Vertical ground displacement at Campi Flegrei(Italy) in the fifth century: Rapid subsidencedriven by pore pressure dropMicol Todesco1, Antonio Costa1, Alberto Comastri1, Florence Colleoni2, Giorgio Spada3,and Francesca Quareni1

1Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Bologna, Italy, 2Centro Euro-Mediterraneo suiCambiamenti Climatici, Bologna, Italy, 3Dipartimento di Scienze di Base e Fondamenti, Università di Urbino “Carlo Bo”,Urbino, Italy

Abstract Campi Flegrei (Italy) caldera has experienced episodes of ground deformation throughout itsgeological history, alternating between uplift and subsidence phases. Although uplift periods are typicallymore alarming, here we focus on subsidence, looking for its driving mechanisms and its role in the calderaevolution. Historical and archaeological records constrain ground deformation over the last two millennia.Here we revise such records and combine themwith published radiometric dating and with the simulation ofsea level change. The resulting analysis highlights for the first time a rapid subsidence during the fifthcentury. We show that rate and magnitude of this subsidence are consistent with the compaction of porousmaterial caused by a pressure drop of ~ 1MPa within the hydrothermal system. We interpret this event as thedecompression of the hydrothermal system following an unrecognized episode of unrest, during Romantimes. These findings redefine the pattern of ground deformation and bear important implications forvolcanic hazard assessment.

1. Introduction

Calderas are volcanic structures that undergo significant ground deformation, with uplift and subsidence thatin case precede eruptive phases. Uplift is usually ascribed to magma ascent or to perturbation of thehydrothermal system [Casertano et al., 1976; Gaeta et al., 1998; Hurwitz et al., 2007; Fournier and Chardot,2012]. Subsidence can be caused by various mechanisms characterized by different timescales, such aseruptions or magma migration [Dzurisin et al., 1994; Kaneko et al., 2005], regional extension [Dzurisin et al.,1994], elastic rebound, compaction of caldera filling deposits [Yokoyama and Nazzaro, 2002], magma coolingand contraction [Kwoun et al., 2006], or the poroelastic deformation associated with hydrothermal fluidcirculation [Castagnolo et al., 2001; Waite and Smith, 2002; Todesco et al., 2004; Hurwitz et al., 2007; Rinaldiet al., 2010]. Campi Flegrei (Figure 1) is an ideal location to study these phenomena, as the presence of Romanruins along the shore provides an unusual benchmark, which recorded the ground displacement with respectto sea level during the last two millennia. The marble pillars of the ancient Roman market in Pozzuoli, knownas Serapeum, have fascinated generations of scientists, from Lyell [1830] on, who contributed to unfold thehistory of the caldera. Reconstruction of ground deformation is important to understand the mechanismsdriving the unrest episodes of this active volcanic system, which recently experienced periods of upliftaccompanied by seismicity. Here we consider all available geological, geochronological, and archaeologicaldata to draw a unique and comprehensive picture of vertical ground displacement at Campi Flegrei. A briefreconstruction of the history of Serapeum is provided in Appendix S1 in the supporting information, togetherwith the information we used to reconstruct its vertical movements (Table S1). Our goal is to assess thetimescale associated with ground deformation and to provide temporal constraints on the verticaldisplacement of the Serapeum pillars. The uncertainties associated with our estimate are discussed in thesupporting information (Appendix S2). The history of vertical ground displacement reveals that a rapidsubsidence (of the order of hundreds of mm/yr) took place during the fifth century A.D. We show thatsubsidence is likely due to compaction of shallow, porous rocks, driven by abundant discharge of volcanicfluids. We speculate that such deflation phase may have been preceded by a volcanic unrest that pressurizedthe hydrothermal system.

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 1

PUBLICATIONSGeophysical Research Letters

RESEARCH LETTER10.1002/2013GL059083

Key Points:• An integrated approach constrainsfast subsidence at Campi Flegrei inthe fifth century

• Fast subsidence can be caused bycompaction due to decompression ofhydrothermal system

• A sequence of unrest events precededlast eruption with no secular subsidence

Supporting Information:• ReadMe• Figure S1• Table S1• Appendix S1• Appendix S2

Correspondence to:M. Todesco,micol.todesco@bo.ingv.it

Citation:Todesco, M., A. Costa, A. Comastri,F. Colleoni, G. Spada, and F. Quareni(2014), Vertical ground displacement atCampi Flegrei (Italy) in the fifth century:Rapid subsidence driven by pore pressuredrop, Geophys. Res. Lett., 41, doi:10.1002/2013GL059083.

Received 18 DEC 2013Accepted 6 FEB 2014Accepted article online 7 FEB 2014

2. Unrest at Campi Flegrei

The Campi Flegrei caldera was formed by two major events: the Campanian Ignimbrite eruption, 39 ka ago[Fedele et al., 2011; Costa et al., 2012], and the Neapolitan Yellow Tuff eruption, 15 ka ago [Deino et al., 2004].Later activity mainly consisted of smaller, intracaldera explosive events (predominantly phreatomagmatic)that were clustered in short periods separated by quiet intervals [Di Vito et al., 1999; Fedele et al., 2011]. Thepresence of raised marine terraces and sediments alternating with beach deposits and subaerial volcanicproducts indicate repeated and widespread episodes of uplift and subsidence [Di Vito et al., 1999; Bellucciet al., 2006; Isaia et al., 2009]. Roman ruins along the coast contain tracks of marine organisms, providingevidence for periods of partial submersion. The Serapeum pillars, in particular, hold biological borings causedby the activity of marine organism (Lithophaga lithophaga, among others) [Morhange et al., 1999, 2006]. Theupper limit of these perforations (presently 7m above sea level) indicates the maximum submersion andcorresponds to the biological sea level during the lifetime of the marine colony. Radiometric dating of thefossil remains collected at this elevation identified three separate age groups: A.D. 400–530, A.D. 700–900,and finally at the end of fifteenth century [Morhange et al., 1999, 2006]. These ages record the death of themarine fauna, which in this case relates to the emersion of the colony (i.e., ground uplift), since no evidencesupports other causes of death (such as changes in water salinity). The onset of Lithophaga on a newsubstratum requires at least a few years, and the shells grow to their maximum size in about 70–80 years[Morhange et al., 1999]. Given the size of the colony, the pillars must have reached their maximumsubmersion several decades before the onset of each uplift phase. Last emersion culminated in the A.D. 1538Monte Nuovo eruption, ending more than 3000 years of eruptive quiescence. The appearance of new

ITALY

Campi Flegrei

MonteNuovo

Pozzuoli

Solfatara

Napoli

Miseno

Serapeum

(a) (b) Monte Nuovo

Pozzuoli

Baia

Bacoli

Miseno

Serapeum

*

Figure 1. (a) The Campi Flegrei caldera, with the position of Serapeum, the Solfatara crater, and Monte Nuovo. (b) Submerged Roman ruins in the Pozzuoli Bay. Contourlines every 5m (modified after Dvorak and Mastrolorenzo [1991]).

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 2

stretches of land preceding this eruption is documented in historical chronicles that confirm a large,widespread ground deformation [Parascandola, 1947; Di Vito et al., 1987; Guidoboni and Ciuccarelli, 2011].

Geodetic measurements of ground elevation are available since 1905 and indicate episodes of ground upliftin 1953, 1969, and 1982, totalling approximately 4m of vertical displacement [Del Gaudio et al., 2010]. Thepattern of uplift and subsidence is symmetrical [Dvorak and Berrino, 1991], with maximum displacement atthe caldera center. A subsidence phase between 1985 and 2005 was periodically interrupted by minor(few centimeters), short-lasting episodes of uplifts. Since 2006, the caldera has been inflating again, at aslow but steady rate [D’Auria et al., 2011].

2.1. Other Evidence of Vertical Ground Displacement

Constrain on past vertical ground displacement derived from the fossil fauna found in the Averno lake(Figure 1b). The lake was connected to the sea by a canal, built in 37 B.C., and abandoned 10–20 years later.The presence of marine species within the fossil record of the lake testifies to increasingly frequent marinetransgression between the years 50 and 400A.D. [Welter-Schultes and Richling, 2000]. In particular, asubstantial increase in the proportion of marine species was observed after A.D. 350 and was interpreted interms of ground subsidence. The marine contribution declined between A.D. 500 and 700.

Further information comes from the other archaeological remains on the western side of the bay. The town ofBaia became an appreciated location for Roman notables since the beginning of the first century B.C. Evidenceof restoration andwritten accounts by Symmachus (A.D. 340–402) suggests that the area was still above the sealevel at the beginning of the fourth century. This is supported by similar evidence found at other locations likePunta Epitaffio, and Misenum, only abandoned at the end of the fourth century [Maniscalco, 1998]. Here tracksofmarine organisms on ancient buildings, as well as the presence ofmarine sediments within the ruins, suggestthat subsidence of the coastline took place between the fourth and fifth century. The presence of graves datingto the Late Antiquity (250–750A.D.) within the submerged Roman ruins in Baia suggests that these were againabove sea level in the sixth century. In Misenum, the town abandonment was followed by new constructions inthe sixth to seventh century, which were later buried under beach and marine deposits, suggesting a newsinking phase between seventh and eighth century [Maniscalco, 1998].

3. Reconstruction of Late Antique Subsidence

Our reconstruction of vertical ground displacement at Campi Flegrei is based on the assessment of the positionof Serapeum pillars with respect to sea level. A brief summary of the available archeological evidence on thebuilding of Serapeum is provided in the supporting information (Appendix S1). To estimate its elevationthrough time, we combined this information with an estimate of the sea level variations along this stretch ofcoast. In the Mediterranean region abundant archaeological markers allow a detailed reconstruction of ancientcoastlines [Auriemma and Solinas, 2009]. In the NWMediterranean Sea, the sea level rise since Roman times wasevaluated to be ~50 cm [Pirazzoli, 1976]. Estimates carried out for central Tyrrhenian Sea range from 1 to 1.4m[Lambeck et al., 2004]. This corresponds to rates of relative sea level rise ranging from 0.25 to 0.7mm/yr,consistent with predicted, present-day isostatic rates [Lambeck et al., 2011]. These sea level changes primarilyresult from glacial isostatic adjustment (GIA), due to the melting of the polar ice sheets since the Last GlacialMaximum ~21,000 years ago. We address the process with a model that solves the sea level equation [Farrelland Clark, 1976; Spada and Stocchi, 2007]: two numerical simulations were carried out considering different icemelting scenarios and different viscosity profiles for the Earth’s mantle. The first is ICE-5G (VM2) model (heremodel P) [Peltier 2004], while the second was developed at the Research School of Earth Science of theAustralian National University by K. Lambeck and coworkers [Lambeck et al., 1998, and references therein](model L). The obtained sea level changes (Figure 2) bound the upper and lower values proposed in theliterature for the last 2000 years. Model L implies larger isostatic disequilibrium hence greater relative sea levelvariations. Our results are largely unaffected by the choice of the GIA model: in the following, we consider themodel P, but the subsidence rate computed with model L does not differ significantly. Both curves are shownin Figure 2 for completeness, but the rate for model L is not used.

In reconstructing the vertical ground displacement we take as a reference level the upper limit of perforationon Serapeum pillars (Figure S1). At the time of the market construction this level was ~13m above present-day sea level (Table S1). The construction of a second floor suggests that subsidence had already begun at the

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 3

time of the Severian restorations(during the third century, Appendix S1).We assume that the 2m difference infloor elevations reflects the magnitudeof the actual ground displacement. Thecolumns were still out of the water atthe time of last restoration (A.D. 394).Assuming that the second floor was atthe sea level of the time, the completecolumn submersion required at least7m of subsidence (Appendix S2).

To infer the timing of such subsidence,we rely on the radiometric dating offossil remains from Serapeum pillars(Figure 2, bottom). The first group ofradiometric dates ranges in time fromA.D. 378 to 540, with large errors,spanning from ±114 to±138 [Morhangeet al., 1999, 2006]. Their average value,and its associated corrected standarddeviation is A.D. 450±70. Ages olderthan A.D. 394 must be discarded, as weknow that the pillars were above thewater at the time of last restoration. Theaverage value (A.D. 450) is also

questionable, as it implies very fast rates of subsidence (see discussion in Appendix S2). For this reason the youngerage (540±125) obtained by Morhange et al. [1999] is considered here as a reference for the estimation ofsubsidence rate. This age is also in better agreement with archaeological evidence suggesting that uplift didnot begin before the sixth century. Taking this age as the beginning of the uplift and considering a time lapseof a century for the growth of the Lithophaga colony, the complete columns submersion occurs in A.D. 440,i.e., in 46 years, corresponding to an average subsidence rate of the order of 160mm/yr.

The estimate above is affected by large uncertainties, being based on indirect evidence and a fewassumptions. An assessment of the resulting error is carried out in Appendix S2, where each source ofuncertainty is carefully examined to constrain its reasonable bounds. The resulting admissible range ofsubsidence rates varies between 50 and 500mm/yr, with our best guess being of the order of 200mm/yr.Further radiometric ages established for fossil remains on the pillars indicate more periods of subsidence anduplift [Morhange et al., 1999, 2006], culminating in the A.D. 1538 Monte Nuovo eruption. Unfortunately, pillarelevation during the Middle Ages is not constrained by historical or geological data, and rates of grounddeformation cannot be estimated for that period.

4. The Mechanism of Subsidence

Ground deformation in active calderas is commonly interpreted in terms of pressure or volume changes of asource of deformation at depth, typically a magma chamber [Waite and Smith, 2002; Farrell et al., 2010; Shellyet al., 2013], a hydrothermal system, or both [Casertano et al., 1976; Gaeta et al., 1998; Amoruso et al., 2014].Volume or pressure changes at the source act on the surrounding rocks causing a deformation, whosemagnitude and temporal evolution depend on the rheological properties of the subsurface rock. In thisframework, subsidence is caused by a pressure drop or a volume contraction of the source, possibly due tomagma cooling and crystallization or to the migration of magma or hydrothermal fluids. The timescale ofdeformation and the magnitude of the resulting surface displacement may help in discriminating amongpossible driving mechanisms. Typical subsidence rates of calderas range from a few to a few tens ofmillimeters per year [Newhall and Dzurisin, 1988; Hurwitz et al., 2007]. Vertical ground displacement at CampiFlegrei was first described as a long-term, secular subsidence with an estimated, constant rate spanning from

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Year

0 200 400 600 800 1000 1200 1400 1600 1800 2000−8

−6

−4

−2

0

2

4

6

8

10

12

14

Ele

vatio

n ab

ove

pres

ent s

ea le

vel (

m)

Arecheological data (Parascandola, 1947)Ground level (Del Gaudio et al., 2010)Time of lithodom emersionSea level (Model P)Sea level (Model L)

Figure 2. Reconstructed vertical motion of the Roman pillars of Serapeum.Green dots: elevation of the reference line (i.e., the upper limit of the perfo-rated sector, Figure S1), above present sea level. The elevations are based onsea level change corresponding to model P (blue, solid line) although bothsea level lines are shown for completeness. Black dots: ground elevationmeasurements [after Del Gaudio et al., 2010]. Red arrows: times of pillaremersion, according to (bottom) radiometric dating of fossil marine organ-isms. Shaded areas: time intervals associatedwith subsidence (blue) or uplift(yellow) based on other evidence (see text for discussion).

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 4

15 to 20mm/yr and periodically interrupted by uplift or eruption [Dvorak and Mastrolorenzo, 1991; Bellucci et al.,2006]. Our reconstruction does not reveal such a monotonic, long-term trend and rather highlights a sequenceof deformation episodes, with a rapid subsidence between A.D. 400 and 500, when meters of displacementtook place within decades. Hydrothermal fluid flow can cause pressure changes within the consideredtimescale and is in good agreement with the available data on the caldera. The onset of subsidence wasexplained assuming a decompression of the large hydrothermal system that feeds the Solfatara emissions[Bonafede, 1991; Gaeta et al., 1998, 2003; Gottsmann et al., 2006; D’Auria et al., 2011; Amoruso et al., 2014]. Thesubsidence rate reached a maximum value of ~150mm/yr between January 1985 and October 1986 [DelGaudio et al., 2010], which is the same order ofmagnitude of the subsidence ratewe have estimated for the fifthcentury. A decompression of the hydrothermal system can be accompanied by compaction: when the porespace is saturated with fluids, pore pressure mitigates the effects of lithostatic load. If the pore pressure drops,the confining pressure acting on rock grains correspondingly increases, eventually leading to compaction asshown in the theory of soil consolidation [Biot, 1941]. This mechanism is known to generate ground subsidenceduring reservoir exploitation in geothermal power plants [Zoback, 2008; Allis et al., 2009].

To evaluate whether a pore pressure drop can account for subsidence at Campi Flegrei, we performed somesimple calculations. We assume that the subsidence is entirely due to the loss of pore volume, without deformationof the solid grains, and we consider that for a given cross section, such volume loss only occurs along the verticaldirection. Under these assumptions, the observed subsidence can be expressed in terms of porosity loss as

Δh ¼ h0ϕ0 � ϕ1� ϕ

(1)

where h0 is the initial thickness of the compacting layer, ϕ0 is the initial porosity, and ϕ is the porosity aftercompaction. In the case of Campi Flegrei, the compacting layer may range from thousands of meters (thickness

of the caldera fill deposits) to afew hundred meters (the singletuff layer within the caldera), whiletypical porosity values rangebetween 0.3 and 0.5 [Vanorio et al.,2002]. For any combination ofthese values, the porosity lossnecessary to achieve 10m ofsubsidence is within a fewpercent. Rock compaction cantherefore account for themagnitude of the observeddeformation. The compaction ratedepends on the hydraulic andelastic properties of the porousmaterial and on the magnitude ofthe pore pressure change, asdescribed by the general theory ofconsolidation [Biot, 1941].

For the sake of simplicity andbecause limited information are

Δh (

m)

Figure 3. Observed ground subsidence (crosses) from January 1985 to June 1988,after the last large uplift episode (1982–1984) and before the onset of the so-called“mini” uplifts (in 1988). The different lines show the fits obtained with equation (4)for different choices of the initial layer thickness. The corresponding parametersare listed in Table 1.

Table 1. Results From Fitting Equation (4) With Campi Flegrei Subsidence Data for the Period 1985–1988a

Parameter (Dimension) h0 = 500m h0 = 600m h0 = 1000m

ϕ0 (�) 0.48 0.41 0.32Pc (MPa) 0.67 0.64 0.44A (MPa db) 488,785 488,785 488,785b (�) 0.82 0.82 0.82

aFor comparison, values proposed by Zoback [2008] for uncemented sedimentary rocks are for A=5410MPa db,b=0.164, and ϕ0 = 0.27.

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 5

available on the system, here weinfer the timescale for thecompaction based on an empiricalrelation established in reservoirgeomechanics [Zoback, 2008].Laboratory experiments showedthat uncemented rock samplessubjected to increments ofconfining pressure, after a firstinstantaneous deformation, keepstraining at constant pressure[Chang et al., 1997]. This creepingbehavior is related to the motion ofparticles adjusting to the newconfining pressure. Thecorresponding evolution of porositycan be described as [Zoback, 2008]

ϕ Pc; tð Þ ¼ ϕi �PcAtb (2)

where ϕi is the porosity after the instantaneous compaction, and the second term describes the time-dependent evolution, where Pc is the confining pressure (MPa) (i.e., the applied pressure change), t is time(days), A and b are empirical parameters which express the rheological properties of the compacting layer: A is ascalar associated with the magnitude of creep compaction and b is related to the apparent viscosity of thesystem (smaller b for greater viscosities). Note that relationship (2) is valid for small porosity variations, i.e.,ϕi�ϕ(t) =Δϕ<< 1. The instantaneous term can be expressed as ϕi=ϕ0(Pc/P*)

d, where P* is a referencepressure (i.e., P* = 1MPa) and d is an empirical parameter, related to the rock compliance (smaller d for stifferrocks) [Zoback, 2008]. Experiments carried out with uncemented sand provided small d values, implying anegligible instantaneous compaction, with ϕi close to ϕ0 [Zoback, 2008]. Geodetic measurements at CampiFlegrei after January 1985 confirm a time-dependent response to the depressurization of system, withoutinstantaneous compaction [Del Gaudio et al., 2010]. In the following, we assume that vertical displacement isentirely driven by creep compaction with a negligible instantaneous component, i.e.,

ϕi ¼ ϕ0 (3)

Combining equations (1), (2),and (3), we obtain a relation describing the temporal evolution of caldera subsidence:

Δh ¼ h0PcA�1tb

1� ϕ0 þ PcA�1tb(4)

The thickness of the compacting layer and its initial porosity can be constrained by geological information,and we can estimate the magnitude of pressure changes in our system. On the contrary, values for theempirical parameters A and b in equation (4) are not available for Campi Flegrei. To constrain their values, weused equation (4) to fit the subsidence data for the period 1985–1988 (Figure 3). In this time span, the datashow a clear trend, and the available measurements are just enough to estimate all the equation parameters (A,b, Pc, andϕ0) assuming different thicknesses of the compacting layer, h0. Keeping in mind the limitations due tothe paucity of data, we obtained a satisfactory fit with the empirical relation. The values of initial porosity andpressure inferred for different thicknesses of the compacting layer are in good agreement with available dataand suggest that the 1985–88 subsidence was driven by the compaction of a shallow and porous layer. Theretrieved values of the empirical parameters A and bwere then assigned to equation (4) to quantify the amountof subsidence due to different pore pressure drops. Figure 4 shows the temporal evolution of subsidencecalculated for different initial thickness of the compacting layer, (i.e., 500, 600, and 1000m) and displays theeffect of different values of initial porosity (from 0.3 to 0.5). Figure 4 also shows the magnitude and timingof the fifth century subsidence, with associated error bars. The large uncertainties that affect our estimate

Δh (

m)

φ = 0.5

φ = 0.3

Figure 4. Temporal evolution of ground subsidence resulting from the compac-tion of a 600m thick layer, for different values of confining pressure (dashed lines)and initial porosities (dotted lines). Where not otherwise specified, the initialporosity of 0.4 and the confining pressure is 2MPa. The star shows the durationand magnitude of the fifth century subsidence with the related uncertainties.The choice of best guess and errors are fully discussed in Appendix S2.

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 6

are discussed in greater detail in Appendix S2, together with our choice of the best guess. A comparisonbetween such a best guess and the curves plotted in Figure 4 suggest that a pressure drop in the rangeof 1–2MPa could have caused a complete column submersion. These values are consistent with the estimatedpressure decline that caused more than 10m of subsidence in Wairakei, New Zealand [Allis et al., 2009].

Numerical simulations show that pore pressure changes of this order of magnitude easily result from changesin the feeding rate of magmatic gases at the base of the hydrothermal system or may arise from changes insurface degassing due to permeability transients [Todesco et al., 2003, 2004; Rinaldi et al., 2010].

The subsidence during the fifth century is consistent with the time-dependent compaction due to deflationof the hydrothermal system. While pore pressure changes of the order of a few megapascal easilycharacterize hydrothermal circulation, the occurrence of a pressure drop usually premises a previous phase ofpore pressure buildup. We speculate that an unrest phase during Roman times could have provided theexcess of magmatic fluids then discharged during Late Antiquity. Our reconstruction depicts a lively history ofvertical ground displacement, in which the long repose time before eruptive activity appears to befragmented into several, connected episodes of uplift and subsidence. In particular, the Monte Nuovoeruption, in 1538, was preceded by more than a thousand years of ground deformation, with repeatedtransitions between unrest and quiet phases. Our reconstruction provides a new perspective on the history ofvolcanic activity at Campi Flegrei.

ReferencesAllis, R., C. Bromley, and S. Currie (2009), Update on subsidence at the Wairakei-Tauhara geothermal system, New Zealand, Geothermics, 38,

169–180.Amoruso, A., L. Crescentini, and I. Sabbetta (2014), Paired deformation sources of the Campi Flegrei caldera (Italy) required by recent

(1980–2010) deformation history, J. Geophys. Res. Solid Earth, doi:10.1002/2013JB010392, in press.Auriemma, R., and E. Solinas (2009), Archaeological remains as sea level change markers: A review, Quat. Int., 206, 134–146, doi:10.1016/j.

quaint.2008.11.012.Bellucci, F., J. Woo, C. Kilburn, and G. Rolandi (2006), Ground deformation at Campi Flegrei, Italy: Implications for hazard assessment, in

Mechanisms of Activity and Unrest at Large Calderas, Special Publications, vol. 269, edited by C. Troise, G. De Natale, and C. Kilburn,pp. 141–157, Geol. Soc., London.

Biot, M. A. (1941), General theory of three-dimensional consolidation, J. Appl. Phys., 12(2), 155–164.Bonafede, M. (1991), Hot fluid migration: An efficient source of ground deformation: Application to the 1982–1985 crisis at Campi Flegrei-

Italy, J. Volcanol. Geotherm. Res., 48(1–2), 187–198.Casertano, L., A. Olivieri del Castillo, and M. T. Quagliariello (1976), Hydrodynamics and geodynamics in the Phlegraean fields area of Italy,

Nature, 264(5582), 161–164, doi:10.1038/264161a0.Castagnolo, D., F. S. Gaeta, G. De Natale, F. Peluso, G. Mastrolorenzo, C. Troise, F. Pingue, and D. G. Mita (2001), Campi Flegrei unrest

episodes and possible evolution towards critical phenomena, J. Volcanol. Geotherm. Res., 109(1–3), 13–40, doi:10.1016/S0377-0273(00)00302-4.

Chang, C., D. Moos, and M. D. Zoback (1997), Anelasticity and dispersion in dry unconsolidated sands, Int. J. Rock Mech. Min. Sci., 34(3–4),48.e1–48.e12, doi:10.1016/S1365-1609(97)00204-9.

Costa, A., A. Folch, G. Macedonio, B. Giaccio, R. Isaia, and V. C. Smith (2012), Quantifying volcanic ash dispersal and impact of the CampanianIgnimbrite super-eruption, Geophys. Res. Lett., 39, L10310, doi:10.1029/2012GL051605.

D’Auria, L., F. Giudicepietro, I. Aquino, G. Borriello, C. Del Gaudio, D. Lo Bascio, M. Martini, G. P. Ricciardi, P. Ricciolino, and C. Ricco (2011),Repeated fluid-transfer episodes as a mechanism for the recent dynamics of Campi Flegrei caldera (1989–2010), J. Geophys. Res., 116,B04313, doi:10.1029/2010JB007837.

Deino, A. L., G. Orsi, S. de Vita, and M. Piochi (2004), The age of the Neapolitan Yellow Tuff caldera-forming eruption (Campi Flegrei caldera–Italy) assessed by 40 Ar/39 Ar dating method, J. Volcanol. Geotherm. Res., 133(1), 157–170.

Del Gaudio, C., I. Aquino, G. P. Ricciardi, C. Ricco, and R. Scandone (2010), Unrest episodes at Campi Flegrei: A reconstruction of verticalground movements during 1905–2009, J. Volcanol. Geotherm. Res., 195(1), 48–56, doi:10.1016/j.jvolgeores.2010.05.014.

Di Vito, M., L. Lirer, G. Mastrolorenzo, and G. Rolandi (1987), The 1538 Monte Nuovo eruption (Campi Flegrei, Italy), Bull. Volcanol., 49(4),608–615, doi:10.1007/BF01079966.

Di Vito, M. A., R. Isaia, G. Orsi, J. Southon, S. De Vita, M. D’Antonio, L. Pappalardo, and M. Piochi (1999), Volcanism and deformation since12,000 years at the Campi Flegrei caldera (Italy), J. Volcanol. Geotherm. Res., 91(2), 221–246.

Dvorak, J. J., and G. Berrino (1991), Recent ground movement and seismic activity in Campi Flegrei, southern Italy: Episodic growth of aresurgent dome, J. Geophys. Res., 96(B2), 2309–2323, doi:10.1029/90JB02225.

Dvorak, J. J., and G. Mastrolorenzo (1991), The mechanisms of recent vertical crustal movements in Campi Flegrei caldera, southern Italy,Spec. Pap. Geol. Soc. Am., 263, 49.

Dzurisin, D., K. M. Yamashita, and J. W. Kleinman (1994), Mechanisms of crustal uplift and subsidence at the Yellowstone caldera, Wyoming,Bull. Volcanol., 56(4), 261–270, doi:10.1007/BF00302079.

Farrell, W. E., and J. A. Clark (1976), On postglacial sea level, Geophys. J. R. Astron. Soc., 46(3), 647–667, doi:10.1111/j.1365-246X.1976.tb01252.x.Farrell, J., R. B. Smith, T. Taira, W.-L. Chang, and C. M. Puskas (2010), Dynamics and rapid migration of the energetic 2008–2009 Yellow- stone

Lake earthquake swarm, Geophys. Res. Lett., 37, L19305, doi:10.1029/2010GL044605.Fedele, L., D. D. Insinga, A. T. Calvert, V. Morra, A. Perrotta, and C. Scarpati (2011), 40Ar/39Ar dating of tuff vents in the Campi Flegrei caldera

(southern Italy): Toward a new chronostratigraphic reconstruction of the Holocene volcanic activity, Bull. Volcanol., 73(9), 1323–1336,doi:10.1007/s00445-011-0478-8.

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 7

AcknowledgmentsThe work was partially supported by theDPC-INGV project V2—“Precursori.” Wewish to thank Shaul Hurwitz and ananonymous reviewer for the usefulcomments that greatly improved theoriginal manuscript.

The Editor thanks Shaul Hurwitz andJoachim Gottsmann for their assistancein evaluating this paper.

Fournier, N., and L. Chardot (2012), Understanding volcano hydrothermal unrest from geodetic observations: Insights from numericalmodeling and application to White Island volcano, New Zealand, J. Geophys. Res., 117, B11208, doi:10.1029/2012JB009469.

Gaeta, F. S., G. De Natale, F. Peluso, G. Mastrolorenzo, D. Castagnolo, C. Troise, F. Pingue, D. G. Mita, and S. Rossano (1998), Genesis andevolution of unrest episodes at Campi Flegrei caldera: The role of thermal fluid-dynamical processes in the geothermal system, J. Geophys.Res., 103(B9), 20,921–20,933, doi:10.1029/97JB03294.

Gaeta, F. S., F. Peluso, I. Arienzo, D. Castagnolo, G. De Natale, G. Milano, C. Albanese, and D. Mita (2003), A physical appraisal of a new aspect ofbradyseism: The miniuplifts, J. Geophys. Res., 108(B8), 2363, doi:10.1029/2002JB001913.

Gottsmann, J., A. G. Camacho, K. F. Tiampo, and J. Fernández (2006), Spatiotemporal variations in vertical gravity gradients at the CampiFlegrei caldera (Italy): A case for source multiplicity during unrest?, Geophys. J. Int., 167, 1089–1096, doi:10.1111/j.1365-246X.2006.03157.x.

Guidoboni, E., and C. Ciuccarelli (2011), The Campi Flegrei caldera: Historical revision and new data on seismic crises, bradyseisms, theMonteNuovo eruption and ensuing earthquakes (twelfth century 1582 AD), Bull. Volcanol., 73(6), 655–677, doi:10.1007/s00445-010-0430-3.

Hurwitz, S., L. B. Christiansen, and P. A. Hsieh (2007), Hydrothermal fluid flow and deformation in large calderas: Inferences from numericalsimulations, J. Geophys. Res., 112, B02206, doi:10.1029/2006JB004689.

Isaia, R., P. Marianelli, and A. Sbrana (2009), Caldera unrest prior to intense volcanism in Campi Flegrei (Italy) at 4.0 ka B.P.: Implications forcaldera dynamics and future eruptive scenarios, Geophys. Res. Lett., 36, L21303, doi:10.1029/2009GL040513.

Kaneko, T., A. Yasuda, T. Shimano, S. Nakada, T. Fujii, T. Kanazawa, A. Nishizawa, and Y. Matsumoto (2005), Submarine flank eruptionpreceding caldera subsidence during the 2000 eruption of Miyakejima Volcano, Japan, Bull. Volcanol., 67(3), 243–253, doi:10.1007/s00445-004-0407-1.

Kwoun, O.-I., Z. Lu, C. Neal, and C. Wicks (2006), Quiescent deformation of the Aniakchak Caldera, Alaska, mapped by InSAR, Geology, 34(1),5–8, doi:10.1130/G22015.1.

Lambeck, K., C. Smither, and P. Johnston (1998), Sea-level change, glacial rebound and mantle viscosity for northern Europe, Geophys. J. Int.,134, 102–144, doi:10.1046/j.1365-246x.1998.00541.x.

Lambeck, K., F. Antonioli, A. Purcell, and S. Silenzi (2004), Sea-level change along the Italian coast for the past 10,000 yr, Quat. Sci. Rev.,23(14–15), 1567–1598, doi:10.1016/j.quascirev.2004.02.009.

Lambeck, K., F. Antonioli, M. Anzidei, L. Ferranti, G. Leoni, G. Scicchitano, and S. Silenzi (2011), Sea level change along the Italian coast duringthe Holocene and projections for the future, Quat. Int., 232(1), 250–257.

Lyell, C. (1830), Principles of Geology, Being an Attempt to Explain the Former Changes of the Earth’s Surface, by Reference to Causes Now inOperation, J. Murray, London.

Maniscalco, F. (1998), Gli edifici sommersi del Palatium dei Severi a Baia: Nuovi dati per la definizione del bradisismo flegreo, in Forma Maris,Forum internazionale di Archeologia Subacquea, edited by P. A. Gianfrotta and F. Maniscalco, Massa Editore, Pozzuoli, Italy.

Morhange, C., M. Bourcier, J. Laborel, C. Giallanella, J. Goiran, L. Crimaco, and L. Vecchi (1999), New data on historical relative sea levelmovements in Pozzuoli, Phlaegrean Fields, southern Italy, Phys. Chem. Earth Part A., 24(4), 349–354.

Morhange, C., N. Marriner, J. Laborel, M. Todesco, and C. Oberlin (2006), Rapid sea-level movements and noneruptive crustal deformations inthe Phlegrean Fields caldera, Italy, Geology, 34(2), 93–96.

Newhall, C. A., and D. Dzurisin (1988), Historical Unrest at Large Calderas of the World, vol. 2, U.S. Geol. Sur. Bull. 1855, 1108 pp.Parascandola, A. (1947), I Fenomeni Bradisismici del Serapeo di Pozzuoli, Stabilimento Tipografico G. Genovese, Naples, Italy.Peltier, W. R. (2004), Global glacial isostasy and the surface of the ice-age Earth: The ICE-5G (VM2) model and GRACE, Annu. Rev. Earth Planet.

Sci., 32, 111–149.Pirazzoli, P. A. (1976), Sea level variations in the northwest Mediterranean during Roman times, Science, 194(4264), 519–521.Rinaldi, A., M. Todesco, and M. Bonafede (2010), Hydrothermal instability and ground displacement at the Campi Flegrei caldera, Phys. Earth

Planet. Inter., 178(3–4), 155–161.Shelly, D. R., D. P. Hill, F. Massin, J. Farrell, R. B. Smith, and T. Taira (2013), A fluid-driven earthquake swarm on the margin of the Yellowstone

caldera, J. Geophys. Res. Solid Earth, 118, 4872–4886, doi:10.1002/jgrb.50362.Spada, G., and P. Stocchi (2007), SELEN: A Fortran 90 program for solving the “sea-level equation”, Comput. Geosci., 33(4), 538–562.Todesco, M., G. Chiodini, and G. Macedonio (2003), Monitoring and modelling hydrothermal fluid emission at La Solfatara (Phlegrean Fields,

Italy). An interdisciplinary approach to the study of diffuse degassing, J. Volcanol. Geotherm. Res., 125(1–2), 57–79.Todesco, M., J. Rutqvist, G. Chiodini, K. Pruess, and C. M. Oldenburg (2004), Modeling of recent volcanic episodes at Phlegrean Fields (Italy):

Geochemical variations and ground deformation, Geothermics, 33(4), 531–547.Vanorio, T., M. Prasad, and A. Nur (2002), Petrophysical properties of volcanic samples: A contribution to studying ground deformation in

Campi Flegrei volcanic area, Italy, EGS General Assembly Conference Abstracts, 27, 3685.Waite, G., and R. Smith (2002), Seismic evidence for fluid migration accompanying subsidence of the Yellowstone caldera, J. Geophys. Res.,

107(B9), 2177, doi:10.1029/2001JB000586.Welter-Schultes, F., and I. Richling (2000), Palaeoenvironmental history of the Holocene volcanic crater lake Lago d’Averno (central southern

Italy) inferred from aquatic mollusc deposits, J. Quat. Sci., 15(8), 805–812.Yokoyama, I., and A. Nazzaro (2002), Anomalous crustal movements with low seismic efficiency-Campi Flegrei, Italy, and some examples in

Japan, Ann. Geophys., 45(6), 709–722.Zoback, M. D. (2008), Reservoir Geomechanics, Cambridge Univ. Press, Cambridge, U. K.

Geophysical Research Letters 10.1002/2013GL059083

TODESCO ET AL. ©2014. American Geophysical Union. All Rights Reserved. 8