants from NSF. We thank Chris Cline, Gary Egbert, Don Forsyth, Ian Jackson, Cin‐Ty Lee, Tolulope Olugboji, Jarka Plomerová, and Akira Takeuchi for helpful discussions. The comments by Barbara Romanowicz and two anony-mous reviewers were helpful in improving the paper. Thank you all.

REFERENCES

Abt, D. L., Fischer, K. M., French, S. W., Ford, H. A., Yuan, H., and Romanowicz, B. (2010), North American lithosphere discontinuity structure imaged by Ps and Sp receiver functions, J. Geophys. Res., 115, 2009JB00691.

Aizawa, Y., Barnhoon, A., Faul, U. A., Fitz Gerald, J. D., Jackson, I., and Kovács, I. (2008), Seismic properties of Anita Bay dunite: An exploratory study of the influence of water, J. Petrol., 49, 841–855.

Amin, K. E., Mukherjee, A. K., and Dorn, J. E. (1970), A  universal law for high‐temperature diffusion controlled transient creep, J. Mech. Phys. Solids, 18, 413–426.

Anderson, D. L. and Minster, J. B. (1979), The frequency dependence of Q in the Earth and implications for mantle rheology and Chandler wobble, Geophys. J. R. Astron. Soc., 58, 431–440.

Anderson, D. L. and Spetzler, H. (1970), Partial melting and the low‐velocity zone, Phys. Earth Planet. Inter., 4, 62–64.

Artemieva, I. M. (2011), The Lithosphere, Cambridge University Press, Cambridge, 773 pp.

Babuška, V. and Plomerová, J. (1989), Seismic anisotropy of the subcrustal lithosphere in Europe: Another clue to recognition of accreted terranes?, in J. W. Hillhouse (ed.), Deep Structure and Past Kinematics of Accreted Terranes, pp. 209–217, American Geophysical Union, Washington, DC.

Babuška, V., Montagner, J.‐P., Plomerová, J., and Girardin, N. (1998), Age‐dependent large‐scale fabric of the mantle litho-sphere as derived from surface‐wave velocity anisotropy, Pure Appl. Geophys., 151, 257–280.

Bagley, B. and Revenaugh, J. (2008), Upper mantle seismic shear discontinuities of the Pacific, J. Geophys. Res., 113, 10.1029/2008JB005692.

Bai, Q., Mackwell, S. J., and Kohlstedt, D. L. (1991), High tem-perature creep of olivine single crystals: 1. Mechanical results for buffered samples, J. Geophys. Res., 96, 2441–2463.

Beghein, C., Yuan, K., Schmerr, N., and Xing, Z. (2014), Changes in seismic anisotropy shed light on the nature of the Gutenberg discontinuity, Science, 343, 1237–1240.

0004167983.INDD 29 10/11/2018 8:51:10 PM

30 LITHOSPHERIC DISCONTINUITIES

Behn, M. D. and Kelemen, P. B. (2006), Stability of arc lower crust: Insights from the Talkeetna arc section, south central Alaska, and the seismic structure of modern arcs, J. Geophys. Res., 111, 2006JB004327.

Ben Ismail, W. and Mainprice, D. (1998), An olivine fabric data-base: An overview of upper mantle fabrics and seismic anisot-ropy, Tectonophysics, 296, 145–157.

Benz, H. M. and McCarthy, J. (1994), Evidence for an upper‐mantle low‐veloicty zone beneath the southern Basin and Range–Colorado Plateau transition zone, Geophys. Res. Lett., 21, 509–512.

Birch, F. (1952), Elasticity and constitution of the Earth’s interior, J. Geophys. Res., 57, 227–286.

Booth, C. M., Forsyth, D. W., and Weeraratne, D. S. (2014), Upper mantle Q structure beneath old seafloor in the western Pacific, J. Geophys. Res., 119, 3448–3461. doi:10.1002/2013JB010589.

Budianski, B. and O’Connell, J. R. (1980), Bulk dissipation in heterogeneous media, in S. N. Nasser (ed.), Solid Earth Geophysics and Geotechnology, pp. 1–10, American Society of Mechanical Engineering, New York.

Burdick, L. J. and Helmberger, D. V. (1978), The upper mantle P velocity structure of the Western United States, J. Geophys. Res., 83, 1699–1712.

Calò, M., Bodin, T., and Romanowicz, B. (2016), Layered struc-ture in the upper mantle across Northern American from joint inversion of long and short period seismic data, Earth Planet. Sci. Lett., 449, 164–175.

Carlson, R. W., Pearson, D. G., and James, D. E. (2005), Physical, chemical, and chronological characteristics of continental mantle, Rev. Geophys., 43, 2004RG000156.

Chen, C., Gilbert, H., Fischer, K. M., Andronicos, C. L., Pavlis, G. L., Hamburger, M. W., Marshak, S., Larson, T., and Yang, X. (2018), Lithospheric discontinuities beneath the U.S. Midcontinent—significance of Proterozoic terrane accretion and failed rifting, Earth Planet. Sci. Lett., 481, 223–235.

Chen, L.,.Jiang, M., Yang, J., Wei, Z., Liu, C., and Ling, Y. (2014), Presence of an intralithospheric discontinuity in the central and western North China Craton: Implications for destruction of the craton, Geology, 42, 223–226.

Cline, C. J., II. and Jackson, I. (2016), Relaxation of the bulk modulus in partially molten dunite?, Geophys. Res. Lett., 43, 11644–11651.

Cline, C. J., II., Faul, U, David, E. C., Berry, A. J., and Jackson, I. (2018), Redox‐influenced seismic properties of upper‐mantle olivine, Nature, 555, 355–358.

Collerson, K. D., Hapugoda, S., Kamber, B. S., and Williams, Q. (2000), Rocks from the mantle transition zone: Majorite‐bearing xenoliths from Malaita, southwest Pacific, Science, 288, 1215–1223.

Dai, L. and Karato, S. (2009), Electrical conductivity of orthopyroxene: Implications for the water content of the asthenosphere, Proc. Jpn Acad., 85, 466–475.

Dalton, C. A., Ekström, G., and Dziewonski, A. M. (2009), Global seismological shear velocity and attenuation: A comparison with experimental observations, Earth Planet. Sci. Lett., 284, 65–75.

Dalton, C. A., Bao, X., and Ma, Z. (2017), The thermal struc-ture of cratonic lithosphere from global Rayleigh wave atten-uation, Earth Planet. Sci. Lett., 457, 250–262.

Darbyshire, F. A., Eaton, D. W., Frederiksen, A. W., and Ertolahti, L. (2007), New insights into the lithosphere beneath the Superior Province from Rayleigh wave dispersion and receiver function analysis, Geophys. J. Int., 169, 1043–1068.

Dziewonski, A. M. and Anderson, D. L. (1981), Preliminary ref-erence Earth model, Phys. Earth Planet. Inter., 25, 297–356.

Eaton, D. W., Darbyshire, F., Evans, R. L., Grütter, H., Jones, A. G., and Yuan, X. (2009), The elusive lithosphere–astheno-sphere boundary (LAB) beneath cratons, Lithos, 109, 1–22.

Evans, R. L., Hirth, G., Baba, K., Forsyth, D. W., Chave, A., and Makie, R. (2005), Geophysical evidence from the MELT area for compositional control on oceanic plates, Nature, 437, 249–252.

Faul, U. H. and Jackson, I. (2005), The seismological signature of temperature and grain size variations in the upper mantle, Earth Planet. Sci. Lett., 234, 119–134.

Faul, U. H. and Jackson, I. (2015), Transient creep and strain energy dissipation: An experimental perspective, Annu. Rev. Earth Planet. Sci., 43, 541–569.

Fehler, M., Hoshiba, M., Sato, H., and Obara, K. (1992), Separation of scattering and intrinsic attenuation for the Kanto‐Tokai region, Japan, using measurements of S‐wave energy versus hypocentral distance, Geophys. J. Int., 108, 787–800.

Ford, H. A., Fischer, K. M., Abt, D. L., Rychert, C. A., and Elkins‐Tanton, L. T. (2010), The lithosphere–asthenosphere boundary and cratonic lithospheric layering beneath Australia from Sp wave imaging, Earth Planet. Sci. Lett., 300, 299–310.

Forsyth, D. W. (1975), The early structural evolution and anisotropy of the oceanic upper mantle, Geophys. J. R. Astron. Soc., 43, 103–162.

Forsyth, D. W. (1977), Evolution of the upper mantle beneath the mid‐ocean ridges, Tectonophysics, 38, 89–118.

Frassetto, A. and Thybo, H. (2013), Receiver function analysis of the crust and upper mantle in Fennoscandia—isostatic implications, Earth Planet. Sci. Lett., 381, 234–246.

Gaherty, J. B., Kato, M., and Jordan, T. H. (1999), Seismological structure of the upper mantle: a regional comparison of seismic layering, Phys. Earth Planet. Inter., 110, 21–41.

Ghahremani, F. (1980), Effect of grain boundary sliding on anelasticity of polycrystals, Int. J. Solid Struct., 16, 825–845.

Given, J. W. and Helmberger, D. V. (1980), Upper mantle structure of northwestern Eurasia, J. Geophys. Res., 85, 7183–7194.

Grand, S. and Helmberger, D. V. (1984), Upper mantle shear structure of North America, Geophys. J. R. Astron. Soc., 76, 399–438.

Gueguen, Y. and Mercier, J. M. (1973), High attenuation and low velocity zone, Phys. Earth Planet. Inter., 7, 39–46.

Gung, Y., Romanowicz, B., and Panning, M. (2003), Global anisotropy and the thickness of continents, Nature, 422, 707–711.

Gutenberg, B. (1926), Untersuchen zur Frage, bis zu welcher Tiefe die Erde kristallin ist, Z. Geophis., 2, 24–29.

Hirschmann, M. M. (2006), Water, melting, and the deep Earth H

2O cycle, Annu. Rev. Earth Planet. Sci., 34, 629–653.Hirschmann, M. M. (2010), Partial melt in the oceanic low

velocity zone, Phys. Earth Planet. Inter., 179, 60–71.

0004167983.INDD 30 10/11/2018 8:51:10 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 31

Hirth, G. and Kohlstedt, D. L. (1996), Water in the oceanic upper mantle—implications for rheology, melt extraction and the evolution of the lithosphere, Earth Planet. Sci. Lett., 144, 93–108.

Holtzman, B. K. (2016), Questions on the existence, persistence, and mechanical effects of a very small melt fraction in the asthenosphere, Geochem. Geophys. Geosyst., 17, 470–484.

Holtzman, B. K., Kohlstedt, D. L., Zimmerman, M. E., Heidelbach, F., Hiraga, K., and Hustoft, J. (2003), Melt segregation and strain partitioning: Implications for seismic anisotropy and mantle flow, Science, 301, 1227–1230.

Hopper, E. and Fischer, K. M. (2015), The meaning of mid-lithospheric discontinuities: A case study in the northern U.S. craton, Geochem. Geophys. Geosyst., 16, 4057–4083.

Jackson, I. (2009), Properties of rocks and minerals—physical origins of anelasticity and attenuation in rock, in G. Schubert (ed.), Treatise on Geophysics, pp. 493–525, Elsevier, Amsterdam.

Jackson, I. (2015), Properties of rocks and minerals—physical origins of anelasticity and attenuation in rock, in G. Schubert (ed.), Treatise on Geophysics, Elsevier, Amsterdam.

Jackson, I. and Faul, U. A. (2010), Grainsize‐sensitive visco-elastic relaxation in olivine: Towards a robust laboratory‐based model for seismological application, Phys. Earth Planet. Inter., 183, 151–163.

Jackson, I., Paterson, M. S., and Fitz Gerald, J. D. (1992), Seismic wave dispersion and attenuation in Åheim dunite, Geophys. J. Int., 108, 517–534.

Jackson, I., Faul, U. A., Fitz Gerald, J. D., and Tan, B. (2004), Shear wave attenuation and dispersion in melt‐bearing olivine polycrystals: 1. Specimen fabrication and mechanical testing, J. Geophys. Res., 109, 10.1029/2003JB002406.

Jackson, I., Faul, U. A., and Skelton, R. (2014), Elastically accommodated grain‐boundary sliding: New insights from experiment and modeling, Phys. Earth Planet. Inter., 228, 203–210.

Jin, A., Mayeda, K., Adams, D., and Aki, K. (1994), Separation of intrinsic and scattering attenuation in southern California using TERRAscope data, J. Geophys. Res., 99, 17835–17848.

Kaminsky, F. V., Wirth, R., and Schreiber, A. (2015), A micro-inclusion of lower‐mantle rock and other minerals and nitrogen lower mantle inclusions in a diamond, Can. Mineral., 53, 83–104.

Kanamori, H. and Press, F. (1970), How thick is the litho-sphere?, Nature, 226, 330–331.

Karato, S. (1986), Does partial melting reduce the creep strength of the upper mantle?, Nature, 319, 309–310.

Karato, S. (1990), The role of hydrogen in the electrical conduc-tivity of the upper mantle, Nature, 347, 272–273.

Karato, S. (1993), Importance of anelasticity in the interpreta-tion of seismic tomography, Geophys. Res. Lett., 20, 1623–1626.

Karato, S. (1995), Effects of water on seismic wave velocities in the upper mantle, Proc. Jpn Acad., 71, 61–66.

Karato, S. (2003), Mapping water content in Earth’s upper mantle, in J. E. Eiler (ed.), Inside the Subduction Factory, pp. 135–152, American Geophysical Union, Washington, DC.

Karato, S. (2008a), Deformation of Earth Materials: Introduction to the Rheology of the Solid Earth, Cambridge University Press, Cambridge, 463 pp.

Karato, S. (2008b), Insights into the nature of plume–astheno-sphere interaction from central Pacific geophysical anom-alies, Earth Planet. Sci. Lett., 274, 234–240.

Karato, S. (2010), Rheology of the Earth’s mantle: A historical review, Gondwana Res., 18, 17–45.

Karato, S. (2012), On the origin of the asthenosphere, Earth Planet. Sci. Lett., 321/322, 95–103.

Karato, S. (2014), Does partial melting explain geophysical anomalies?, Phys. Earth Planet. Inter., 228, 300–306.

Karato, S. (2018), Some remarks on hydrogen‐assisted electrical conductivity in olivine and other minerals, in K. Mibe and J. Kasahara (eds), Water in the Earth, Springer, Tokyo in press.

Karato, S. and Jung, H. (1998), Water, partial melting and the origin of seismic low velocity and high attenuation zone in the upper mantle, Earth Planet. Sci. Lett., 157, 193–207.

Karato, S. and Jung, H. (2003), Effects of pressure on high‐temperature dislocation creep in olivine polycrystals, Philos. Mag., A83, 401–414.

Karato, S. and Spetzler, H. A. (1990), Defect microdynamics in min-erals and solid state mechanisms of seismic wave attenuation and velocity dispersion in the mantle, Rev. Geophys., 28, 399–421.

Karato, S., Paterson, M. S., and Fitz Gerald, J. D. (1986), Rheology of synthetic olivine aggregates: influence of grain‐size and water, J. Geophys. Res., 91, 8151–8176.

Karato, S., Jung, H., Katayama, I., and Skemer, P. A. (2008), Geodynamic significance of seismic anisotropy of the upper mantle: New insights from laboratory studies, Annu. Rev. Earth Planet. Sci., 36, 59–95.

Karato, S., Olugboji, T. M., and Park, J. (2015), Mechanisms and geologic significance of the mid‐lithosphere discontinuity in the continents, Nature Geosci., 8, 509–514.

Kawakatsu, H., Kumar, P., Takei, Y., Shinohara, M., Kanazawa, T., Araki, E., and Suyehiro, K. (2009), Seismic evidence for sharp lithosphere–asthenosphere boundaries of oceanic plates, Science, 324, 499–502.

Kê, T. S. (1947), Experimental evidence of the viscous behavior of grain boundaries in metals, Phys. Rev., 71, 533–546.

Knapmeyer‐Endrun, B., Krüger, F., Geissler, W. H., and Group, P. W. (2017), Upper mantle structure across the Trans‐European Suture Zone imaged by S‐receiver functions, Earth Planet. Sci. Lett., 458, 429–441.

Kumar, P. and Kawakatsu, H. (2011), Imaging the seismic lith-osphere–asthenosphere boundary of the oceanic plate, Geochem. Geophys. Geosyst., 12, 10.1029/2010GC003358.

Lambert, I. B. and Wyllie, P. J. (1970), Low‐velocity zone of the Earth’s mantle: Incipient melting caused by water, Science, 169, 764–766.

Lay, T. and Wallace, T. C. (1995), Modern Global Seismology, Academic Press, San Diego, CA, 521 pp.

Lee, C.‐T. A. (2003), Compositional variation of density and seismic velocities in natural peridotites at STP conditions: Implications for seismic imaging of compositional heterogene-ities in the upper mantle, J. Geophys. Res., 108, 2003JB0002413.

Levin, V. and Park, J. (2000), Shear zones in the Proterozoic lithosphere of the Arabian Shield and the nature of the Hales discontinuity, Tectonophysics, 323, 131–148.

0004167983.INDD 31 10/11/2018 8:51:10 PM

32 LITHOSPHERIC DISCONTINUITIES

Lin, P.‐Y. P., Gaherty, J. B., Jin, G., Collins, J. A., Lizarralde, D., Evans, R. L., and Hirth, G. (2016), High‐resolution seismic constraints on flow dynamics in the oceanic asthenosphere, Nature, 535, 538–541.

Mandal, P. (2017), Lithospheric thinning in the Eastern Indian Craton: Evidence for lithospheric delamination below the Archean Singhbhum Craton?, Tectonophysics, 698, 91–108.

Matsukage, K. N., Nishihara, Y., and Karato, S. (2005), Seismological signature of chemical evolution of Earth’s upper mantle, J. Geophys. Res., 110, 2004JB003504.

Mavko, G. M. (1980), Velocity and attenuation in partially molten rocks, J. Geophys. Res., 85, 5173–5189.

McCarthy, C., Takei, Y., and Hiraga, T. (2011), Experimental study of attenuation and dispersion over a broad frequency range: 2. The universal scaling of polycrystalline materials, J. Geophys. Res., 116, 2011JB008384.

McKenzie, D., Jackson, J. A., and Priestley, K. (2005), Thermal structure of oceanic and continental lithosphere, Earth Planet. Sci. Lett., 233, 337–349.

McLachlin, D. C. (1987), An equation for the conductivity of binary mixtures with anisotropic grain structures, J. Phys. C Solid State Phys., 20, 865–877.

Mehouachi, F. and Singh, S. C. (2018), Water‐rich sublitho-spheric melt channel in the equatorial Atlantic Ocean, Nature Geosci., 11, 65–69.

Mei, S. and Kohlstedt, D. L. (2000a), Influence of water on plastic deformation of olivine aggregates: 1. Diffusion creep regime, J. Geophys. Res., 105, 21457–21469.

Mei, S. and Kohlstedt, D. L. (2000b), Influence of water on plastic deformation of olivine aggregates: 2. Dislocation creep regime, J. Geophys. Res., 105, 21471–21481.

Meqbel, N. M., Egbert, G., Wannamaker, P. E., Kelbert, A., and Schultz, A. (2014), Deep electrical resistivity structure of the northwestern U.S. derived from 3‐D inversion of USArray magnetotelluric data, Earth Planet. Sci. Lett., 402, 290–304.

Minster, J. B. and Anderson, D. L. (1981), A model of disloca-tion‐controlled rheology for the mantle, Philos. Trans. R. Soc. Lond., A299, 319–356.

Mitchell, B. J. (1995), Anelastic structure and evolution of the continental crust and upper mantle from seismic surface wave attenuation, Rev. Geophys., 33, 441–462.

Mizutani, H. and Kanamori, K. (1964), Variation in elastic wave velocity and attenuative property near the melting tem-perature, J. Phys. Earth, 12, 43–49.

Montagner, J.‐P. and Tanimoto, T. (1990), Global anisotropy in the upper mantle inferred from the regionalization of phase velocities, J. Geophys. Res., 95, 4797–4819.

Montagner, J.‐P. and Tanimoto, T. (1991), Global upper mantle tomography of seismic wave velocities and anisotropies, J. Geophys. Res., 96, 20337–20351.

Naif, S., Key, K., Constable, S., and Evans, R. L. (2013), Melt‐rich channel observed at the lithosphere–asthenosphere boundary, Nature, 495, 356–359.

Nestola, F. (2017), Inclusions in super‐deep diamonds: Windows on the very deep Earth, Rend. Fis. Acc. Lincei, 28, 595–604.

Nicolas, A. (1989), Structures of Ophiolites and Dynamics of Oceanic Lithosphere (Petrology and Structural Geology), Springer, Berlin, 380 pp.

Nishimura, C. E. and Forsyth, D. W. (1989), The anisotropic structure of the upper mantle in the Pacific, Geophys. J., 96, 203–229.

Nowick, A. S. and Berry, B. S. (1972), Anelastic Relaxation in Crystalline Solids, Academic Press, San Diego, CA, 677 pp.

Ohira, A., Kodaira, S., Nakamura, Y., Fujie, G., Arai, S., and Miura, S. (2017), Evidence for frozen melts in the mid‐litho-sphere detected from active‐source seismic data, Nature Sci. Rep., 7, 15770.

Olugboji, T. M., Karato, S., and Park, J. (2013), Structure of the oceanic lithosphere–asthenosphere boundary: Mineral‐physics modeling and seismological signatures, Geochem. Geophys. Geosyst., 14, 880–901.

Olugboji, T. M., Park, J., Karato, S., and Shinohara, M. (2016), Nature of the seismic lithosphere–asthenosphere boundary within normal oceanic mantle from high‐resolution receiver functions, Geochem. Geophys. Geosyst., 17, 1265–1282.

Park, J. and Levin, V. (2016), Anisotropic shear zones revealed by back‐azimuthal harmonics of teleseismic receiver functions, Geophys. J. Int., 207, 1216–1243.

Parsons, B. and McKenzie, D. (1978), Mantle convection and  the thermal structure of the plates, J. Geophys. Res., 83, 4485–4496.

Pearson, D. G. (1999), The age of continental roots, Lithos., 24, 171–194.

Pearson, D. G., Canil, D., and Shirey, S. B. (2003), Mantle sam-ples included in volcanic rocks: Xenoliths and diamonds, in R. W. Carlson, H. D. Holland and K. K. Turekian (eds), Treatise on Geochemistry, pp. 171–275, Elsevier, Amsterdam.

Pearson, D. G., Brenker, F. E., Nestola, F., McNeill, J. C. R., Nasdala, L., Hutchison, M. T., Matveev, S., Mather, K., Silversmit, G., Schmitz, S., Vekemans, B., and Vincze, L. (2014), Hydrous mantle transition zone indicated by ring-woodite included within diamond, Nature, 507, 221–224.

Pedersen, H. A., Fishwick, S., and Snyder, D. B. (2009), A comparison of cratonic roots through consistent analysis of seismic surface waves, Lithos, 109, 81–95.

Peslier, A. H., Woodland, A. B., Bell, D. R., and Lazarov, M. (2010), Olivine water contents in the continental lithosphere and the longivity of cratons, Nature, 467, 78–83.

Pezzotti, G. (1999), Internal friction of polycrystalline ceramic oxides, Phys. Rev., B60, 4018–4029.

Plomerová, J. and Babuška, V. (2010), Long memory of mantle lithosphere fabric—European LAB constrained from seismic anisotropy, Lithos, 120, 131–143.

Plomerová, J., Kouba, D., and Babuška, V. (2002), Mapping the lithosphere–asthenosphere boundary through changes in sur-face‐wave anisotropy, Tectonophyiscs, 358, 175–185.

Priestley, K. and McKenzie, D. (2006), The thermal structure of the lithosphere from shear wave velocities, Earth Planet. Sci. Lett., 244, 285–301.

Priestley, K. and McKenzie, D. (2013), The relationship bet-ween shear wave velocity, temperature, attenuation and vis-cosity in the shallow part of the mantle, Earth Planet. Sci. Lett., 381, 78–91.

Rader, E., Emry, E. L., Schmerr, N., Frost, D., Cheng, C., Menard, J., Yu, C.‐Q., and Geist, D. (2015), Characterization and petrological constraints of the midlithospheric discontinuity, Geochem. Geophys. Geosyst., 16, 3484–3504.

0004167983.INDD 32 10/11/2018 8:51:10 PM

ON THE ORIgIN Of THE UPPER MaNTLE SEISMIC DISCONTINUITIES 33

Raj, R. and Ashby, M. F. (1971), On grain boundary sliding and diffusional creep, Metall. Trans., 2, 1113–1127.

Revenaugh, J. and Jordan, T. H. (1991a), Mantle layering from ScS reverberations: 1. Waveform inversion of zeroth‐order reverberations, J. Geophys. Res., 96, 19749–19762.

Revenaugh, J. and Jordan, T. H. (1991b), Mantle layering from ScS reverberations: 3. Upper mantle, J. Geophys. Res., 96, 19781–19810.

Revenaugh, J. and Jordan, T. H. (1991c), Mantle layering from ScS revervarations: 2. The transition zone, J. Geophys. Res., 96, 19763–19780.

Ribe, N. M. (1985), The deformation and compaction of partial molten zones, Geophys. J. R. Astron. Soc., 83, 487–501.

Ringwood, A. E. (1975), Composition and Structure of the Earth’s Mantle, McGraw‐Hill, New York, 618 pp.

Ringwood, A. E. (1991), Phase transformations and their bear-ings on the constitution and dynamics of the mantle, Geochim. Cosmochim. Acta, 55, 2083–2110.

Romanowicz, B. and Mitchell, B. J. (2007), Deep Earth struc-ture—Q of the Earth from crust to core, in B. Romanowicz and A. Dziewonski (eds), Seismology and Structure of the Earth, pp. 731–774, Elsevier, Amsterdam.

Rychert, C. A. and Shearer, P. M. (2009), A global view of the lithosphere–asthenosphere boundary, Science, 324, 495–498.

Rychert, C. A. and Shearer, P. M. (2011), Imaging the litho-sphere–asthenosphere boundary beneath the Pacific using SS waveform modeling, J. Geophys. Res., 116, 1–15. doi:10.1029/ 2010JB008070.

Rychert, C. A., Fischer, K. M., and Rodenay, S. (2005), A sharp lithosphere–asthenosphere boundary imaged beneath eastern North America, Nature, 434, 542–545.

Rychert, C. A., Schmerr, N., and Harmon, N. (2012), The Pacific lithosphere–asthenosphere boundary: Seismic imaging and anisotropic constraints from SS waveforms, Geochem. Geophys. Geosyst., 13, 10.1029/2012GC004194.

Sarafian, E., Evans, R. L., Collins, J. A., Elsenbeck, J., Gaetani, G. A., Gaherty, J. B., Hirth, G., and Lizarralde, D. (2015), The electrical structure of the central Pacific upper mantle constrained by the NoMELT experiment, Geochem. Geophys. Geosyst., 16, 10.1002/2014GC005709.

Schmerr, N. (2012), The Gutenberg discontinuity: Melt at the lithosphere–asthenosphere boundary, Science, 335, 1480–1483. doi:10.1126/science.1215433.

Schubert, G., Froidevaux, C., and Yuen, D. A. (1976), Oceanic lithosphere and asthenosphere: Thermal and mechanical structure, J. Geophys. Res., 81, 3525–3540.

Schutt, D. L. and Lesher, C. E. (2006), Effects of melt depletion on the density and seismic velocity of garnet and spinel lher-zolite, J. Geophys. Res., 111, 2003JB002950.

Sclater, J. C., Parsons, B., and Jaupart, C. (1981), Oceans and continents: Similarities and differences in the mechanisms of heat loss, J. Geophys. Res., 86, 11535–11552.

Selway, K., Ford, H. A., and Kelemen, P. (2015), The seismic mid‐lithosphere discontinuity, Earth Planet. Sci. Lett., 414, 45–57.

Shen, X., Liu, M., Gao, Y., Wang, W., Shi, Y., An, M., Zhang, Y., and Liu, X. (2017), Lithospheric structure across the north-eastern margin of the Tibetan Plateau: Implications for the plateau’s lateral growth, Earth Planet. Sci. Lett., 459, 80–92.

Shimamura, H. and Asada, T. (1976), Apparent velocity mea-surements on an oceanic lithosphere, Phys. Earth Planet. Inter., 13, P15–P22.

Shito, A., Karato, S., and Park, J. (2004), Frequency dependence of Q in Earth’s upper mantle inferred from continuous spectra of body wave, Geophys. Res. Lett., 31, 10.1029/ 2004GL019582.

Sodoudi, F., Yuan, X., Kind, R., Lebedev, S., Adam. J. M.‐C., Kästle, E., and Tilmann, F. (2013), Seismic evidence for strat-ification in composition and anisotropic fabric within the thick lithosphere of Kalahari Craton, Geochem. Geophys. Geosyst., 14, 2013GC004955.

Spetzler, H. A. and Anderson, D. L. (1968), The effect of tem-perature and partial melting on velocity and attenuation in a simple binary system, J. Geophys. Res., 73, 6051–6060.

Spiegelman, M. and Elliott, T. (1993), Consequences of melt transport for uranium series disequilibrium in young lavas, Earth Planet. Sci. Lett., 118, 1–20.

Stocker, R. L. and Gordon, R. B. (1975), Velocity and internal friction in partial melts, J. Geophys. Res., 80, 4828–4836.

Sundberg, M. and Cooper, R. F. (2010), A composite visco-elastic model for incorporating grain boundary sliding and transient diffusion creep: Correlating creep and attenuation responses for materials with a fine grain size, Philos. Mag., 90, 2817–2840.

Takei, Y. (2002), Effect of pore geometry on Vp/Vs: From equilibrium geometry to crack, J. Geophys. Res., 107, 10.1029/2001JB000522.

Takei, Y. and Holtzman, B. (2009), Viscous constitutive rela-tions of solid–liquid composites in terms of grain boundary contiguity: 1. Grain boundary diffusion control model, J. Geophys. Res., 114, 10.1029/2008JB005850.

Takeuchi, N., Kawakatsu, H., Shiobara, H., Isse, T., Sugioka, H., Ito, A., and Utada, H. (2017), Determination of intrinsic attenuation in the lithosphere–asthenosphere system, Science, 358, 1593–1596.

Tan, B., Jackson, I., and Fitz Gerald, J. D. (2001), High‐temper-ature viscoelasticity of fine‐grained polycrystalline olivine, Phys. Chem. Miner., 28, 641–664.

Tan, Y. and Helmberger, D. V. (2007), Trans‐Pacific upper mantle shear velocity structure, J. Geophys. Res., 112, 10.1029/2006JB004853.

Tauzin, B., Debayle, E., and Wittingger, C. (2010), Seismic evi-dence for a global low‐velocity layer within the Earth’s upper mantle, Nature Geosci., 3, 718–721.

Tharimena, S., Rychert, C. A., and Harmon, N. (2016), Seismic imaging of a mid‐lithospheric discontinuity beneath Ontong Java Plateau, Earth Planet. Sci. Lett., 450, 62–70.

Tharimena, S., Rychert, C., Harmon, N., and White, P. (2017a), Imaging Pacific lithosphere seismic discontinuities—insight from SS precursor modelling, J. Geophys. Res., 122, 2131–2152.

Tharimena, S., Rychert, C. A., and Harmon, N. (2017b), A unified continental thickness from seismology and diamonds suggests a melt‐defined plate, Science, 357, 580–583.

Thybo, H. (2006), The heterogeneous upper mantle low velocity zone, Tectonophysics, 416, 53–79.

Thybo, H. and Perchuć, E. (1997), The seismic 8° discontinuity and partial melting in continental mantle, Science, 275, 1626–1629.

0004167983.INDD 33 10/11/2018 8:51:10 PM

34 LITHOSPHERIC DISCONTINUITIES

Toramaru, A. and Fujii, N. (1986), Connectivity of melt phase in a partially molten peridotite, J. Geophys. Res., 91, 9239–9252.

Wang, D., Mookherjee, M., Xu, Y., and Karato, S. (2006), The effect of water on the electrical conductivity in olivine, Nature, 443, 977–980.

Wang, D., Guo, Y., Yu, Y., and Karato, S. (2012), Electrical con-ductivity of amphibole‐bearing rocks: Influence of dehydra-tion, Contrib. Mineral. Petrol., 164, 17–25.

Webb, S. and Jackson, I. (2003), Anelasticity and microcreep in polycrystalline MgO at high temperature: An exploratory study, Phys. Chem. Miner., 30, 157–166.

Webb, S., Jackson, I., and Fitz Gerald, J. (1999), Visoelasticity of the titanate perovskite CaTiO3 and SrTiO3 at high temper-ature, Phys. Earth Planet. Inter., 115, 259–291.

Wei, S. S. and Shearer, P. M. (2017), A sporadic low‐velocity layer atop the 410‐km discontinuity beneath the Pacific Ocean, J. Geophys. Res., 122, 5144–5159.

Wirth, R. and Long, M. D. (2014), A constrast in anisotropy across mid‐lithospheric discontinuities beneath the central United States—a relic of craton formation, Geology, 42, 851–854.

Wölbern, I., Rümpker, G., Link, K., and Sodoudi, F. (2012), Melt infiltration of the lower lithosphere beneath the Tanzania craton and the Albertine rift inferred from S recever functions, Geochem. Geophys. Geosyst., 13, 2012GC004167.

Yamamoto, J., Korenaga, J., Hirano, N., and Kagi, H. (2014), Melt‐rich lithosphere–asthenosphere boundary inferred from petit‐spot volcanoes, Geology, 42, 967–970.

Yang, Y., Forsyth, D. W., and Weeraratne, D. S. (2007), Seismic attenuation near the East Pacific Rise and the origin of the

low‐velocity zone, Earth Planet. Sci. Lett., 258, 260–268. doi:10.1016/j.epsl.2007.03.040.

Yoshii, T., Kono, Y., and Ito, K. (1976), Thickening of the oce-anic lithosphere, in G. H. Sutton, M. H. Manghnani, and R. Moberly (eds), The Geophysics of the Pacific Ocean Basin and its Margin, pp. 423–430, American Geophysical Union, Washington, DC.

Yoshino, T., Yamazaki, D., and Mibe, K. (2009), Well‐wet-ted olivine grain boundaries in partially molten peridotite in the asthenosphere, Earth Planet. Sci. Lett., 283, 167–173.

Yuan, H. and Levin, V. (2014), Stratified seismic anisotropy and the lithosphere asthenosphere boundary beneath eastern North America, J. Geophys. Res., 119, 3096–3114.

Yuan, H. and Romanowicz, B. (2010a), Depth dependent azimuthal anisotropy in the western US upper mantle, Earth Planet. Sci. Lett., 300, 385–394.

Yuan, H. and Romanowicz, B. (2010b), Lithospheric layering in the North American craton, Nature, 466, 1063–1066.

Yuan, X., Heit, B., Brune, S., Steinberger, S., Geissler, W. H., Jokat, W., and Weber, M. (2017), Seismic structure of the lithosphere beneath NW Namibia: Impact of the Tristan da  Cunha mantle plume, Geochem. Geophys. Geosyst., 18, 125–141.

Zener, C. (1941), Theory of the elasticity of polycrystals with viscous grain boundaries, Phys. Rev., 60, 906–908.

Zhu, L. and Kanamori, K. (2000), Moho depth variations in southern California from teleseismic receiver functions, J. Geophys. Res., 105, 2969–2980.

0004167983.INDD 34 10/11/2018 8:51:10 PM

3

Bioenergy and Land Use Change, Geophysical Monograph 231, First Edition. Edited by Zhangcai Qin, Umakant Mishra, and Astley Hastings. © 2018 American Geophysical Union. Published 2018 by John Wiley & Sons, Inc.

1Bioenergy and Land Use Change: An Overview

Pankaj Lal1, Aditi Ranjan2, Bernabas Wolde1, Pralhad Burli1, and Renata Blumberg3

1.1. INTRODUCTION

The United States is the largest consumer of petroleum products in the world, consuming around 7.45 million barrels per day in 2012 [Energy Information Administration (EIA), 2014]. A significant share of these petroleum products is imported from politically unstable regions of the world. This reliance on fossil fuels has led to economic, social, and environmental concerns that have gained public attention. Bioenergy appears to offer hope by reducing the gap between domestic energy supply and demand, diversifying energy sources, reducing greenhouse gas (GHG) emissions, and by providing socioeconomic benefits in the form of additional income and new jobs.

Bioenergy encompasses energy produced from biomass and includes fuels such as sugarcane‐ or corn‐based eth­anol, biofuels produced from energy grasses, farm residue, and woody materials, as well as energy obtained from other plant‐based sources. Agricultural and forested biomass‐based energy is considered an option to reduce dependency on fossil fuels, increase the current share of the nation’s renewable energy, and improve the sustain­ability of forests and marginal lands. Cellulosic biomass‐based energy or second‐generation fuels, for example, fuels produced from energy grasses or woody biomass, have certain advantages over other energy sources, such as first‐generation fuels like corn ethanol, because they limit competition between agricultural food crops and those destined for fuel production [Hill et  al., 2006]. While the development of cellulosic biofuels could result in competition for use of resources, including water, labor, carbon storage, and financial resources, one of the most important issues surrounding bioenergy mar­kets is land use impact [Searchinger and Heimlich, 2015].

ABSTRACT

Bioenergy production can have direct and indirect land use impacts. These impacts have varied implications, ranging from land tenure, commodity production, urbanization, carbon sequestration, and energy independence to several others. In recognition of its broad and intricate impacts, a growing amount of research focuses on this area, hoping to address the controversies and inform the relevant policies in a way that ensures more sustainable outcomes. In this chapter, we provide a summary of the research around land use change economics and modeling. We examine various concerns, as well as their empirical evidences, and outline the conceptual opportu­nities and challenges involved in measuring both direct and indirect land use change. We also describe a number of modeling methods that have been used in previous studies, including spatially disaggregated modeling approaches, econometric land use change approaches, and integrated environmental economic approaches. The chapter concludes with an analysis of policy imperatives and suggestions that could form the foundation of a more sustainable bioenergy development pathway.

1 Department of Earth and Environmental Studies, Montclair State University, Montclair, New Jersey, USA

2 MYMA Solutions LLC, Lincoln Park, New Jersey, USA3 Department of Nutrition and Food Studies, Montclair State

University, Montclair, New Jersey, USA

0003312060.INDD 3 10/23/2017 3:19:15 PM

COPY

RIGH

TED

MATER

IAL

4 BIOENERGY AND LAND USE CHANGE

There are varied answers to the following question: What is the land use impact of using biomass‐based energy, and how does it change over time? Arable land is a scarce resource that is already under pressure from food agri­culture, forestry, industry, urban, and other demands. The situation is further complicated by the fact that the land use change (LUC) associated with increased bioen­ergy production can negate the potential benefits of production and ultimately degrade the environment. The bioenergy market is largely encouraged on the national/supranational scale by government initiatives; countries with notable biofuel initiatives include the member states of the European Union, the United States, and Brazil. Without taking into account both the economic and envi­ronmental factors that surround the bioenergy market, these initiatives can lead to poor land use decisions at the local level [Baker et al., 2010; de Oliveira Bordonal et al., 2015; Vasile et al., 2016].

Economic methods of biomass supply estimation are based on the interplay of demand and supply markets of bioenergy. The bioenergy market is shaped by the inter­play between global oil fuel, other types for energy sources, price of substitutes, bioenergy production costs, and alternative uses for arable land. A high fuel price not only increases the demand for biofuels by increasing the incentive for the production of alternative sources but also increases the cost of cultivation or harvesting. Bioenergy production also affects potential profit mar­gins for the sale of biofuels, and if costs of production are high, actual production will necessarily drop [Rajcaniova et al., 2014].

As fossil fuel prices rise, the production of bioenergy becomes more profitable and therefore leads to the conversion, or construction, of agricultural land for bio­fuel production. For example, Piroli and Ciaian [2012] showed that a one‐dollar increase in per‐barrel price of oil could encourage the planting of between 54,000 and 68,000 ha of bioenergy cropland globally. The same study found that increasing oil prices increased agricultural area globally by 35.5 million ha, out of which biofuel feedstocks were 12.12 million ha/year. Volatility in the oil market encourages the use of first‐generation biofuel crops, such as maize, wheat, and soybean, over second‐generation biofuels, like perennial grasses, because of the comparatively low turnaround time of first‐generation bioenergy crops. This variable demand for bioenergy, met by quick switches to energy crops over food, as prices fluctuate, can lead to not only fluctuations in supply for conversion plants but also unsustainable biomass cultiva­tion and/or harvesting practices resulting in modified land use, deforestation, soil degradation, and GHG emis­sions, among others. Often the first lands to be developed are pasturelands and other croplands. However, other land uses are converted to biofuel production with an

increasing demand for biofuels, including forest and swampland [Rajcaniova et al., 2014].

Uncontrolled bioenergy transition can result in increased agricultural runoff because of improper culti­vation and harvest methods. This has been observed in the red river basin in the Dakotas. When the primary crops in the area changed to corn with increased biofuel demand, sediments increased by 2.6%, phosphorous by 14.1%, and nitrogen by 9.1% [Lin et al., 2015]. Second‐generation biofuels tend to increase soil organic carbon (SOC) when planted on cropland but tend to decrease SOC when planted on their native counterparts: forests and grasslands [Harris et al., 2015].

There is also apprehension that the conversion of land to bioenergy can lead to more arable land being opened up to raise supply. This increase in agricultural land has the potential of being environmentally deleterious, as more sensitive natural environments may be repurposed. Land use change to agricultural row systems can also cause habitat loss [Jonsell, 2007]. Land use change from natural forests to forest plantations, including short‐rota­tion woody crops, is an important area of concern from an ecological point of view [Wear et al., 2010]. Fargione et al. [2008] contend that the conversion of lands, such as rainforests, peatlands, and grasslands, to produce crop‐based biofuels in Brazil, Southeast Asia, and the United States could potentially release 17–420 times more CO2 than the annual GHG reductions from the use of these biofuels. Meanwhile, Plevin et  al. [2010] estimate emis­sions associated with indirect land use change for U.S. corn ethanol ranging between 10 and 340 g CO2e MJ−1 for a variety of modeling scenarios and assumptions. Similarly, using a spatially explicit model to project land use changes, Lapola et al. [2010] suggest that indirect land use changes resulting from expansion of biofuel planta­tions in Brazil could create a carbon debt that would take about 250 years to be repaid. Biomass production might also have negative consequences unless coordinated with breeding and nesting seasons and maintaining cover for overwintering small mammal species [Bies, 2006]. However, interventions focused on ecological res­toration or fuel‐reduction activities associated with woody biomass can also benefit wildlife habitat [Janowiak and Webster, 2010].

In the face of a growing bioenergy sector and associ­ated policy incentives, land use analysis is considered critical for the future of bioenergy markets. Many authors have explored this issue; what has been lacking is a systematic analysis of trends, evidence, and complexity in assessing bioenergy market growth and associated land use impacts. Toward this goal, a comprehensive literature review was undertaken. We review the problems, applica­tions of economic techniques, methodological complex­ities, and certification efforts from the literature, focusing

0003312060.INDD 4 10/23/2017 3:19:15 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 5

more on the ways in which bioenergy and land use issues have been approached by economists. We focus less on the methodological aspects and rely more on compara­tive results and outcomes in order to provide the reader a broad understanding of current research in this area.

The rest of the chapter is organized as follows. In the following section, we discuss forest biomass supply assessments, focusing on some of their differences and similarities. We then turn to the modeling efforts and methodologies that have been used to estimate land use change impacts associated with bioenergy markets. In the fourth section, we focus on the empirical evidence of land use change analysis and challenges such as uncer­tainty and modeling challenges. We discuss technological and policy imperatives and bioenergy certification in the fifth section. Finally, we summarize observations and provide perspectives on the future of bioenergy markets and land use impacts.

1.2. LAND USE CHANGE AND CURRENT RESEARCH

Land use change is an important component of the use of biofuels, and it is important in terms of delineating effect of biofuels on GHG emissions and carbon seques­tration by considering the lifetime GHG effects of the fuels as opposed to their effects when they are only burning. It is increasingly being recognized that land use effects of bioenergy production are linked to net GHG reductions. Assuming that energy crops do not lead to land use changes, life cycle analyses of different biofuels (including woody biomass) suggest overall GHG reduc­tions [Birdsey et  al., 2006; Blottnitz and Curran, 2007; Eriksson et al., 2007; Gustavsson et al., 2007]. However, Searchinger et al. [2008] argue that life cycle studies have failed to factor in indirect land use change effects and suggest that using U.S. croplands or forestlands for biofu­els results in adverse land use effects elsewhere, thus harming the environment rather than helping it.

To this end, researchers and organizations, including the Intergovernmental Panel for Climate Change [Watson et  al., 2000], the National Wildlife Federation, and the Union of Concerned Scientists have put considerable effort into defining and studying the effects of land use change in biofuels [Watson et al., 2000; Searchinger et al., 2008; Union of Concerned Scientists, 2008; Plevin et al., 2010; National Wildlife Federation, 2014].

1.2.1. Direct Land Use Change

Direct land use change (DLUC), the direct change of land usage due to increased biofuel production, has the most obvious and measurable effect on the land and surrounding areas. DLUC occurs most commonly when

uncultivated areas, such as forests or grasslands, are converted into farmland for the production of biofuel crops. Direct land use change can have considerable con­sequences for GHG emissions and other environmental concerns [Union of Concerned Scientists, 2008; National Wildlife Federation, 2014]. Destruction of forest ecosys­tems, for example, causes a significant amount of carbon stored in the forests to be released, which can offset many of the carbon advantages of using biofuels [National Wildlife Federation, 2014]. In addition to this, DLUC can also impact the environmental benefits that these areas could provide, including biodiversity and ecosystem ser­vices such as water filtration, erosion control, and ground water recharge. DLUC, when considering these factors, can result in adverse ecosystem trade‐offs.

DLUC is an important consideration for any bioen­ergy project. The environmental and economic cost of clearing land, planting, and growing the bioenergy plants can also influence net GHG emissions. The emissions vary considerably, depending on the crop and the area in which they are planted, so bioenergy crops should be carefully considered for the potential plantation area before planting to ensure that emissions do not outweigh the benefit of the crop. Fueled by some of the above mentioned concerns, DLUC impacts arising because of biofuel production have come under scrutiny in recent years. Sustainability initiatives, both in the United States and abroad, have made it a priority to understand the consequences of DLUC [Stappen et  al., 2011; Jones et al., 2013]. Because of the nature of land use change and the differing nature of bioenergy crops and cultiva­tion practices, these result in different economic and environmental cost and benefits [Natural Resource Defense Council (NRDC), 2014].

Sugarcane, for example, is a bioenergy crop that has gained popularity over the years and is currently a staple in the United States and Brazil, two countries that together account for around 90% of the world’s ethanol production. In addition to potentially reducing GHG emissions by nearly 85% in comparison to fossil fuels, in areas of Brazil, converting land to sugarcane production can potentially lead to positive benefits, such as local climate cooling [de Oliveira Bordonal et al., 2015]. However, sugarcane cultivation practices for bioenergy production, particularly the burning of plant residues during harvest time, can contribute to GHG emissions. De Oliveira Bordonal et  al. [2015] sought to quantify some of the effects of DLUC and provide an analysis of the effects of sugarcane production on GHG emission. They found that the expansion of sugarcane plantations contributed to significant GHG emissions from agricultural produc­tion, but around 57% of this was offset through carbon uptake in the new biomass. However, calculations of GHG emissions can differ depending on the system

0003312060.INDD 5 10/23/2017 3:19:15 PM

6 BIOENERGY AND LAND USE CHANGE

boundaries of the assessment and key drivers of the land use change, coupled with factors such as ecosystem services and carbon pools and sinks, among others.

It is also important to note that the GHG emissions from land use can differ wildly, depending on the loca­tion in which they are grown. A study by Bailis and McCarthy [2011] found that there was a considerable difference in the carbon debt between jatropha planta­tions in India and Brazil as a result of the differing cli­mates and soil types. The study by Stappen et al. [2011] confirms this by finding a considerable difference in net GHG emissions between soy grown in the United States (27%) and Argentina (−568%). This has important implications as it indicates that there is no one‐size‐fits‐all bioenergy crop that has superior performance com­pared to all others in all places and at all times. As such, bioenergy should be considered for each area individu­ally to fully understand how it will perform and how much work in the form of emissions and labor will be produced in their growth.

1.2.2. Indirect Land Use Change

Indirect land use change (ILUC), a secondary consideration in land use change for bioenergy crops, includes the effects that are related to, but not immedi­ately caused by, the cultivation of bioenergy crops. Because it is difficult to define where ILUC begins and ends, it is far more difficult to calculate than DLUC. Though some studies attempt to approximate ILUC, it is perhaps more important to understand some of the dynamic factors that influence the proliferation of ILUC and get a more complete understanding as to how a bioenergy operation will affect the surrounding land. Understanding the possible extent of ILUC is critical for making management decisions that can accurately and successfully reduce carbon emissions while simulta­neously protecting local ecosystems and ensuring that the operation does not produce an unhealthy amount of GHGs.

The ILUC impact of biofuel feedstocks was brought to the forefront during the “food versus fuel” debate. This common criticism of bioenergy refers to the conflict between using food crops for fuel instead of food [Naylor et  al., 2007]. The price of crops, such as corn, that are used for both food and fuel increased in some instances because more of it was being used for fuel. This can often extend to land use change elsewhere, as displaced food‐growing operations in one part of the world may result in land being diverted for growing bio­energy crop elsewhere [Doornbosch and Steenblik, 2008]. Such a change may happen in geographically discon­nected areas, as it is largely a result of market mechanisms,

and therefore can encourage a significant change in land use that cannot be traced firmly back to any singular bioenergy operation. This proliferation of changing land use as a result of bioenergy operations can result in increased GHG emissions if carbon‐rich ecosystems are converted to farmlands [Fritsche et al., 2010]. What makes the problem “wicked” is the fact that ILUC is more difficult to quantify, and it is more difficult to identify the drivers resulting in adverse bioenergy‐based land use change.

Because ILUC is difficult to quantify, there is some confusion and even skepticism in the scientific community over its relevance. Finkbeiner [2014] posits that ILUC quantification methods are still in their infancy and that there exists no proven method to accurately convey how ILUC affects GHG emissions. He cites wildly varying estimates (from −200% to over 1700%) among studies that try to quantify the effects of this change, far more than other scientific studies of its type, and notes that there are no relevant standards for this type of study at this time. He notes that major international standards, such as the EU Product Environmental Footprint Methodology and ILCD Handbook, do not include ILUC in their calculations, bringing their relevance and reli­ability into question. He also points out that adding ILUC emissions into the emissions of major biofuels may be misleading and may result in a lopsided comparison with fossil fuels; ILUC can also occur with the produc­tion of fossil fuels, but they are never included in fossil fuel emission calculations. Moreover, methodologies including comprehensive life cycle assessment (LCA) and input‐output models can potentially address these challenges by accounting for direct and indirect land use change more precisely [Liang et  al., 2012; Marvuglia et al., 2013; Dilekli and Duchin, 2016]. Though ILUC can represent an important piece of the land use change emis­sions, it is important to consider the concerns brought up by this study and others as this field of study matures. Only then can the indirect effects of biofuels be accurately represented. Though the study of its effects is imperfect, ILUC can be deemed important to under­standing the complete effect of biofuel production on the environment.

1.2.3. Current Research Trends in Bioenergy and Land Use Literature

Though the production of biofuels has advanced considerably in recent years, it has also led to some con­cerns. This is evident from our word‐cloud analysis whereby we quantitatively analyzed large collections of textual information to evaluate some of the most widely cited publications of the previous decade. Text mining

0003312060.INDD 6 10/23/2017 3:19:15 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 7

has become more sophisticated and has benefitted from computational and technological advances in linguistics, computer science, and statistics [Meyer et al., 2008]. We analyzed 46 papers published since January 2006 to August 2016, encompassing the broad theme of bioen­ergy and land use. These publications were cumulatively cited over 3000 times according to citation statistics com­piled using Google Scholar. The analysis was performed using text‐mining features in the programming language R [Williams, 2016].

The word cloud provides a visual of the 50 most fre­quently used words in the papers analyzed in the text analysis and illustrates some of the important focus areas in previously published land use and bioenergy literature (Figure 1.1). Within the text‐mining algorithm, we exclude numbers, as well as commonly used conjunc­tions, prepositions, and “stopwords” such as “all,” “almost,” and “largely,” among others. It is interesting to note that along with the obvious focus on biomass, land, and bioenergy, the text analysis identifies “food,” “crops,” “agriculture,” “forest,” and “feedstock” as keywords with a relatively high frequency. This suggests that the food versus fuel argument and conversion of forestland for cultivating bioenergy feedstocks emerge as an impor­tant story line. Finally, “oil,” “gas,” “ethanol,” and “fuel” also feature in many publications. While it is plausible to infer that the frequency of words is correlated with the number of publications included in the analysis, enhanced computational and analytical capacities allow us to decipher important trends in the literature in relatively less amount of time.

1.3. MODELING EFFORTS AND METHODOLOGIES ESTIMATING LAND USE CHANGE

AND BIOENERGY MARKETS

Several approaches, including land tenure, urbaniza­tion, energy production and consumption, climate change, economic growth, and population growth, have been used to model land use [Irwin and Geoghegan, 2001; Lambin et  al., 2001]. These approaches are crucial in helping us understand, quantify, and predict likely social, economic, and environmental outcomes, all of which are valuable in informing relevant decisions and in mini­mizing potential adverse outcomes while maximizing the potential positive outcomes. The results are useful for making informed decisions regarding land use planning and policy.

In order to quantify land use changes, LCA is often used. LCA is a methodology that attempts to bring all the factors of a crop’s life cycle into the equation, from its conception to its disposal, to fully understand their effects. Though many different crops have different carbon‐emission savings when compared to fossil fuels and this number can be calculated [Stappen et al., 2011], large differences in the LCA of these crops generally come from differences in how direct and indirect land use changes are assessed [Marvuglia et al., 2013]. For example, generally more land use contributes to higher GHG emis­sions and therefore can cause a net GHG increase; studies that assess a larger area of land use, likely by including more land from indirect use, may appear to have higher emissions and therefore lower net savings [Finkbeiner, 2014]. Furthermore, different types of LCA may change the net balance; Marvuglia et  al. [2013], for example, describe the difference between LCA, which aims to describe the impacts of the human economy on the global environment, and consequential LCA (CLCA), which aims to show how the environment will respond to pos­sible decisions. All of these factors can ultimately lead to considerably different results, and thus the major factors, namely, land use change, must be assessed in order to understand how they change the models.

The way land use change is modeled depends partly on the drivers, including social, economic, technolog­ical, biophysical, political, and demographic, and the implications of land use change one aims to model. The subject matter, actual use, applicability, and type of information available to the model developer also affect the way one models land use change [Adams et al., 1999]. On the basis of the techniques adopted and end use, these approaches can be sorted into various groups, the operational classification for this study being (i) spa­tially disaggregated approach, (ii) economic approach, and (iii) integrated environmental economic approach.

Figure  1.1 Word cloud representing the 50 most frequent words in the text analysis.

0003312060.INDD 7 10/23/2017 3:19:15 PM

8 BIOENERGY AND LAND USE CHANGE

Each of these approaches is best suited to handle a specific set of land use change drivers and relevant impli­cations by addressing a given problem from different specialized points of view. Given their unique perspec­tive, these approaches differ in their unit of analyses, land use of interest, intended user, contagion criteria, temporal and spatial considerations, and their ability to model emergent behavior [Agarwal et al., 2002]. Their respec­tive scopes can be local, regional, national, or even inter­national, and they can either project future trends or describe the processes that resulted in the change that has already occurred [Drummond and Loveland, 2010]. The given approaches and the specific models used can follow a probabilistic or deterministic transition rule for conver­sions among given land use classes [Agarwal et al., 2002].

While the specific perspective may differ, the models are not always mutually exclusive. For instance, the statistical models and transitional matrices are comparable, whereas hybrid models build on the strengths of one model and overcome inherent weaknesses [Briassoulis, 2000; Gibson et al., 2000]. Quality control of the results can be assessed through the realism in capturing relevant underlying processes, precision of results in describing data, and the generalizability or replicability of the results in other set­tings [Grimm et al., 2005].

The models also share some common constraints, including the uncertainty and limited availability of his­torical land use information, which is compounded while modeling feedback loops [Kim and Dale, 2009]. Other factors being the same, variations in input data and related assumptions may affect the estimated land use and cover change and associated impacts [Center for BioEnergy Sustainability (CBES), 2009]. In what follows, we describe each approach and give examples of specific cases where the said approaches are used.

1.3.1. Spatially Disaggregated Modeling Approaches

This approach explicitly accounts for the spatial hetero­geneity of the area of study. It can also account for the various land use options available to a given land, allowing for the inclusion of relevant neighborhood conditions such as presence and proximity of developed sites, roads, and land features. This is important in assessing correlating land uses and changes, and it increases the chances of cor­rectly predicting the amount, probability of conversion between different land use classes, and type and stability of land use change over time [Brown, 2002]. Moreover, current and historical land use patterns and the contagion specification help to validate model prediction and check the reliability of the results [Clarke et  al., 1997; Pontius and Schneider, 2001]. The unit of analyses affects the data requirement, precision of results, and relevance of the results to varying stakeholders [Walsh et al., 2001].

In the context of bioenergy, regional differences exist in the biomass yield along with the corresponding biofuel yield, both positive and negative. Moreover, different feedstocks have varying agronomic conditions and input requirements [Varvel et  al., 2008]. This approach can prove useful in accounting for such variations while modeling the extent and outcomes of land use change associated with feedstock production. However, the scale and quality of remote‐sensing data and other types of data have to be uniform, making it challenging to model land cover and land use change effects from biomass regrowth [Schulze, 2000]. The reliance on existing and historic land uses is also intractable for emerging land use develop­ments, distant future projections magnifying the problem and showing the need for reasonable and flexible thresh­olds systems for given land uses [Agarwal et al., 2002].

The more complex spatial models, including spatially representative and spatially interactive models, can incor­porate or produce data at up to three spatial dimensions. The area base model, for instance, predicts land use proportions among farmland, forests, and urban or other types of land uses [Hardie and Parks, 1997]. Using counties as units of analyses, it attempts to predict and explain the coexistence of several land uses and conversion among them by using their respective heterogeneous attributes.

Alternatively, the spatial dynamic model predicts shifts in cultivation for given topographies on the basis of their proximity to urban centers by using site productivity, ease of clearing, and erosion hazard, among others, as variables [Gilruth et al., 1995]. However, it has a relatively large unit of analyses, 6 km2. Similarly, Chomitz and Gray [1996] use spatially disaggregated information such as wetness, road, and slope to determine land use among natural vegetation and farming. Although it implicitly accounts for human decision making, a more explicit accounting of the human dimension could increase its application. In addition, a longitudinal analysis, in lieu of a cross‐sectional analysis, should be included. The challenge for this approach is that it does not always account for economic agents and policy changes, a challenge that can be addressed by employing econo­metric approaches.

1.3.2. Economic Land Use Change Approaches

This approach identifies and explicitly accounts for multiple economic actors, their interaction, and the drivers they respond to, including tax and energy policies, to determine the probability and magnitude of land use change [Walker et al., 2000; Brown, 2002]. While in some cases the econometric model uses the land’s exogenously determined market price, or an equivalent thereof, selling price, and conversion cost between alternative uses for

0003312060.INDD 8 10/23/2017 3:19:15 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 9

given land use class [Bockstael, 1996], prices can be modeled endogenously as well. Thus, the land and the physical processes that involve land are treated in economic terms, not in physical terms as in the spatially disaggregated approach.

One can also assess consumer and producer surpluses and carbon path under varying policies over time and at the regional level [Adams et al., 1996]. It allows for scenario analyses, where varying policy options are evalu­ated for potential land use impact along with forcing factors such as population growth and commodity demand. The ability of such approaches to handle differ­ent policy scenarios makes it useful in assessing and quan­tifying trade‐offs associated with varying options available to decision makers. Given the temporal dynamics of the said economic variables, the time span between different land uses can also be determined by using hazard and survival models, respective modeling varying from one another in the smallest temporal unit of analyses for a given change to occur [Irwin and Geoghegan, 2001]. Such approaches combine the probability of land use change with information on the timing of such change and how long given changes last, the time dimension providing an additional contextual layer.

Moreover, this approach can account for individuals’ socioeconomic attributes and motivation [Pfaff, 1999]. Given their stake in land ownership and biomass supply decision, the ability to account for such stakeholders is crucial in ensuring more accurate estimations and policy‐relevant results [Nelson et  al., 1999; Ahn et  al., 2000]. Using agricultural inputs such as agricultural prices and profitability, timber age, site condition, and the lagged values associated with land uses, for instance, such an approach can allocate land for competing uses in a way that maximizes a predefined objective function [Chomitz and Gray, 1996].

Similarly, the production of a given feedstock affects its other uses, the resource allocation to the other types of feedstock, and how that reflects on the relevant market prices. These intricate relationships are important in understanding the dynamics within the bioenergy feed­stock production system, and the econometric approach is suited to handle such relationships in modeling land use change [Perlack et al., 2011].

Using a timber model, forage production, and nontim­ber benefits, Swallow et  al. [1997] simulate the optimal harvest sequence. By including the discounted values of future cash flow from alternative uses for the land, this approach provides a decision support tool. The NELUP extension model [O’Callaghan, 1995], on the other hand, uses linear programming at the farm level to determine the optimal ways to maximize profit for a given set of resources and farm activity. It also explicitly accounts for individual’s risk taking or aversion behavior.

Similarly, the forest and agricultural sector optimiza­tion model uses a dynamic nonlinear model to assess the relationship between forestry, agriculture, and terrestrial carbon [Adams et  al., 1996]. It maximizes economic welfare of the decision maker and determines optimal allocation of land to alternative uses. Moreover, it allows for policy effects and has a feedback loop for inter­temporal price dynamics. Besides the data intensity, this approach does not fully account for parcel contagion criteria and potential candidates for conversion where all plausible land use options are not accounted for, a challenge addressed by spatially explicit approaches.

1.3.3. Integrated Environmental Economic Approaches

This approach couples landscape attributes with the relevant economic agent or economic drivers to deter­mine land use change in a dynamic setting. Its ultimate goal is to understand human‐environment dynamics across space, time, and decision‐making process [Grimm et al., 2000; Agarwal et al., 2002]. It can feature data on relevant development and household. Such models can also predict feedback effects between land use decisions, environmental outcomes, and relevant policies. It is also crucial in handling emerging land use developments, including energy production and progress on the relevant technologies.

The way land is managed, before and after the change, may be as important as the amount of land use and cover change, having notable effects, both in the short term and long term, on ecological, biophysical, economic and social, and climatic outcomes [Fargione et  al., 2008; Searchinger et al., 2008]. Thus, interest exists in tracking, evaluating, and monitoring its consequences in a way that accounts for the complex relationship that exists between the physical and the socioeconomic systems [Tesfatsion, 2001].

The integrated environmental economic approach is suited to handle such relationships in modeling land use change. In the case of bioenergy, for instance, the original use of the land can affect the overall energetic and GHG performance of the bioenergy produced. While some feedstocks compete with prime agricultural land, others do not. Instead, they grow on marginal and even contam­inated sites, which are otherwise unusable, and in the pro­cess, they help restore its functional capacity [Liebig et al., 2005; McLaughlin and Kszos, 2005; Mitchell et al., 2010]. Similarly, woody bioenergy can also improve forest conditions and reduce the fire and disease outbreak risk associated with overstocked forests [Polagye et al., 2007]. These benefits need to be considered for the purpose of comprehensiveness. By using this approach, one can inte­grate physical information on land use change with policy and economic information, including sustainable land

0003312060.INDD 9 10/23/2017 3:19:15 PM

10 BIOENERGY AND LAND USE CHANGE

management practices and upcoming technologies, such as carbon capture and storage [Barkley, 2007]. However, since they account for varying drivers and impacts simul­taneously, such models can be taxing in both data and computation needs and usability by varied stakeholders.

The general systems framework uses a recursive linear programming model at the regional and catchment levels and takes into account hydrological and ecological models and inputs, including soil characteristics, species, meteo­rological data, and input/output farm data [O’Callaghan, 1995; Alberti and Waddell, 2000]. It helps one determine the dynamics between market forces, hydrology, and ecological outcomes. Such frameworks explicitly model choices of decision makers using probability functions and include relevant socioeconomic determinants while having an ecological analysis unit of 1 km2.

Similarly, the land use change analyses system model integrates socioeconomic module with landscape and impact modules [Berry et  al., 1996]. It models the transition probability matrix, develops new land use maps, and shows corresponding impacts on species hab­itat, using 90 m2 as its unit of analyses.

1.4. EMPIRICAL EVIDENCE AND MODELING CHALLENGES

Given the contextual considerations, differences in feedstock and technology, and the varied working assumptions used by the studies, direct comparison of the previous studies on land use change is not always possible. However, researchers have attempted to not only evaluate the physical aspects of land use change but also study effects on a variety of variables ranging from GHG emission to soil quality and biodiversity. Moreover, whether or not the switch to biofuels will result in carbon savings could be significantly influenced by the types of land that are used to produce them [Searchinger et  al., 2008; Lapola et al., 2010].

Andersen [1996] analyzed the determinants of defores­tation in the Brazilian Amazon using two different mea­sures. The first measure of deforestation was based on satellite photos, whereas the second was based on land surveys. Using regression analysis, the author concluded that local economic forces were more important factors than the government’s development policy at explaining deforestation across 316 municipalities during 1975–1985. More recent studies, which use spatially disaggre­gated data, also combine other modeling techniques in an attempt to provide a more in‐depth analysis. Tompkins et al. [2015] present a deforestation and land use change scenario generator model that interfaces with dynamic vegetation models. The model named FOREST‐SAGE disaggregates the regional‐scale scenario to the local grid‐ scale scenario using certain risk rules based on physical

and socioeconomic attributes. Their model successfully reproduced spatial patterns of forest‐cover change as recorded by the Moderate Resolution Imaging Spectroradiometer Vegetation Continuous Field data.

Since the expansion of biofuels is a relatively recent phenomenon, there has been limited data and evidence to show that biofuels have resulted in large‐scale changes to land use [Taheripour and Tyner, 2013]. Yet in many cases, the relationship between biofuel production and land use change is a contentious issue. In the United States, Piroli and Ciaian [2012] employed time series analysis to eval­uate the relationships between fuel‐price changes and land use change, both direct and indirect, to test for inter­dependencies and the role of biofuels. They analyzed data for the 1950–2007 time period for five majorly traded agricultural commodities, the cultivated agricultural land area, and crude‐oil prices. The study concludes that the expansion of the bioenergy sector has indeed accelerated land use change in the United States wherein bioenergy crops are being cultivated on lands previously used for food crops. Meanwhile, Taheripour and Tyner [2013] state that the allocation of cropland in the United States has witnessed significant changes over the past two decades, in response to a variety of market conditions and policies. However, the total amount of harvested land has remained relatively unchanged.

Rajcaniova et  al. [2014] also estimated the impact of bioenergy on global land use change using econometric techniques. They concluded that increasing energy prices and biofuels’ production had a significant impact on global land use change through both direct and indirect pathways. They estimate a 12.12 million ha increase in global agricultural area due to higher biofuel production, which translates to 0.25% of the total worldwide agricul­tural area. Furthermore, the results for region‐specific estimates indicate a yearly total land use change increase due to biofuel production in Asia, South America, and North America. Similarly, Al‐Riffai et al. [2010] found that biofuel policies in the European Union resulted in indirect land use changes through deforestation in other parts of the world.

In a simulation‐based study, Lal [2011]has applied a modified version of Subregional Timber Supply model [Abt et  al., 2000] spanning 13 southern states in the United States and found that bioenergy markets arrest the decline in private forest lands and in fact lead to an increase if higher opportunity costs gets translated to high bioenergy demand. The results of moderate bioen­ergy demand scenario suggest that the Southern private forestland acreage increases by 4.6% from 175.39 million acres in 2010 to 183.47 million acres in 2050 (Figure 1.2). This acreage is 18% higher than no woody bioenergy sce­nario results. This is largely due to the increase in planted pine acreage (20% from 2007 levels), which offsets the

0003312060.INDD 10 10/23/2017 3:19:15 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 11

acreage decline from other forest types. The percentage decrease in the acreage of other four forest types (natural pine, oak pine, upland hardwood, and lowland hard­wood) is also much smaller than the low woody biomass scenario or where woody bioenergy markets do not exist.

Finally, we consider an example of integrated environ­mental economic models. Over the years, Brazil has witnessed a substantial increase in acreage allocated for sugarcane cultivation, a large portion of which is used for biofuel production [de Souza Ferreira Filho and Horridge, 2014]. De Sá et al. [2013] empirically evaluated the poten­tial indirect effects of sugarcane cultivation on forest conversion in the Amazon region in Brazil. The impact of factors such as road density, access to markets, credit availability, soil quality, precipitation on deforestation in the Brazilian Amazon was evaluated within the modeling framework. Furthermore, the study also sheds light on the potential shift of cattle ranching from the Sao Paulo region to the Amazon. The authors claim that sugarcane expansion has led to the movement of cattle‐ranching activities form the Center‐South region toward the

Amazon and that the indirect effect owing to displace­ment is both significant and nonnegligible. They also suggest that the results of their study are consistent with those obtained through studies relying on remote‐sensing evidence. Meanwhile, Elobeid et al. [2011] found that the expansion of sugarcane cultivation came at the expense of other crops and pastures in Brazil.

1.4.1. Modeling Challenges

Land use change analyses often focus on the under­lying processes that result in the alteration of landscapes and try to evaluate the implications of such changes [USDOE, 2011a]. Reports, including USDOE [2011b], state that while land use models focus on specific sec­tors of the economy, the complexity of real‐world inter­actions, such as the effects of biofuel policies in one part of the world on land use influences in another part, is difficult to capture accurately. Furthermore, collabo­rations between researchers are required to improve methodological uncertainties as well as for improved data

(a)

(b)

Mill

ion

acre

s

Year

200

180

160

140

120

100

80

60

40

20

02010 2015 2020 2025 2030 2035 2040 2045 2050

Mill

ion

acre

s

Year

200

180

160

140

120

100

80

60

40

20

02010 2015 2020 2025 2030 2035 2040 2045 2050

Planted pine

Natural pine

Oak pine

Upland hardwood

Lowland hardwood

Planted pine

Natural pine

Oak pine

Upland hardwood

Lowland hardwood

Figure 1.2 Private forest acreage change for five forest types in Southern United States comprising 13 states under (a) no biomass diverted to energy scenario and (b) moderate consumption of woody biomass for energy scenario. (See insert for color representation of the figure.)

0003312060.INDD 11 10/23/2017 3:19:16 PM

12 BIOENERGY AND LAND USE CHANGE

collection and integration. Some of the most pressing challenges from a data perspective include the need for better quantification of land cover, accurate measurement of environmental impact of land use change, model validation through comparison with empirical data, and accurate evaluation against other energy alternatives. More refined modeling approaches that document and quantify the uncertainty in land use change models and allow for improved integration with economic models are also required.

Nassar et al. [2011] note that agricultural activities tend to be one of the primary drivers of land use change at a global level. The authors discuss the most common methodologies used by policymakers to quantify the land use change impacts of biofuel expansion. The paper highlights some of the limitations of current economic models and suggests that the calibration of economic parameters, for example, price elasticity and land supply elasticity in accordance with historical land use data and satellite imagery, could be a step in the direction of improving model accuracy. The authors also conclude that the evaluation of land use change ought to consider the agricultural sector as a whole rather than focusing primarily on the biofuel sector.

Although models that predict land use and land cover change are routinely used to perform environmental assessments and inform policy decisions, inherent in these projections are model uncertainties that can influence model results significantly [Prestele et al., 2016]. Prestele et al. [2016] identify hot spots of uncertainty in land use/land cover models using a regression analysis of 11 global‐scale models. They find that model uncertainty is higher at the edges of “globally important biomes” such as the boreal and tropical forests. The diversity of land use models and the lack of consistent observational data, coupled with multiple definitions of land use and land cover across models, also introduce model uncertainty.

Meanwhile, De Rosa et al. [2016] evaluated six land use change models for LCA, which they classified into three main categories: (i) economic, (ii) causal‐descriptive, and (iii) normative models. They contend that the causal‐descriptive models are better suited for long‐term assess­ments, whereas the economic models perform better for short‐term assessments of the impacts of land use change. The authors also make important suggestions regarding the need to incorporate economic, biophysical, and statistical information to account for the inherent com­plexity of land use change models and achieve robust results. Furthermore, improving the precision of data, inclusion of more impact categories, identification of land use, specifically marginal land, are identified as some of the factors that can improve the quality of assessments of land use models. Finally, the authors point out that, in order to perform more holistic assessments of land use

change, models should extend beyond measuring only GHG impacts and include impact categories such as nutrient leaching, water‐resource impacts, biodiversity impacts, and socioeconomic analyses.

1.5. TECHNOLOGICAL AND POLICY IMPERATIVES AND CERTIFICATION

Land use change is guided by policy imperatives, and government policies and programs can not only influence the economic market for bioenergy but also significantly influence diversion of feedstocks for energy use. Perhaps the most significant drivers of land use change are green‐energy initiatives; bioenergy is generally considered as a source of such energy, and therefore, bioenergy initiatives may increase and grow in size.

In the United States, a range of policies were intro­duced and are being implemented to support research and development in bioenergy with a view to enhance the share of biofuels in the overall energy mix. These policies include the Biomass Research and Development Act of 2000, the 2002 Farm Bill, the Energy Policy Act of 2005, and the Energy Independence and Security Act (EISA) of 2007, among others. The 2002 Farm Bill included provisions and incentives to promote the development of biorefineries and financial support to feedstock producers and engaged in educational out­reach to highlight the benefits of biofuels. The Energy Policy Act of 2005 established the Renewable Fuels Standard (RFS) and introduced quantitative targets, while EISA 2007 aimed to achieve an increase in biofuel output to the tune of 36 billion gallons by 2022 of which 21 billion gallons would come from advanced biofuels [Food and Agricultural Organization (FAO), 2008]. In addition to these targets, a variety of research grants, tax credits, and other fiscal incentives were implemented to support feedstock producers, biofuel processors, and the end consumers. Meanwhile, several other countries, including Brazil, China, India, as well as several European Union member states, have instituted both mandatory and voluntary targets [Global Bioenergy Partnership (GBEP), 2007]. Other countries, particularly European ones, also have green‐energy or specifically bioenergy‐incentivizing initiatives.

The U.S. Department of Agriculture (USDA) has implemented the Conservation Reserve Program (CRP) since 1985 to protect the quality of soil, stabilize land prices, and ensure steady agricultural production by providing technical and financial assistance [Dale et al., 2010]. In Iowa, temporary waivers were granted for har­vesting biomass grown on CRP lands for use in research activities while allowing farmers to take the full benefit of CRP payments to encourage switchgrass establishment [Hipple and Duffy, 2002]. Similarly, under the aegis of the

0003312060.INDD 12 10/23/2017 3:19:16 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 13

Biomass Crop Assistance Program (BCAP), the USDA provides “financial assistance to owners and operators of agricultural and non‐industrial private forest land who wish to establish, produce, and deliver biomass feed­stocks” [USDA, 2016]. This support comes in the form of matching payments for eligible materials supplied to specified biomass‐conversion facilities or through annual payments for producers who enter into contracts with the Commodity Credit Corporation on contract acres within the BCAP project areas [USDA, 2016].

Some of these policies often offer subsidies or payments for farmers who produce bioenergy or can provide a market where none truly existed before. For example, most cars in the United States can run on 10% ethanol fuel, and others produced in the United States can run on nearly 85% ethanol fuel, which has created a large market for producers of ethanol. This, in turn, has created a new job market in the United States for bioenergy, which has positively affected the economy and encouraged further growth. However, as this causes the industry to grow, it can have a detrimental effect on land use change in some of the negative ways outlined earlier.

However, the efficacy of policy incentives is a subject of much debate. In its statement to the House Agriculture Subcommittee on Conservation, Energy, and Forestry in 2012, the Wood Fiber Coalition claimed that the Biomass Crop Assistance Program negatively affected long‐standing industries as several traditional wood products were diverted toward the bioenergy industry [Wood Fiber Coalition, 2012]. The biofuel industry has come under similar criticism whereby it is claimed that the increased production of biofuels comes at the expense of food grains being used to produce bioenergy, resulting in higher food prices. While such ramifications were unin­tended, it is crucial for biofuel policies to safeguard against the direct and indirect consequences of higher bioenergy mandates. As a result, one of most important challenges confronting policymakers is to ensure effective policy design and incentives that eliminate or limit any adverse impacts of policies intended to support the bio­fuel industry.

One of the other challenges with regard to the development of biofuels, specifically cellulosic and other advanced biofuels, is that technological innovations have not been attained at a level to make biofuel production economically viable. Mitchell et al. [2008] stated that sub­stantial advancements in agronomics and genetics were needed to enhance feedstock production, particularly for switchgrass. The underlying challenges are similar for other cellulosic feedstocks as well. Meanwhile, Chung [2013] highlights the requirement for advancements in conversion technologies, along with collaboration among stakeholders, to overcome the obstacles and challenges that inhibit the growth of the industry.

1.5.1. Bioenergy Certification and Land Use

Where bioenergy crops are grown, different policy criteria may work to ensure that they are produced in a responsible way. Regulatory initiatives such as the European Renewable Energy Directive and German Biofuels Sustainability ordinance work toward such a goal. These include suggestions on quotas on ethanol for fuel, GHG emission saving, and conservation of biodi­versity [Stappen et al., 2011]. Furthermore, policies and economics supporting nontraditional biofuels that do not need to replace farm crops and can be grown on marginal lands, such as switchgrass, may ultimately propel the industry to create a more sustainable biofuel‐based economy.

Firbank [2008] highlighted the possible adverse impacts of bioenergy production on land use and biodiversity, while others have emphasized the GHG, socioeconomic, and water‐resource ramifications [Berndes, 2002; Searchinger et al., 2008; German et al., 2011]. It is important to note that the development of biofuels on commercial scales will necessitate the allocation of substantial land area for the cultivation of feedstocks as well as the setting up of processing and conversion facilities [Rinehart, 2006; Mitchell et al., 2012; NRDC, 2014]. In this regard, it will be crucial to evaluate the former land use and land cover of the cultivation and processing sites to ensure that the long‐term benefits outweigh costs and the land use changes are not detrimental from environmental, social, and economic perspectives. As such, the certification criteria are usually not mutually exclusive, thereby ensuring multiple safe­guards from different aspects of biofuel production.

Adverse direct land use change effects caused by bioen­ergy production can be enforced and are already incorpo­rated by some European certifiers of green electricity, such as Eugene (Europe), Bra Miljoval (Sweden), Ok‐Power (Germany), and Naturemade Star (Switzerland), who specify that biomass used for energy must come from Forest Stewarship Council (FSC) certified forests. Some other certifiers have developed their own land use change indicators. Milieukeur (the Netherlands) and Green Power (Australia) insist that biomass not be sourced from plantations that have been planted after clearing existing old‐growth or native forests. However, North American certifiers such as Green‐e (USA) and Environmental Choice (Canada) do not consider land use changes at all in their certification initiatives. In order to minimize the negative effects of direct land use changes, a cutoff year for conversion of natural forests to plantations could be used (such as 1994, as outlined in FSC 2002 guidelines) to ensure that natural, multiple‐species forests are not con­verted to energy crops or monocultures [Fritsche et  al., 2006]. Furthermore, standards could incorporate net changes in aboveground carbon, soil carbon stocks, and

0003312060.INDD 13 10/23/2017 3:19:16 PM

14 BIOENERGY AND LAND USE CHANGE

wood products. The NRDC [2014] delineates specific protective performance guidelines for feedstock producers. First, the feedstock producers should not divert productive croplands, or their operations should not result in the conversion of critical habitats that adversely affect vulner­able species. Additionally, feedstock producers should not convert native forests and grasslands for feedstock cultiva­tion as well as not divert federal or  protected lands. Furthermore, feedstock producers should maximize productivity and minimize land use subject to constraints pertaining to sustainable yields. The guidelines also recom­mend collection of evidence for previous and existing land use and product/land assessments.

However, it is surprising to note that most of the bio­energy‐certification systems lack GHG‐reduction indi­cators. The Roundtable on Sustainable Palm Oil (RSPO) [2007], as well as recent suggestions on bioenergy certification [Richardson et al., 2005; Lewandowski and Faaij, 2006; Van Dam et al., 2008], suggests having GHG emission reduction but does not elaborate as to how much this reduction should be. The EISA 2007 specifies that a minimum of 50% reduction in GHG emissions from the use of advanced biofuels (such as cellulosic biofuels) should take place compared to fossil fuels sold or distributed as transportation fuel in 2005.

From a certification perspective, Burli et al. [2016] list nine impact categories including land use, agrochemical application, waste management, air quality, and moni­toring, among others. The stated impact categories can safeguard from potential risks associated with large‐scale adoption of switchgrass‐based biofuels; however, the guidelines can be adapted for other feedstocks as well. Building on the protection of productive farmlands and limiting the impact on food production, the sustainability criteria ensure that the allocation of land to feedstock cultivation does not result in the propagation of mono­cultures and irrecoverable losses to above‐ or below­ground vegetation and carbon sinks.

The Roundtable on Sustainable Biomaterials (RSB) [2013] provides specific guidelines for biofuel‐feedstock producers and processers pertaining to the safeguarding of land rights and land use rights. The guidelines require the operator to conduct a Land Rights Assessment in cases where the initial screenings and assessments indicate a neg­ative impact to existing land rights and land use rights as a result of biofuel operations. The guidelines require the operators to settle any disputes through free and informed consent on the basis of negotiated agreements with the affected parties. In addition, the guidelines state that bio­fuel operations should not be conducted on lands that were acquired as a result of involuntary resettlements and that if there are any disputes about the tenure agreements among the stakeholders, the biofuel operations shall not be approved under the RSB guidelines [RSB, 2013].

In addition, guidelines and biofuel certification criteria also require thorough impact assessments and documenta­tion that show positive net benefits for GHG emissions, favorable crop mix vis‐à‐vis prior land use, and evaluation of local/regional land and food prices and social indicators such as income and employment compared to past trends, among other factors.

1.6. MOVING FORWARD

In this chapter, we have provided a summary of the land use change issues arising because of bioenergy market development. On the basis of systematic litera­ture review, we discuss direct and indirect land use change impacts, modeling approaches used, and ensuing chal­lenges. Empirical evidence from the United States and elsewhere has also been outlined. We discuss policy imperatives and bioenergy‐certification initiatives as well. Our analysis suggests that bioenergy production must ensure an efficient allocation of land while limiting both direct and indirect impacts on land, water, and other natural resources as well as cascading impacts on food and feed for local and regional populations.

A realistic framework to model such impact requires a considerable integration of concepts, methods, and disci­plines. Although recent efforts have shown progress in this area, more needs to be done. Methodologically integrated models can offer new perspective, strengthen the reliability of the results, better capture the real‐world complexities, and build on the strength of each individual model and make up for their respective weaknesses. Moreover, the geographic focus of studies can strive to include previously unexplored areas and topics including urban‐rural interactions. This can reduce replication of efforts that only produce marginally new insights instead of addressing needs that have not been met to date. This can also free up resources that can be allocated to other priorities. Standardizing methods, data collection, analyses, and presentation of findings will also help researchers and users to better focus on the results instead of having to sort out the effects that are of interest and those having to do with the way the results are produced or presented. Ensuring that such models are also avail­able to all interested users can lead to greater use out of such resources, create synergistic benefits, and lead to a quicker collection and processing of data. Results based on such approaches will better serve decision‐making process and contribute toward better land use plans and policies.

There are also advancements in LCA and means to estimate net GHG emissions arising from bioenergy pro­duction. It is being realized that LCAs include not only direct emissions but also significant indirect emissions from processes such as land use changes. The emission

0003312060.INDD 14 10/23/2017 3:19:16 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 15

reductions from biomass production and harvesting play a significant part in this calculation, and sustainability indicators should ensure that these processes allow for emission reductions across the entire system. Because LCA and GHG accounting for each biomass production site entail significant costs, the challenge is to develop har­vesting standards for small‐producer compliance. The ILUC effects are also emphasized in Low Carbon Fuel Standards, which have already been enacted by California [Devereaux and Lee, 2009]. Many states in the northeast are also working toward developing such standards [Sperling and Yeh, 2009]. However, ILUCs are much dif­ficult to assess, and today there is no generally accepted methodology for determining such effects. Fritsche et al. [2006] argue for assessing indirect influence of bioenergy on land use change through measures such as land prices and rents. However, conducting such assessments at site level and translating these to operational constraints is quite resource intensive. A satisfactory methodology might be incorporated into the life cycle GHG emissions of fuels at a later date to account for indirect land use change impacts.

It is also worthwhile to note that policy incentives and support for bioenergy can result in not only diversion of feedstocks for energy use but also unintended con­sequences. Identifying such potential unintended conse­quences is important toward determining the net effects associated with incentive programs. It is also important to determine if multiple programs can be integrated for synergistic purposes. This may reduce transaction cost, ease the application process, and ensure that all potential enrollees and eligible applicants know about and take advantage of the programs available to them.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the support of the National Science Foundation CAREER Award 1555123, partial support for this study from U.S. Department of Energy’s International Affairs under award number DE‐PI0000031 and Department of Agriculture National Institute of Food and Agriculture Grant 2012‐67009‐19742. Thanks go to Taylor Wieczerak, Erik Lyttek, and Mike Fowler (Montclair State University) for their help in data collection and review.

REFERENCES

Abt, R., F. Cubbage, and G. Pacheco (2000), Southern forest resource assessment using the Subregional Timber Supply (SRTS) model, For. Prod. J., 50(4), 25–33.

Adams, D. M., R. J. Alig, B. A. McCarl, J.M Callaway, and S.M Winnett (1999). Minimum cost strategies for sequestering carbon in forests. Land Econ., 360–374.

Adams, D. M., R. J. Alig, J. M. Callaway, S. M. Winnett, and B. A. McCarl (1996), The Forest and Agricultural Sector Optimization Model (FASOM): Model Structure and Policy Applications, DIANE Publishing, Darby, PA.

Agarwal, C., G. M. Green, J. M. Grove, T. P. Evans, and C. M. Schweik (2002), A review and assessment of land‐use change models: Dynamics of space, time, and human choice, General Technical Report NE‐297, U.S. Department of Agriculture, Forest Service, Northeastern Research Station, Newtown Square, PA

Ahn, S., A. J. Plantinga, and R. J. Alig (2000), Predicting future forestland area: A comparison of econometric approaches, For. Sci., 46(3), 363–376.

Alberti, M., and P. Waddell (2000), An integrated urban development and ecological simulation model, Integr. Assess., 1(3), 215–227.

Al‐Riffai, P., B. Dimaranan, and D. Laborde (2010), Global Trade and Environmental Impact Study of the EU Biofuels Mandate, vol. 125, IFPRI, Washington, DC.

Andersen, L. E. (1996), The causes of deforestation in the Brazilian Amazon, J. Environ. Dev., 5(3), 309–328.

Bailis, R., and H. Mccarthy (2011), Carbon impacts of direct land use change in semiarid woodlands converted to biofuel plantations in India and Brazil, Glob. Change Biol. Bioenergy, 3(6), 449–460.

Baker, J. S., B. A. McCarl, B. C. Murray, S. K. Rose, R. J. Alig, D. Adams, G. Latta, R. Beach, and A. Daigneault (2010), Net farm income and land use under a U.S. greenhouse gas cap and trade, Policy Issues, 7, 1–5.

Barkley, A. (2007), Ethanol: The good, the bad, and the ugly, prepared for presentation at 2007 Risk and Profit Conference, 16–17 August 2007, Kansas State Alumni Center, Kansas State University, Manhattan, KS.

Berndes, G. (2002), Bioenergy and water—The implications of large‐scale bioenergy production for water use and supply, Glob. Environ. Chang., 12(4), 253–271.

Berry, M. W., B. C. Hazen, R. L. MacIntyre, and R. O. Flamm (1996), LUCAS: A system for modeling land‐use change, IEEE Comput. Sci. Eng., 3(1), 24.

Bies, L. (2006), The biofuels explosion: Is green energy good for wildlife? Wildl. Soc. Bull., 34, 1203–1205.

Birdsey, R. A., K. Pregitzer, and A. Lucier (2006), Forest carbon management in the United States: 1600–2100, J. Environ. Qual., 35(4), 1461–1469.

Blottnitz, V. H., and M. A. Curran (2007), A review of assess­ments conducted on bio‐ethanol as a transportation fuel from a net energy, greenhouse gas, and environmental life‐cycle perspectives, J. Clean. Prod., 15(7), 607–619.

Bockstael, N. E. (1996), Modeling economics and ecology: The importance of a spatial perspective, Am. J. Agric. Econ., 78(5), 1168–1180.

Briassoulis, H. (2000), Analysis of Land Use Change: Theoretical and Modeling Approaches, Regional Research Institute, West Virginia University, Morgantown, WV.

Brown, S. (2002), Measuring carbon in forests: Current status and future challenges, Environ. Pollut., 116(3), 363–372.

Burli, P., P. Lal, B. Wolde, and J. Alavalapati (2016), Sustainability protocols and certification criteria for switch­grass‐based bioenergy, BioResources, 11(3), 7102–7123.

0003312060.INDD 15 10/23/2017 3:19:16 PM

16 BIOENERGY AND LAND USE CHANGE

Center for BioEnergy Sustainability, Oak Ridge National Laboratory (CBES) (2009), Land‐Use Change and Bioenergy: Report from the 2009 workshop, ORNL/CBES‐001, U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy and Oak Ridge National Laboratory, Center for Bioenergy Sustainability. http://www.ornl.gov/sci/ees/cbes (accessed 22 August 2017).

Chomitz, K. M., and D. A. Gray (1996), Roads, land use, and deforestation: A spatial model applied to Belize, World Bank Econ. Rev., 10(3), 487–512.

Chung, J. N. (2013), Grand challenges in bioenergy and biofuel research: Engineering and technology development, environ­mental impact, and sustainability, Front. Energy Res., 1, 4.

Clarke, K. C., S. Hoppen, and L. Gaydos (1997), A self‐modi­fying cellular automaton model of historical urbanization in the San Francisco Bay area, Environ. Plann. B, 24(2), 247–261.

Dale, V. H., K. L. Kline, J. Wiens, and J. Fargione (2010), Biofuels: Implications for Land Use and Biodiversity, Ecological Society of America, Washington, DC.

De Rosa, M., M. T. Knudsen, and J. E. Hermansen (2016), A  comparison of land use change models: Challenges and future developments, J. Clean. Prod., 113, 183–193.

De Sá, S. A., C. Palmer, and S. Di Falco (2013), Dynamics of indirect land‐use change: Empirical evidence from Brazil, J. Environ. Econ. Manag., 65(3), 377–393.

Devereaux, C. and H. Lee (2009), Biofuels and certification, a workshop at Harvard Kennedy School, Discussion Paper 2009‐07, Belfer Center for Science and International Affairs, Cambridge, MA.

Dilekli, N., and F. Duchin (2016), Prospects for cellulosic bio­fuel production in the northeastern United States: A scenario analysis, J. Ind. Ecol., 20(1), 120–131.

Doornbosch, R. and R. Steenblik (2008), Biofuels: Is the cure worse than the disease? Revista Virtual REDESMA, 2, 63.

Drummond, M. A., and T. R. Loveland (2010), Land‐use pressure and a transition to forest‐cover loss in the eastern United States, BioScience, 60(4), 286–298.

Elobeid, A. E., M. A. Carriquiry, and J. F. Fabiosa (2011), Global biofuel expansion and the demand for Brazilian land: Intensification versus expansion, paper presented at 2011 Annual Meeting, 24–26 July 2011, Agricultural and Applied Economics Association, Pittsburgh, PA.

Energy Information Administration (EIA) (2014), Annual Energy Outlook 2014 with projections to 2040. http://www.eia.gov/forecasts/aeo/pdf/0383(2014).pdf (accessed 15 July 2017).

Eriksson, E., A. Gillespie, L. Gustavsson, O. Langvall, M. Olsson, R. Sathre, and J. Stendahl (2007), Integrated carbon analysis of forest management practices and wood substitution, Can. J. For. Res., 37(3), 671–681.

Fargione, J., J. Hill, D. Tilman, S. Polasky, and P. Hawthorne (2008), Land clearing and the biofuel carbon debt, Science, 319(5867), 1235–1238.

Finkbeiner, M. (2014), Indirect land use change – Help beyond the hype? Biomass Bioenergy, 62, 218–221.

Firbank, L. G. (2008), Assessing the ecological impacts of bio­energy projects, BioEnerg. Res., 1(1), 12–19.

Food and Agricultural Organization (FAO) (2008), The State of Food and Agriculture 2008: Biofuels: Prospects, Risks and Opportunities, Food and Agricultural Organization, Rome.

Fritsche, U. R., K. Hunecke, F. Schulze, and K. Wiegman (2006), Sustainability Standards for Bioenergy, WWF Germany, Oeko‐Institut, Darmstadt. www.wwf.de/fileadmin/fm‐wwf/pdf_neu/Sustainability_Standards_for_Bioenergy.pdf (accessed 15 July 2017).

Fritsche, U. R., R. E. Sims, and A. Monti (2010), Direct and indirect land‐use competition issues for energy crops and their sustainable production—An overview, Biofuels Bioprod. Biorefin., 4(6), 692–704.

German, L., G. Schoneveld, and P. Pacheco (2011), The social and environmental impacts of biofuel feedstock cultivation: Evidence from multi‐site research in the forest frontier, Ecol. Soc., 16(3).

Gibson, C. C., E. Ostrom, and T.‐K. Ahn (2000), The concept of scale and the human dimensions of global change: A survey, Ecol. Econ., 32(2), 217–239.

Gilruth, P. T., S. E. Marsh, and R. Itami (1995), A dynamic spatial model of shifting cultivation in the highlands of Guinea, West Africa, Ecol. Modell., 79(1), 179–197.

Global Bioenergy Partnership (GBEP) (2007), A Review of the Current State of Bioenergy Development in G8+5 Countries, GBEP Secretariat, FAO, Rome.

Grimm, N. B., J. G. Grove, S. T. Pickett, and C. L. Redman (2000), Integrated approaches to long‐term studies of urban ecological systems: Urban ecological systems present mul­tiple challenges to ecologists—Pervasive human impact and extreme heterogeneity of cities, and the need to integrate social and ecological approaches, concepts, and theory, AIBS Bull., 50(7), 571–584.

Grimm, V., E. Revilla, U. Berger, F. Jeltsch, W. M. Mooij, S. F. Railsback, H.‐H. Thulke, J. Weiner, T. Wiegand, and D. L. DeAngelis (2005), Pattern‐oriented modeling of agent‐based complex systems: Lessons from ecology, Science, 310(5750), 987–991.

Gustavsson, L., J. Holmberg, V. Dornburg, R. Sathre, T. Eggers, K. Mahapatraa, and G. Marland (2007), Using biomass for climate change mitigation and oil reduction, Energy Policy, 35(11), 5671–5691.

Hardie, I. W., and P. J. Parks (1997), Land use with heteroge­neous land quality: An application of an area base model, Am. J. Agric. Econ., 79(2), 299–310.

Harris, Z., R. Spake, and G. Taylor (2015), Land use change to bioenergy: A meta‐analysis of soil carbon and GHG emis­sions, Biomass Bioenergy, 82, 27–39.

Hill, J., E. Nelson, D. Tilman, S. Polasky, and D. Tiffany (2006), Environmental, economic, and energetic costs and benefits of biodiesel and ethanol biofuels, Proc. Natl. Acad. Sci. U. S. A., 103(30), 11206–11210.

Hipple, P. C., and M. D. Duffy (2002), Farmers’ motivations for adoption of switchgrass, in Trends in New Crops and New Uses, edited by J. Janick and A. Whipkey, pp. 256–266, ASHS Press, Alexandria, VA.

Irwin, E. G., and J. Geoghegan (2001), Theory, data, methods: Developing spatially explicit economic models of land use change, Agric. Ecosyst. Environ., 85(1), 7–24.

0003312060.INDD 16 10/23/2017 3:19:16 PM

BIOENERGY AND LAND USE CHANGE: AN OVERVIEW 17

Janowiak, M., and C. Webster (2010), Promoting ecological sus­tainability in woody biomass harvesting, J. For., 108(1), 16–23.

Jones, A. D., W. D. Collins, J. Edmonds, M. S. Torn, A. Janetos, K. V. Calvin, A. Thomson, L. P. Chini, J. Mao, X. Shi, and P. Thornton (2013), Greenhouse gas policy influences cli­mate via direct effects of land‐use change, J. Clim., 26(11), 3657–3670.

Jonsell, M. (2007), Effects on biodiversity of forest fuel extrac­tion, governed by processes working on a large scale, Biomass Bioenergy, 31(10), 726–732.

Kim, S., and B. E. Dale (2009), Regional variations in greenhouse gas emissions of biobased products in the United States—Corn‐based ethanol and soybean oil, Int. J. Life Cycle Assess., 14(6), 540–546.

Lal, P. (2011), Economic modeling of forest biomass supply for bioenergy production in Southern United States, Ph.D. thesis, Department of Forest Resources & Conservation, University of Florida, Gainseville, FL.

Lambin, E. F., B. L. Turner, H. J. Geist, S. B. Agbola, A. Angelsen, J. W. Bruce, O. T. Coomes, R. Dirzo, G. Fischer, and C. Folke (2001), The causes of land‐use and land‐cover change: Moving beyond the myths, Glob. Environ. Chang., 11(4), 261–269.

Lapola, D. M., R. Schaldach, J. Alcamo, A. Bondeau, J. Koch, C. Koelking, and J. A. Priess (2010), Indirect land‐use changes can overcome carbon savings from biofuels in Brazil, Proc. Natl. Acad. Sci. U. S. A., 107(8), 3388–3393.

Lewandowski, I., and A. Faaij (2006), Steps towards the development of a certification system for sustainable bio‐energy trade, Biomass Bioenergy, 30(2), 83–104.

Liang, S., M. Xu, and T. Zhang (2012), Unintended conse­quences of bioethanol feedstock choice in China, Bioresour. Technol., 125, 312–317.

Liebig, M., H. Johnson, J. Hanson, and A. Frank (2005), Soil carbon under switchgrass stands and cultivated cropland, Biomass Bioenergy, 28(4), 347–354.

Lin, Z., M. J. Anar, and H. Zheng (2015), Hydrologic and water‐quality impacts of agricultural land use changes incurred from bioenergy policies, J. Hydrol., 525, 429–440.

Marvuglia, A., E. Benetto, S. Rege, and C. Jury (2013), Modelling approaches for consequential life‐cycle assessment (C‐LCA) of bioenergy: Critical review and proposed framework for biogas production, Renew. Sust. Energ. Rev., 25, 768–781.

McLaughlin, S. B., and L. A. Kszos (2005), Development of switchgrass (Panicum virgatum) as a bioenergy feedstock in the United States, Biomass Bioenergy, 28(6), 515–535.

Meyer, D., K. Hornik, and I. Feinerer (2008), Text mining infra­structure in R, J. Stat. Softw., 25(5), 1–54.

Mitchell, R., K. P. Vogel, and G. Sarath (2008), Managing and enhancing switchgrass as a bioenergy feedstock, Biofuels Bioprod. Biorefin., 2(6), 530–539.

Mitchell, R. B., K. P. Vogel, J. Berdahl, and R. A. Masters (2010), Herbicides for establishing switchgrass in the central and northern Great Plains, Bioenergy Res., 3(4), 321–327.

Mitchell, R., K. Vogel, and M. Schmer (2012), Switchgrass (Panicum virgatum) for biofuel production, edited, Exten­sion Publications. http://www.extension.org/pages/26635/switchgrass‐panicum‐virgatum‐for‐biofuel‐production (accessed 15 July 2017).

Nassar, A. M., L. Harfuch, L. C. Bachion, and M. R. Moreira (2011), Biofuels and land‐use changes: Searching for the top model, Interface Focus, 1(2), 224–232.

National Wildlife Federation (2014), Land use change and bio­fuels, Factsheet, National Wildlife Federation. http://www.nwf.org/~/media/PDFs/Wildlife/Biofuels_Land_Use_Change_factsheet_v1_09‐04‐14.pdf (accessed 15 July 2017).

Natural Resource Defense Council (2014), Biofuel Sustainability Performance Guidelines, NRDC, New York.

Naylor, R. L., A. J. Liska, M. B. Burke, W. P. Falcon, J. C. Gaskell, S. D. Rozelle, and K. G. Cassman (2007), The ripple effect: Biofuels, food security, and the environment, Environ. Sci. Policy Sust. Dev., 49(9), 30–43.

Nelson, G. C., V. Harris, and S. Stone (1999), Spatial Econometric Analysis and Project Evaluation: Modeling Land Use Change in the Darién, Inter‐American Development Bank, Washington, DC.

O’Callaghan, J. (1995), NELUP: An introduction, J. Environ. Plan. Manag., 38(1), 5–20.

de Oliveira Bordonal, R., R. Lal, D. A. Aguiar, E. B. de Figueiredo, L. I. Perillo, M. Adami, B. F. T. Rudorff, and N. La Scala (2015), Greenhouse gas balance from cultivation and direct land use change of recently established sugarcane (Saccharum officinarum) plantation in south‐central Brazil, Renew. Sust. Energy Rev., 52, 547–556.

Perlack, R. D., L. M. Eaton, A. F. Turhollow Jr., M. H. Langholtz, C. C. Brandt, M. E. Downing, R. Graham, L. Wright, J. Kavkewitz, A. Shamey, R. Nelson, et  al. (2011), U.S. Billion‐Ton Update: Biomass Supply for a Bioenergy and Bioproducts Industry, Oak Ridge National Laboratory, Oak Ridge, TN.

Pfaff, A. S. (1999), What drives deforestation in the Brazilian Amazon? Evidence from satellite and socioeconomic data, J. Environ. Econ. Manag., 37(1), 26–43.

Piroli, G., and P. Ciaian (2012), Land use change impacts of biofuels: Near‐VAR evidence from the U.S., Ecol. Econ., 84, 98–109.

Plevin, R. J., A. D. Jones, M. S. Torn, and H. K. Gibbs (2010), Greenhouse gas emissions from biofuels’ indirect land use change are uncertain but may be much greater than previously estimated, Environ. Sci. Technol., 44(21), 8015–8021.

Polagye, B. L., K. T. Hodgson, and P. C. Malte (2007), An  economic analysis of bio‐energy options using thin­nings from overstocked forests, Biomass Bioenergy, 31(2), 105–125.

Pontius, R. G., and L. C. Schneider (2001), Land‐cover change model validation by an ROC method for the Ipswich water­shed, Massachusetts, USA, Agric. Ecosyst. Environ., 85(1), 239–248.

Prestele, R., P. Alexander, M. D. Rounsevell, A. Arneth, K. Calvin, J. Doelman, D. A. Eitelberg, K. Engström, S. Fujimori, and T. Hasegawa (2016), Hotspots of uncertainty in land‐use and land‐cover change projections: A global‐scale model comparison, Glob. Chang. Biol., 22(12), 3967–3983.

Rajcaniova, M., d. A. Kancs, and P. Ciaian (2014), Bioenergy and global land‐use change, Appl. Econ., 46(26), 3163–3179.

Richardson, J., R. Björheden, and C. T. Smith (2005), Certification of forest fuel production systems: A solution for

0003312060.INDD 17 10/23/2017 3:19:16 PM

18 BIOENERGY AND LAND USE CHANGE

sustainable use of biomass from forest residues for energy, Technology Progress Report for IEA Bioenergy Task 31 Biomass production for energy from Sustainable Forestry, Dublin.

Rinehart, L. (2006), Switchgrass as a bioenergy crop, National Center for Appropriate Technology. http://attra.ncat.org/attra‐pub/PDF/switchgrass.pdf (accessed 15 July 2017).

Roundtable on Sustainable Biomaterials (RSB) (2013), Roundtable on sustainable biomaterials principles and criteria for sustainable biofuel production. http://rsb.org/pdfs/standards/11‐03‐08%20RSB%20PCs%20Version%202.1.pdf (accessed 15 July 2017).

Roundtable Sustainable Palm Oil (2007), RSPO Principles and Criteria for Sustainable Palm Oil Production Including Indicators and Guidance. RSPO, Kuala Lumpur, 53pp.

Schulze, R. (2000), Transcending scales of space and time in impact studies of climate and climate change on agrohydro­logical responses, Agric. Ecosyst. Environ., 82(1), 185–212.

Searchinger, T. and R. Heimlich (2015), Avoiding bioenergy competition for food crops and land, Working Paper, Installment 9 of Creating a Sustainable Food Future, World Resources Institute, Washington, DC. http://www.worldresourcesreport.org (accessed 15 July 2017).

Searchinger, T., R. Heimlich, R. A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, S. Tokgoz, D. Hayes, and T.‐H. Yu (2008), Use of U.S. croplands for biofuels increases greenhouse gases through emissions from land‐use change, Science, 319(5867), 1238–1240.

de Souza Ferreira Filho, J. B., and M. Horridge (2014), Ethanol expansion and indirect land use change in Brazil, Land Use Policy, 36, 595–604.

Sperling, D., and S. Yeh (2009), Low carbon fuel standards, Issues Sci. Technol., 2009(2), 57–66.

Stappen, F. V., I. Brose, and Y. Schenkel (2011), Direct and indirect land use changes issues in European sustainability initiatives: State‐of‐the‐art, Biomass Bioenergy, 35(12), 4824–4834.

Swallow, S. K., P. Talukdar, and D. N. Wear (1997), Spatial and temporal specialization in forest ecosystem management under sole ownership, Am. J. Agric. Econ., 79(2), 311–326.

Taheripour, F., and W. E. Tyner (2013), Biofuels and land use change: Applying recent evidence to model estimates, Appl. Sci., 3(1), 14–38.

Tesfatsion, L. (2001), Introduction to the special issue on agent‐based computational economics, J. Econ. Dyn. Control., 25(3), 281–293.

Tompkins, A. M., L. Caporaso, R. Biondi, and J. P. Bell (2015), A generalized deforestation and land‐use change scenario generator for use in climate modelling studies, PLoS One, 10(9), e0136154.

Union of Concerned Scientists (2008), Land Use and Biofuels, Factsheet, Union of Concerned Scientists Fact Sheet. http://www.ucsusa.org/sites/default/files/legacy/assets/documents/clean_vehicles/Indirect‐Land‐Use‐Factsheet.pdf (accessed 15 July 2017).

U.S. Department of Agriculture Farm Service Agency (USDA) (2016), Energy programs. http://www.fsa.usda.gov/programs‐and‐services/energy‐programs/ (accessed 10 September 2016).

U.S. Department of Energy (USDOE) (2011a), U.S. billion‐ton update: Biomass supply for a Bioenergy and Bioproducts Industry ORNL/TM‐2011/224, pp. 227, Oak Ridge National Laboratory, Oak Ridge, TN.

U.S. Department of Energy (USDOE) (2011b), Land‐use change and bioenergy. http://www1.eere.energy.gov/bioenergy/pdfs/land_use_change.pdf (accessed 15 July 2017).

Van Dam, J. V., M. Junginger, A. Faaij, I. Jürgens, G. Best, and U. Fritsche (2008), Overview of recent developments in sustainable biomass certification, Biomass Bioenergy, 32(8), 749–780.

Varvel, G. E., K. P. Vogel, R. B. Mitchell, R. Follett, and J. Kimble (2008), Comparison of corn and switchgrass on marginal soils for bioenergy, Biomass Bioenergy, 32(1), 18–21.

Vasile, A. J., I. R. Andreea, G. H. Popescu, N. Elvira, and Z.  Marian (2016), Implications of agricultural bioenergy crop production and prices in changing the land use paradigm—The case of Romania, Land Use Policy, 50, 399–407.

Walker, R., E. Moran, and L. Anselin (2000), Deforestation and cattle ranching in the Brazilian Amazon: External capital and household processes, World Dev., 28(4), 683–699.

Walsh, S. J., T. W. Crawford, W. F. Welsh, and K. A. Crews‐Meyer (2001), A multiscale analysis of LULC and NDVI var­iation in Nang Rong district, northeast Thailand, Agric. Ecosyst. Environ., 85(1), 47–64.

Watson, R. T., I. R. Noble, B. Bolin, N. H. Ravindranath, D. J. Verardo, and D. J. Dokken (Eds.) (2000), Land Use, Land‐Use Change, and Forestry, Cambridge University Press, Cambridge, UK.

Wear, D., R. C. Abt, J. Alavalapati, G. Comatas, M. Countess, and W. McDow (2010), The South’s Outlook for Sustainable Forest Bioenergy and Biofuels Production, The Pinchot Institute for Conservation, Washington, DC.

Williams, G (2016), Hands‐on data science with R: Text mining. http://handsondatascience.com/TextMiningO.pdf (accessed 15 July 2017).

Wood Fiber Coalition (2012), Testimony to the subcommittee on conservation, energy and forestry. http://www.compositepanel.org/userfiles/filemanager/878/ (accessed 10 September 2016).

0003312060.INDD 18 10/23/2017 3:19:16 PM

3

Chemostratigraphy Across Major Chronological Boundaries, Geophysical Monograph 240, First Edition. Edited by Alcides N. Sial, Claudio Gaucher, Muthuvairavasamy Ramkumar, and Valderez Pinto Ferreira. © 2019 the American Geophysical Union. Published 2019 by John Wiley & Sons, Inc.

1.1.  INTRODUCTION

The use of elemental and isotope chemostratigraphy in interpretation and correlation of global events was established with the pioneer work of Emiliani [1955] on oxygen isotope composition of foraminifers from deep‐sea cores. Shackleton and Opdyke [1973] established the first 22 oxygen isotope stages, which was effectively the first formal application of chemostratigraphy. Williams et al. [1988] extended the oxygen isotope stage zonation to the rest of the Quaternary and Lisiecki and Raymo [2005] to the whole Pliocene. The success of oxygen isotope chemostratigraphy encouraged researchers to use stable isotope stratigraphy in ancient sedimentary successions.

Precambrian chemostratigraphy followed the pioneer research by William T. Holser on ancient ocean water chemistry [Kaufman et al., 2007a]. Long‐term fluctuations

in the chemistry of the seawater have been examined from the C isotope record across thick successions [e.g., Veizer et al., 1980; Magaritz et al., 1986], and, in spite of poten­tial effects of late diagenesis on isotope record, important isotope events were demonstrated on a global scale [e.g., Knoll et  al., 1986; Magaritz, 1989; Holser, 1997]. Since then, it became evident that contemporaneous, geograph­ically widely separated marine strata registered similar isotopic compositions. Thereafter, chemostratigraphy became an important technique/tool of intrabasinal and interbasinal stratigraphic correlation to help assemble Precambrian stratigraphic record from fragments pre­served in different successions [Kaufman et al., 2007b; Karhu et al., 2010; Sial et al., 2010a], compensating for poor biostratigraphic resolution of Precambrian fossils [Veizer et  al., 1980; Knoll et  al., 1986; Magaritz et  al., 1986; Knoll and Walter, 1992; Kaufman et  al., 1997; Corsetti and Kaufman, 2003; Halverson et  al., 2005]. Correlations established through chemostratigraphy can be used to comment on biogeochemical and climate changes through time although the paucity of radiometric con­straints on the absolute age of few of the extreme isotope excursions have led to debates on their temporal equivalence [e.g., Kaufman et al., 1997; Kennedy et al., 1998; Calver et al., 2004; Allen and Etienne, 2008].

Chemostratigraphy as a Formal Stratigraphic Method

Alcides Nobrega Sial1, Claudio Gaucher2, Muthuvairavasamy Ramkumar3, and Valderez Pinto Ferreira1

ABSTRACT

Elemental and isotope chemostratigraphies are used as tracers for glacial events, buildup of volcanic gases during glaciations (e.g., CO2), role of volcanism in mass extinction, salinity variation, redox state of the ocean and atmosphere, and provenance, among other applications. The use of isotope systems (C, O, S, N, Sr, Nd, Os), nontraditional stable isotope systems (e.g., Ca, Mg, B, Mo, Fe, Cr, Li), and elemental composition or elemental ratio (e.g., V, Ir, Mo, P, Ni, Cu, Hg, Rb/K, V/Cr, Zr/Ti, Li/Ca, B/Ca, Mg/Ca, I/Ca, Sr/Ca, Mn/Sr, Mo/Al, U/Mo, Th/U) in chemostratigraphy, especially across major chronological boundaries, are reviewed in this chapter. Furthermore, it is discussed what validates chemostratigraphy as a formal stratigraphic method.

1

1 NEG–LABISE, Department of Geology, Federal University of Pernambuco, Recife, PE, Brazil

2 Instituto de Ciencias Geológicas, Facultad de Ciencias, Universidad de la República, Montevideo, Uruguay

3 Deparment of Geology, Periyar University, Salem, TN, India

0004203144.INDD 3 11/9/2018 4:52:36 PM

COPY

RIGH

TED

MATER

IAL

4 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

1.2.  BASIS AND DEVELOPMENT OF CHEMOSTRATIGRAPHY

High‐resolution chemostratigraphy provides records that are multidimensional and that may yield climatic, stratigraphic, biologic, environmental, oceanographic, and, last but not least, tectonic information. Hence, the number of studies relying on isotope stratigraphy has grown substantially. In the case of C isotope stratigraphy, it can be even applied to sedimentary rocks diagenetically altered or that have undergone up to amphibolite facies metamorphism but that may have retained the original isotope signal [Melezhik et al., 2005; Nascimento et al., 2007; Kaufman et al., 2007b; Chiglino et al., 2010].

There are a myriad of isotope systems that have been successfully used in chemostratigraphy: carbon, oxygen, sulfur, nitrogen, calcium, boron, chromium, molybdenum,

lithium, strontium, neodymium, osmium, iron, and zinc. In order to apply the isotope record of any of these systems for chemostratigraphy of sedimentary sequences, it is essential to have good knowledge of the secular and other variations of marine isotope ratios. As carbon isotopes have higher resilience against postdepositional alteration, they are measured in carbonates and organic matter that led to the establishment of a larger database than other isotope systems. Therefore, δ13C on carbonates are more widely used in chemostratigraphy, except in carbonate‐poor successions characterized by black shales [e.g., Johnston et  al., 2010] in which one can measure organic carbon isotopes or carbonate carbon isotopes on fossils (bivalves, ammonites, belemnites, ostracods, etc.). An attempt to compile carbon isotope data to determine a secular variation curve of δ13C has revealed remarkable δ13C anomalies in the Proterozoic and Phanerozoic

0004203144.INDD 4 11/9/2018 4:52:36 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 5

[e.g., Veizer et  al., 1980, 1999; Karhu and Holland, 1996;  Hoffman et  al., 1998b; Kah et  al., 1999; Melezhik et al., 1999, 2007; Zachos et al., 2001; Lindsay and Brasier, 2002; Halverson et al., 2005, 2010a, 2010b; Saltzman, 2005; Bekker et al., 2006; Saltzman and Thomas, 2012], and it became apparent that δl3C minima, perhaps, follow main extinction events [e.g., Magaritz, 1989]. The Hirnantian and Frasnian‐Famennian episodes, however, are charac­terized by a positive excursion, and negative excursions are known where extinction was only minor (e.g., early Aptian). A compilation of global secular variation curves of δ13C, δ18O, δ34S, and 87Sr/86Sr, together with major anoxic events, glaciations, and sea‐level variation, can be found in Sial et al. [2015a].

The use of chemostratigraphy as a stratigraphic tool requires a careful examination of the diagenetic history of rocks. Petrographic, elemental (e.g., Mn/Sr, Sr, and Rb/Sr vs. δl3C), and isotopic (δl8O vs. δl3C) proxies are fundamental for the assessment of the nature of C iso­tope signals [e.g., Marshall, 1992; Jacobsen and Kaufman, 1999; Melezhik et al., 2001]. In doing so, dolostones and limestones have to be dealt with separately due to their different capacity to retain primary isotopic compositions [e.g., Kah et al., 1999; Gaucher et al., 2007].

Two special issues focusing Precambrian chemostratig­raphy were published in Chemical Geology [Kaufman et  al., 2007a] and Precambrian Research [Karhu et  al., 2010]. In these special issues, results of some cutting‐edge research on traditional (C, Sr, S) isotope chemostratigra­phy, few nontraditional isotope systems (Ca), and Hg chemostratigraphy have been reported. These publications encompass studies that highlighted chemical events from the Paleoproterozoic (Africa, South America, Europe, and India), Mesoproterozoic (South America), and Cryogenian‐Ediacaran (North America, South America, and India) and a special focus to the atmospheric, climatic, and bio­geochemical changes in both ends of the Proterozoic eon. In addition, a comprehensive synthesis on the basis and use of chemostratigraphy is presented in the book by Ramkumar [2015].

1.2.1. Hydrogen Isotopes

Hydrogen isotopes are relatively little used in chemo­stratigraphy except in studies of ice and snow stratigraphy, but deuterium has proved to be important isotope in defining the Holocene Global Stratotype Section and Point (GSSP) [Walker et al., 2009]. Quaternary scientists have always sought a boundary stratotype for the Holocene in terrestrial sedimentary records, but it was within the NorthGRIP (NGRIP) ice core, Greenland, that the Holocene GSSP at 1492.45 m depth has been ratified by the International Union of Geological Sciences (IUGS). Physical and chemical parameters within the ice

enable the base of the Holocene, marked by the first signs of climatic warming at the end of the Younger Dryas/Greenland Stadial 1 cold phase, located with a high degree of precision [Walker et  al., 2009]. This climatic event is reflected in an abrupt shift in deuterium excess values, accompanied by more gradual changes in δ18O, dust concentration, a range of chemical species, and annual layer thickness.

1.2.2. Carbon Isotopes

Carbon isotope investigation on Paleoproterozoic carbonate rocks of the Lomagundi province in Africa revealed much larger δ13C variation [Schidlowski et  al., 1983] than previously known from the Phanerozoic car­bonate successions [Veizer et al., 1980]. This observation led to the assumption that δ13C stratigraphic variation could be a tool in stratigraphic correlation. The pioneer work of Scholle and Arthur [1980] is one of the first to use carbon isotopes as stratigraphic tool, and Berger and Vincent [1981] recognized chemostratigraphy as a valid stratigraphic method. The potential use of δ13C trends and excursions of marine carbonates to date and corre­late rocks relies on the fact that their 13C/12C ratios varied over time as the result of partitioning of carbon between Corg and Ccarb reservoirs in the lithosphere [e.g., Shackleton and Hall, 1984; Berner, 1990; Kump and Arthur, 1999; Falkowski, 2003; Sundquist and Visser, 2004; Saltzman and Thomas, 2012]. The knowledge of the C isotope record is very important not only in stratigraphic correla­tion but also because of its potential to help understand the development of Earth’s climate, evolution of its biota, and CO2 levels in the atmosphere.

The compilations of the secular δ13Ccarb variation for the entire Phanerozoic [Veizer et al., 1999] and the Cenozoic [Zachos et al., 2001] were important steps to enable carbon isotope chemostratigraphy to be routinely used as a strati­graphic tool. Currently, the most complete available curve on the δ13Ccarb fluctuations through geologic time has been compiled from multiple literature sources by Saltzman and Thomas [2012]. Difficulties faced in con­structing such a curve reside on the fact that materials analyzed for curve construction, available in the litera­ture, differ between authors and geological time periods, as cautioned by Saltzman and Thomas [2012]. In an attempt to use these compiled curves, one should care­fully consider whether skeletal carbonate secreted by specific organisms or bulk carbonate has been used in evaluating or comparing C isotope stratigraphic records. Apparently, the most accepted carbonate δ13Ccarb record spanning the Neoproterozoic era is found in Halverson et al. [2010a, 2010b].

Covariation between δ13Ccarb and δ13Corg helps find out whether variations in the δ13Ccarb record reflect changes in

0004203144.INDD 5 11/9/2018 4:52:36 PM

6 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

the isotopic composition of the ancient dissolved inor­ganic carbon (DIC) pool [e.g., Oehlert and Swart, 2014]. Covariant δ13Ccarb and δ13Corg records attest that both carbonate and organic matter were originally produced in the ocean surface waters and have retained their original δ13C composition [e.g., Korte and Kozur, 2010; Meyer et al., 2013] as no secondary process is able to shift δ13Ccarb and δ13Corg in the same direction at the same rate [Knoll et al., 1986]. Conversely, the decoupled δ13Ccarb and δ13Corg records point to diagenetic alteration [e.g., Grotzinger et al., 2011; Meyer et al., 2013] or denounce that noise in the δ13Corg record resulted from local syn‐sedimentary processes [Maloof et  al., 2010]. One should remember, however, that the organic carbon isotope record is very much dependent on the source of the organic matter (terrestrial vs. marine) and terrestrial records may retain the secular variations known from the marine records.

Carbon isotopes can also be used as a pCO2 proxy. Stratigraphic variation in the offset between the δ13Ccarb and δ13Corg expressed by Δ13C offers a potential tool for tracing paleo‐pCO2 change [Kump and Arthur, 1999; Jarvis et al., 2011]. Increased burial of organic carbon leads to a fall in atmospheric pCO2 and a positive excursion in both inorganic and organic carbon. The peak in δ13Corg may postdate that of inorganic carbon and may be larger in magnitude, because Δ13C decreases as atmospheric pCO2 falls. This difference in response is tied to a draw­down in atmospheric pCO2 [Kump and Arthur, 1999]. The “robust voice” of carbon isotopes has the potential to tell us about Earth’s history [Knauth and Kennedy, 2009], but some postdepositional alteration of carbonate rocks may alter the story [Bristow and Kennedy, 2008]. However, indiscriminate use of C isotope stratigraphy to correlate Neoproterozoic carbonates (“blind dating”) has been cautioned by Frimmel [2008, 2009, 2010] from his studies on REE + Y distribution in Neoproterozoic carbonates from different settings in Africa. These studies have raised some doubt on the usefulness of cap carbonates for strati­graphic correlation of Neoproterozoic sediment successions based on carbon isotopes. They deserve further investiga­tion, although one can argue that rare earth elements (REEs) and DIC behave differently in seawater and are affected by diagenesis in a complete different way.

The application of carbon isotope chemostratigraphy to the study of oceanic anoxic events (OAEs) which record profound global climatic and paleoceanographic changes and disturbance of the carbon cycle, is one of the best examples of use of chemostratigraphy as a stratigraphic tool. The OAEs resulted from abrupt global warming induced by rapid influx of CO2 into the atmosphere from volcanogenic or methanogenic sources and were accompanied by accelerated hydrological cycle, increased weathering, nutrient discharge to oceans, intensified upwelling, and increase in organic productivity [Jenkyns,

2010]. Nine major OAEs are known, the oldest in the Jurassic (Toarcian, called T‐OAE, around 183 Ma), seven in the Cretaceous, and the youngest one in the Cenozoic (corresponding to the Paleocene‐Eocene Thermal Maximum (PETM), around 55.8 Ma).

An OAE event implies very high burial rates of marine organic carbon (12C), resulting in an increase in δ13C values of marine and atmospheric carbon, as observed in the pronounced regionally developed positive carbon isotope excursion in δ13Ccarb across the Cenomanian‐Turonian boundary [Scholle and Arthur, 1980]. However, the carbon isotope signatures of the early Toarcian, early Albian, and early Aptian OAEs are more complicated as signals from δ13Ccarb, δ

13Corg, and specific biomarkers exhibit both positive and pronounced negative excursions [Jenkyns and Clayton, 1986; Herrle et al., 2003; Jenkyns, 2003, 2010]. This observation suggests that besides carbon burial driving to global δ13C heavier values, input of light carbon implies movement in the opposite direction.

The selection of a section at El Kef, Tunisia, to be the GSSP for the Cretaceous‐Paleogene boundary (K‐Pg; 66.02; Molina et al., 2006, 2009), and of one at Dababiya, Egypt, to be the one for the Paleocene‐Eocene boundary (PETM; 58.8 ± 0.2 Ma; Aubry et  al., 2007), is the best example of use of carbon isotope chemostratigraphy in boundary definition. A δ13C negative shift in the section at El Kerf was one of the five marker criteria to define the K/Pg boundary, while the Paleocene‐Eocene boundary was defined based on global δ13Corg and δ13Ccarb isotope excursions (CIE).

1.2.3. Nitrogen Isotopes

The use of δ15N variations in organic matter (kerogen, δ15Norg) has proved to be a valuable tool in the investiga­tion of the evolution of the ocean chemistry, bioproduc­tivity, and chemostratigraphic correlation, especially where biostratigraphy is of limited usefulness [Beaumont and Robert, 1999; Papineau et al., 2005; Algeo et al., 2008; Cremonese et al., 2009]. Nitrogen isotope values for bulk samples (δ15Nbulk) from sections across the Ediacaran‐Cambrian boundary in South China display positive values in the uppermost Ediacaran strata and strong negative shift in the Cambrian strata, especially in black shales, testifying to the changes in the biogeochemical cycle of the ancient ocean [Cremonese et al., 2009, 2013, 2014]. Nitrate and nitrite are reduced to nitrogen gas by denitrification, as part of the global nitrogen cycle in modern oceans [Algeo et al., 2008].

The hypothesis that transition from anoxic to oxygenated deep ocean took place at the end of the Neoproterozoic era (Neoproterozoic Oxygenation Event) is relatively well accepted [e.g., Canfield et  al., 2008; Och and Shields‐Zhou, 2012]. Some of the available geochemical data for

0004203144.INDD 6 11/9/2018 4:52:36 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 7

the age interval of this transition, however, allow the interpretation of possibly full oxygenation in the early Ediacaran and preservation of deep ocean anoxia up to as late as the Early Cambrian [Ader et al., 2014].

Changes in marine redox structure are related to changes in the nitrogen nutrient cycling in the global ocean, implying that δ15Nsed probably reflects deep ocean redox transition [Ader et  al., 2014]. Nitrogen isotope data from Canada, Svalbard, Amazonia, and China, spanning the 750–580 Ma interval, together with other available δ15Nsed data, show no apparent change between the Cryogenian and Ediacaran, revealing a δ15Nsed distri­bution that closely resembles modern marine sediments, ranging from −4 to +11, with a δ15N mode close to +4 [Ader et al., 2014]. δ15N data from the earlier Proterozoic show distribution relatively similar to this, but shifted slightly toward more negative δ15N values and with a wider range. A possible explanation for similarity of this δ15N distributions is that as in the modern ocean, nitrate (and hence O2) was stable in most of the middle to late Neoproterozoic ocean and possibly much of the Proterozoic eon [Ader et al., 2014].

Global climate over Quaternary glacial‐interglacial time scales may have affected fluctuations of denitrification intensity whose rates varied over time, especially during OAEs (e.g., T‐OAE; Jenkyns et al., 2001). Some Upper Carboniferous black shales display Corg/N ratios and nitrogen isotope data that attest to fluctuations in the intensity of denitrification associated with glacially driven sea‐level changes [Algeo et al., 2008]. Sedimentary δ15N increases during rapid sea‐level rise in each cycle, with intensified denitrification, returning to background levels as sea level stabilized during the interglacial phase.

Bulk 15Ntot data from early Toarcian black carbon‐rich shales from British Isles and northern Italy (T‐OAE; Jenkyns et al., 2001, 2010) and from the Toarcian‐Turonian OAE [Jenkyns et al., 2007] have revealed a pronounced positive δ15Ntot excursion that broadly correlates with a relative maximum in weight percent TOC and, in some sections, with a negative δ13Corg excursion. Perhaps, the upwelling of a partially denitrified, oxygenated water mass is the explanation for the relative enrichment of δ15Ntot, and the development of early Toarcian suboxic water masses and partial denitrification is attributed to increases in organic productivity [Jenkyns et al., 2001]. A negative δ15Norg peak to near 0‰ air/N2 occurs at the Permian‐Triassic (P‐T) boundary parallel to a negative δ13C excursion. It has been interpreted as the result of a diminished biomass of eukaryotic algae due to mass extinction, which were replaced by microbial N2 fixers such as cyanobacteria [Fio et  al., 2010]. An analogous negative δ15Norg and δ15Nbulk excursion has been reported from the Ordovician‐Silurian boundary [Luo et al., 2016] and from the Ediacaran‐Cambrian boundary [Kikumoto

et al., 2014]. Thus, nitrogen isotopes are valuable for the definition of major chronostratigraphic boundaries.

1.2.4. Oxygen Isotopes

Oxygen isotope chemostratigraphy has become an impor tant tool for Mesozoic and Cenozoic stratigraphic correlation of marine sediments [e.g., Friedrich et al., 2012]. For such studies, δ18O is usually measured on benthic fora­minifera to avoid isotopic gradient effects [e.g., Emiliani, 1955; Shackleton and Opdyke, 1973; Lisiecki and Raymo, 2005]. The demonstration of primary nature of δ18O values in older successions, however, is often difficult, although oxygen isotopes have been successfully used in carbonates from belemnites and brachiopods and phosphates from shark teeth and conodonts [e.g., Vennemann and Hegner, 1998; Joachimski and Buggisch, 2002; Puceat et al., 2003; Price and Mutterlose, 2004; Bodin et al., 2009; Dera et al., 2009; Van de Schootbrugge et al., 2013].

Oxygen isotope ratios in foraminifera from deep‐sea cores have shown a consistent pattern representing changes in the ocean‐atmosphere system through time. Emiliani [1955], based on the major swings in his data, has recog­nized the “marine isotope stages” (MIS). Shackleton [1969] has subdivided Emiliani’s stage 5 into lettered substages, and since then, Quaternary time is divided into marine iso­tope stages and substages. The MIS scheme was the first attempt to use oxygen isotope chemostratigraphy in the Quaternary. Railsback et  al. [2015] have proposed the scheme of marine isotope substages currently in use.

A general increase from −8 to 0‰ VPDB in the Phanerozoic, punctuated by positive excursions coinci­dent with cold intervals, has been recognized by Veizer et al. [1999] who have suggested that δ18O analyses of care­fully screened, well‐preserved brachiopods and mollusks can still retain a primary signal even in Paleozoic samples. Nevertheless, similar consideration is not possible for the Precambrian due to the absence of calcified metazoans, except for the Ediacaran. δ18O analyses of whole rock samples of Precambrian successions usually reflect diage­netic conditions, although primary trends have been reported in rare/limited occasions [Tahata et al., 2012].

According to Bao et al. [2008, 2009], triple oxygen iso­tope evidence proved to be an important tool in the discrimination of early‐Cryogenian from end‐Cryogenian cap carbonates. Sulfate from ancient evaporites and barite shows variable negative 17O isotope anomalies over the past 750 million years. An important difference in 17O isotope anomalies of barite at top of the dolostones from the Marinoan cap carbonates (negative spike ∼ −0.70‰) suggests that by the time this mineral was precipitated, PCO2 was highest for the past 750 million years (CO2 levels reached 0.01–0.08 bar during and just after ∼635 Ma glacial event; Bao et al., 2008, 2009].

0004203144.INDD 7 11/9/2018 4:52:36 PM

8 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

Oxygen isotopes of dissolved inorganic phosphate (δ18Op) are a powerful stable isotope tracer for biogeo­chemical research, offering insights into the relative importance of different sources of phosphorus within natural ecosystems [Davies et  al., 2014]. Besides, the isotope fractionations alongside the metabolism of phosphorus allow δ18Op to be used to better understand intracellular/extracellular reaction mechanisms that control phosphorus cycling.

An organic paleothermometer based upon the membrane lipids of mesophilic marine Thaumarchaeota, the tetraether index of lipids, with 86 carbon atoms (TEX86) has been used for over a decade when attempt­ing to reconstruct sea surface temperatures (SSTs). This thermometer is particularly useful when other SST proxies are diagenetically altered (e.g., planktic forami­nifera; Pearson et al., 2007) or absent (e.g., alkenones; Bijl et al., 2009).

The oldest TEX86 record is from the Middle Jurassic (~160Ma) and indicates relatively warm SST [Jenkyns et  al., 2012]. It has been also used to reconstruct SST throughout the Cenozoic era (66–0 Ma) [e.g., Sluijs et al., 2009; Zachos et al., 2006] and particularly to reconstruct the Eocene (55.8–34 Ma) SST. During the early Eocene, TEX86 values indicate warm high southern hemisphere latitude SSTs (20–25 °C) in agreement with other indepen­dently derived proxies (e.g., alkenones, Mg/Ca). During the middle and late Eocene, high southern latitude sites cooled, while the tropics remained stable and warm.

The field of clumped isotopes is concerned with how the various isotopes of carbon and oxygen are distributed in the lattice of the carbonate crystal, allowing distinc­tion of the “isotopologues,” that is, molecules of similar chemical composition but different isotopic composition [Eiler, 2007]. This field is concerned with measuring an isotopologue of CO2 gas with a mass of 47, that is, where the two “heavy” rare isotopes (13C and 18O) are substi­tuted in the CO2 molecule. This is representative of the amount of “clumping” of the heavy isotopes in the crystal lattice of the carbonate. As Δ47 is measured, the amount of clumping at a known temperature can be determined [e.g., Ghosh et  al., 2006]. Guo et  al. (2009b) provided a theoretical Δ47 calibration for a number of different miner­alogies, making clumped isotopes to be one of the most promising paleothermometer for paleoclimate and dia­genesis [e.g., Eagle et al., 2010; Tripati et al., 2010; Petrizzo et al., 2014]. The great advantage is that it is unnecessary to know the oxygen isotope composition of the water with which carbonates have isotopically equilibrated. The growing interest on use of this technique is reflected in a rapid increase in the number of laboratories equipped to perform routine analyses of clumped isotope and by the organization of a series of international workshops focusing on its development and general applications.

1.2.5. Sulfur Isotopes

A secular δ34S variation curve for evaporites (1.0 Ga to present) was reported by Claypool et al. [1980], and since then sulfur isotope chemostratigraphy has been largely used for marine evaporite sulfate, in terrains ranging from 1.0 Ga to recent. Extensive critical review on sedimentary sulfur through time and on potential use of sulfur isotopes in the investigation of time boundaries is found in Strauss [1997], while detailed discussion on the use of sulfur isotopes on Neoproterozoic chemostratigraphy can be found in Halverson et al. [2010a]. Halverson et al. [2010b] have subdivided Neoproterozoic sulfur isotope data into two kinds: one recording seawater sulfate (δ34Ssulph) and the other recording epigenic or authigenic pyrite (δ34Spyr). The former is recovered from evaporites, barites, phospho­rites, and carbonates (as carbonate‐associated sulfate (CAS)). Fractionation that occurs during bacterial sulfate reduction (BSR) plus additional fractionation effects of reactions during oxidative recycling of sulfides is recorded by the pyrite data [Canfield and Teske, 1996], while the sulfur isotope data from barite, phosphorite, and CAS depict seawater sulfate (δ34Ssulph). Due to BSR, δ34Spyr is usually lower (lighter) than δ34Ssulph. Two important exceptions to this rule have been reported [Ries et  al., 2009]: (i) Archean successions usually yield similar values for pyrite and CAS, because the ocean was anoxic, and therefore BSR was negligible. (ii) Superheavy pyrites, that is, with δ34S values exceeding that of coeval sulfides, occur in late Neoproterozoic successions and were interpreted as the result of very low sulfate concentrations and fer­ruginous conditions in the ocean and intense aerobic reoxidation of pyrite [Ries et al., 2009].

Mass‐independent fractionation (MIF) is observed in O, S, and Hg, linked to photochemical reactions in the atmosphere, and in the case of sulfur, it can be observed in ancient sediments [Farquhar et  al., 2000; Guo et  al., 2009b] where it preserves a signal of the prevailing envi­ronmental conditions which makes sulfur isotopes as a tracer of early atmospheric oxygenation up to the formation of the ozone shield. The method implies mea­surements of multiple sulfur isotopes (δ33S, δ34S, and δ36S) on CAS and sulfides. The creation and transfer of the mass‐independent (MI) signature into minerals would be unlikely in an atmosphere containing abundant oxygen, constraining the Great Oxygenation Event (GOE) and the establishment of an ozone shield to sometime after 2.45  Ga ago. Prior to this time, the MI sulfur record implies that sulfate‐reducing bacteria did not play a significant role in the global sulfur cycle and that the MI sulfur signal is due primarily to changes in volcanic activity [Halevy et al., 2010]. After 2.3 Ga, the MIF signal disappears, attesting to the continued existence of an ozone layer since the Paleoproterozoic [Guo et al., 2009a].

0004203144.INDD 8 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 9

Therefore, sulfur isotopes are important in the study of the Archean‐Paleoproterozoic boundary and the fundamental biotic and environmental changes that took place during the GOE.

Biological and abiotic reactions in the sulfur biogeo­chemical cycle show distinctive stable isotopic fraction­ation and are important in regulating the Earth’s surface redox state [Pasquier et al., 2017]. The δ34S composition of sedimentary sulfate‐bearing phases reflects temporal changes in the global sulfur cycle and can be used to infer major changes in the Earth’s surface environment, including rise of atmospheric oxygen.

Sulfur isotope pyrite‐based records have been less explored. Pasquier et al. [2017] have studied Mediterranean sediments deposited over 500,000 y which exhibit strati­graphic variations >76‰ in the δ34Spyr data. These authors have demonstrated the relationship between the strati­graphic isotopic variation and phases of glacial‐interglacial sedimentation rates. Their results suggest that the control of the sulfur isotope record can be associated with strong sea‐level variations. Besides, they provided an important perspective on the origin of variability in such records and suggested that meaningful paleoenvironmental information can be derived from pyrite δ34S records.

1.2.6. Calcium, Magnesium, and Boron Isotopes

Technological advances in analytical procedures and sophistication of equipment (e.g., micro‐SIMS, nano‐SIMS, MC‐ICPMS) for few nontraditional stable isotopes, mainly Li, B, Mg, Cl, Ca, Cr, Fe, Ni, Cu, Zn, Ge, Se, Mo, Os, Hg, and Th [Johnson et  al., 2004; Baskaran, 2012; Teng et al., 2017], have opened new avenues, some still to be explored in terms of isotope chemostratigraphy. In particular, Ca, Mo, and Fe have received more attention in Precambrian isotope chemostratigraphy [Kasemann et  al., 2005; Arnold et  al., 2004; Siebert et  al., 2003; Johnson and Beard, 2006; Staubwasser et al., 2006, among others], and Cr isotopes have proven to be an important tool in this regard [Frei et al., 2009, 2011, among others].

It is not known exactly how Ca isotopes work in modern carbonate rocks or the extension on how diagen­esis affects them. A fairly updated review on the global calcium cycle is found in Fantle and Tipper [2014] and Gussone et al. [2016].

The global Ca isotope signal from end‐Cryogenian carbonate successions suggests that Ca isotope chemo­stratigraphy can be an additional tool for the correlation of postglacial Neoproterozoic carbonate successions [Higgins and Schrag, 2010; Kasemann et al., 2005; Silva Tamayo et al., 2007, 2010a, 2010b]. These authors have claimed that the Neoproterozoic Ca isotopic record is, perhaps, an archive of changes in the oceanic Ca isotopic composition.

Rapid glacier melting and significant increase in the Ca input to the ocean immediately after deglaciation, followed by progressive increase in carbonate precipita­tion and burial compensating for the large initial Ca input, have been depicted from Ca isotope behavior. Post‐Sturtian and post‐Marinoan global δ44/40Ca patterns seem to differ from each other, probably because of the difference in Ca mass balance evolution among these two deglaciation events as a consequence of contrasting gla­cier melting regimes [Silva Tamayo et al., 2010a, 2010b]. This divergent behavior of the Ca isotopic evolution makes Ca isotope stratigraphy a promise, perhaps, to discriminate and correlate Neoproterozoic postglacial carbonate successions. Possibly, there is a close connection between Ca isotopic cycling in the Phanerozoic, seawater chemistry, carbonate sedimentation, and evolutionary trends [Blättler et  al., 2012]. MI isotope fractionation effects as observed in O, S, and Hg isotopes were not so far observed in Ca isotopes [Gussone et al., 2016].

Use of magnesium isotope to understand geological phenomenon/processes has been on the rise during recent times [e.g., Tipper et al., 2006a, 2006b, 2006c; Higgins and Schrag, 2010; Wombacher et al., 2011; Azmy et al., 2013; Geske et al., 2015]. Chang et al. [2003], Tipper et al. [2008], and Wombacher et  al. (2009) presented the systematics and analytical protocols in Mg isotope analyses, and accuracy of Mg isotope determination in MC‐ICPMS was discussed by Tipper et al. [2008]. Brenot et al. (2008) examined the Mg isotope variability within a lithologi­cally diverse river basin. The relationships between continental weathering, riverine influx of Mg into the oceans, and global Mg isotope budgets of modern oceans were examined by Tipper et  al. [2006a, 2006b, 2006c]. Higgins and Schrag [2010] demonstrated the utility of constraining Mg cycle in marine sediments through the use of Mg isotope. As magnesium is part of the C cycle and dolomite is a major sink for Mg and a main control for δ26Mgseawater, Geske et  al. [2015] studied Mg isotope and suggested its use as a vital proxy. Azmy et al. [2013] are also of the similar opinion. Nevertheless, use of Mg isotopes in truly stratigraphic context has been scarce, for example, Strandmann et  al. [2014] and Pokrovsky et  al. [2011], to name a few. Despite this scarcity, the information that the Mg isotope system follows that of Sr and Ca isotopic systems [Fantle and Tipper, 2014] and the fact that the Mg isotopic composition of the oceans is relatively constant (δ26Mgseawater  = −0.82 ± 0.01‰, Foster et  al., 2010) and Mg has a long residence time in the ocean (≈10 Myr; Berner and Berner, 1987; 14–16 Myr, Lécuyer et al., 1990) could suggest its utility in establish­ing chemostratigraphic curve similar to that of Sr isotopic curve; however, the potential remains yet to be tapped and tested. It was Galy et al. [2002] who have reported a latitudinal gradient of Mg isotopic fractionation in

0004203144.INDD 9 11/9/2018 4:52:37 PM

10 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

calcites of speleothems. Li et al. [2012] precipitated calcite in a wide range of temperature (4–45 °C) and reported a feeble gradient between δ26Mgcalcite in solution and temperature (0.011 ± 0.002‰  °C−1). This finding could help establish Mg isotope as a proxy to temporal trends of paleotem­perature and paleolatitudinal variations.

There is fair agreement on that the aftermath of the Cryogenian glaciations has been marked by cap dolos­tone deposition that have followed intense continental chemical weathering. Huang et  al. [2016] have explored the behavior of Mg isotopes to demonstrate that this was the picture in the deposition of the terminal Cryogenian‐age Nantuo Formation and the overlying cap carbonate of the basal Doushantuo Formation, South China. They observed a δ26Mg positive excursion, with values ranging from +0.56 to +0.95‰, in the top of the Nantuo Formation that likely resulted from an episode of intense chemical weathering. The siliciclastic component of the overlying Doushantuo cap carbonate, on the contrary, has yielded much lower δ26Mg values (<+0.40‰), sug­gesting low‐intensity chemical weathering during the cap carbonate deposition. Huang et al. [2016] concluded that such a behavior of Mg isotopes confirms an intense chemical weathering at the onset of deglaciation and that it has reached its maximum before the cap carbonate deposition.

There are a growing number of publications that have applied boron isotopes as a paleo‐pH proxy although boron isotope analyses are complex [e.g., Palmer et  al., 1998; Sanyal et al., 2001; Joachimski et al., 2005; Hemming and Hönisch, 2007; Hönisch et al., 2012; Foster and Rae, 2016]. A secular change in the boron isotope geochemistry of seawater over the Phanerozoic is found in Joachimski et al. [2005], based on the boron isotope geochemistry of brachiopod calcite.

It is known that oceanic uptake of CO2 decreases ocean pH [Kasemann et al., 2005]. Calcium and boron isotopes have been used to estimate paleoenvironmental condi­tions in the aftermath of the two major Neoproterozoic glaciations in Namibia. Kasemann et al. [2005] presented a record of Cryogenian interglacial ocean pH based on boron isotopes in marine carbonates. Their B isotope data suggest a largely constant ocean pH and no critically elevated pCO2 throughout the older postglacial and interglacial periods. Marked ocean acidification event, in contrast, marks the younger deglaciation period and is compatible with elevated postglacial pCO2 concentration. Negative δ11B excursions in postglacial carbonates have been interpreted as an indication of temporary decrease in seawater pH.

It has been proposed that during the PETM, thousands of petagrams of carbon (Pg C) were released as methane or CO2 into the ocean‐atmosphere system for about 10 kyr, concomitant to a carbon isotope excursion, widespread

dissolution of deep‐sea carbonates, and global warming, leading to possible severe acidification of the ocean surface [Penman et  al., 2014]. Using boron‐based proxies for ocean carbonate chemistry, these authors demonstrated that there is evidence for a pH drop of surface and sea­water thermocline during the PETM. They have observed a decrease of 0.8‰ in δ11B at the onset of the PETM event and a reduction of almost 40% in shell B/Ca, at a drill site in the North Pacific and similar trends in the South Atlantic and Equatorial Pacific, consistent with global acidification of the surface of the ocean.

1.2.7. Chromium, Iron, Molybdenum, and Thallium Isotopes

Widespread deepwater anoxia predominated in the Archean and Paleoproterozoic oceans, while the Neo­proterozoic was transitional between anoxic and largely oxygenated Phanerozoic oceans. Stratified, ferruginous oceans have characterized the Archean‐Paleoproterozoic and Neoproterozoic ocean chemistries, while during the Mesoproterozoic, sulfidic (euxinic) marine conditions prevailed in contrast with Phanerozoic oxygenated con­ditions [Canfield et  al., 2008]. Investigation on the Fe, Cr, and Mo isotope behavior has provided further insights into the question of surface ocean oxygenation [Scott et al., 2008; Frei et al., 2009].

It is well known that Cr is very sensitive to the redox state of the surface environment, oxidative weathering processes producing the oxidized hexavalent Cr. Positive isotopic fractionation of up to 5‰ accompanies the oxidation of the reduced Cr(III) on land [Frei et al., 2009 and references therein]. Lyons and Reinhard [2009] and Døssing et al. [2011] have discussed in detail the isotopic systematic of the Cr cycle, including incorporation into banded iron formation (BIF). From Cr isotopes in BIFs, one can track the presence of hexavalent Cr in Precambrian oceans to understand the oxygenation his­tory of the Earth’s atmosphere‐hydrosphere system [Frei et al., 2009]. Frei et al. [2011] applied for the first time Cr isotope systematics to ancient carbonates, representing a useful tracer for climate change and for reconstructing the redox state of ancient seawater and atmosphere. Cr and C isotope curves in carbonates are virtually parallel [Frei et  al., 2011], and therefore, coupled δ13C‐δ53Cr che­mostratigraphy of mixed BIF/carbonate/chert succes­sions may provide more continuous curves than C isotopes alone. This method is suitable for chemical sediments (BIF, chert, and carbonates), with low amounts of terrig­enous material; otherwise Cr isotopic composition of the rock will predominate [Frei et  al., 2013]. Cr isotopes enabled the detection of early oxygenation pulses at 2.7 Ga [Frei et al., 2009] and even at 2.95 Ga [Crowe et al., 2013], long before the GOE. If one compares the δ53Cr

0004203144.INDD 10 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 11

values of BIF of different ages, Archean BIFs are the less fractionated and Neoproterozoic BIFs show the largest positive values, in accordance with the progres­sive oxygenation of surface environments [Frei et  al., 2009, 2013, 2017].

Iron (Fe) isotopes are a tool in the study of iron cycling due to its large isotopic fractionation attending to redox transformations in near‐surface environment. Before the impossibility of applying the traditional stable or radio­genic isotope systems, the Fe isotope system has been largely applied to BIF [Halverson et al., 2011]. Archean and Paleoproterozoic BIFs have revealed an extraordi­nary variability in Fe isotope compositions, from the stratigraphic [e.g., Beard et al., 2003; Johnson et al., 2008; Heimann et al., 2010] to the mineral [e.g., Johnson et al., 2003; Frost et al., 2007] and microscale [e.g., Steinhoefel et al., 2010]. These variations are usually ascribed to the large fractionation resulting from reduction/oxidation of Fe and the isotopic differences between mineral phases [e.g., Johnson et al., 2008].

In Neoproterozoic iron formations (IF), Fe occurs almost predominantly as hematite [Klein and Beukes, 1993] in contrast to some Archean‐Paleoproterozoic BIFs in which iron occurs as both Fe2+ and Fe3+ in a range of different minerals [Klein and Beukes, 1993]. Therefore, primary isotope signatures are easier to obtain from the Neoproterozoic BIFs which are usually associated with episodes of global glaciation (Rapitan‐type BIF) as their Fe isotope composition reflects the chemistry of the glacial ocean [Halverson et al., 2011].

The evolution of the redox state of the oceans can be also investigated using Mo concentrations in black shales [Scott et al., 2008]. Its isotopic composition, in turn, allows differentiation between euxinic (i.e., sulfidic) and oxygen­ated environments [Arnold et al., 2004]. Three oxygenation events at 2.65 Ga, ca. 2.5 Ga, and 550 Ma were recognized, with the late Paleoproterozoic and Mesoproterozoic (1.8–1.0 Ga) being characterized by euxinic conditions (“Canfield Ocean”; Canfield, 1998; Arnold et  al., 2004; Scott et  al., 2008). This is consistent with other proxies, such as MIF of sulfur and chromium isotopes.

High‐precision measurements of thallium (Tl) isotope ratios were only made possible in the late 1990s, and, therefore, one has only limited knowledge of its isotopic behavior. Despite of their heavy masses of 203 and 205 a.m.u., it is known that thallium isotopes can be frac­tionated substantially in the marine environment [Nielsen et al., 2017].

Thallium isotopes have been applied to investigate paleoceanographic processes in the Cenozoic, and a com­pilation of the Tl (ε205Tlsw) isotope composition of sea­water over the last 75 Myrs is found in Nielsen et al. [2009, 2017], together with contemporaneous δ34Ssw variation curve. These two curves show relatively similar behavior,

with the lowest values within the 55–70 Ma range, the δ34Ssw curve displaying minimum values around 55 Ma and ε205Tlsw around 66 Ma. Thallium isotopes may be utilized as a proxy for changes in Fe and Mn supply to the water column over million year time scales according to Nielsen and Rehkämper [2012] to monitor changes in marine Mn sources and/or Mn oxide precipitation rates back in time.

1.2.8. Strontium and Neodymium Isotopes

As 87Sr/86Sr ratios and δ13C fluctuate independently from each other, their combined use through the appli­cation of high‐resolution chemostratigraphy represents a powerful tool to resolve geological problems. The radio­genic nature of 87Sr, which forms as a result of radioac­tive decay of 87Rb, implies that the 87Sr/86Sr ratio of the mantle, crust, and surface environments rises with time [e.g., Shields, 2007a, 2007b]. The 87Sr/86Sr seawater variation curve is better known for the Phanerozoic [e.g., Burke et al., 1982; Veizer et al., 1999; McArthur et al., 2001; Leckie et al., 2002; McArthur, 2010] showing long‐term varia­tions of about 500–550 Ma from the Upper Cambrian (0.709; Montañez et al., 2000) gradually decreasing with a nadir of 0.7068 at 250 Ma and rising again to values of 0.7092 in the present‐day ocean [Macdougall, 1991; McArthur et al., 2001]. This makes Sr isotopes a pretty straightforward and precise method for dating marine carbonates and calcareous fossils in the upper half of the Cenozoic, for example, because 87Sr/86Sr ratios rise con­tinuously from 0.7077 in the Bartonian (ca. 40 Ma) to 0.7092 in the Holocene [McArthur et al., 2001]. Sr isotope chemostratigraphy is equally well feasible in the other periods of the Phanerozoic, depending on the morphology of the Sr isotope record. The close correlation in time between the strontium isotope excursions and the major OAEs (Jurassic and Cretaceous) is compatible with a causal linkage [e.g., Jones and Jenkyns, 2001].

The main source of 87Sr is the weathering of Rb‐rich granitic rocks. Hydrothermal vents near mid‐ocean ridges are enriched in non‐radiogenic 86Sr [Shields, 2007a], and therefore, high 87Sr/86Sr ratios are considered as an indi­cation of periods of enhanced orogenesis, while low ratios characterize periods of continental breakup and enhanced hydrothermal activity. Flament et  al. [2011], however, have pointed out that 87Sr/86Sr is influenced by the area of emerged land rather than by orogenic processes alone, something especially important for calculations of continental growth in the Archean, when maybe <4% of Earth’s area was emerged [Shields, 2007b; Flament et al., 2011].

Efforts to compile Sr isotope data aimed at determining the secular 87Sr/86Sr seawater curve for the Proterozoic have been made [e.g., Jacobsen and Kaufman, 1999;

0004203144.INDD 11 11/9/2018 4:52:37 PM

12 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

Melezhik et  al., 2001; Halverson et  al., 2007, 2010a, 2010b; Kuznetsov et  al., 2010]. The use of strontium isotopes in chemostratigraphy, however, is limited by the paucity of limestone in many successions. Another diffi­culty is posed by the likelihood of alteration in samples with low strontium contents through the incorporation of 87Sr from the decay of 87Rb in coexisting clay minerals [Kaufman et al., 2009]. Therefore, it is advisable to con­sider only analyses of high‐Sr limestones, less prone to postdepositional alteration. Geochemical screens (Rb/Sr, Mn/Sr, Sr concentration, and δ18O) have been widely adopted to evaluate the degree of postdepositional alter­ation of strontium isotope ratios [Veizer et  al., 1983; Kaufman et al., 1992, 1993; Marshall, 1992; Jacobsen and Kaufman, 1999; Melezhik et  al., 2001]. Dolostones are usually not suitable for Sr isotope studies due to the lower Sr concentrations of usually a few tens of ppm [Kah et al., 1999; Gaucher et al., 2007], although a few exceptions have been reported [Sawaki et al., 2010].

Another problem of the method is the differing labora­tory procedures, which yield different results for the same samples. The use of pre‐leaching with ammonium acetate removes adsorbed Sr and yields lower 87Sr/86Sr ratios than a more aggressive one‐step HCl leaching method [Melezhik et al., 2001; Rodler et al., 2017]. An intermediate approach for limestones is the use of 0.5 M acetic acid for a short time (5–10 min), which predominantly liberates calcite‐associated Sr, thereby yielding lower Sr isotope ratios [e.g., Frei et al., 2011].

Details on the neodymium isotope geochemistry are found in DePaolo [1988]. A “global average” εNd curve for the oceans since 800 Ma has been constructed [Keto and Jacobsen, 1988; Macdougall, 1991], although neodymium isotopes have been seldom used as a chemostratigraphic tool. Similar to Sr isotope secular curve, this curve shows εNd values at the end of the Precambrian oceans not substantially different from those at present ones. There is a remarkable decrease of the average εNd values (−5 to −15) in the time interval between 700 and 550 Ma. Despite the precision of modern instruments, the scatter in measured values is substantial, limiting the use of εNd in chemostrati­graphic studies. Even so, secular εNd variations coupled with δ13C and δ53Cr have been reported from Ediacaran rocks from Uruguay [Frei et  al., 2011, 2013] and Brazil [Dantas et  al., 2009], yielding valuable information regarding the tectonic evolution of the basin.

1.2.9. Osmium and Lithium Isotopes

The temporal variations of the 187Os/188Os ratio are preserved in several marine depositional environments, where osmium is an ultra trace element [Peckeur‐Ehrenbrink and Ravizza, 2000]. Several developments over the last three decades have allowed direct measuring of 187Os/188Os

ratio and osmium concentration in seawater, river water, and rain, improving the knowledge on the surficial cycle of osmium [Sharma et al., 1997; Peckeur‐Ehrenbrink and Ravizza, 2000]. Ravizza and Peucker‐Ehrenbrink [2003] have observed a decline of about 25% in the marine 187Os/188Os record that predated the Cretaceous‐Paleocene transition (K‐Pg) and that coincides with a warming in the late Maastrichtian. They have interpreted this osmium isotope ratio decline as a chemostratigraphic marker of the Deccan volcanism which was responsible for a transient global warming event (3–5 °C) and likely one of the causes of the K‐Pg mass extinction.

Precambrian‐to‐Pleistocene marine osmium isotope records, particularly the Cenozoic and Mesozoic ones, and interpretations of their temporal variations have been reviewed by Peckeur‐Ehrenbrink and Ravizza [2012]. Although the Cenozoic seawater 187Os/188Os mimics the marine 87Sr/86Sr record and suggests that both reflect continental weathering linked to climatic or tectonic processes, these two marine isotope systems differ fundamentally from each other [Peckeur‐Ehrenbrink and Ravizza, 2000]. The marine residence time of osmium is distinctly shorter, allowing to record short‐term fluctua­tions (e.g., glacial‐interglacial periods), something that escapes to the buffered marine strontium isotope system. This difference between these two systems allows discrimination between climatic and tectonic forcings. Besides, large‐amplitude changes in the marine 187Os/188Os record can be useful as chemostratigraphic event markers [Peckeur‐Ehrenbrink and Ravizza, 2012].

The decline of atmospheric CO2 has a potential role in initiating glaciation and its increase of terminating it [Vandenbroucke et al., 2010]. Both cases involve changes in silicate weathering rates [Lenton et al., 2012; Ghienne et al., 2014]. The change of 187Os/188Os ratios during glacial periods may represent a response to change in silicate weathering, but does not help in tracing the weathering rate or processes involved [Finlay et  al., 2010]. The behavior of Li isotopes, however, is solely controlled by silicate weathering processes and, therefore, gives a unique insight into CO2 drawdown and climate stabilization [Pogge von Strandmann et al., 2017].

Biological processes do not lead to lithium isotope fractionation [Pogge von Strandmann et al., 2017], and carbonate weathering does not affect Li isotope signals [Dellinger et al., 2017]. The δ7Li of primary silicate rocks have a narrow range [Sauzeat et al., 2015] if compared to the high variability of modern rivers which reflects weathering processes, particularly the extent of preferen­tial uptake of 6Li into secondary minerals [Dellinger et al., 2017]. Marine carbonates have a negligible sink of Li [Marriott et al., 2004; Pogge von Strandmann et al., 2013].

A comprehensive review on lithium isotope geochemistry is found in Tomascak et al. [2016] and Penniston‐Dorland

0004203144.INDD 12 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 13

et al. [2017] in which the possibility of use of Li isotope in chemostratigraphy has been overlooked. Lithium isotope chemostratigraphy of Late Ordovician bulk carbonate sections and brachiopods in Anticosti Island, Canada [Achab et al., 2013] (Pointe Laframboise Ellis Bay West), and of an equivalent shale section at Dob’s Linn, United Kingdom [Finlay et al., 2010; Melchin et al., 2013], was presented by Pogge von Strandmann et al. [2017]. In all sections in that study, the relative timings of δ7Li and the Hirnantian carbon isotope excursion (HICE) are similar, suggesting that Li isotope excursions occur contempo­raneously, consistent with the Li residence time in the ocean (1 Myr). The positive δ7Li excursion during the Hirnantian cooling event compares well to negative δ7Li during warming events [Pogge von Strandmann et  al., 2013; Lechler et al., 2015].

1.2.10. Elemental Chemostratigraphy

Elemental chemostratigraphy (element and element ratios) is a supplementary, useful tool in stratigraphy, and Mo, Ir, V, Ni, Cu, P, Hg, REEs, and Fe are among the most used elements, while Mo/Al, U/Mo, Rb/K, V/Cr, Zr/Ti, I/Ca, Li/Ca, B/Ca, Sr/Ca, Mg/Ca, Mo/Th, V/Th, and Th/U ratios seem to be particularly interesting. Paleoceanographic applications of trace‐metal concentration data have been reviewed by Algeo and Rowe [2012].

Iron speciation has been widely used to determine the redox state of ancient basins. The method involves sequential extraction procedures to extract highly reac­tive iron (oxide, carbonates, and sulfide) and compare their concentration to total iron (FeHR/FeT; Canfield, 1989; Shen et al., 2003; Poulton and Canfield, 2005). Sediments deposited in an oxygenated water column yield FeHR/FeT lower than 0.38 [Canfield, 1989]. Furthermore, the sulfide‐bound iron (FeP) can be compared to highly reactive iron (FeP/FeHR), with values higher than 0.8 characterizing sulfidic (euxinic) basins [Canfield et al., 2008]. Iron speci­ation chemostratigraphy has been applied successfully to sedimentary units of different ages, from the Archean to recent [e.g., Shen et al., 2003; Poulton et al., 2004; Canfield et al., 2008; Lyons et al., 2009; Johnston et al., 2010; Scott et al., 2011; Hammarlund et al., 2012; Frei et al., 2013].

In sediments deposited immediately after major glacial events, Hg tends to concentrate as a result from leaching of volcanogenic Hg from land surface and accumulation along argillaceous sediments [Santos et  al., 2001]. This element is usually found in low geological background concentrations, and this makes this trace element suitable for identifying accumulation pulses in sediments that can be tentatively related to weathering processes and thus to climatic changes.

Carbon dioxide buildup in the atmosphere during the Neoproterozoic glacial events resulted from volcanism

that led to enhanced greenhouse effect, ice melting, and cap carbonate deposition [e.g., Hoffman et  al., 1998a; Hoffman, 2011]. Besides, intense volcanism may have witnessed the P‐T and Cretaceous‐Paleogene transition (K‐Pg) and was, perhaps, co‐responsible for dramatic climatic changes and thus for the decrease in biodiversity and mass extinction [e.g., Keller, 2005; Archibald et al., 2010]. Sial et  al. [2010b] demonstrated the use of Hg chemostratigraphy of the cap carbonates to document intense volcanism and resultant CO2 buildup in the atmosphere, following the Neoproterozoic snowball events. Moreover, Hg chemostratigraphy was applied to investigate the relationships between large igneous prov­ince (LIP) activity, abrupt environmental changes, and mass extinctions [e.g., Nascimento‐Silva et al., 2011, 2013; Sanei et al., 2012; Sial et al., 2013a, 2014, 2016, 2017, this volume; Adatte et  al., 2015; Grasby et  al., 2013, 2015, 2017; Percival et al., 2015, 2017; Font et al., 2016, 2018; Thibodeau et  al., 2016; Charbonnier et  al., 2017; Jones et al., 2017; Thibodeau and Bergquist, 2017; Keller et al., 2018). To assure that the measured Hg contents result from true Hg loading to the environment, it is necessary to examine Hg/TOC ratios for chemostratigraphy [e.g., Grasby et al., 2015; Percival et al., 2015]. Mercury enrich­ments in sedimentary successions that recorded the mid‐Cenomanian Event and Oceanic Anoxic Event 2 (OAE2) in the Late Cretaceous have been regarded by Scaife et al. [2017] as a marker for submarine LIP volcanism, and Hg enrichment recorded in the PETM is assumed to be related to volcanic activity of the North Atlantic Igneous Province (NAIP) (e.g., Keller et al., 2018). Hg is doubtless a good benchmark for high volcanic activity, but normal­ization by TOC is in some cases problematic if TOC values are <0.2, leading to exaggerated peaks.

Mo and V chemostratigraphy may be useful in the investigation of the redox state of deep ocean water. Mo is a redox‐sensitive element, scavenged from seawater into sediments in the form of MoSxO4−x2−, under anoxic conditions [Wen et  al., 2015]. The transfer of aqueous Mo to the sediment can be increased by means of metal‐oxyhydroxide particulate shuttles, but aqueous U is not affected by this process [Tribovillard et al., 2012]. There­fore, an increase in U/Mo ratio may suggest oxic condi­tions [e.g., Sosa‐Montes et al., 2017]. According to Scheffler et al. [2003], certain elemental ratios can be useful as prox­ies for investigation of salinity variation (Rb/K), redox state (V/Cr), or provenance (Zr/Ti). Mo and V can be normalized with Th (Mo/Th and V/Th) and, together with other redox‐sensitive trace elements such as Ni, Zn, and Pb, can be used to determine redox variations in ancient sedimentary successions [Spangenberg et al., 2014].

REE has been extensively used in different types of sedimentary rocks, often in combination with yttrium (REEY). The most widely used proxies are Ce, Eu, and Pr

0004203144.INDD 13 11/9/2018 4:52:37 PM

14 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

anomalies (Ce/Ce*, Eu/Eu*, Pr/Pr*), Y/Ho, La/Yb, and ΣREE, which can be applied to shales, carbonates, BIF, cherts, phosphorites, and other fine‐grained rocks [Elderfield and Greaves, 1982; Liu et  al., 1988; Bau and Dulski, 1996; Kato et  al., 2006; Lawrence and Kamber, 2006; see Sial et al., 2015b for an overview of proxies]. REE chemostratigraphy has been applied to Archean [Kamber et  al., 2014], Paleoproterozoic [Bau and Dulski, 1996], Mesoproterozoic [Azmy et  al., 2009], Neoproterozoic [Tribovillard et al., 2006; Frimmel, 2009; Sansjofre et al., 2014; Spangenberg et al., 2014; Gaucher et  al., 2015; Sial et  al., 2015b; Hu et  al., 2016; Rodler et al., 2016], and Phanerozoic successions [Schmitz et al., 1988; Lécuyer et al., 2004; Fio et al., 2010].

The redox behavior of iodine is well known [Broecker et al., 1982]. Besides, it is also known that there is a linear covariation between carbonate‐associated iodine (CAI) and IO3 during calcite precipitation, but I− is completely excluded [Lu et  al., 2010]. This trait, coupled with the residence time of iodine in seawater (300 ky; Broecker et al., 1982) and concentration near 450 nM in modern ocean, makes I/Ca (or I/Ca + Mg) ratios in carbonates a robust indicator of the presence of IO3 and hence oxygen in the water column. Therefore, surface ocean oxygena­tion has been investigated using I/Ca ratios as a paleo­redox indicator [e.g., Hardisty et al., 2014].

Li/Ca and B/Ca in carbonates are regarded as proxies for carbonate saturation state [Hall and Chan, 2004; Hall et al., 2005; Lear and Rosenthal, 2006; Yu and Elderfield, 2007; Foster, 2008], and Mg/Ca ratios of foraminiferal shells have been regarded as useful paleothermometer to determine ocean temperature. The difference in the Mg/Ca ratio of the foraminiferal shell and that from a baseline value (defined by the global ocean Mg and Ca concentration) when calibrated for the vital effects of the organism is a function of temperature [e.g., Lea et  al., 2000; Lear et al., 2000]. The baseline composition of sea­water is relatively simple to infer, once both Mg and Ca have long residence times in the oceans (>1 Ma) and are major components of ocean salts.

1.3.  CHEMOSTRATIGRAPHY AND CHRONOSTRATIGRAPHIC BOUNDARIES

The International Commission on Stratigraphy (ICS) recognizes the existence of one hundred fourteen chro­nostratigraphic boundaries. Sixty‐seven sections strad­dling chronostratigraphic boundaries were internationally agreed upon as reference points to define the lower bound­aries of stages on the geologic time scale, the Global Boundary Stratotype Section and Point (GSSP), and a golden spike is placed precisely at the boundary defined. Accessibility and degree of representativity of the same boundary on sections worldwide are among the most

important criteria in the GSSP selection. Since GSSPs require well‐preserved sections of rock without interrup­tions in sedimentation, and since most are defined by dif­ferent biozones, defining them becomes more difficult as one goes further back in time in the Precambrian.

So far, chemostratigraphy has been overlooked as a formal criterion on GSSP selection. Carbon isotope excursions (CIE) have been reported only from seven of the established GSSPs [Cooper et al., 2001; Dupuis et al., 2003; Peng et al., 2004; Knoll et al., 2006; Xu et al., 2006; Aubry et  al., 2007; Goldman et  al., 2007; Schmitz et  al., 2011; Keller et al., 2018], probably due to the absence of carbonate rocks in several chronostratigraphic boundary sections. Only in the selection of the Cretaceous‐Paleogene (K‐Pg; Molina et  al., 2006, 2009) and the Paleocene‐Eocene (PETM; Aubry et al., 2007) GSSPs was carbon isotope chemostratigraphy one of the criteria, and hydrogen isotopes in the Pleistocene–Holocene GSSP [Walker et al., 2009]. In addition, oxygen isotopes have been reported from two other GSSPs [Steininger et  al., 1997; Hilgen et al., 2009]. Heavy element (e.g., Ir, Os, Hg) enrichments at the Cretaceous‐Tertiary boundary are well known since the seminal paper of Alvarez et al. [1980] and later studies [e.g., Schmitz et al., 1988; Frei and Frei, 2002; Sial et al., 2016; Keller et al., 2018, and references therein].

1.4.  CHEMOSTRATIGRAPHY AS FORMAL STRATIGRAPHIC METHOD

The stratigraphic record shows changes of the concen­tration of certain elements with time [Morante et  al., 1994], as a function of geological conditions including, but not limited to, tectonic, climatic, redox, oceano­graphic, biotic, and other processes. Chemostratigraphy enables not only apparently uniform thick successions to be subdivided and correlated with coeval strata located elsewhere [Ramkumar, 1999] but also thinner and more heterogeneous sedimentary records. Initially, chemostratig­raphy was applied to recognize unique geochemical compositions for characterizing depositional units and correlating them with coeval strata elsewhere and found its use in the stratigraphic location of boundaries and later expanded to examination of specific causes to the stratigraphic variations of geochemical compositions [Ramkumar et al., 2010, 2011]. The utility of chemo­stratigraphy for age determination was demonstrated through documentation of stratigraphic variations of isotopic trends, beginning with oxygen isotopes. Linear, secular, cyclic, and perturbed trends have been recog­nized, which are utilized for stratigraphic classification and spatial correlation [e.g., Zachos et  al., 2001; Ramkumar, 2014, 2015]. In addition, the chemozones, calibrated with absolute time, are in use as chemochrons. Although chemostratigraphy is firmly recognized as a

0004203144.INDD 14 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 15

valid stratigraphic method since more than thirty years [e.g., Berger and Vincent, 1981], given to its sensitivity and wide applications, time has come to recognize this technique as an individual method of stratigraphy. The International Stratigraphic Commission (ISC) defines stratigraphy as “the description of all rock bodies form­ing the Earth’s crust and their organization into distinctive, useful, mappable units based on their inherent properties or attributes… in order to establish their distribution and relationship in space and their succession in time, and to interpret geologic history” [Salvador, 1994]. Chemo­stratigraphy “recognizes and organizes” rock bodies into useful units based on their inherent chemical properties. Most importantly, it helps in establishing the distribution and relationships of these units in space and time and, espe­cially, interpreting geological history. In many cases, the res­olution of chemostratigraphy proves to be better than other conventional methods of stratigraphy and can be refined by improving sampling resolution, although cyclostratigraphy can be even better, providing a resolution on Milankovitch time scales, reaching back far into the Cenozoic. In this regard, it can be stated that chemostratigraphy as an independent stratigraphic method can serve well even where other methods fail or have limitations. Chemostratigraphy is complementary to other types of stratigraphic units, such as lithostratigraphy, biozones, and magnetostratigraphy. With these attributes in mind, we are of the opinion that chemostratigraphy can be recognized as an independent standard method of stratigraphic classification.

ACKNOWLEDGMENTS

We express our thankfulness to the American Geophysical Union (AGU)/Wiley for the invitation to compile a special publication on “chemostratigraphy across major chronological eras.” Three anonymous reviewers are thanked for thorough, critical analysis of the original manuscript. This is the contribution n. 287 from the Nucleus for Geochemical Studies–Stable Isotope Laboratory (NEG‐LABISE), Department of Geology, Federal University of Pernambuco, Brazil.

REFERENCES

Achab, A., Asselin, E., Desrochers, A., Riva, J.F., 2013. The end‐Ordovician chitinozoan zones of Anticosti Island, Quebec: Definition and stratigraphic position. Review of Palaeobotany and Palynology 198, 92–109.

Adatte, T., Keller, G., Schoene, B., Samperton, K.M., Font, E., Sial, A.N., Lacerda, L.D., Punekar, J., Fantasia, A., Khadri, S., 2015. Paleoenvironmental influence of Deccan volcanism relative to the KT extinction. Geological Society of America Abstracts with Programs 47(7), 210, Baltimore.

Ader, A., Sansjofre, P., Halverson, G.P., Busigny, V., Trindade, R.L.F., Kunzmann, M., Nogueira, A.C.R., 2014. Ocean

redox structure across the Late Neoproterozoic Oxygenation Event: A nitrogen isotope perspective. Earth and Planetary Science Letters 396, 1–13.

Algeo, T.J., Rowe, H., 2012. Paleoceanographic applications of trace‐metal concentration data. Chemical Geology 324–325, 6–18.

Algeo, T., Rowe, H., Hower, J.C., Schwark, L., Herrmann, A., Heckel, P., 2008. Changes in ocean denitrification during late carboniferous glacial interglacial cycles. Nature Geoscience 1, 709–714.

Allen, P.A., Etienne, J.L., 2008. Sedimentary challenge to snowball Earth. Nature Geoscience 1, 817–825.

Alvarez, L.W., Alvarez, W., Asaro, F., Michel, H.V., 1980. Extraterrestrial cause for the Cretaceous–Tertiary extinction. Science 208, 1095–1108.

Archibald, J.D., Clemens, W.A., Padian, K., Rowe, T., Macleod, N., Barrett, P.M., Gale, A., Holroyd, P., Sues, H.D., Arens, N.C., Horner, J.R., Wilson, G.P., Goodwin, M.B., Brochu, C.A., Lofgren, D.L., Hurlbert, S.H., Hartman, J.H., Eberth, D.A., Wignall, P.B., Currie, P.J., Weil, A., Prasad, G.V., Dingus, L., Courtillot, V., Milner, A., Milner, A., Bajpai, S., Ward, D.J., Sahni, A., 2010. Cretaceous extinctions: Multiple causes. Science 328, 973–976.

Arnold, G.L., Anbar, A.D., Barling, J., Lyons, T.W., 2004. Molybdenum isotope evidence for widespread anoxia in mid‐Proterozoic oceans. Science 304, 87–90.

Aubry, M.‐P., Ouda, K., Dupuis, C., Berggren, W.A., Van Couvering, J.A., Working Group on the Paleocene/Eocene Boundary, 2007. The Global Standard Stratotype‐section and Point (GSSP) for the base of the Eocene Series in the Dababiya section (Egypt). Episodes 30, 271–286.

Azmy, K., Sylvester, P., de Oliveira, T.F., 2009. Oceanic redox conditions in the Late Mesoproterozoic recorded in the upper Vazante Group carbonates of São Francisco Basin, Brazil: Evidence from stable isotopes and REEs. Precambrian Research 168(3), 259–270.

Azmy, K., Lavoie, D., Wang, Z., Brand, U., Al‐Aasm, I., Jackson, S., Girard, I., 2013. Magnesium isotope and REE composition of Lower Ordovician carbonates from eastern Laurentia: Implications for the origin of dolomites and limestones. Chemical Geology 356, 64–75.

Bao, H., Lyons, J.R., Zhou, C., 2008. Triple oxygen isotope evi­dence for elevated CO

2 levels after a Neoproterozoic glacia­tions. Nature 453, 504–506.

Bao, H., Fairchild, I.J., Wynn, P.M., Spötl, C., 2009. Stretching the envelope of past surface environments: Neoproterozoic glacial lakes from Svalbard. Science 323, 119–122.

Baskaran, M., 2012. Handbook of Environmental Isotope Geochemistry. Advances in Isotope Geochemistry. Springer‐Verlag, Berlin, 951 pages.

Bau, M., Dulski, P., 1996. Distribution of yttrium and rare‐earth elements in the Penge and Kuruman iron‐formations, Transvaal Supergroup, South Africa. Precambrian Research 79(1–2), 37–55.

Beard, B.L., Johnson, C.M., Skulan, J.L., Nealson, K.H., Cox, L., Sun, H., 2003. Application of Fe isotopes to tracing the geochem­ical and biological cycling of Fe. Chemical Geology 195, 87–117.

Beaumont, V., Robert, F., 1999. Nitrogen isotope ratios of kero­gens in Precambrian cherts: A record of the evolution of atmosphere chemistry? Precambrian Research 96, 63–82.

0004203144.INDD 15 11/9/2018 4:52:37 PM

16 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

Bekker, A., Karhu, J.A., Kaufman, A.J., 2006. Carbon isotope record for the onset of the Lomagundi carbon isotope excur­sion in the Great Lakes area, North America. Precambrian Research 148, 145–180.

Berner, R.A., 1990. Atmospheric carbon dioxide levels over Phanerozoic time. Science 249, 1382–1386.

Berger, W.H., Vincent, E., 1981. Chemostratigraphy and bio­stratigraphic correlation: Exercises in systemic stratigraphy. Oceanologica Acta 1981(SP), 115–127.

Berner, E.K., Berner, R.A., 1987. The Global Water Cycle: Geochemistry and Environment. Prentice Hall, New York, 397 p.

Bijl, P.K., Schouten, S., Sluijs, A., Reichart, G.‐J., Zachos, J.C., Brinkhuis, H., 2009. Early Palaeogene temperature evolution of the southwest Pacific Ocean. Nature 461, 776–779.

Blättler, C.L., Henderson, G.M., Jenkys, H.C., 2012. Explaining the Phanerozoic Ca isotope history of seawater. Geology 40, 843–846.

Bodin, S., Fiet, N, Godet, A., Matera, V., Westermann, S., Clement, A., Janssen, N.M.M., Stille, P, Foellmi, K.B., 2009. Early Cretaceous (late Berriasian to early Aptian) pal­aeoceanographic change along the northwestern Tethyan margin (Vocontian Trough, southeastern France): δ13C, δ18O and Sr‐isotope and whole‐rock records. Cretaceous Research 30, 1247–1262.

Brenot, A., Cloquet, C., Vigier, N., Carignan, J., France‐Lanord, C., 2008. Magnesium isotope systematics of the lith­ologically varied Moselle river basin, France. Geochimica et Cosmochim Acta 72, 5070–5089.

Bristow, T.F., Kennedy, M.J., 2008. Carbon isotope excursions and the oxidant budget of the Ediacaran atmosphere and ocean. Geology 36, 863–866.

Broecker, W.S., Peng, T.H., Beng, Z., 1982. Tracers in the Sea. Lamont‐Doherty Geological Observatory, Columbia University, New York, 690 p.

Burke, W.H., Denison, R.E., Hetherington, E.A., Koepnick, R.B.,Nelson, H.F., Otto, J.B., 1982. Variation of seawater 87Sr/86Sr throughout Phanerozoic time. Geology 10, 516–519.

Calver, C.R., Black, L.P., Everard, J.L., Seymour, D.B., 2004. U‐Pb zircon age constraints on late Neoproterozoic glacia­tion in Tasmania. Geology 32, 892–896.

Canfield, D.E., 1989. Reactive iron in marine sediments. Geochimica et Cosmochimica Acta 53, 619–632.

Canfield, D.E., 1998. A new model for Proterozoic ocean chemistry. Nature 396, 450–453

Canfield, D.E., Teske, A., 1996. Late Proterozoic rise in atmo­spheric oxygen concentration inferred from phylogenetic and sulphur‐isotope studies. Nature 381, 127–132.

Canfield, D.E., Poulton, S.P., Knoll, H., Narbonne, G.M., Ross, G., Goldberg, T., Strauss, H., 2008. Ferruginous conditions dominated later Neoproterozoic deep‐water chemistry. Science 321, 949–952.

Chang, V.T.‐C., Masishima, A., Belshaw, N.S., Onions, R.K., 2003. Purification of Mg from low Mg biogenic carbonates for isotope ratio determination using multi‐collector ICP‐MS. Journal of Analytical Atomic Spectrometry 18, 296–301.

Charbonnier, G., Morales, C., Duchamp‐Alphonse, S., Westermann, S., Adatte, T., Föllmi, K.B., 2017. Mercury enrichment indicates volcanic triggering of Valanginian

environmental change. Scientific Reports 7, 40808, January 2017. doi:10.1038/srep40808.

Chiglino, L., Gaucher, C., Sial, A.N., Bossi, J., Ferreira, V.P., Pimentel, M.M., 2010. Chemostratigraphy of Mesoproterozoic and Neoproterozoic carbonates of the Nico Pérez Terrane, Río de la Plata Craton, Uruguay. Precambrian Research 182, 313–336.

Claypool, G.E., Holser, W.T., Kaplan, I.R., Sakai, H., Zak, I., 1980. The ages, curves of sulfur and oxygen isotopes in marine sulfate and their mutual interpretation. Chemical Geology 28, 199–260.

Cooper, R.A., Nowlan, G.S., Williams, S.H., 2001. Global Stratotype Section and Point for base of the Ordovician System. Episodes 24, 19–28.

Corsetti, F.A., Kaufman, A.J., 2003. Stratigraphic investiga­tions of carbon isotope anomalies and Neoproterozoic ice ages in Death Valley, California. Geological Society of America Bulletin 115, 916–932.

Cremonese, L., Struck, U., Shields‐Zhou, G., Ling, H., Och, L., 2009. δ15N chemostratigraphy of Ediacaran–Cambrian sections of South China. Supplement Geochimica et Cosmochimica Acta 73(13S), A‐251.

Cremonese, L., Shields‐Zhou, G., Struck, U., Ling, H.F., Och, L., Chen, X., Li, D., 2013. Marine biogeochemical cycling during the early Cambrian constrained by a nitrogen and organic carbon isotope study of the Xiaotan section, South China. Precambrian Research 225, 148–165.

Cremonese, L., Shields‐Zhou, G., Struck, U., Ling, H.F., Och, L., 2014. Nitrogen and organic carbon isotope stratigraphy of the  Yangtze Platform during the Ediacaran–Cambrian transition in South China. Palaeogeography, Palaeoclimatology, Palaeoecology 398, 165–186.

Crowe, S.A., Døssing, L.N., Beukes, N.J., Bau, M., Kruger, S.J., Frei, R., Canfield, D.E., 2013. Atmospheric oxygenation three billion years ago. Nature 501, 535–538.

Dantas, E.L., Alvarenga, C.J.S., Santos, R.V., Pimentel, M.M., 2009. Using Nd isotopes to understand the provenance of sedimentary rocks from a continental margin to a foreland basin in the Neoproterozoic Paraguay Belt, Central Brazil. Precambrian Research 170, 1–12.

Davies, C.L., Surridge, B.W.J., Gooddy, D.C., 2014. Phosphate oxygen isotopes within aquatic ecosystems: Global data syn­thesis and future research priorities. Science of the Total Environment 496, 563–575.

Dellinger, M., Bouchez, J., Gaillardet, J., Faure, L., Moureau, J., 2017. Tracing weathering regimes using the lithium isotope composition of detrital sediments. Geology 45, 411–414.

DePaolo, D.J., 1988. Neodymium Isotope Geochemistry: An Introduction. Springer‐Verlag, Berlin, New York, 187 p.

Dera, G., Puceat, E., Pellenard, P., Neige, P., Delsate, D., Joachimski, M.M., Reisberg, L., Martinez, M., 2009. Water mass exchange and variations in seawater temperature in the NW Tethys during the Early Jurassic; evidence from neo­dymium and oxygen isotopes of fish teeth and belemnites. Earth and Planetary Science Letters 286, 198–207.

Døssing, L.N., Dideriksen, K., Stipp, S.L.S., Frei, R., 2011. Reduction of hexavalent chromium by ferrous iron: A pro­cess of chromium isotope fractionation and its relevance to natural environments. Chemical Geology 285, 157–166.

0004203144.INDD 16 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 17

Dupuis, C., Aubry, M.‐P., Steurbaut, E., Berggren, W.A., Ouda, K., Magioncalda, R., Cramer, B.S., Kent, D.V., Speijer, R.P., Heilmann‐Clausen, C., 2003. The Dababiya Quarry section: Lithostratigraphy, clay mineralogy, geochemistry and pale­ontology. Micropaleontology 49, 41–59.

Eagle, R.A., Schauble, E.A., Tripati, A.K., Tutken, T., Hulbert, R.C., Eiler, J.M., 2010. Body temperatures of modern and extinct vertebrates from 13C–18O bond abundances in bioa­petite. Proceedings of the National Academy of Sciences 107, 10377–10382.

Eiler, J.M., 2007. “Clumped‐isotope” geochemistry: The study of naturally‐occurring, multiply‐substituted isotopologues. Earth and Planetary Science Letters 262, 309–327. doi:10.1016/ j.epsl.2007.08.020.

Elderfield, H., Greaves, M.J., 1982. The rare earth elements in seawater. Nature 296, 214–219.

Emiliani, C. 1955. Pleistocene temperatures. The Journal of Geology 63(6), 538–578.

Falkowski, P., 2003. Biogeochemistry of primary production in the sea. Treatise on Geochemistry 8, 185–213.

Fantle, M.S., Tipper, E.T., 2014. Calcium isotopes in the global biogeochemical Ca cycle: Implications for development of a Ca isotope proxy. Earth‐Science Review 129, 148–177.

Farquhar, J., Bao, H., Thiemens, M., 2000. Atmospheric influence of Earth’s earliest sulfur cycle. Science 289, 756–758.

Finlay, A.J., Selby, D., Grocke, D.R., 2010. Tracking the Hirnantian glaciation using Os isotopes. Earth and Planetary Science Letters 293, 339–348.

Fio, K., Spangenberg, J.E., Vlahovic, I., Sremac, J., Velic, I., Mrinjek, E., 2010. Stable isotope and trace element stratig­raphy across the Permian–Triassic transition: A redefinition of the boundary in the Velebit Mountain, Croatia. Chemical Geology 278(1), 38–57.

Flament, N., Coltice, N., Rey, P.F., 2011. The evolution of 87Sr/86Sr of marine carbonates does not constrain continental growth. Precambrian Research 229, 177–188. doi:10.1016/j.precamres.2011.10.009.

Font, F., Adatte, T., Sial, A.N., Lacerda, L.D., Keller, G., Punekar, J., 2016. Mercury anomaly, Deccan Volcanism and the end‐Cretaceous Mass Extinction. Geology 44, 171–174. doi:10.1130/G37451.1.

Font, E., Adatte, T., Andrade, M., Keller, G., Mbabi Bitchong, A., Carvallo, C., Ferreira, J., Diogo, Z., Mirão, J., 2018. Deccan volcanism induced high‐stress environment during the Cretaceous–Paleogene transition at Zumaia, Spain: Evidence from magnetic, mineralogical and biostratigraphic records. Earth and Planetary Science Letters 484, 53–66.

Foster, G.L., 2008. Seawater pH, pCO2 and [CO3 2−] variations in the Caribbean Sea over the last 130 kyr: A boron isotope and B/Ca study of planktic foraminifera. Earth Planetary Science Letters 271, 254–66.

Foster, G.L., Rae, J.W.B., 2016. Reconstructing ocean pH with boron isotopes in foraminifera. Annual Reviews Earth Planetary Science 44, 207–37.

Foster, G.L., Pogge von Strandmann, P.A.E., Rae, J.W.B., 2010 Boron and magnesium isotopic composition of seawater. Geochemistry, Geophysics, Geosystems 11, 1–10.

Frei, R., Frei, K.M., 2002. A multi‐isotopic and trace element investigation of the Cretaceous–Tertiary boundary layer at

Stevns Klint, Denmark–inferences for the origin and nature of siderophile and lithophile element geochemical anomalies. Earth and Planetary Science Letters 203(2), 691–708.

Frei, R., Gaucher, C., Poulton, S.W., Canfield, D.E., 2009. Fluctuations in Precambrian atmospheric oxygenation recorded by chromium isotopes. Nature 46, 250–254.

Frei, R., Gaucher, C., Døssing, L.N., Sial, A.N., 2011. Chromium isotopes in carbonates: A tracer for climate change and for reconstructing the redox state of ancient sea­water. Earth Planetary Sciences Letters 312, 114–125.

Frei, R., Gaucher, C., Stolper, D., Canfield, D.E., 2013. Fluctuations in late Neoproterozoic atmospheric oxidation: Cr isotope chemostratigraphy and iron speciation of the late Ediacaran lower Arroyo del Soldado Group (Uruguay). Gondwana Research 23, 797–811.

Frei, R., Døssing, L.N., Gaucher, C., Boggiani, P.C., Frei, K.M., Bech Árting, T., Crowe, S.A., Freitas, B.T., 2017. Extensive oxidative weathering in the aftermath of a late Neoproterozoic glaciation: Evidence from trace element and chromium isotope records in the Urucum district (Jacadigo Group) and Puga iron formations (Mato Grosso do Sul, Brazil). Gondwana Research 49, 1–20.

Friedrich, O., Norris, R.D., Erbacher, J., 2012. Evolution of middle to Late Cretaceous oceans: A 55 m.y. record of Earth’s temperature and carbon cycle. Geology 40, 107–110.

Frimmel, H.E., 2008. REE geochemistry of Neoproterozoic carbonates: Deviations from normal marine signatures. Abstract, 33 International Geological Congress, Oslo, Norway.

Frimmel, H.E., 2009. Trace element distribution in Neopro­terozoic carbonates as palaeoenvironmental indicator. Chemical Geology 258, 338–353.

Frimmel, H.E., 2010. On the reliability of stable carbon iso­topes for Neoproterozoic chemostratigraphic correlation. Precambrian Research 182, 239–252.

Frost, C.D., von Blanckenburg, F., Schoenberg, R., Frost, B.R., Swapp, S.M., 2007. Preservation of Fe isotope heterogene­ities during diagenesis and metamorphism of banded iron formation. Contributions Mineralogy Petrology 153, 211–235.

Galy, A., Bar‐Matthews, M., Halicz, L., O’Nions, R.K., 2002. Mg isotopic composition of carbonate: Insight from speleo­them formation. Earth Planetary Science Letters 201, 105–115.

Gaucher, C., Sial, A.N., Ferreira, V.P., Pimentel, M.M., Chiglino, L., Sprechmann, P., 2007. Chemostratigraphy of the Cerro Victoria Formation (Lower Cambrian, Uruguay): Evidence for progressive climate stabilization across the Precambrian–Cambrian boundary. Chemical Geology 237, 28–46.

Gaucher, C., Sial, A.N., Frei, R., 2015. Chemostratigraphy of Neoproterozoic banded iron formation (BIF): Types, age and origin. In: Ramkumar, M. (Ed.), Chemostratigraphy: Concepts Techniques and Applications. Elsevier, Amsterdam, pp. 433–449.

Geske, A., Goldstein, R.H., Mavromatis, V., Richter, D.K., Buhl, D., Kluge, T., John, C.M., Immenhauser, A., 2015. The magnesium isotope (δ26Mg) signature of dolomites. Geochimica et Cosmochimica Acta 149, 131–151.

0004203144.INDD 17 11/9/2018 4:52:37 PM

18 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

Ghienne, J.‐F., Desrochers, A., Vandenbroucke, T.R.A., Achab, A., Asselin, E., Dabard, M.‐P., Farley, C., Loi, A., Paris, F., Wickson, S., Veizer, J., 2014. A Cenozoic‐style scenario for the end‐Ordovician glaciation. Nature Communications 5, 4485. doi:10.1038/ncomms5485.

Ghosh, P., Adkins, J., Affek, H., Balta, B., Guo, W., Schauble, E.A., Schrag, D., Eiler, J.M., 2006. 13C‐18O bonds in carbonate minerals: A new kind of paleothermometer. Geochimica et Cosmochimica Acta 70, 1439–1456.

Goldman, D., Leslie, S.A., Nõlvak, J., Young, S., Bergström, S.M., Huff, W.D., 2007. The Global Stratotype Section and Point (GSSP) for the base of the Katian Stage of the Upper Ordovician Series at Black Knob Ridge, Southeastern Oklahoma, USA. Episodes 30, 258–270.

Grasby, S.E., Sanei, H., Beauchamp, B., Chen, Z., 2013. Mercury deposition through the Permo–Triassic Biotic Crisis. Chemical Geology 351, 209–16.

Grasby, S.E., Beauchamp, B., Bond, D.P.G., Wignall, P.B., Sanei, H., 2015. Mercury anomalies associated with three extinction events (Capitanian Crisis, Latest Permian Extinction and the Smithian/Spathian Extinction) in NW Pangea. Geological Magazine 153, 285–297.

Grasby, S.E., Shen, W., Yin, R., Gleason, J.D., Blum, J.D., Lepak, R.F., Hurley, J.P., Beauchamp, B., 2017. Isotopic sig­natures of mercury contamination in latest Permian oceans. Geology 45, 55–58.

Grotzinger, J.P., Fike, D.A., Fischer, W.W., 2011. Enigmatic origin of the largest‐known carbon isotope excursion in Earth’s history. Nature Geoscience 4, 285–291.

Guo, Q., Strauss, H., Kaufman, A.J., Schröder, S., Gutzmer, J., Wing, B., Baker, M.A., Bekker, A., Jin, Q., Kim, S.‐T., Farquhar, J., 2009a. Reconstructing Earth’s surface oxidation across the Archean–Proterozoic transition. Geology 3, 399–402.

Guo, W., Mosenfelder, J.L., Goddard, W.A., III, Eiler, J.M., 2009b. Isotopic fractionations associated with phosphoric acid digestion of carbonate minerals: Insights from first‐principles theoretical modeling and clumped isotope measurements. Geochimica et Cosmochimica Acta 73, 7203–7225.

Gussone, N., Schmitt, A.‐D., Heuser, A., Wmbacher, F., Dietzel, M., Tipper, E., Schiller, M., 2016. Calcium Isotope Geochemistry. Advances in Isotope Geochemistry. Springer, Berlin, Heidelberg, 260 p. doi:10.1007/978‐3‐540‐68953‐9.

Halevy, I., Johnston, D., Schrag, D., 2010. Explaining the struc­ture of the Archean mass‐independent sulfur isotope record. Science 329, 204–207

Hall, J.M., Chan, L.H., 2004. Li/Ca in multiple species of ben­thic and planktonic foraminifera: Thermocline, latitudinal, and glacial–interglacial variation. Geochimica Cosmochimica Acta 68, 529–545.

Hall, J.M., Chanb, T.L.‐H., McDonough, W.F., Turekian, K.K., 2005. Determination of the lithium isotopic composition of planktic foraminifera and its application as a paleo‐seawater proxy. Marine Geology 217, 255–265.

Halverson, G.P., Hoffman, P.F., Schrag, D.P., Maloof, A.C., Hugh, A., Rice, N., 2005. Towards a Neoproterozoic composite carbon‐isotope record. Geological Society of America Bulletin 117, 1181–1207.

Halverson, G.P., Dudás, F.Ö., Maloof, A., Bowring, S.A., 2007. Evolution of the 87Sr/86Sr composition of Neoproterozoic seawater. Palaeogeography, Palaeoclimatology, Palaeoecology 256, 103–129.

Halverson, G.P., Hurtgen, M.T., Porter, S.M., Collins, A.S., 2010a. Neoproterozoic–Cambrian biogeochemical evolution. In: Gaucher, C., Sial, A.N., Halverson, G.P., Frimmel, H. (Eds.), Neoproterozoic‐Cambrian Tectonics, Global Change and Evolution: A Focus on South Western Gondwana. Developments in Precambrian Geology 16. Elsevier, Amsterdam, Boston, pp. 351–365.

Halverson, G.P., Wade, B.P., Hurtgen, M.T., Barovich, K., 2010b. Neoproterozoic chemostratigraphy. In: Karhu, J., Sial, A.N., Ferreira, V.P. (Eds.), Precambrian Isotope Stratigraphy, special issue. Precambrian Research 182. Elsevier, Amsterdam, pp. 337–350.

Halverson, G.P., Poitrasson, F., Hoffman, P.F., Nedelec, A., Montel, J.M., Kirby, J., 2011. Fe isotope and trace element geochemistry of the Neoproterozoic syn‐glacial Rapitan iron formation. Earth Planetary Science Letters 309, 100–112.

Hammarlund, E.U., Dahl, T.W., Harper, D.A., Bond, D.P., Nielsen, A.T., Bjerrum, C.J., Schovsbo, N.H., Schönlaub, H.P., Zalasiewicz, J.A., Canfield, D.E., 2012. A sulfidic driver for the end‐Ordovician mass extinction. Earth and Planetary Science Letters 331, 128–139.

Hardisty, D.S., Lu, Z., Planavsky, N.J., Bekker, A., Philippot, P., Zhou, Z., Lyons, T.W., 2014. An iodine record of Paleo­proterozoic surface ocean oxygenation. Geology 42, 619–622.

Heimann, A., Johnson, C.M., Beard, B.L., Valley, J.W., Roden, E.E., Spicuzza, M.J., Beukes, N.J., 2010. Fe, C, and O iso­tope compositions of banded iron formation carbonates demonstrate a major role for dissimilatory iron reduction in 2.5 Ga marine environments. Earth Planetary Science Letters 294, 8–18.

Hemming, N.G., Hönisch, B., 2007. Boron isotopes in marine carbonate sediments and the pH of the Ocean. Developments in Marine Geology 1, 717–734.

Herrle, J.O., Pross, J., Friedrich, O., Koßler, P., Hemleben, C., 2003. Forcing mechanisms for mid‐Cretaceous black shale formation: Evidence from the Upper Aptian and Lower Albian of the Vocontian Basin (SE France). Palaeogeography, Palaeoclimatology, Palaeoecology 190, 399–426.

Higgins, J.A., Schrag, D.P., 2010. Constraining magnesium cycling in marine sediments using magnesium isotopes. Geochimica Cosmochimica Acta 74, 5039–5053.

Hilgen, F.J., Abels, H.A., Iaccarino, S., Krijgsman, H., Raffi, I., Sprovieri, R., Turco, E., Zachariasse, E.W., 2009. The Global Stratotype Section and Point (GSSP) of the Serravallian Stage (Middle Miocene). Episodes 32, 152–166.

Hoffman, P.F., 2011. Strange bedfellows: Glacial diamictite and cap carbonate from the Marinoan (635 Ma) glaciation in Namibia. Sedimentology 58, 57–119.

Hoffman, P.F., Kaufman, A.J., Halverson, G.P., Schrag, D.P., 1998a. A Neoproterozoic snowball Earth. Science 281, 1342–1346.

Hoffman, P.F., Kaufman, A.J., Halverson, G.P., 1998b. Comings and goings of global glaciations on a Neoproterozoic tropical platform in Namibia. GSA Today 8, 1–9.

0004203144.INDD 18 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 19

Holser, W.T., 1997. Geochemical events documented in inor­ganic carbon isotopes. Palaeogeography, Palaeoclimatology, Palaeoecology 132, 173–182.

Hönisch, B., Ridgwell, A., Schmidt, D.N., Thomas, E., Gibbs, S.J., Sluijs, A., Zeebe, R., Kump, L., Martindale, R.C., Greene, S.E., Kiessling, W., Ries, J., Zachos, J.C., Royer, D.L., Barker, S., Marchitto, T.M., Jr., Moyer, R., Pelejero, C., Ziveri, P., Foster, G.L., Williams, B., 2012. The geological record of ocean acidification. Science 335, 1058. doi:10.1126/science.1208277.

Hu, R., Wang, W., Li, S.‐K.,Yang, Y.‐Z., Chen, F., 2016. Sedimentary environment of Ediacaran sequences of South China: Trace element and Sr‐Nd isotope constraints. The Journal of Geology 124, 769–789.

Huang, K.‐J., Teng, F.‐Z., Shen, B., Xiao, S., Lang, X., Ma, H.‐R., Fu, Y., Peng, Y., 2016. Episode of intense chemical weathering during the termination of the 635 Ma Marinoan glaciation. Proceedings of the National Academy of Science 113, 14904–14909.

Jacobsen, S.B., Kaufman, A.J., 1999. The Sr, C and O isotopic evolution of Neoproterozoic seawater. Chemical Geology 161, 37–57.

Jarvis, I., Lignum, J.S., Gröcke, D.R., Jenkyns, H.C., Pearce, M.A., 2011. Black shale deposition, atmospheric CO

2 drawdown, and cooling during the Cenomanian‐Turonian Oceanic Anoxic Event. Paleoceanography 26, PA3201. doi:10.1029/2010PA002081.

Jenkyns, H.C., 2003. Evidence for rapid climate change in the  Mesozoic–Palaeogene greenhouse world. Philosophical Transactions of the Royal Society of London 361, 1885–1916.

Jenkyns, H.C., 2010. Geochemistry of oceanic anoxic events. Geochemistry, Geophysics, Geosystems 11(3). doi:10.1029/ 2009GC002788.

Jenkyns, H.C., Clayton, C.J., 1986, Black shales and carbon isotopes from the Tethyan Lower Jurassic. Sedimentology 33, 87–106.

Jenkyns, H.C., Gröcke, D.R., Hesselbo, S.P., 2001. Nitrogen iso­tope evidence for water mass denitrification during the early Toarcian (Jurassic) oceanic anoxic event. Paleoceanography 16(6), 593–603. doi:10.1029/2000PA000558.

Jenkyns, H.C., Matthews, A., Tsikos, H., Erel, Y., 2007. Nitrate reduction, sulfate reduction, and sedimentary iron isotope evolution during the Cenomanian–Turonian oceanic anoxic event. Paleoceanography 22, PA3208. doi:10.1029/ 2006PA001355.

Jenkyns, H., Schouten‐Huibers, L., Schouten S., Sinninghe‐Damste, J.S., 2012. Warm Middle Jurassic‐early Cretaceous high‐latitude sea surface temperature from the Southern Ocean. Climate of the Past 8, 215–226.

Joachimski, M.M., Buggisch, W. 2002. Conodont apatite δ18O signatures indicate climatic cooling as a trigger of the Late Devonian mass extinction. Geology 30, 711–714.

Joachimski, M.M., Simon, L., Van Geldern, R., Lécuyer, C., 2005. Boron isotope geochemistry of Paleozoic brachiopod calcite: Implications for a secular change in the boron isotope geochemistry of seawater over the Phanerozoic. Geochimica et Cosmochimica Acta 69, 4035–4044.

Johnson, C., Beard, L., 2006. Fe isotopes: An emerging tech­nique for understanding modern and ancient biogeochemical cycles. GSA Today 16, 4–10.

Johnson, C.M., Beard, B.L., Beukes, N.J., Klein, C., O’Leary, J.M., 2003. Ancient geochemical cycling in the Earth as inferred from Fe isotope studies of banded iron formations from the Transvaal Craton. Contributions to Mineralogy and Petrology 114, 523–547.

Johnson, C.M., Beard, B.L., Albarede, F., 2004. Geochemistry of non‐traditional stable isotopes. Reviews in Mineralogy and Geochemistry 55, 454 pp.

Johnson, C.M., Beard, B.L., Roden, E.E., 2008. The iron iso­tope fingerprints of redox and biogeochemical cycling in modern and ancient oceans. Annual Reviews Earth Planetary Science 36, 457–493.

Johnston, D.T., Poulton, S.W., Dehler, C., Porter, S., Husson, J., Canfield, D.E., Knoll, A.H., 2010. An emerging picture of Neoproterozoic ocean chemistry: Insights from the Chuar Group, Grand Canyon, USA. Earth and Planetary Science Letters 290, 64–73.

Jones, C.E., Jenkyns, H.C., 2001. Seawater strontium isotopes, oceanic anoxic events, and seafloor hydrothermal activity in the Jurassic and Cretaceous. American Journal of Science 301, 112–149.

Jones, D.S., Martini, A.M., Fike, A., Kaiho, K., 2017. A volcanic trigger for the Late Ordovician mass extinction? Mercury data from south China and Laurentia. Geology 45(7), 631–634. doi:10.1130/G38940.1.

Kah, L.C., Sherman, A.G., Narbonne, G.M., Knoll, A.H., Kaufman, A.J., 1999. δ13C stratigraphy of the Proterozoic by lot supergroup, Baffin Islands, Canada: Implications for regional lithostratigraphy correlations. Canadian Journal of Earth Sciences 36, 313–332.

Kamber, B.S., Webb, G.E., Gallagher, M., 2014. The rare earth element signal in Archaean microbial carbonate: Information on ocean redox and biogenicity. Journal of the Geological Society 171(6), 745–763.

Karhu, J.A., Holland, H.D., 1996. Carbon isotopes and the rise of atmospheric oxygen. Geology 24, 867–870.

Karhu, J., Sial, A.N., Ferreira, V.P., 2010. Insights from pre­cambrian isotope stratigraphy. Precambrian Research 182, 239–412.

Kasemann, S.A., Hawkesworth, C.J., Prave, A.R., Fallick, A.E., Pearson, P.N., 2005. Boron and calcium isotope composition in Neoproterozoic carbonate rocks from Namibia: Evidence for extreme environmental change. Earth Planetary Sciences Letters 23, 73–86.

Kato, Y., Yamaguchi, K.E., Ohmoto, H., 2006. Rare earth elements in Precambrian banded iron formations: Secular changes of Ce and Eu anomalies and evolution of atmospheric oxygen. Geological Society of America Memoirs 198, 269–289.

Kaufman, A.J., Knoll, A.H., Awramik, S.M., 1992. Biostratigraphic and chemostratigraphic correlation of Neoproterozoic sedimentary successions: Upper Tindir Group, northwestern Canada, as a test case. Geology 20, 181–185.

Kaufman, A.J., Jacobsen, S.B., Knoll, A.H., 1993. The Vendian record of C‐ and Sr‐isotopic variations: Implications for tec­tonics and paleoclimate. Earth and Planetary Science Letters 120, 409–430.

Kaufman, A.J., Knoll, A.H., Narbonne, G.M., 1997. Isotopes, ice ages, and terminal Proterozoic Earth history. Proceedings of the National Academy Science 94, 6600–6605.

0004203144.INDD 19 11/9/2018 4:52:37 PM

20 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

Kaufman, A.J., Sial, A.N., Ferreira, V.P., (guest editors), 2007a.Precambrian isotope chemostratigraphy. Chemical Geology 237(1/2), special Issue, 232 p.

Kaufman, A.J., Sial, A.N., Ferreira, V.P., 2007b. Preface to special issue of chemical geology on precambrian chemostratigraphy in honor of the late William T. Holser. Chemical Geology 237, 1–4.

Kaufman, A.J., Sial, A.N., Frimmel, H.E., Misi, A., 2009. Neoproterozoic to Cambrian Palaeoclimatic events in Southwestern Gondwana. In: Gaucher, C., Sial, A.N., Halverson, G.P., Frimmel, H. (Eds.), Neoproterozoic–Cambrian Tectonics, Global Change and Evolution: A Focus on Southwestern Gondwana. Developments in Precambrian Geology 16. Elsevier, Amsterdam, pp. 369–388.

Keller, G., 2005. Impacts, volcanism and mass extinction: Random coincidence or cause and effect? Australian Journal of Earth Science 52, 725–757.

Keller, G., Mateo, P., Punekar, J., Khozyem, H., Gertsch, B., Spangenberg, J., Bitchong, A.M., Adatte, A., 2018. Environmental changes during the Cretaceous–Paleogene mass extinction and Paleocene–Eocene Thermal Maximum: Implications for the Anthropocene. Gondwana Research 56, 69–89.

Kennedy, M.J., Runnegar, B., Prave, A.R., Hoffmann, K.H., Arthur, M.A., 1998.Two or four Neoproterozoic glaciations? Geology 26, 1059–1063.

Keto, L.S., Jacobsen, S.B., 1988. Nd isotopic variations of early Paleozoic oceans. Earth Planetary Science Letters 84, 27–41.

Kikumoto, R., Tahata, M., Nishizawa, M., Sawaki, Y., Maruyama, S., Shu, D., Han, J., Komiya, T., Takai, K., Ueno, Y., 2014. Nitrogen isotope chemostratigraphy of the Ediacaran and Early Cambrian platform sequence at Three Gorges, South China. Gondwana Research 25(3), 1057–1069.

Klein, C., Beukes, N.J., 1993. Sedimentology and geochemistry of the glaciogenic Late Proterozoic Rapitan Iron Formation in Canada. Economic Geology 88, 542–565.

Knauth, L.P., Kennedy, M.J., 2009. The late Precambrian greening of the Earth. Nature 460, 728–732.

Knoll, A.H., Walter, M.R., 1992. Latest Proterozoic stratig­raphy and Earth history. Nature 356, 673–678.

Knoll, A.H., Hayes, J.M., Kaufman, A.J., Swett, K., Lambert, I.B., 1986. Secular variation in carbon isotope ratios from Upper Proterozoic successions of Svalbard and East Greenland. Nature 321, 832–838.

Knoll, A., Walter, M., Narbonne, G., Christie‐Blick, N., 2006. The Ediacaran Period: A new addition to the geologic time scale. Lethaia 39, 13–30.

Korte, C., Kozur, H.W, 2010. Carbon‐isotope stratigraphy across the Permian–Triassic boundary: A review. Journal of Asian Earth Sciences 39, 215–235.

Kump, L.R., Arthur, M.A., 1999. Interpreting carbon‐isotope excursions: Carbonates and organic matter. Chemical Geology 161, 181–198.

Kuznetsov, A., Melezhik, V., Gorokhov, I., Melnikov, N., Konstantinova, G., Kutyavin, E., Turchenko, T. 2010. Sr isotopic composition of Paleoproterozoic 13C‐rich carbonate rocks: The Tulomozero Formation, SE Fennoscandian Shield. Precambrian Research 182, 300–312.

Lawrence, M.G., Kamber, B.S., 2006. The behaviour of the rare earth elements during estuarine mixing: Revisited. Marine Chemistry 100(1), 147–161.

Lea, D.W., Pak, D.K., Spero, H.J., 2000. Climate impact of late Quaternary equatorial Pacific sea surface temperature varia­tions. Science 289, 1719–1724.

Lear, C.H., Y. Rosenthal, 2006. Benthic foraminiferal Li/Ca: Insights into Cenozoic seawater carbonate saturation state, Geology 34, 985–988. doi:10.1130/G22792A.1.

Lear, C.H., Elderfield, H., Wilson, P.A., 2000. Cenozoic deep‐sea temperatures and global ice volumes from Mg/Ca in benthic foraminiferal calcite. Science 287, 269–272.

Lechler, M., Pogge von Strandmann, P.A.E., Jenkyns, H.C., Prosser, G., Parente, M., 2015. Lithium‐isotope evidence for enhanced silicate weathering during OAE 1a (Early Aptian Selli event). Earth and Planetary Science Letters 432, 210–222.

Leckie, R.M., Bralower, T.J., Cashman, R., 2002, Oceanic anoxic events and plankton evolution: Biotic response to tec­tonic forcing during the mid‐Cretaceous. Paleoceanography 17, 1041. doi:10.1029/2001PA000623.

Lécuyer, C., Brouxel, M., Albarède, F., 1990 Elemental fluxes during hydrothermal alteration of the Trinity ophiolite (California, USA) by seawater. Chemical Geology 89, 87–115.

Lécuyer, C., Reynard, B., Grandjean, P., 2004. Rare earth element evolution of Phanerozoic seawater recorded in bio­genic apatites. Chemical Geology 204(1), 63–102.

Lenton, T.M., Crouch, M., Johson, M., Pires, N., Dolan, L., 2012. First plants cooled the Ordovician. Nature Geoscience 5, 86–89.

Li, W., Chakraborty, S., Beard, B.L., Romanek, C.S., Johnson, C.M., 2012. Magnesium isotope fractionation during precip­itation of inorganic calcite under laboratory conditions. Earth Planetary Science Letters 333–334, 304–316.

Lindsay, J.F., Brasier, M.D., 2002. Did global tectonics drive early biosphere evolution? Carbon isotope record from 2.6 to1.9 Ga carbonates of Western Australian basins. Precambrian Research 114, 1–34.

Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene‐Pleistocene stack of 57 globally distributed benthic δ18O records. Paleoceanography 20(1), PA1003.

Liu, Y.G., Miah, M.R.U., Schmitt, R.A., 1988. Cerium: A chemical tracer for paleo‐oceanic redox conditions. Geochimica et Cosmochimica Acta 52(6), 1361–1371.

Lu, Z., Jenkyns, H.C., Rickaby, R.E.M., 2010. Iodine to calcium ratios in marine carbonates as a paleo‐redox proxy during oceanic anoxic events. Geology 38, 1107–1110.

Luo, G., Algeo, T.J., Zhan, R., Yan, D., Huang, J., Liu, J., Xie, S., 2016. Perturbation of the marine nitrogen cycle during the Late Ordovician glaciation and mass extinction. Palaeogeography, Palaeoclimatology, Palaeoecology 448, 339–348.

Lyons, T.W., Reinhard, C.T., 2009. Oxygen for heavy‐metal fans. Nature 461,179–181.

Lyons, T.W., Anbar, A.D., Severmann, S., Scott, C., Gill, B.C., 2009. Tracking euxinia in the ancient ocean: A multiproxy perspective and Proterozoic case study. Annual Review of Earth and Planetary Sciences 37, 507–534.

Macdougall, J.D., 1991. Radiogenic isotopes in seawater and sedimentary systems In: Heaman, L., Ludden, J.N. (Eds.), Applications of Radiogenic Isotope Systems to Problems in Geology. NMCA Short Course Handbook. Mineralogical Association of Canada, Toronto, ON, pp. 337–364, (Chapter10).

0004203144.INDD 20 11/9/2018 4:52:37 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 21

Magaritz, M., 1989. δl3C minima follow extinction events: A clue to faunal radiation. Geology 17, 337–340.

Magaritz, M., Holser, W.T., Kirschvink, J.L., 1986. Carbon‐isotope events across the Precambrian/Cambrian boundary on the Siberian Platform. Nature 320, 258–259.

Maloof, A.C., Porter, S.M., Moore, J.H., Dudás, F.O., Bowring, S.A., Higgins, J.A., Fike, D.A., Eddy, M.P., 2010. The earliest Cambrian record of animals and ocean geo­chemical change. Geological Society of America Bulletin 122, 1731–1774.

Marriott, C.S., Henderson, G.M., Crompton, R., Staubwasser, M., Shaw, S., 2004. Effect of mineralogy, salinity, and tempera­ture on Li/Ca and Li isotope composition of calcium carbonate. Chemical Geology 212, 5–15.

Marshall, J.D., 1992. Climatic and oceanographic isotopic sig­nals from the carbonate rock record and their preservation. Geological Magazine 129, 143–160.

McArthur, J.M., 2010. Strontium isotope stratigraphy. In: Ratcliffe K.T., Zaitlin B.A. (Eds.), Application of Modern Stratigraphic Techniques: Theory and Case Histories. SEPM Special Publication 94. SEPM (Society for Sedimentary Geology), Tulsa, OK, pp. 129–142.

McArthur, J.M., Howarth, R.J., Bailey, T.R., 2001. Strontium isotope stratigraphy: Lowess version 3: Best fit to the marine Sr‐isotope curve for 0–509 Ma and accompanying look‐up table for deriving numerical age. Journal of Geology 109, 155–170.

Melchin, M.J., Mitchell, C.E., Holmden, C., Storch, P., 2013. Environmental changes in the Late Ordovician‐early Silurian: Review and new insights from black shales and nitrogen isotopes. Geological Society of America Bulletin 125, 1635–1670.

Melezhik, V.A., Fallick, A.E., Medvedev, P.V., Marakarikhin, V.V., 1999. Extreme 13Ccarb enrichment in ca. 2.0Ga magne­site–stromatolite–dolomite‐‘redbeds’ association in a global context: A case for the world‐wide signal enhanced by a local environment. Earth Science Review 48, 71–120.

Melezhik, V.A., Gorokov, I.M., Kuznetsov, A.B., Fallick, A.E., 2001. Chemostratigraphy of Neoproterozoic carbonates: Implications for “blind dating”. Terra Nova 13, 1–11.

Melezhik, V.A., Roberts, D., Fallick, A.E., Gorokhov, I.M., Kusnetzov, A.B., 2005. Geochemical preservation potential of high‐grade calcite marble versus dolomite marble: Implication for isotope chemostratigraphy. Chemical Geology 216, 203–224.

Melezhik, V.A., Huhma, H., Fallick, A.E., Whitehouse, M.J., 2007.Temporal constraints on the Palaeoproterozoic Lomagundi‐Jatuli carbon isotope event. Geology 35, 655–658.

Meyer, K.M., Yu, M., Lehrmann, D., van de Schootbrugge, B., Payne, J.L., 2013. Constraints on early Triassic carbon cycle dynamics from paired organic and inorganic carbon isotope records. Earth Planetary Science Letters 361, 429–435.

Molina, E., Alegret, L., Arenillas, I., Arz, J.A., Gallala, N., Hardenbol, J., Von Salis, K., Steurbaut, E., Vandenberghe, N., Zaghbib‐Turki, D., 2006. The global boundary Section and Point for the base of the Danian Stage (Paleocene, Paleogene, “Tertiary”, Cenozoic) at El Kef, Tunisia: Original definition and revision. Episodes 29, 263–273.

Molina, E., Alegret, L., Arenillas, E., Arz, J.A., Gallala, N., Grajales‐Nishimura, J.M., Murillo‐Muñetón, G., Zaghbib‐Turki, D., 2009. The Global Boundary Stratotype Section and Point for the base of the Danian Stage (Paleocene, Paleogene, “Tertiary”, Cenozoic): Auxiliary sections and correlation. Episodes 32, 84–95.

Montañez, I.P., Osleger, D.A., Banner, J., Mack, L.E., Musgrove, M., 2000. Evolution of the Sr and C isotope com­position of Cambrian Oceans. GSA Today 10, 1–7.

Morante, R., Veevers, J.J., Andrew, A.S., Hamilton, P.J., 1994. Determination of the Permian–Triassic boundary in Australia. APEA Journal 34, 330–336.

Nascimento, R.S.C., Sial, A.N., Pimentel, M.M., 2007. C‐and Sr‐isotope systematics applied to Neoproterozoic marbles of the Seridó Belt, northeastern Brazil. Chemical Geology 237, 209–228.

Nascimento‐Silva, V.M., Sial, A.N., Ferreira, V.P., Neumann, V.H., Barbosa, J.A., Pimentel, M.M., Lacerda, L.D., 2011. Cretaceous–Paleogene transition at the Paraíba Basin, Northeastern, Brazil: Carbon‐isotope and mercury subsur­face stratigraphies. Journal of South American Earth Sciences 32, 379–392.

Nascimento‐Silva, M.V., Sial, A.N., Ferreira, V.P., Barbosa, J.A., Neumann, V.H., Pimentel, M.M., Lacerda, L.D., 2013. Carbon Isotopes, rare‐earth elements and mercury behavior of Maastrichtian–Danian carbonate succession of the Paraíba Basin, Northeastern Brazil. In: Bojar, A.V., Melinte‐Dobrinescu, M.C., Smit, J. (Eds.), Isotopic Studies in Cretaceous Research. Geological Society, London, Special Publications 382. Geological Society of London, London, pp. 85–104.

Nielsen, S.G., Rehkämper, M., 2012. Thallium isotopes and their application to problems in Earth and environmental sci­ence. In: Baskaran, M. (Ed.), Handbook of Environmental Isotope Geochemistry. Advances in Isotope Geochemistry. Springer‐Verlag, Berlin, pp. 247–269.

Nielsen, S.G., Mar‐Gerrison, S., Gannoun, A., LaRowe, D.E., Klemm, V., Halliday, A., Burton, K.W., Hein, J.R., 2009. Thallium isotope evidence for increased marine organic carbon export in the early Eocene. Earth Planetary Science Letters 278, 297–307.

Nielsen, S.G., Rehkämper, M., Prytulak, J., 2017. Investigation and application of thallium isotope fractionation. Reviews in Mineralogy and Geochemistry 82, 759–798.

Och, L.M., Shields‐Zhou, G.A., 2012. The Neoproterozoic oxygenation event: Environmental perturbations and biogeo­chemical cycling. Earth‐Science Reviews 110, 26–57.

Oehlert, A.M., Swart, P.K., 2014. Interpreting carbonate and organic carbon isotope covariance in the sedimentary record. Nature Communications 5, 4672. doi:10.1038/ncomms5672.

Palmer, M.R., Pearson, P.N., Cobb, S.J., 1998. Reconstructing past ocean pH‐depth profiles. Science 282(5393), 1468–1471. doi:10.1126/science.282.5393.1468.

Papineau, D., Mojzsis, S.J., Karhu, J.A., Marty, B., 2005. Nitrogen isotopic composition of ammoniated phyllosilicates: Case studies from Precambrian metamorphosed sedimentary rocks. Chemical Geology 216, 37–58.

Pasquier, V., Sansjofre, P., Rabineau, M., Revillon, S., Houghton, J., Fike, D.A., 2017. Pyrite sulfur isotopes reveal

0004203144.INDD 21 11/9/2018 4:52:38 PM

22 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

glacial−interglacial environmental changes. Proceedings of  the National Academy of Science 114, 5941–5945. doi:10.1073/pnas.1618245114.

Pearson, P.N., van Dongen, B.E., Nicholas, C.J., Pancost, R.D., Schouten, S., Singano, J.M., Wade, B.S., 2007. Stable warm tropical climate through the Eocene Epoch. Geology 35, 211–214.

Peckeur‐Ehrenbrink, B., Ravizza, G., 2000. The marine osmium isotope record. Terra Nova 12, 205–219.

Peckeur‐Ehrenbrink, B., Ravizza, G., 2012. Osmium isotope stratigraphy. In: Gradstein, F.M., Ogg, J.G., Schmitz, M.D., Ogg, G.M. (Eds.), The Geologic Time Scale 2012. Elsevier, Amsterdam, Boston, vol. 1, chapter 8, pp. 145–166.

Peng, S., Babcock, L., Robison, R., Lin, H., Rees, M., Saltzman, M., 2004. Global Standard Stratotype‐section and Point (GSSP) of the Furongian Series and Paibian Stage (Cambrian). Lethaia 37, 365–379.

Penman, D.E., Hönisch, B., Zeebe, R.E., Thomas, E., Zachos, J.C., 2014. Rapid and sustained surface ocean acidification during the Paleocene–Eocene Thermal Maximum. Paleoceanography 29, 357–369. doi:10.1002/2014PA002621.

Penniston‐Dorland, S., Liu, X.M., Rudnick, R.L., 2017. Lithium isotope geochemistry. Reviews in Mineralogy and Geochemistry 82, 165–217.

Percival, L.M.E., Witt, M.L.I., Mather, T.A., Hermoso, M., Jenkyns, H.C., Hesselbo, S.P., Al‐Suwaidi, A.H., Storm, M.S., Xu, W., Ruhl, M., 2015. Globally enhanced mercury deposition during the end‐Pliensbachian extinction and Toarcian OAE: A link to the Karoo–Ferrar Large Igneous Province. Earth and Planetary Science Letters 428, 267–280.

Percival, L.M.E., Ruhl, M., Hesselbo, S.P., Jenkyns, H.C., Mather, T.A., Whiteside, J.H., 2017. Mercury evidence for pulsed volcanism during the end‐Triassic mass extinction. Proceedings of the National Academy of Sciences of the United States of America 114(30), 7929–7934. doi:10.1073/pnas.1705378114.

Petrizzo, D.A., Young, E.D., Runnegar, B.N., 2014. Implications of high‐precision measurements of 13C–18O bond ordering in CO2 for thermometry in modern bivalved mollusc shells. Geochimica et Cosmochimica Acta 142, 400–410.

Pogge von Strandmann, P.A.E., Jenkyns, H.C., Woodfine, R.G., 2013. Lithium isotope evidence for enhanced weathering dur­ing Oceanic Anoxic Event 2. Nature Geoscience 6, 668–672.

Pogge von Strandmann, P.A.E., Desrochers, A., Murphy, M.J., Finlay, A.J., Selby, D., Lenton, T.M., 2017. Global climate stabilisation by chemical weathering during the Hirnantian glaciation. Geochemical Perspectives Letters 3(2), 230–237.

Pokrovsky, B.G., Mavromatis, V., Pokrovsky, O.S., 2011. Covariation of Mg and C isotopes in late Precambrian car­bonates of the Siberian Platform: A new tool for tracing the change in weathering regime? Chemical Geology 290, 67–74.

Poulton, S.W., Canfield, D.E., 2005. Development of a sequential extraction procedure for iron: Implications for iron partitioning in continentally derived particulates. Chemical Geology 214, 209–221.

Poulton, S.W., Fralick, P.W., Canfield, D.E., 2004. The transition to a sulphidic ocean ~1.84 billion years ago. Nature 431, 173–77.

Price, G.D., Mutterlose, J., 2004. Isotopic signals from late Jurassic–early Cretaceous (Volgian–Valanginian)sub‐Arctic belemnites, Yatria River, Western Siberia. Journal of the Geological Society 161, 959–968.

Puceat, E., Lecuyer, C., Sheppard, S.M.F., Dromart, G., Reboulet, S., Grandjean, P., 2003. Thermal evolution of Cretaceous Tethyan marine waters inferred from oxygen iso­tope composition of fish tooth enamels. Paleoceanography 18(2), 1029. doi:10.1029/2002PA000823.

Railsback, L.B., Gibbard, P.L., Head, M.J., Voarintsoa, N.R.G., Toucanne, S., 2015. An optimized scheme of lettered marine isotope substages for the last 1.0 million years, and the clima­tostratigraphic nature of isotope stages and substages. Quaternary Science Reviews 111, 94–106.

Ramkumar, M., 1999. Role of chemostratigraphic technique in reservoir characterization and global stratigraphic correla­tion. Indian Journal of Geochemistry 14, 33–45.

Ramkumar, M., 2014. Characterization of depositional units for stratigraphic correlation, petroleum exploration and reservoir characterization. In: Sinha, S. (Ed.), Advances in  Petroleum Engineering. Studium Press L.L.C, Houston, pp. 1–13.

Ramkumar, M., 2015. Toward standardization of terminologies and recognition of chemostratigraphy as a formal strati­graphic method. In: Ramkumar, M. (Ed.), Chemostratigraphy: Concepts, Techniques and Applications. Elsevier, Amsterdam, pp. 1–21. doi:10.1016/ B978‐0‐12‐419968‐2.00001‐7.

Ramkumar, M., Stüben, D., Berner, Z., 2010. Hierarchical delineation and multivariate statistical discrimination of che­mozones of the Cauvery Basin, South India: Implications on Spatio‐temporal scales of stratigraphic correlation. Petroleum Science 7, 435–447.

Ramkumar, M., Stüben, D., Berner, Z., 2011. Barremian–Danian chemostratigraphic sequences of the Cauvery Basin, South India: Implications on scales of stratigraphic correla­tion. Gondwana Research 19, 291–309.

Ravizza, G., Peucker‐Ehrenbrink, B., 2003. Chemostratigraphic evidence of Deccan Volcanism from the marine osmium iso­tope record. Science 302, 1392–1395.

Ries, J.B., Fike, D.A., Pratt, L.M., Lyons, T.W., Grotzinger, J.P., 2009. Superheavy pyrite (δ34Spyr > δ34SCAS) in the terminal Proterozoic Nama Group, southern Namibia: A consequence of low seawater sulfate at the dawn of animal life. Geology 37, 743–746.

Rodler, A.S., Frei, R., Gaucher, C., Germs, G.J.B., 2016. Chromium isotope, REE and redox‐sensitive trace element chemostratigraphy across the late Neoproterozoic Ghaub glaciation, Otavi Group, Namibia. Precambrian Research 286, 234–249.

Rodler, A., Frei, R., Gaucher, C., Korte, C., Rosing, S.A., Germs, G.J.B. 2017. Multiproxy isotope constraints on ocean compositional changes across the late Neoproterozoic Ghaub glaciation, Otavi Group, Namibia. Precambrian Research 298, 306–324.

Saltzman, M.R., 2005. Phosphorus, nitrogen, and the redox evolution of the Paleozoic oceans. Geology 33, 573–576.

Saltzman, M.R., Thomas, E., 2012. Carbon isotope stratig­raphy. In: Gradstein, F.M., Ogg, J.G., Schmitz, M., Ogg, G.

0004203144.INDD 22 11/9/2018 4:52:38 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 23

(Eds.), The Geologic Time Scale. Elsevier, Amsterdam, Heidelberg. doi:10.1016/B978‐0‐444‐59425‐9.00011‐1.

Salvador, A., 1994. International Stratigraphic Guide: A Guide to Stratigraphic Classification, Terminology and Procedure, 2nd Edition. IUGS‐GSA, Boulder, pp. 1–214.

Sanei, H., Grasby, S.E., Beauchamp, B., 2012. Latest Permian mercury anomalies. Geology 40, 63–66.

Sansjofre, P., Trindade R.I.F., Ader M., Soares, J.L., Nogueira A.C.R., Tribovillard, N., 2014. Paleoenvironment recon­struction of the Ediacaran Araras platform (Western Brazil) from the sedimentary and trace metals records. Precambrian Research 241, 185–202.

Santos, G.M., Cordeiro, R.C., Silva Filho, E.V., Turcq, B., Lacerda, L.D., Fifield, L.K., Gomes, P.R.S., Hauscaden, P.A., Sifeddine, A., Albuquerque, A.L.S., 2001. Chronology of the atmospheric mercury in Lagoa da Pata Basin, Upper Rio Negro of Brazilian Amazon. Radiocarbon 43, 801–808.

Sanyal, A., Bijma, J., Spero, H., David, H., Lea, W., 2001. Empirical relationship between pH and the boron isotopic composition of Globigerinoides sacculifer implications for the boron isotope paleo‐pH proxy. Paleoceanography 16, 515–519.

Sauzeat, L., Rudnick, R.L., Chauvel, C., Garcon, M., Tang, M., 2015. New perspectives on the Li isotopic composition of the upper continental crust and its weathering signature. Earth and Planetary Science Letters 428, 181–192.

Sawaki, Y., Ohno, T., Tahata, M., Komiya, T., Hirata, T., Maruyama, S., Windley, B.F., Han, J., Shud, D., Li, Y., 2010. The Ediacaran radiogenic Sr isotope excursion in the Doushantuo Formation in the Three Gorges area, South China. Precambrian Research 176, 46–64.

Scaife, J.D., Ruhl, M., Dickson, A.J., Mather, T.A., Jenkyns, H.C., Percival, L.M.E., Hesselbo, S.P., Cartwright, J., Eldrett, J.S., Bergman, S.C., Minisini, D., 2017. Sedimentary mercury enrichments as a marker for submarine Large Igneous Province volcanism? Evidence from the Mid‐Cenomanian Event and Oceanic Anoxic Event 2 (Late Cretaceous). Geochemistry, Geophysics, Geosystems 18(12), 4253–4275. doi:10.1002/2017GC007153.

Scheffler, K., Hoernes, S., Schwark, L., 2003. Global changes during Carboniferous–Permian glaciation of Gondwana: Linking polar and equatorial climate evolution by geochemical proxies. Geology 31, 505–608.

Schidlowski, M., Hayes, J.M., Kaplan, I.R., 1983. Isotopic infer­ences of ancient biochemistries: Carbon, sulfur, hydrogen and nitrogen. In: Schopf, J.W. (Ed.), Earth’s Earliest Biosphere: Its Origin and Evolution. Princeton University Press, Princeton, NJ, pp. 149–186.

Schmitz, B., Andersson, P., Dahl, J., 1988. Iridium, sulfur isotopes and rare earth elements in the Cretaceous‐Tertiary boundary clay at Stevns Klint, Denmark. Geochimica et Cosmochimica Acta 52(1), 229–236.

Schmitz, B., Pujalte, V., Molina, E., Monechi, S., Orue‐Etxebarria, X., Speijer, R.P., Alegret, L., Apellaniz, E., Arenillas, I., Aubry, M.‐P., Baceta, J.‐I., Berggren, W.A., Bernaola, G., Caballero, F., Clemmensen, A., Dinarès‐Turell, J., Dupuis, C., Heilmann‐Clausen, C., Orús, A.H., Knox, R., Martín‐Rubio, M., Ortiz, S., Payros, A., Petrizzo, M.R., von Salis, K., Sprong, J., Steurbaut, E., Thomsen, E., 2011. The

global stratotype sections and points for the bases of the Selandian (Middle Paleocene) and Thanetian (Upper Paleocene Paleocene) stages at Zumaia, Spain. Episodes 34, 220–243.

Scholle, P.A., Arthur, M.A. 1980. Carbon isotope fluctuations in Cretaceous pelagic limestones: Potential stratigraphic and petroleum exploration tool. American Association of Petroleum Geologists, Bulletin 64, 67–87.

Scott, C., Lyons, T.W., Bekker, A., Shen, Y., Pultron, S.W., Chu, X., Anbar, A.D., 2008. Tracing the stepwise oxygenation of the Proterozoic ocean. Nature 452, 456–459.

Scott, C.T., Bekker, A., Reinhard, C.T., Schnetger, B., Krapež, B., Rumble, D., Lyons, T.W., 2011. Late Archean euxinic condi­tions before the rise of atmospheric oxygen. Geology 39(2), 119–122.

Shackleton, N.J., 1969. The last interglacial in the marine and terrestrial record. Proceedings of the Royal Society of London 174, 135–154.

Shackleton, N.J., Opdyke, N.D., 1973. Oxygen isotope and pal­aeomagnetic stratigraphy of Equatorial Pacific core V28‐238: Oxygen isotope temperatures and ice volumes on a 105 year and 106 year scale. Quaternary research 3(1), 39–55.

Shackleton, N.J., Hall, M.A., 1984. Carbon isotope data from Leg 74 sediments. Initial Reports of the Deep Sea Drilling Project 74, 613–619.

Sharma, M., Papanastassiou, D.A., Wasseburg, J., 1997. The concentration and isotopic composition of osmium in the oceans. Geochimica et Cosmochimica Acta 61, 3287–3299.

Shen, Y., Knoll, A.H., Walter, M.R., 2003. Evidence for low sulphate and anoxia in a mid‐Proterozoic marine basin. Nature 423, 632.

Shields, G.A., 2007a. A normalised seawater strontium isotope curve and the Neoproterozoic‐Cambrian chemical weathering event. Earth Discussions 2, 69–84.

Shields, G., 2007b. The marine carbonate and chert isotope records and their implications for tectonics, life and climate on the early Earth. In: van Kranendonk, M.J., Smithies, R.H., Bennett, V.C. (Eds.), Earth’s Oldest Rocks. Develop­ments in Precambrian Geology 15. Elsevier, Amsterdam, pp. 971–983.

Sial, A.N., Karhu, J., Ferreira, V.P., 2010a. Insights from iso­tope stratigraphy. Preface to the special issue on precambrian isotope stratigraphy. Precambrian Research 182(4), v–viii.

Sial, A.N., Gaucher, C., Silva Filho, M.A., Ferreira, V.P., Pimentel, M.M., Lacerda, L.D., Silva Filho, E.V., Cezario, W., 2010b. C‐, Sr‐isotope and Hg chemostratigraphy of Neoproterozoic cap carbonates of the Sergipano Belt, Northeastern Brazil. In: Karhu, J., Sial, A.N., Ferreira, V.P. (Eds.), Precambrian Isotope Stratigraphy. Precambrian Research 182. Elsevier, Amsterdam, pp. 351–372.

Sial, A.N., Lacerda, L.D., Ferreira, V.P., Frei, R., Marquillas, R.A., Barbosa, J.A., Gaucher, C., Windmöller, C.C., Pereira, N.S., 2013a. Mercury as a proxy for volcanic activity during extreme environmental turnover: The Cretaceous–Paleogene transition. Palaeogeography, Palaeoclimatology, Palaeoecology 387, 153–164.

Sial, A.N., Chen, J.‐B., Lacerda, L.D., Peralta, S., Gaucher, C., Frei, R., Cirilli, S., Ferreira, V.P., Marquillas, R.A., Barbosa, J.A., Pereira, N.S., Belmino, I.K.C., 2014. High‐resolution

0004203144.INDD 23 11/9/2018 4:52:38 PM

24 CHEMOSTRATIGRAPHY ACROSS MAJOR CHRONOLOGICAL BOUNDARIES

Hg chemostratigraphy: A contribution to the distinction of chemical fingerprints of the Deccan volcanism and Cretaceous−Paleogene Boundary impact event. Palaeogeography, Palaeoclimatology, Palaeoecology 414, 98–115. doi:10.1016/j.palaeo.2014.08.013.

Sial, A.N., Gaucher, G., Ferreira, V.P., Pereira, N.S., Cezario, W.S., Chiglino, L., Monteiro, H., 2015a. Isotope and ele­mental chemostratigraphy. In: M. Ramkumar (Ed.), Chemostratigraphy, Concepts, Techniques and Applications. Elsevier, Amsterdam, pp. 23–64.

Sial, A.N., Campos, M.S., Gaucher, C., Frei, R., Ferreira, V.P., Nascimento, R.C., Pimentel, M.M., Pereira, N.S., Rodler, A., 2015b. Algoma‐type Neoproterozoic BIFs and related mar­bles in the Seridó Belt (NE Brazil): REE, C, O, Cr and Sr isotope evidence. Journal of South American Earth Sciences 61, 33–52.

Sial, A.N., Chen, J., Lacerda, L.D., Frei, R., Tewari, V.C., Pandit, M.K., Gaucher, C., Ferreira, V.P., Cirilli, S., Peralta, S., Korte, C., Barbosa, J.A., Pereira, N.S., 2016. Mercury enrichment and mercury isotopes in Cretaceous–Paleogene boundary successions: Links to volcanism and palaeoen­vironmental impacts. Cretaceous Research 66, 60–81. doi:10.1016/j.cretres.2016.05.006.

Sial, A.N., Chen, J., Lacerda, L.D., Frei, R., Tewari, V.C., Pandit, M.K., Gaucher, C., Ferreira, V.P., Cirilli, S., Peralta, S., Korte, C., Barbosa, J.A., Pereira, N.S., 2017. Reply to comments by Sanjay K. Mukhopadhyay, Sucharita Pal, J. P. Shrivastava on the paper by Sial et al. (2016) Mercury enrichments and Hg isotopes in Cretaceous–Paleogene boundary successions: Links to volcanism and palaeoenviron­mental impacts. Cretaceous Research 66, 60–81. Cretaceous Research 78, 84–88.

Sial, A.N., Chen, J., Lacerda, L.D., Frei, R., Higgins, J., Tewari, V.C., Gaucher, C., Ferreira, V.P., Cirilli, S., Peralta, S., Korte, C., Barbosa, J.A., Pereira, N.S., Ramos, D.S., this volume. Chemostratigraphy across the Cretaceous–Paleogene boundary. In: Sial, A.N., Gaucher, C., Ramkumar, M., Ferreira, V.P. (guest editors), Chemostratigraphy Across Major Chronological Eras. AGU/Wiley, United States.

Siebert, C., Nägler, T.F., von Blanckenburg, F., Kramers, J.D., 2003. Molybdenum isotope records as a potential new proxy for paleoceanography. Earth Planetary Science Letters 211, 159–171.

Silva‐Tamayo, J.C., Nägler, T.F., Villa, I.M., Kyser, K., Narbonne, G., James, N.P., Sial, A.N., Silva Filho, M.A., 2007. The aftermath of Snowball Earth: Ca‐ and Mo‐ isotope constraintson post‐glacial ocean conditions. Geophysical Research Abstracts, 9, 01980.

Silva Tamayo, J.C., Nägler, T., Villa, I.M., Kyser, K., Vieira, L.C., Sial, A.N., Narbonne, G.M., James, N.P., 2010a. Global Ca isotope variations in Post‐Sturtian carbonate successions. Terra Nova 22, 188–194.

Silva Tamayo, J.C., Nägler, T.F., Nogueira, A., Kyser, K., Villa, I., Riccomini, C., Sial, A.N., Narbonne, G.M., James, N.P., 2010b. Global perturbation of the marine Ca‐isotopic com­position in the aftermath of the Marinoan global glaciations. Precambrian Research 182(4), 373–381.

Sluijs, A., Schouten, S., Donders, T.H., Schoon, P.L., Rohl, U., Reichart, G.‐J., Sangiorgi, F., Kim, J.‐H., Sinninghe Damste,

J.S., Brinkhuis, H., 2009. Warm and wet conditions in the Arctic region during Eocene Thermal Maximum 2. Nature Geosciences 2, 777–780.

Sosa‐Montes, C., Rodríguez‐Tovar, F.J., Martínez‐Ruiz, F., Monaco, C.P., 2017. Paleoenvironmental conditions across the Cretaceous–Paleogene transition at the Apennines sec­tions (Italy): An integrated geochemical and ichnological approach. Cretaceous Research 71, 1–13. doi:10.1016/j.cretres.2016.11.005.

Spangenberg, J.E., Bagnoud‐Velásquez, M., Boggiani, P.C., Gaucher, C., 2014. Redox variations and bioproductivity in the Ediacaran: Evidence from inorganic and organic geochemistry of the Corumbá Group, Brazil. Gondwana Research 26, 1186–1207.

Staubwasser, M., Von Blackenburg, F., Shoenberg, R., 2006. Iron isotopes in the early marine diagenetic iron cycle. Geology 34, 629–632.

Steinhoefel, G., von Blanckenburg, F., Horn, I., Konhauser, K.O., Beukes, N.J., Gutzmer, J., 2010. Deciphering formation processes of banded iron formations from the Transvaal and the Hamersley successions by combined Si and Fe isotope analysis using UV femtosecond laser ablation. Geochimica et Cosmochimica Acta 74, 2677–2696.

Steininger, F.F., Aubry, M.P., Berggren, W.A., Biolzi, M., Borsetti, A.M., Cartlidge, J.E., Cati, F., Corfield, R., Gelati, R., Iaccarino, S., Napoleone, C., Ottner, F., Rögl, F., Roetzel, R., Spezzaferri, S., Tateo, F., Villa, G., Zevenboom, D., 1997. The Global Stratotype Section and Point (GSSP) for the base of the Neogene. Episodes 20, 23–28.

Strandmann, P.A.E., Forshaw, J., Schmidt, D.N., 2014. Modern and Cenozoic records of seawater magnesium from forami­niferal Mg isotopes. Biogeosciences 11, 5155–5168.

Strauss, H., 1997. The isotopic composition of sedimentary sulfur through time. Paleogeography, Palaeoclimatology, Palaeoecology 132, 97–118.

Sundquist, E.T., Visser, K., 2004. The geologic history of the carbon cycle. Treatise on Geochemistry 8, 425–472.

Tahata, M., Ueno, Y., Ishikawa, T., Sawaki, Y., Murakami, K., Han, J., Shu, D., Li, Y., Guo, J., Yoshida, N., Komiya, T., 2013. Carbon and oxygen isotope chemostratigraphies of the Yangtze platform, South China: Decoding temperature and environmental changes through the Ediacaran. Gondwana Research 23(1), 333–356. doi:10.1016/j.gr.2012.04.005.

Teng, F.‐Z., Watkins, J.M., Dauphas, N., 2017. Non‐traditional stable isotopes. Reviews in Mineralogy and Geochemistry 82, 885 pp.

Thibodeau, A.M., Bergquist, B.A., 2017. Do mercury isotopes record the signature of massive volcanism in marine sedimen­tary records? Geology 45, 95–96.

Thibodeau, A.M., Ritterbush, K., Yager, J.A., West, A.J., Ibarra, Y., Bottjer, D.J., Berelson, W.M., Bergquist, B.A., Corsetti, F.A., 2016. Mercury anomalies and the timing of biotic recovery following the end‐Triassic mass extinction. Nature Communications 7, 1–8.

Tipper, E.T., Bickle, M.J., Galy, A., West, J., Pomiés, C., Chapman, H.J., 2006a. The short term sensitivity of carbon­ates and silicate weathering fluxes: Insight from seasonal variations in river chemistry. Geochimica et Cosmochimica Acta 70, 2737–2754.

0004203144.INDD 24 11/9/2018 4:52:38 PM

CHEMOSTRATIGRAPHY AS A FORMAL STRATIGRAPHIC METHOD 25

Tipper, E.T., Galy, A., Bickle, M.J., 2006b. Riverine evidence for a fractionated reservoir of Ca and Mg on the continents: Implications for the oceanic Ca cycle. Earth Planetary Science Letters 247, 267–279.

Tipper, E.T., Galy, A., Gaillardet, J., Bickle, M.J., Elderfield, H., Carder, E.A., 2006c. The magnesium isotope budget of the modern ocean: Constraints from riverine magnesium isotope ratios. Earth Planetary Science Letters 250, 241–253.

Tipper, E.T., Louvat, P., Capmas, F., Galy, A., Gaillardet, J., 2008. Accuracy of stable Mg and Ca isotope data obtained by MC‐ICP‐MS using the standard addition method. Chemical Geology 257, 65–75.

Tomascak, P.B., Magna, T.S., Dohmen, R., 2016. Advances in Lithium Isotope Geochemistry. Springer International Publishing, Cham, 195 pp.

Tribovillard, N., Algeo, T.J., Lyons, T., Riboulleau, A., 2006. Trace metals as paleoredox and paleoproductivity proxies: An update. Chemical Geology 232, 12–32.

Tribovillard, N., Algeo, T.J., Baudin, F., Riboulleau, A., 2012. Analysis of marine environmental conditions based on molybdenum‐uranium covariation‐applications to Mesozoic paleoceanography. Chemical Geology 324/325, 46–58.

Tripati, A., Eagle, R., Thiagarajan, N., Gagnon, A., Bauch, H., Halloran, P., Eiler, J., 2010. 13C–18O isotope signatures and “clumped isotope” thermometry in foraminifera and cocco­liths. Geochimica et Cosmochimica Acta 74, 5697–5717.

Vandenbroucke, T.R.A., Armstrong, H.A., Williams, M., Paris, F., Zalasiewicz, J.A., Sabbe, K., Nolvak, J., Challandsa, T.J., Verniers, J., Servais, T., 2010. Polar front shift and atmo­spheric CO2 during the glacial maximum of the Early Paleozoic Icehouse. Proceedings of the National Academy of Sciences of the United States of America 107, 14983–14986.

Van de Schootbrugge, B., Bachan, A., Suan, G., Richoz, S., Payne, J.L., 2013. Microbes, mud and methane: Cause and consequence of recurrent Early Jurassic anoxia following the end‐Triassic mass extinction. Paleontology 56, 1–25. doi:10.1111/pala.120341.

Veizer, J., Holser, W.T., Wilgus, C.K., 1980. Correlation of 13C/12C and 34S/32S secular variations. Geochimica Cosmochimica Acta 44, 579–588.

Veizer, J., Compston, W., Clauer, N., Schidlowski, M., 1983. 87Sr/86Sr in late Proterozoic carbonates: Evidence for a “mantle” event at approximately 900 Ma ago. Geochimica et Cosmochimica Acta 47, 295–302.

Veizer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, P., Bruhn, F., Carden, G.A.F., Diener, A., Ebneth, S., Godderis, Y., Jasper, T., Korte, C., Pawellek, F., Podlaha, O.G., Strauss, H., 1999. 87Sr/86Sr, δ13C and δ18O evolution of Phanerozoic seawater. Chemical Geology 161, 59–88.

Vennemann, T.W., Hegner, E., 1998. Oxygen, strontium, and neodymium isotope composition of fossil shark teeth as a proxy for the palaeoceanography and palaeoclimatology of the Miocene northern Alpine Paratethys. Palaeogeography, Palaeoclimatology, Palaeoecology 142, 107–121.

Walker, M., Johnsen, S., Rasmussen, S.O., Popp, T., Steffensen, J.‐P., Gibbard, P., Hoek, W., Lowe, J., Andrews, J., Rck, S.B., Cwynar, L.C., Hughen, K., Kershaw, P., Kromer, B., Litt, T., Lowe, D.J., Nakagawa, T., Newnham, R., Schwander, J., 2009. Formal definition and dating of the GSSP (Global Stratotype Section and Point) for the base of the Holocene using the Greenland NGRIP ice core, and selected auxiliary records. Journal of Quaternary Science 24, 3–17.

Wen, H., Fan, F., Zhang, Y., Cloquet, C.,Carignan, J., 2015. Reconstruction of early Cambrian ocean chemistry from Mo isotopes. Geochimica et Cosmochimica Acta 164, 1–16.

Williams, D.F., Thunell, R.C., Tappa, E., Rio, D., Raffi, I. 1988. Chronology of the Pleistocene oxygen isotope record: 0–1.88 my BP. Palaeogeography, Palaeoclimatology, Palaeoecology 64(3–4), 221–240.

Wombacher, F., Eisenhauer, A., Heuser, A., Weyer, S., 2009. Separation of Mg, Ca and Fe from geological reference materials for stable isotope ratio analyses by MC‐ICP‐MS and double‐spike TIMS. Journal Analytical Atomic Spectrometry 24, 627–636.

Wombacher, F., Eisenhauer, A., Böhm, F., Gussone, N., Regenberg, M., Dullo, W.‐C., Rüggeberg, A., 2011. Magnesium stable isotope fractionation in marine biogenic calcite and aragonite. Geochimica Cosmochica Acta 75, 5797–5818.

Xu, C., Rong, J., Fan, J., Zhan, R., Mitchell, C.E., Harper, D.A.T., Melchin, M.J., Peng, P., Finney, S.C., Wang, X., 2006. The Global Boundary Stratotype Section and Point (GSSP) for the base of the Hirnantian Stage (the uppermost of the Ordovician System). Episodes 2, 183–196.

Yu, J., Elderfield, H., 2007. Benthic foraminiferal B/Ca ratios reflect deep water carbonate saturation state. Earth and Planetary Science Letters 258, 73–86.

Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–693.

Zachos, J.C., Schouten, S., Bohaty, S., Quattlebaum, T., Sluijs, A., Brinkhuis, H., Gibbs, S.J., Bralower, T.J., 2006. Extreme warming of mid‐latitude coastal ocean during the Paleocene‐Eocene Thermal Maximum: Inferences from TEX86 and isotope data. Geology 34, 737–740.

0004203144.INDD 25 11/9/2018 4:52:38 PM