lable at ScienceDirect

Cretaceous Research 46 (2013) 257e266

Contents lists avai

Cretaceous Research

journal homepage: www.elsevier .com/locate/CretRes

Constraining the colouration mechanisms of Cretaceous Oceanic RedBeds using diffuse reflectance spectroscopy

Xiang Li a,b, Yuanfeng Cai b,c,*aWuhan Institute of Geology and Mineral Resources, China Geological Survey, 69 Optic Valley Road, Wuhan 430223, Chinab School of Earth Sciences and Engineering, Nanjing University, 163 Xianlin Avenue, Nanjing 210046, Chinac State Key Laboratory for Mineral Deposits Research, Nanjing University, Nanjing 210093, China

a r t i c l e i n f o

Article history:Received 30 April 2013Accepted in revised form 27 September2013Available online 13 November 2013

Keywords:Cretaceous Oceanic Red BedsHematiteDiffuse reflectance spectroscopyColouration mechanisms

* Corresponding author. State Key Laboratory ofNanjing University, 163 Xianlin Avenue, Nanjing 283597197; fax: þ86 25 83680616.

E-mail addresses: caiyf@nju.edu.cn, caiyf1912@hot

0195-6671/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.cretres.2013.09.009

a b s t r a c t

We have used diffuse reflectance spectroscopy to investigate the colouration mechanisms of hematite inCretaceous Oceanic Red Beds (CORBs). Data for samples of CORBs from the Chuangde section in Tibet,Vispi Quarry section in Italy, and Core 12X of Ocean Drilling Program Hole 1049C in the North Atlanticwere compared with calibration datasets obtained for hematite in different crystalline forms (kidney andspecular hematite) and calcite matrix. Spectra for hematite in either pure form or in calibration datasetsshow that the centre of the reflection peak shifts to a longer wavelength and depth (D) decreases as thecrystallinity of the hematite increases. Compared with specular hematite, the presence of just 0.5% ofkidney hematite can cause a much deeper absorption peak and greater redness value, which indicatesthat kidney hematite has a higher colouration capacity than specular hematite. However, both kidneyand specular hematite exhibit a good correlation between the redness value for each calibration datasetand the absorption peak depth. In all three studied sections, hematite is the main iron oxide mineralresponsible for colouration. Spectral features such as absorption peak depth and peak centre reveal thathematite crystallinity gradually decreases from red shale to limestone to marl. Based on a spectralcomparison of red shale in the Chuangde section before and after citrateebicarbonateedithionite (CBD)treatment, we found that two forms of hematite are present: a fine-grained and dispersed form, and adetrital form. The former is relatively poorly crystalline hematite, which has a much stronger colourationcapacity than the detrital form. In the Vispi Quarry section and Core 12X of ODP Hole 1049C, a goodcorrelation between the absorption peak depth of hematite and redness value indicates that the redcolouration is caused by hematite of similar crystallinity in each section.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Sediment colour is one of the main physical properties that caneasily be observedwith the naked eye. It is generally acknowledgedthat sediment colour is controlled by iron-bearingminerals, such asiron oxides or oxyhydroxides, iron-bearing sulphides, and iron-richclay minerals (Potter et al., 1980). Other minor constituents mayalso have a secondary or local influence on sediment colour (Balsamand Deaton, 1996; Giosan et al., 2002a; Li et al., 2011; Lyle, 1983;Mix et al., 1995). In some cases, sediment colour has been linkedto a prominent geological event, such as the organic-rich blackshale that was widely distributed in the Cretaceous global ocean

Mineral Deposits Research,10046, China. Tel.: þ86 25

mail.com (Y. Cai).

All rights reserved.

and has been attributed to Oceanic Anoxic Events (OAEs) (Arthuret al., 1990; Jenkyns, 1980, 2010). The origin of CretaceousOceanic Red Beds (CORBs) has also been deduced from the redcolour of the beds (Cai et al., 2012; Hu et al., 2005, 2012; Wagreichet al., 2011; Wang et al., 2005).

The term CORBs was defined byWang et al. (2004, 2005). CORBsare red coloured sedimentary rocks of Cretaceous age, which weremainly deposited in a (hemi-) pelagic environment (Hu et al., 2005,2012; Scott, 2009). The nature of CORBs has been attributed to thepresence of iron oxides and, in particular, hematite in these rocks(Cai et al., 2009, 2012; Eren and Kadir, 2001; Hu et al., 2006a, 2009,2012; Li et al., 2009, 2011). However, the colouration mechanismsof hematite in CORBs have not been studied in detail.

Iron oxides in CORBs, soils, and loessepaleosol sequences areoften fine-grained, relatively poorly crystalline, and present at lowconcentrations. Hence, the techniques commonly used to studyiron oxides, such as X-ray diffraction (XRD) and Mössbauer

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266258

spectroscopy, are constrained by their respective detection limits.For example, in the Upper Cretaceous Scaglia Rossa limestone fromthe Vispi Quarry section in central Italy, the average Fe2O3 contentis only 0.22 wt.% (Hu et al., 2009) and no hematite was detected byXRD prior to the elimination of calcite (Cai et al., 2009).

Compared with other analytical techniques, diffuse reflectancespectroscopy (DRS) is a rapid and non-intrusive tool for identifyingthe constituents of sedimentary rocks and their abundances.Recently, DRS in the visible band (Vis-DRS) has increasingly beenused for the identification of iron oxides in soils (Fontes andCarvalho, 2005; Scheinost et al., 1998; Scheinost andSchwertmann, 1999; Torrent et al., 1983), deep sea deposits(Balsam and Deaton,1996; Balsam and Damuth, 2000; Giosan et al.,2002a,b; Harris and Mix, 1999; Mix et al., 1995; Zhang et al., 2007),aerosol particles (Arimoto et al., 2002; Shen et al., 2004, 2006), andloessepaleosol sequences (Balsam et al., 2004; Ji et al., 2001, 2002,2004; Zhou et al., 2010). Vis-DRS is extremely sensitive to thepresence of iron oxides and is approximately one order of magni-tude more sensitive than XRD, with a limit of detection of ca.0.01 wt.% for free iron oxides in soils and sediments (Barranco et al.,1989; Deaton and Balsam, 1991). The successful application of DRSto identify iron oxides makes it feasible for identifying the mineralsresponsible for the colouration of CORBs (Hu et al., 2009, 2012; Liet al., 2011). In previous studies of CORBs, we have reported theapplication of DRS to the detection of iron oxides (particularly he-matite and goethite) in CORBs and performed a quantitative anal-ysis of iron oxides in red beds of ODP Hole 1049C in the NorthAtlantic using multiple linear regression techniques based on DRS(e.g., Li et al., 2011). Compared with hematite, goethite is brightyellow (8.1YRe1.6Y, Munsell hue) and imparts a bright yellowcolour to sediments (Deaton and Balsam,1991; Torrent et al., 2006).The yellow colour of goethite is easily masked by the presence ofred hematite (Torrent et al., 1983). Furthermore, goethite is unsta-ble andmay gradually transform to hematite during late diagenesis.Therefore, in this paper, we focused on the methodology explora-tion using DRS combined with CR during the study the colourationmechanisms of hematite in CORBs. Kidney and specular hematitewere taken as examples of relatively poorly and well-crystallinehematite, respectively, and were mixed with calcite to obtain cali-bration sample sets. CORBs from the Chuangde section in Tibet,Vispi Quarry section in Italy, and Core 12X of ODP Hole 1049C in theNorth Atlantic were chosen for study, as these CORBs representthree different lithologies (shale, limestone, andmarl, respectively).

2. Geological setting and sampling

During the Cretaceous, the Chuangde section, Vispi Quarrysection, and site of ODP Hole 1049C were located in the easternTethys, western Tethys, and Atlantic Tethys, respectively (Fig. 1).

The Chuangde section is situated near Chuangde Village, whichis ca. 10 km to the east of Gyangze City, southern Tibet, China. Thisareawas tectonically part of the northern subzone of the HimalayanTethys zone and contains widely developed Cretaceous oceanicstrata. Wang et al. (2000) revised the stratigraphy in this sectionand divided it in ascending order into the Gyabula, Chuangde, andZongzhuo formations. The BerriasianeConiacian Gyabula Forma-tion comprises black shale with pyrite nodules that is intercalatedwith sandstone beds. The overlying Chuangde Formation com-prises violetered shale that is intercalated with thinly beddedmarlstone. The overlying upper Campanian to Palaeocene Zongz-huo Formation predominantly comprises dark grey to black shalesthat enclose various olistoliths of sandstone, limestone, and beddedchert (Hu et al., 2006a; Liu and Aitchison, 2002). Thirty red shalesamples collected from the Chuangde Formation were studied(Fig. 1). According to Geological Society of America (GSA) Rock-

Colour Chart, their Munsell colour were 10R2/2, 5YR2/2, 5Y3/2,respectively (Table 1).

The Vispi Quarry section crops out on the eastern slope of theContessa Valley about 2 kmwest of the town of Gubbio in Italy. Thissection is composed of the Scaglia Bianca and Scaglia Rossa for-mations (Hu et al., 2006b). The Scaglia Bianca Formation is mainlyyellowish to greyish limestones that are intercalated with sparsepink to reddish limestone beds and several green to greyeblackmarlstones and shales (Bonarelli Level). The Scaglia Rossa Forma-tion predominantly comprises pink to reddish marly limestonebeds that generally contain 65e92 wt.% CaCO3, with the exceptionof rare shale interbeds. The details of this section have beendescribed by Arthur and Fischer (1977) and Alvarez and Montanari(1988). We studied the section above the Bonarelli (OAE2) Levelfrom 9.8 m in the Scaglia Bianca Formation to 10.9 m in the ScagliaRossa Formation. Up to the point that the limestone successionbecomes completely reddish, two transitional beds (at 10.35e10.55and 10.63e10.83 m) are present. Both of the transitional beds are awhitish colour at the base with a gradually increasing pinkishcolour towards the top. Above these two transitional beds(>10.83 m), the limestone is completely pink to red in colour (e.g.,Hu et al., 2009; Fig. 2). We collected 10 red limestone samples fromthe two transitional beds for DRS analysis (Fig. 1). According toGeological Society of America (GSA) Rock-Colour Chart, theirMunsell colour were 10GY7/2, 5Y8/1, respectively.

ODP Leg 171B drilled mid-Cretaceous unconsolidated sedimentsin the western North Atlantic off the coast of Florida. According to acompilation of magnetic poles, Site 1049 was located at 23�Nduring the Cretaceous and represents a pelagic sedimentary envi-ronment above the carbonate compensation depth (CCD) (Norriset al., 1998). Core 12X of Hole 1049C was drilled at depths of139.3e148.1 m below sea floor (mbsf). The sediments are lateAptian to early Albian in age and clayey, calcareous, nannofossil-bearing chalk and claystone, which are rich in planktic foraminif-eral assemblages and have high-frequency colour variations be-tween red (brown/orange), white, and green beds. These rhythmiccolour changes are interrupted by a 46-cm-thick layer of laminatedblack shale that correlates with OAE 1b, which is a black shalesequence identified in equivalent European sections (Erbacheret al., 2001). In a previous study, we divided these sediments intoeight cycles of redewhite beds, based on colour changes andmagnetic susceptibility (Li et al., 2011). In this study, only 24 redmarl samples from Core 12X of Hole 1049C were chosen for DRSanalysis (Fig. 1). According to Geological Society of America (GSA)Rock-Colour Chart, their Munsell colour were 5YR5/6, 10YR 7/4,10YR 7/5, 10YR 8/2, respectively.

3. Sample treatment and reflectance measurements

Samples were analysed using a PerkineElmer Lambda 6 spec-trophotometer with a diffuse reflectance attachment, which iscapable of measuring sample reflectance in the near-ultraviolet(190e400 nm), visible (400e700 nm), and near-infrared (700e2500 nm) bandwidths, at the Institute of Surficial Geochemistry atNanjing University, China. Sample preparation and analysis fol-lowed the procedures described in Balsam and Deaton (1991) and Jiet al. (2002). Powder specimens were made into slurries withdistilled water on glass slides, smoothed, and dried slowly at<40 �C. Data are presented as the percent reflectance relative to theSpectralon� (reflectance ¼ 100%). Data processing was restrictedto the visible spectrum (400e700 nm), which is the region mostsensitive to iron oxide minerals (Deaton and Balsam, 1991). Tobetter understand the colouration mechanisms of hematite, kidneyhematite (as an example of relatively poorly crystalline hematite)was mixed with calcite matrix to obtain a calibration sample set

Fig. 1. Geological sketch map showing the localities and lithological logs of studied CORBs sections in Tibet of China, Gubbio of Italy and ODP Hole 1049C, North Atlantic (modifiedfrom Cai et al., 2009; Hu et al., 2006a; Li et al., 2011). Also shown are the localities of the studied CORBs sections in palaeogeographic map of 105 Ma. The palaeogeographic base mapwas downloaded from the website www.odsn.de.

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266 259

Table 1The redness value (RV), absorption band depth (D) and position of both calibration sample sets and red samples from the Chuangde section, Vispi Quarry section and 12X Coreof Hole 1049C. The Munsell colour of the three studied sections was also included.

Sample RV (%) Depth (D) Position (nm) Munsell colour Sample RV (%) Depth (D) Position (nm) Munsell colour

Calibration sample set containing specular 12X Core of Hole 1049CS-0.5% 23.63 0.063 544 1w19-21 35.90 0.286 518 5YR5/6S-1% 24.26 0.079 544 1w59-61 36.21 0.307 518 5YR5/6S-2% 25.07 0.096 544 1w69-71 35.96 0.301 522 5YR5/6S-3% 25.69 0.106 544 1w104-106 35.72 0.295 522 5YR5/6S-4% 26.02 0.120 544 1w119-121 35.67 0.290 522 5YR5/6S-5% 26.59 0.128 544 1w129-131 37.20 0.341 520 5YR5/6Calibration sample set containing kidney hematite 2w19-21 34.25 0.240 520 5YR5/6K-0.5% 29.62 0.236 534 2w29-31 29.86 0.107 522 10YR 7/4K-1% 33.92 0.352 538 2w49-51 29.27 0.088 522 10YR 7/4K-2% 38.20 0.458 534 2w59-61 29.97 0.119 522 10YR 7/4K-3% 40.16 0.504 534 5w71.5-73.5 31.86 0.160 520 10YR 8/2K-4% 41.47 0.535 534 5w79-81 31.32 0.147 524 10YR 8/2K-5% 42.14 0.551 534 5w109-111 32.97 0.192 524 10YR 7/4Vispi Quarry section 5w119-121 33.84 0.219 522 10YR 7/40924-1-2 24.61 0.028 530 10GY7/2 5w146-148 30.38 0.120 522 10YR 7/40924-2-2 24.52 0.027 526 10GY7/2 6w9-11 33.72 0.217 520 10YR 7/4BO-69 25.56 0.066 532 5Y8/1 6w34-36 30.75 0.130 522 10YR 7/4CO-37 24.86 0.036 522 5Y8/1 6w44-46 32.65 0.183 522 10YR 7/4CO-49 25.13 0.047 524 5Y8/1 6w54-56 30.32 0.117 524 10YR 7/4CQ-09 23.80 0.009 528 5Y8/1 6w69-71 31.65 0.154 524 10YR 7/4CQ-10 24.31 0.022 530 10GY7/2 6w84-86 33.78 0.222 524 10YR 7/4CQ-14 24.15 0.015 530 5Y8/1 6w99-101 32.28 0.170 520 10YR 7/4CQ-15 24.66 0.030 530 5Y8/1 6w111-113 34.33 0.234 520 10YR 7/4CQ-16 25.09 0.043 526 5Y8/1 6w124-126 34.15 0.231 520 10YR 7/4Chuangde section before-CBD Chuangde section after-CBD06cd-007 38.40 0.448 532 10R2/2 06cd-007 30.63 0.252 53406cd-010 38.76 0.451 530 10R2/2 06cd-010 29.98 0.241 53206cd-017 38.45 0.456 530 10R2/2 06cd-017 30.69 0.265 53406cd-026 34.04 0.259 534 10R2/2 06cd-026 30.09 0.200 53406cd-028 35.05 0.306 534 10R2/2 06cd-028 30.46 0.223 53406cd-032 34.86 0.346 534 10R2/2 06cd-032 28.02 0.171 53406cd-034 37.56 0.432 530 10R2/2 06cd-034 27.73 0.175 53206cd-041 36.51 0.387 534 5YR2/2 06cd-041 30.40 0.244 53606cd-044 35.78 0.403 534 5YR2/2 06cd-044 27.23 0.153 53606cd-046 34.71 0.380 536 5YR2/2 06cd-046 27.20 0.152 53406cd-047 32.18 0.182 530 5YR2/2 06cd-047 26.90 0.102 54006cd-049 35.69 0.398 536 5YR2/2 06cd-049 28.40 0.198 53406cd-051 34.26 0.367 536 5YR2/2 06cd-051 26.93 0.146 53406cd-056 33.62 0.308 532 5YR2/2 06cd-056 27.76 0.163 53606cd-059 36.00 0.380 536 5Y3/2 06cd-059 28.70 0.193 53406cd-062 34.93 0.380 534 5Y3/2 06cd-062 26.52 0.128 53406cd-066 32.48 0.130 526 5Y3/2 06cd-066 30.33 0.199 53406cd-068 33.80 0.372 538 5YR2/2 06cd-068 26.46 0.128 53606cd-070 31.28 0.218 538 5YR2/2 06cd-070 26.98 0.128 53406cd-073 33.48 0.364 538 5YR2/2 06cd-073 26.24 0.119 53606cd-074 35.00 0.384 536 5YR2/2 06cd-074 26.80 0.133 53606cd-075 33.36 0.349 538 5YR2/2 06cd-075 24.54 0.080 54206cd-078 31.77 0.288 536 5YR2/2 06cd-078 26.87 0.134 53606cd-080 33.09 0.344 538 5YR2/2 06cd-080 26.16 0.123 53606cd-083 33.55 0.351 536 5Y3/2 06cd-083 27.29 0.154 53606cd-086 34.69 0.398 538 10R2/2 06cd-086 28.20 0.170 53606cd-091 34.44 0.387 438 10R2/2 06cd-091 28.53 0.192 53406cd-097 31.88 0.323 538 5YR2/2 06cd-097 27.29 0.156 53606cd-100 34.19 0.387 538 5YR2/2 06cd-100 26.59 0.142 53606cd-105 35.74 0.407 534 5YR2/2 06cd-105 28.03 0.192 536

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266260

comprising 0.5%, 1%, 2%, 3%, 4%, and 5% kidney hematite. Similarly,specular hematite as an example of crystalline hematite was pre-pared as a calibration sample set at 0.5%, 1%, 2%, 3%, 4%, and 5%specular hematite. The kidney and specular hematite used inmaking these calibration sample sets are mineralogical specimenscollected by the School of Earth Sciences and Engineering, NanjingUniversity, China. The kidney hematite was collected from Pan-gjiabao Town (Xuanhua County, Hebei Province, China) and thespecular hematite was taken from Rio Marina on Elba Island, Italy(Cai et al., 2008). The kidney hematite is poorly crystalline with adark red colour, whereas the specular hematite is crystalline with abrownish-black colour and a red streak. The calibration sample

sets, samples from the three stratigraphic sections, as well as redshale samples after treated with CBD procedures (Mehra andJackson, 1960) from the Chuangde section were analysed with thePerkineElmer Lambda 6 spectrophotometer.

Spectral data were processed by the following methodology:

(1) Each DRS analysis was converted into the spectral libraryformat of the ENVI software package (Research Systems Inc.,2003) for continuing removal (CR) analysis. Previous studieshave shown that the spectral parameters obtained by the CRlogarithm can be used to identify the sediment constituents(Clark and Roush, 1984; Clark, 1999; Duke, 1994; Gaffey,

Fig. 2. The diffuse reflectance spectra of kidney and specular hematite. (A) and (B) show the calculation process of continuum removal logarithm of pure kidney and specularhematite, respectively and; (C) and (D) are for calibration sample sets containing kidney and specular hematite, respectively.

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266 261

1986; Richter et al., 2009; Viscarra Rossel et al., 2009). Asshown in Fig. 2A and B, the CR logarithm was used to ratiothe spectrum with the lowest convex curve lying above thespectrum (Clark and Roush, 1984; Clark, 1999). In our study,the absorption peak depth (D) was defined as follows:

D ¼ 1� Rb=Rc

where Rb is the reflectance at the band centre and Rc is thereflectance of the continuum at the same wavelength as Rb. In theresulting CR spectrum, weak absorption features are enhanced, andhence band depths of absorption features can be quantitativelycompared (Wu et al., 2005). In the CR spectrum, the absorptionpeak centred near 540 nmwas assigned to hematite (Ben-Dor et al.,2006; Crowley et al., 2003).

(2) Histograms were compiled showing the distribution of theabsorption peak centres of hematite in the CR spectra forsamples from each section.

(3) The redness values of samples, as defined by Ji et al. (2002),were calculated by summing the reflectance values from 630to 700 nm and dividing by the summed reflectance valuesfrom 400 to 700 nm (the entire visible range), and thenmultiplied by 100. Sediment colour can be quantified by thisredness value, with higher values denoting a stronger redcolouration.

4. Results

The original and CR spectra of kidney and specular hematite intheir pure forms are shown in Fig. 2A and B. Original reflectancespectra show that the absorption edge of pure kidney hematiteoccurs at ca. 540 nm, and is deep and sharp in nature. In contrast,specular hematite has a relatively weak absorption edge at a longer

wavelength (ca. 570 nm). After applying the CR logarithm, the ab-sorption peak of hematite in the CR spectra is more enhanced andmore clearly reveals the presence of hematite. The pure kidneyhematite CR spectra has a deep absorption peak centred at 534 nmwith a D ¼ 0.61, whereas the absorption peak centre of purespecular hematite shifts to a longer wavelength (ca. 548 nm) andD ¼ 0.38 (Fig. 2A and B).

CR spectra of the calibration sample sets (Figs 2C, D and 3A)show that D increases with hematite content for both kidney andspecular hematite. For the calibration sample set containing kidneyhematite, D increases from 0.236 to 0.551, whereas D increasesfrom 0.063 to 0.128 for specular hematite (hematite content¼ 0.5e5%). It is important to note that the calibration sample set con-taining kidney hematite has a considerably higher D than thatcontaining specular hematite, even when the former only contains0.05% of hematite. The corresponding absorption peaks in thecalibration sample sets are centred at 534 and 544 nm for kidneyand specular hematite, respectively. However, the calibrationsample containing 1% of kidney hematite has an absorption peakcentre at 538 nm.

The redness and D values exhibit a similar trend withincreasing hematite content in both calibration sample sets. Asshown in Fig. 3A, increasing hematite content is associated withan increase in the redness value of both calibration sample sets.When the hematite content increases from 0.5% to 5%, theredness value increases from 29.6% to 42.1% (kidney hematite)and from 23.6% to 26.6% (specular hematite). Kidney hematite hasa much higher redness value and exhibits a greater increase inredness value with hematite content than does specular hema-tite. Furthermore, a similar positive linear relationship betweenredness value and D characterises both calibration sample sets.The equations obtained by linear regression analysis of these dataare shown in Fig. 3B, and both have good fit with squared cor-relation coefficients (R2) of 0.993 (kidney hematite) and 0.999(specular hematite).

Fig. 3. The variation of redness value and absorption band depth D with increasing hematite contents in calibration sample sets (A) and the correlation of redness value withabsorption band depth D for both calibration sample sets (B). (For interpretation of the references to color in this figure legend,the reader referred to the web version of this article.)

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266262

The CR spectra of red shale from the Chuangde section, redlimestone from the Vispi Quarry section, and red marl from Core12X of ODP Hole 1049C show an obvious hematite absorptionpeak near 540 nm (Fig. 4). Before CBD treatment, red shale fromthe Chuangde section has CR spectral D values that range from0.130 to 0.456 and band centres that vary from 526 to 538 nm,with a normally distributed mean value at 534.8 nm. After CBDtreatment, D values decrease to a range between 0.080 and 0.265,and the band centres shift to longer wavelengths between 530and 542 nm, with a normally distributed mean value at 535.1 nm(Fig. 5). Red limestone from the Vispi Quarry section has D valuesthat range from 0.009 to 0.076 and band centres that vary be-tween 522 and 532 nm, with a normally distributed mean valueat 528.3 nm. Red marl from Core 12X of ODP Hole 1049C has Dvalues that range from 0.088 to 0.341 and band centres between518 and 524 nm, with a normally distributed mean value at527.1 nm. As is evident from the calibration sample sets, a goodlinear relationship also exists between the redness values and Dfor the red limestone from the Vispi Quarry section and red marlfrom Core 12X of ODP Hole 1049C. In these two sections, the R2

coefficients were 0.970 and 0.993, respectively. The correlationbetween redness values and D is poor for the red shale samplesfrom the Chuangde section prior to CBD treatment. After CBDtreatment, both the redness values and D were significantlyreduced and the correlation between them improved to an R2 of0.903 (Fig. 6).

5. Discussion

5.1. Spectral responses of different hematite crystalline forms

Hematite has been considered the most common pigmentmineral for all red coloured sediments or rocks. As shown in ourprevious studies, hematite is also responsible for the red col-ouration of CORBs (Cai et al., 2009, 2012; Li et al., 2009, 2011).Although many studies have been conducted on CORBs, relativelyfew have focused on the colouration mechanisms of hematite. Aprevious investigation of Martian hematite showed that hematitein nanophase and crystalline forms has different physico-chemical and spectral properties (Christensen et al., 2000,2001; Lane et al., 1999, 2001). Nanophase hematite is <10 nmin diameter and is X-ray amorphous. In contrast, crystalline he-matite is >10 nm in diameter, X-ray crystalline, and has a

thermally stable magnetic moment (Morris et al., 1985, 1989). Inaddition, crystalline hematite has well-defined reflectivity max-ima and minima at ca. 750 and 860 nm, respectively, in thevisible/near-infrared (Christensen et al., 2000). Crystalline he-matite can be further subdivided into red and grey (specular ormicaceous) hematite according to grain size. The <10 mm fractionis red (orangeepurple) with a shallow absorption edge startingnear 520 nm, whereas grey hematite particles are >10 mm indiameter and have relatively flat reflectivity spectra over thevisible light range (Christensen et al., 2000; Lane et al., 1999;Morris et al., 1989). Compared with published reflectancespectra for hematite, the kidney and specular hematite investi-gated in our study are red and grey hematite, respectively.

In this study, we have shown that poorly crystalline kidneyhematite and crystalline specular hematite have different spectralresponses in the visible light region. The former has a dark redcolour and a strong absorption peak in the visible light region,whereas the latter has a brownish-black colour with only a redstreak and a relatively weak absorption peak (Fig. 2A). CR spectra ofeither pure form of hematite or the calibration sample setsdemonstrate that the absorption peak centre shifts to longerwavelengths and D values reduce with increasing hematite crys-tallinity (Fig. 2C and D). However, the D values for calibrationsamples containing kidney hematite are much higher than thesamples containing specular hematite. As hematite is present ineither kidney or specular hematite form, the D values increase withhematite content without a shift in wavelength. Previous studies ofthe visible absorption spectra of hematite in different crystallineforms have shown similar spectral features (Cai et al., 2008). Whenhematite is in the kidney or specular hematite form, redness valuesof the samples increase with hematite content and are wellcorrelated with D values (Fig. 3).

The influence of hematite crystallinity on its colouration ca-pacity is reflected by the sample redness value, which is largelycaused by hematite. As is evident from the calibration sample sets,the sample redness value increases not only with increasing he-matite content but also decreasing hematite crystallinity. Calibra-tion samples containing specular hematite have much higher andstrongly increasing redness values (even at 0.5% hematite) than thecalibration samples containing kidney hematite (Fig. 3A). There-fore, we speculate that the colouration capacity of relatively poorlycrystalline hematite is considerably higher than that of crystallinehematite.

Fig. 4. The CR spectra for red marl from the 12X Core of ODP Hole 1049C (A), red limestone from the Vispi Quarry section (B), and red shale from the Chuangde section before (C)and after CBD (D) treatment. (For interpretation of the references to color in this figure legend,the reader referred to the web version of this article.)

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266 263

5.2. Colouration mechanisms of hematite in CORBs

For the red shale from the Chuangde section, red limestone fromthe Vispi Quarry section, and red marl from Core 12X of ODP Hole1049C, the CR spectra show an absorption peak at ca. 540 nm,indicating that hematite is the main iron oxide mineral responsiblefor colouration. As mentioned above, hematite in different crys-talline forms has different DRS responses, i.e. the absorption peak

Fig. 5. The distributions of absorption band centre of hematite in

centre of hematite shifting to longer wavelengths with increasinghematite crystallinity. The normally distributed mean values of thehematite absorption peak are centred at 534.8, 528.3, and 521.7 nm(Fig. 5), suggesting that hematite crystallinity gradually decreasesfrom red shale to red limestone to red marl. This is also evidentfrom petrographic and scanning electron microscopy (SEM) ob-servations (Li, 2011). Petrographic observations have identifiedabundant crystalline hematite grains in the red shale samples (Li,

CR spectra for the samples from the three studied sections.

Fig. 6. The redness value vs. absorption band depth in CR spectra plots for samplesfrom the Chuangde section, Vispi Quarry section and 12X Core of Hole 1049C.

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266264

2011). The shale has experienced clearly diagenesis that resulted inthe formation of crystalline hematite. Core 12X of ODP Hole 1049Ccomprises unconsolidated sediments that are largely unaffected bydiagenesis (Erbacher et al., 2001) and, therefore, hematite in redmarl from Core 12X of ODP Hole 1049C is poorly crystalline and canbe easily extracted by CBD treatment (Li et al., 2011). Red limestonefrom the Vispi Quarry section has undergone diagenesis to a certaindegree and has a hematite crystallinity intermediate between thatof the other two sections.

When hematite is only present in kidney or specular hematiteforms, redness values of the samples are well correlated with theabsorption peak depth (Fig. 3B). Both the redmarl from Core 12X ofODP Hole 1049C and red limestone from the Vispi Quarry sectionexhibit good correlations between redness values and D, indicatingthat hematite is present in similar crystallinity forms in each sec-tion. Two crystalline forms of hematite are present in the red shaleof the Chuangde section (Li, 2011; this study). Prior to CBD treat-ment, the correlation between the redness values and D is poor. Therelative proportions of these two forms of hematite in the red shalevary andmay explain the poor correlation. After CBD treatment, thefine-grained and dispersed (<2 mm; Zhang et al., 2003) hematitewas dissolved, leaving behind a residue of crystalline (or detrital)hematite, which improved the correlation between redness valuesand D. After CBD treatment, the D values of hematite in the CRspectra were significant lower and the normally distributed meanvalue of the peak centre migrated to a longer wavelength (Fig. 5),indicating that only crystalline hematite was present.

We now consider to what extent hematite crystallinity in-fluences its colouration capacity, and examine the implications forCORBs. We compared the CBD-extracted poorly crystalline hema-tite content (Fe2O3CBD) with the total Fe2O3 of the bulk red shalesamples from the Chuangde section and found that on average52.4% of the total Fe2O3 was extracted by CBD treatment (Li, 2011).In our previous study, we quantified the hematite contents in thered shale samples as having a mean value of 6.63 wt.%. Assume thatthe residual Fe2O3 is present as a singular form of hematite, theresidual hematite content can be calculated to be ca. 3 wt.%. The redmarl from Core 12X of ODP Hole 1049C has an average hematitecontent of 0.43 wt.% (Li et al., 2011) and its redness value isconsiderably higher than that of the red shale from the Chuangdesection after CBD treatment. These observations further suggestthat relatively poorly crystalline hematite has a much higher col-ouration capacity than crystalline hematite.

6. Conclusions

Diffuse reflectance spectral analysis of calibration sample setscontaining different forms of crystalline hematite have demon-strated that the hematite absorption peak centre shifts to longerwavelengths and that the absorption peak depth decreases withincreasing hematite crystallinity. Sediment colour is mainlycontrolled by hematite and can be quantified by the redness value.Hematite crystallinity exerts a primary control on a sample’sredness value, with relatively poorly crystalline hematite having amarkedly higher redness value than crystalline hematite. This in-dicates that poorly crystalline hematite has a much higher colour-ation capacity than crystalline hematite. When hematite is presentin either its kidney or specular form, absorption peak depth is wellcorrelated with the redness value. In three studied sections,including red shale from the Chuangde section, red limestone fromthe Vispi Quarry section, and red marl from Core 12X of ODP Hole1049C, hematite is the main iron oxide mineral responsible forcolouration and exhibits difference in crystallinity. Two types ofhematite are present in the Chuangde section: a fine-grained anddispersed form, and a detrital form. The fine-grained form is rela-tively poorly crystalline, and has a markedly higher colourationcapacity than the detrital form. In the Vispi Quarry section and Core12X of ODP Hole 1049C, red colouration is caused by hematite of asimilar crystallinity in each section.

Acknowledgements

Samples of the Vispi Quarry section and of Core 12X of Hole1049C were provided by Xiumian Hu and the Ocean Drilling Pro-gram, respectively. We are grateful to Junfeng Ji for assistance withDRS and to Wei Zhou for assistance in preparing samples. We arealso deeply grateful to the reviewers of this paper for theirconstructive, valuable comments. This study was financially sup-ported by the ChineseMOST 973 Project (Grant No. 2012CB822001)and the Natural Science Foundation of China (Grant No. 41302033)

References

Alvarez, W., Montanari, A., 1988. The Scaglia limestones (Late Cretaceous-Oligocene) in the northeastern Apennines carbonate sequences: stratigraphiccontext and geological significance. In: Premoli Silva, I., Coccioni, R.,Montanari, A. (Eds.), The Eocene-Oligocene boundary in the Marche-UmbriaBasin (Italy), Int. Subcomm. Paleog. Strat., E/O meeting, Ancona, IUGS SpecialPublication. F.lli Aniballi Publishers, Ancona, pp. 13e30.

Arimoto, R., Balsam, W.L., Schloesslin, C., 2002. Visible spectroscopy of aerosolparticles collected on filters: iron-oxide minerals. Atmospheric Environment 36,89e96.

Arthur, M.A., Brumsack, H.J., Jenkyns, H.C., Schlanger, S.O., 1990. Stratigraphy,geochemistry and paleoceanography of organic carbon-rich Cretaceous se-quences. In: Ginsburg, R.N., Beaudoin, B. (Eds.), Cretaceous resources, eventsand rhythms-background and plans for research. Kluwer Academic Publica-tions, Dordrecht, pp. 75e119.

Arthur, M.A., Fischer, A.G., 1977. Upper Cretaceous-Paleocene magnetic stratigraphyat Gubbio, Italy: I. Lithostratigraphy and sedimentology. Geological Society ofAmerican Bulletin 88, 367e371.

Balsam, W.L., Damuth, J.E., 2000. Further investigations of shipboard vs. shore-based spectral data: implications for interpreting Leg 164 sediment composi-tion. In: Paull, C.K., Matsumoto, R., Wallace, P.J., Dillon, W.P. (Eds.), Proceedingsof Ocean Drilling Program, Scientific Results 164, College Station, TX. OceanDrilling Program. Texas A & M University, pp. 313e324.

Balsam, W.L., Deaton, B.C., 1991. Sediment dispersal in the Atlantic Ocean: Evalua-tion by visible spectra. Reviews in Aquatic Sciences 4, 411e447.

Balsam, W.L., Deaton, B.C., 1996. Determining the composition of late Quaternarymarine sediments from NUV, VIS, and NIR diffuse reflectance spectra. MarineGeology 134, 31e55.

Balsam, W.L., Ji, J.F., Chen, J., 2004. Climatic interpretation of the Luochuan andLingtai loess sections, China, based on changing iron oxide mineralogy andmagnetic susceptibility. Earth and Planetary Science Letters 223, 335e348.

Barranco, F.T., Balsam, W.L., Deaton, B.C., 1989. Quantitative reassessment of brickred lutites: evidence from reflectance spectrophotometry. Marine Geology 89,299e314.

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266 265

Ben-Dor, E., Levin, N., Singer, A., Karnieli, A., Braun, O., Kidron, G.J., 2006. Quanti-tative mapping of the soil rubification process on sand dunes using an airbornehyperspectral sensor. Geoderma 131, 1e21.

Cai, Y.F., Hu, X.X., Li, X., Pan, Y.M., 2012. Origin of the red colour in a red limestonefrom the Vispi Quarry section (central Italy): A high-resolution transmissionelectron microscopy analysis. Cretaceous Research 38, 97e102.

Cai, Y.F., Li, X., Pan, Y.G., Hu, X.M., 2008. The colour-causing mechanism of Mn2þ andFe3þ: evidence from the Italian Cretaceous pelagic red limestones. Acta Geo-logica Sinica 82, 133e138 (in Chinese with English abstract).

Cai, Y.F., Li, X., Hu, X.M., Chen, X.M., Pan, Y.G., 2009. Paleoclimatic approach to theorigin of the colouring of Turonian pelagic limestones from the Vispi Quarrysection (Cretaceous, central Italy). Cretaceous Research 30, 1205e1216.

Christensen, P.R., Bandfield, J.L., Clark, R.N., Edgett, K.S., Hamilton, V.E., Hoefen, T.,Kieffer, H.H., Kuzmin, R.O., Lane, M.D., Malin, M.C., Morris, R.V., Pearl, J.C.,Pearson, R., Roush, T.L., Ruff, S.W., Smith, M.D., 2000. Detection of crystallinehematite mineralization on Mars by the Thermal Emission Spectrometer: Evi-dence for near-surface water. Journal of Geophysical Research 105 (E4),9623e9642.

Christensen, P.R., Morris, R.V., Lane, M.D., Bandfield, J.L., Malin, M.C., 2001. Globalmapping of Martian hematite mineral deposits: Remnants of water-drivenprocesses on early Mars. Journal of Geophysical Research 106 (E10),23873e23885.

Clark, R.N., 1999. Spectroscopy of rocks and minerals, and principles of spectros-copy. In: Rencz, A.N. (Ed.), Manual of remote sensing, remote sensing for theearth sciences. John Wiley & Sons, New York, pp. 3e58.

Clark, R.N., Roush, T.L., 1984. Reflectance spectroscopy: Quantitative analysis tech-niques for remote sensing applications. Journal of Geophysical Research 89 (B7),6329e6340.

Crowley, J.K., Williams, D.E., Hammarstrom, J.M., Piatak, N., Chou, I.M., Mars, J.C.,2003. Spectral reflectance properties (0.4e2.5 mm) of secondary Fe-oxide, Fe-hydroxide, and Fe-sulphate-hydrate minerals associated with sulphide-bearingmine wastes. Geochemistry: Exploration, Environment, Analysis 3 (3), 219e228.

Deaton, B.C., Balsam, W.L., 1991. Visible spectroscopy: a rapid method for deter-mining hematite and goethite concentration in geological materials. Journal ofSedimentary Petrology 61, 628e632.

Duke, E.F., 1994. Near infrared spectra of muscovite, tschermak substitution, andmetamorphic reaction progress: implications for remote sensing. Geology 22,621e624.

Erbacher, J., Huber, B.T., Norris, R.D., Markey, M., 2001. Increased thermohalinestratification as a possible cause for an ocean anoxic event in the Cretaceousperiod. Nature 409, 325e327.

Eren, M., Kadir, S., 2001. Colour genesis of Upper Cretaceous pelagic red sedimentswithin the Eastern Pontides, northeast Turkey. Yerbilinleri 24, 71e79.

Fontes, M.P.F., Carvalho Jr., I.A., 2005. Colour attributes and mineralogical charac-teristics, evaluated by radiometry of highly weathered tropical soils. Soil Sci-ence Society of America Journal 69, 1162e1172.

Gaffey, S.J., 1986. Spectral reflectance of carbonates minerals in the visible and nearinfrared (0.35e2.55 microns): calcite, aragonite, and dolomite. AmericanMineralogist 71, 151e162.

Giosan, L., Flood, R.D., Aller, R.C., 2002a. Paleoceanographic significance of sedimentcolour on western North Atlantic drifts: I. Origin of colour. Marine Geology 189,25e41.

Giosan, L., Flood, R.D., Grützner, J., Mudie, P., 2002b. Paleoceanographic significanceof sediment colour on western North Atlantic drifts: II. Plio-Pleistocene sedi-mentation. Marine Geology 189, 43e61.

Harris, S.E., Mix, A.C., 1999. Pleistocene precipitation balance in the Amazon basinrecorded in deep sea sediments. Quaternary Research 51, 14e26.

Hu, X.M., Cheng, W.B., Ji, J.F., 2009. Origin of Cretaceous oceanic red beds from theVispi Quarry section, central Italy: visible reflectance and inorganic geochem-istry. In: Hu, X.M., Wang, C.S., Scott, R.W., Wagreich, M., Jansa, L. (Eds.), Creta-ceous oceanic red beds: Stratigraphy, composition, origins andpaleoceanographic and paleoclimatic significance, SEPM Special Publication 91.Society for Sedimentary Geology, Tulsa, Oklahoma, U.S.A., pp. 183e197.

Hu, X.M., Jansa, L., Wang, C.S., Sarti, M., Bak, K., Wagreich, M., Michalik, J., Soták, J.,2005. Upper Cretaceous Oceanic Red Beds (CORB) in the Tethys: occurrence,lithofacies, age and environment. Cretaceous Research 26, 3e20.

Hu, X.M., Wang, C.S., Li, X.H., Luba, J., 2006a. Upper Cretaceous oceanic red beds insouthern Tibet: lithofacies, environments and colour origin. Science in ChinaSeries D-Earth Sciences 49, 785e795.

Hu, X.M., Jansa, L., Sarti, M., 2006b. Mid-Cretaceous oceanic red beds in the Umbria-Marche Basin, central Italy: Constraints on paleoceanography and paleoclimate.Palaeogeography, Palaeoclimatology, Palaeoecology 233, 163e186.

Hu, X.M., Scott, R.W., Cai, Y.F., Wang, C.S., Melinte-Dobrinescu, M.C., 2012. Creta-ceous Oceanic Red Beds (CORBs): different time scales, different models. Earth-Science Reviews 115, 217e248.

Jenkyns, H.C., 1980. Cretaceous anoxic events: from continents to oceans. Journal ofthe Geological Society 137, 171e188.

Jenkyns, H.C., 2010. Geochemistry of oceanic anoxic events. GeochemistryGeophysics Geosystems 11 (Q03004). http://dx.doi.org/10.1029/2009GC002788.

Ji, J.F., Balsam, W.L., Chen, J., 2001. Mineralogical and climate interpretation of theLuochuan loess section (China) based on diffuse reflectance spectrophotometry.Quaternary Research 56, 23e30.

Ji, J.F., Balsam, W.L., Chen, J., Liu, L.W., 2002. Rapid and quantitative measurement ofhematite and goethite in the Chinese loess-paleosol sequence by diffusereflectance spectroscopy. Clays and Clay Minerals 50, 208e216.

Ji, J.F., Chen, J., Balsam, W.L., Lu, H.Y., Sun, Y.B., Xu, H.F., 2004. High resolution he-matite/goethite records from Chinese loess sequences for the last glacial-interglacial cycle: Rapid climatic response of the East Asian Monsoon to thetropical Pacific. Geophysical Research Letters 31, 2e5.

Lane, M.D., Hartmann, W.K., Berman, D.C., 2001. Update on studies of the Martianhematite-rich areas, 32nd Annual Lunar and Planetary Science Conference,Houston, Texas, abstract no. 1984.

Lane, M.D., Morris, R.V., Christensen, P.R., 1999. Spectral behavior of hematite atvisible/ near infrared and midinfrared wavelengths. In: The Fifth InternationalConference on Mars. California Institute of Technology, Pasadena, 6085 pp.

Li, X., 2011. Mineralogical genesis investigation of two types of Cretaceous oceaniciron-bearing sediments. Unpublished PhD thesis. Nanjing University, p. 136.

Li, X., Cai, Y.F., Xue, J.Y., 2009. Colour genesis of red shales in cretaceous marine redbeds from Gyangze, southern Tibet. Acta Mineralogica Sinica 29, 283e290 (inChinese with English abstract).

Li, X., Hu, X.M., Cai, Y.F., Han, Z.Y., 2011. Quantitative analysis of iron oxide con-centrations within Aptian-Albian cyclic oceanic red beds in ODP Hole 1049C,North Atlantic. Sedimentary Geology 235, 91e99.

Liu, J.B., Aitchison, J.C., 2002. Upper Paleocene radiolarians from the Yamdrok mél-ange, south Xizang (Tibet), China. Micropaleontology 48 (Suppl. 1), 145e154.

Lyle, M., 1983. The brown-green color transition in marine sediments: a markerof the Fe (III)-Fe (II) redox boundary. Limnology and Oceanography 28,1026e1033.

Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by adithionate-citrate system buffered with sodium bicarbonate. Clays and ClayMinerals 7, 317e327.

Mix, A.C., Harris, S.E., Janecek, T.R., 1995. Estimating lithology from nonintrusivereflectance spectra: Leg 138. In: Proceedings of the Ocean Drilling Program,Initial Reports 138, College Station, TX, pp. 413e428.

Morris, R.V., Agresti, D.G., Lauer, H.V., Newcomb, J.A., Shelfer, T.D., Murali, A.V., 1989.Evidence for pigmentary hematite on Mars based on optical, magnetic, andMössbauer studies of super paramagnetic (nanocrystalline) hematite. Journal ofGeophysical Research 94, 2760e2778.

Morris, R., Lauer, H., Lawson, C., Gibson, E., Nace, G., Stewart, C., 1985. Spectral andother physicochemical properties of submicron powders of hematite (a-Fe2O3),maghemite (c-Fe2O3), magnetite (Fe3O4), goethite (a-FeOOH), and lepidocrocite(c-FeOOH). Journal of Geophysical Research 90, 3126e3144.

Norris, R.D., Kroon, D., Klaus, A., et al., 1998. Shipboard Scientific Party. In:Norris, R.D., Kroon, D., Klaus, A., et al. (Eds.), Proceedings of the Ocean DrillingProgram, Initial Reports 171B, College Station, TX (Ocean Drilling Program),pp. 1e749.

Potter, P.E., Maynard, B.J., Pryor, W.A., 1980. Sedimentology of Shale. Springer, Berlin,306 pp.

Research Systems, Inc., 2003. ENVI user’s guide. Research Systems, Inc., Boulder,Colo., 1084 pp.

Richter, N., Jarmer, T., Chabrillat, S., Oyonarte, C., Hostert, P., Kaufmann, H., 2009.Free iron oxide determination in Mediterranean soils using diffuse reflectancespectroscopy. Soil Science Society of America Journal 73 (1), 72e81.

Scheinost, A.C., Chavernas, A., Barrón, V., Torrent, J., 1998. Use and limitations ofsecond derivative diffuse reflectance spectroscopy in the visible to near-infrared range to identify and quantify Fe oxide minerals in soils. Clays andClay Minerals 46, 528e536.

Scheinost, A.C., Schwertmann, U., 1999. Colour identification of iron oxides andhydroxysulfates: Use and limitations. Soil Science Society of America Journal 63,1463e1471.

Shen, Z.X., Cao, J.Y., Li, X.X., Tomoaki, O., Wang, Y.Q., Zhang, X.Y., 2006. Mass con-centration and mineralogical characteristics of aerosol particles collected atDunhuang during ACE-Asia. Advances in Atmospheric Sciences 23, 291e298.

Shen, Z.X., Zhang, X.Y., Ji, J.F., Wang, Y.Q., Cao, J.J., Wang, D., Cheng, Y., 2004. Spec-troscopy analysis of iron-oxide minerals in aerosol particles in North China.Progress in Natural Science 14, 910e916 (in Chinese with English abstract).

Scott, R., 2009. Chronostratigraphic database for Upper Cretaceous oceanic red beds(CORBs). In: Hu, X.M., Wang, C.S., Scott, R.W., Wagreich, M., Jansa, L. (Eds.),Cretaceous Oceanic Redbeds: Stratigraphy, Composition, Origins, Paleogeo-graphic, and Paleoclimatic Significance. SEPM Special Publication 91, Society forSedimentary Geology, Tulsa, Oklahoma, U.S.A., pp. 31e53.

Torrent, J., Barrón, V., Liu, Q.S., 2006. Magnetic enhancement is linked to and pre-cedes hematite formation in aerobic soil. Geophysical Research Letters 33,L02401eL02404.

Torrent, J., Schwertmann, U., Fechter, H., Alferez, F., 1983. Quantitative relationshipsbetween soil colour and hematite content. Soil Science 136, 354e358.

Viscarra Rossel, R.A., Cattle, S.R., Ortega, A., Fouad, Y., 2009. In situ measurements ofsoil colour, mineral composition and clay content by vis-NIR spectroscopy.Geoderma 150 (3), 253e266.

Wagreich, M., Hu, X.M., Sageman, B., 2011. Causes of oxic-anoxic changes inCretaceous marine environments and their implications for Earth systems e Anintroduction. Sedimentary Geology 235, 1e4.

Wang, C.S., Li, X.H., Wan, X.Q., Tao, R., 2000. The Cretaceous in Gyangze, southernTibet. Acta Geologica Sinica 74, 97e107 (in Chinese with English abstract).

Wang, C.S., Hu, X.M., Sarti, M., Scott, R.W., Li, X.H., 2005. Upper Cretaceous oceanicred beds in southern Tibet: a major change from anoxic to oxic, deep-sea en-vironments. Cretaceous Research 2, 21e32.

Wang, C.S., Huang, Y.J., Hu, X.M., Li, X.H., 2004. Cretaceous oceanic red beds: im-plications for paleoclimatology and paleoceanography. Acta Geologica Sinica-English Edition 78 (3), 873e877.

X. Li, Y. Cai / Cretaceous Research 46 (2013) 257e266266

Wu, Y.Z., Chen, J., Wu, X.M., Tian, Q.J., Ji, J.F., 2005. Feasibility of reflectance spec-troscopy for the assessment of soil mercury contamination. EnvironmentalScience & Technology 39, 873e878.

Zhang, W.G., Yu, L.Z., Lu, M., 2003. Relationship between iron oxides and magneticproperties in intertidal sediments of the Yangtze estuary, China. Chinese Journalof Geophysics 46, 79e87.

Zhang, Y.G., Ji, J.F., Balsam, W.L., Liu, L.W., Chen, J., 2007. High resolution hematiteand goethite records from ODP 1143, South China Sea: Co-evolution ofmonsoonal precipitation and El Niño over the past 600,000 years. Earth andPlanetary Science Letters 264, 136e150.

Zhou, W., Chen, L.X., Zhou, M., Balsam, W.L., Liu, L.W., Chen, J., 2010. Thermalidentification of goethite in soils and sediments by diffuse reflectance spec-troscopy. Geoderma 155, 419e425.

Appendix A. Supplementary information

Supplementary information associated with this article can be found, in theonline version, at http://dx.doi.org/10.1016/j.cretres.2013.09.009.