Stable Carbon and Oxygen Isotopes in Shell Carbonatesof modern Land Snails in China and TheirRelation to Environment VariablesXu Wang1,2,3 , Jixuan Zhai1,2,3,4 , Linlin Cui3,5, Shuhua Zhang6, and Zhongli Ding1,2,3

1Key Laboratory of Cenozoic Geology and Environment, Institute of Geology and Geophysics, Chinese Academy ofSciences, Beijing, China, 2CAS Center for Excellence in Life and Paleoenvironment, Beijing, China, 3Innovation Academyfor Earth Science, CAS, 4University of Chinese Academy of Sciences, Beijing, China, 5Institute of Geology and Geophysics,Chinese Academy of Sciences, Beijing, China, 6Brigade 606 of Sichuan Metallurgical and Geological Exploration Bureau,Chengdu, China

Abstract Stable isotopes of land snail fossil shell carbonates (δ13Cshell and δ18Oshell) provide theopportunity to decipher paleodietary and paleoenvironmental changes. However, the environmentalmeanings of δ13Cshell and δ18Oshell remain elusive especially in the monsoonal climate region. To elucidatethis, δ13Cshell and δ18Oshell of the two most common land snails in China, Cathaica and Bradybaena,were analyzed in this study. Results show that Bradybaena has a wider distribution than Cathaica andprovides considerable potential for use in paleoenvironmental reconstruction. In addition, δ13Cshell ofBradybaena increases with altitude and temperature, decreases with mean annual precipitation, and reflectsthe isotopic signal of C3 plants. In contrast, Cathaica δ13Cshell documents a C3/C4 mixing signal and exhibitsonly a strong altitude effect. δ18Oshell of Bradybaena decreases with mean annual precipitation in theIndian monsoon and westerlies regions but increases with mean annual temperature in the East Asianmonsoon region; as a strong correlation exists between the two variables, they could thus potentially be usedto reconstruct paleotemperatures. However, uncertainties should be recognized due to the complexity ofthe land‐snails' ecophysiological responses to environmental changes. Cathaica has slightly lower δ18Oshell

than Bradybaena, which suggests that the two land snails have different ecophysiological traits, and theδ18Oshell of Cathaica follows mean annual temperature in northeast China. This study lays a ground forexplaining land snail stable isotopes in paleorecord in China and reasserts the importance of differentecophysiological traits of land snails in paleoclimatic research.

1. Introduction

Fossil land snails are of use as an archive material in paleoclimatic and paleoenvironmental studies becausethey are more sensitive to climatic changes than other nonbiogenic tissues. In addition, they are widelydistributed throughout the world and are very abundant in geological stratigraphy. Taxonomic analyses ofthe land snail assemblages have been extensively conducted and applied in paleoclimate reconstruction(Laurin & Rousseau, 1985; Moine et al., 2002; Wu et al., 1996, 2002, 2006; Wu & Wu, 2008, 2011), and withthe development of stable isotope techniques, the carbon and oxygen isotopic compositions of land snail shellcarbonate (δ13Cshell and δ18Oshell) have been analyzed to determine new proxies for providingpaleoenvironmental and paleoecological information. Studies from different regions have shown that bothδ13Cshell and δ18Oshell values can be used as dietary & environmental indicators (Goodfriend et al., 1989;Goodfriend & Ellis, 2002; Goodfriend & Magaritz, 1987; Magaritz & Heller, 1980; Milano et al., 2018; Yapp,1979). More recently, high‐resolution δ13Cshell and δ18Oshell analyses conducted along the growth band ofindividual fossil shells have been employed to track seasonal climate changes (e.g., Leng et al., 1998;Rangarajan et al., 2013), and the results have been found to be superior to those of other bulk geologicaland biogenic archives. However, as multiple factors influence the stable isotopic composition of land snailshell carbonate, the relationship between these isotopes and environmental variables remains ambiguous.

This paper provides a brief review of the current progresses made in land snail stable isotope studies. Asshown in previous studies, δ13Cshell is mainly governed by carbon isotope composition of snail intake, whichusually incorporates carbon from three sources: food (Baldini et al., 2007; Goodfriend & Ellis, 2002; Metrefet al., 2003; Stott, 2002; Yanes et al., 2008), atmospheric CO2 (Magaritz & Heller, 1980), and that of CO2

©2019. American Geophysical Union.All Rights Reserved.

RESEARCH ARTICLE10.1029/2019JG005255

Key Points:• Stable isotopes and snail

shell‐dimension studies suggestBradybaena and Cathaica havedifferent ecophysiological traits

• δ13Cshell is a tracer of dietaryvegetation and can be used tomonitor rainfall in the regiondominated by C3 plants

• δ18Oshell of Bradybaena rises withtemperature in East Asia monsoonregion but declines with rainfall inIndian monsoon and westerliesareas

Supporting Information:• Supporting Information S1

Correspondence to:X. Wang,xuking@mail.iggcas.ac.cn

Citation:Wang, X., Zhai, J., Cui, L., Zhang, S., &Ding, Z. (2019). Stable carbon andoxygen isotopes in shell carbonates ofmodern land snails in China and theirrelation to environment variables.Journal of Geophysical Research:Biogeosciences, 124,https://doi.org/10.1029/2019JG005255

Received 13 MAY 2019Accepted 23 SEP 2019Accepted article online 16 OCT 2019

WANG ET AL.

3356–3376.

Published online 8 NOV 2019

3356

produced by the reaction between carbonate absorbed by the snail and its gastric acid (Goodfriend & Hood,1983). To partition the different sources of carbon taken in by land snails, radiogenic 14C analyses have beenused in combination with stable isotope measurements. In this respect, Goodfriend and Hood (1983)conducted 13C and 14C analyses on modern land snails in Jamaica and estimated that only 25–40% of shellcarbonate originates from a plant diet but up to 33% originates from limestone. In contrast, a similar studyof land snails from the Chinese Loess Plateau found that the carbon derived from an organic diet accountedfor more than 70% of the total shell carbon (Xu et al., 2010). Furthermore, a comprehensive examination ofthe 14C content of the most common small land snails from habitats on carbonate terrain in North America,which included 247 accelerator mass spectrometry 14C measurements of modern shell materials from 46different species, revealed that ~78% of the shell aliquots did not contain dead carbon from limestone or othercarbonate rocks, even though it was readily available at all sites, and only 3% of the total containedmore than10% dead carbon (Pigati et al., 2010).

Other research has used snail culturing studies to decipher the controlling factor of δ13Cshell. Land snailculture experiments have revealed that plant dietary carbon accounts for almost 100% of shell carbon andthat if the snail is fed an added CaCO3 source, the ingested carbonate makes no contribution to δ13Cshell

values (Metref et al., 2003; Stott, 2002). However, the results of a more recent study showed that land snailsingested a considerable amount of carbonate (up to 26%) and that the corn‐fed (C4 plants) groups ingestedmore carbonate than the cabbage‐fed (C3 plant) groups had, which suggests the potential effect of a foodpreference on δ13Cshell (Zhang, Yamada, et al., 2014).

Despite the complexity of carbon sources ingested by snails, stable isotopic studies conducted onfield‐collected snails from regions that have differing vegetation over a large geographical range have foundthat dietary δ13C is the primary variable affecting δ13Cshell (Colonese et al., 2014; Liu, Gu, et al., 2007;Prendergast et al., 2017; Stott, 2002; Yanes et al., 2009). In the Chinese Loess Plateau, Liu, Gu, et al. (2007)analyzed δ13Cshell of land snails (mainly Cathaica pulveratricula, Cathaica pulveratrix, and Cathaica

gansuica) and recognized that δ13Cshell can reflect the isotopic characteristics of the food and the potentialecological background. This means that δ13Cshell can be used to infer the relative abundance of C3/C4 plantsin a region with C3‐C4 mixed vegetation and can also reveal the environmental effects on C3 plants, such aswater stress and temperature, in regions dominated by C3 vegetation (i.e., Leng & Lewis, 2016; Prendergastet al., 2017). For example, in north China, the δ13C of C3 plant shows a negative correlation withprecipitation (Liu et al., 2005; Wang et al., 2003; Wang et al., 2008) but a positive correlation with tempera-ture (Wang et al., 2013). However, such environmental effects have seldom been examined on δ13Cshell overthe entire geographical range of China.

The influences of environmental variables on δ18Oshell are rather complex, and a variety of controlling

factors have been determined globally, even though δ18Oshell is theoretically governed by δ18O of snail bodywater and the temperature at which shell carbonate precipitates. In this respect, a number of studies indifferent parts of the world have linked δ18Oshell to rain water δ18O (the δ18O of precipitation, δ18OP)and/or the rainfall amount. For example, Leng et al. (1998) conducted a high‐resolution stable isotopeanalysis along the growth band of the shell of a land snail collected in Ethiopia and proposed that seasonalchanges in the amount and isotopic composition of rainfall accounted for the shape and amplitude ofδ18Oshell cycles. In some parts of Europe, the northern part of Africa (the circum‐Mediterranean region),and North America, a strong correlation has been found to exist ubiquitously between δ18Oshell and δ18OP;however, the regression lines constructed in studies have highly variable slopes with respect to the differentregion studied (Colonese et al., 2014; Goodfriend et al., 1989; Goodfriend & Ellis, 2002; Lécolle, 1985;Prendergast et al., 2015; Zanchetta et al., 2005).

Research conducted on the Chinese Loess Plateau proposed that the δ18Oshell of land snails (mainlyC. pulveratricula, C. pulveratrix, and C. gansuica) is mainly controlled by the δ18O of summer monsoonprecipitation and that it exhibits a negative correlation with summer rainfall amount (Liu et al., 2006). Ithas also been proposed that relative humidity (RH) controls δ18Oshell. For instance, Yapp (1979) performeda pioneering stable isotope study onmodern land snail shells in North America and found that enrichment ofδ18Oshell appears to be linearly related to the reciprocal of local RH. Another study conducted on the GreatPlain of North America further confirmed that RH at the time of snail activity is an important control on

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3357

δ18Oshell (Balakrishnan et al., 2005). Following these results, Zaarur et al. (2011) reinforced the key role of RHand local evaporation in modulating δ18Oshell variations based on studies conducted on stable and clumpedisotopes in land snail shells. However, no clear correlation between δ18Oshell and RHwas found in a seasonalstable isotope study conducted on modern land snails on the coastal plain of Israel, but the results showedthat δ18Oshell was well correlated with mean annual temperature (MAT; Magaritz et al., 1981). Similarly,Yanes et al. (2009) also observed a strong correlation between δ18Oshell and MAT on the low‐latitude oceanicisland of Tenerife in Spain. This influence of temperature on δ18Oshell has been also employed to explain thenegative excursion of δ18Oshell during the Older Dryas (ca. 13,857 ~ 14,171 years calibrated before the presentB.P.) at Grotta del Romito in southern Italy (Colonese et al., 2007). The effect of water vapor has also beenfound to be prominent, particularly in some arid lands (e.g., Colonese et al., 2014; Hassan, 2015).

In summary, the main controlling factors relating to the isotopic compositions of land snail shell carbonatemay differ in differing regions due to discrepancies in both climatic seasonality and the ecophysiologicaltraits of land snails. Therefore, it is necessary to conduct a systematic study of typical land snails in diverseenvironments, particularly in monsoonal China where associated research is scarce, with the ultimate aim ofusing δ13Cshell and δ18Oshell of fossil land snails to probe past environmental changes in this region.

In brief, the current studies on the stable isotopes of modern land snails in China have only been con-ducted within the Chinese Loess Plateau. Although δ13Cshell and δ18Oshell are believed to correlate respec-tively with dietary δ13C and rainwater δ18O on a local scale, it is currently unclear which environmentalvariables have the most important role in affecting stable isotope changes within the shells of land snailswithin the vast area of monsoonal China. To conduct such a systematic study over a broad region ofChina, we collected samples from the two most common land snail genera (Bradybaena and Cathaica)within different climatic zones in China and measured the stable carbon and oxygen isotopes in snailshells. Based on modern climate data and oxygen isotopes of precipitation at the studied sites, we then con-ducted statistical analyses with the aim of determining the key climatic factors affecting the carbon andoxygen isotopic compositions of modern snail shells. Moreover, carbon isotopes (δ13CSBOM) were measuredon shell‐bound organic matter, which includes the periostracum and both the acid soluble and insolubleorganic matrix of the shell, for the both species at selected sites, with the aim of determining the carbonisotopic fractionation (CIF) of shell carbonate relative to diet and evaluating any difference between theCIF in Bradybaena and Cathaica land snails. This study aims to assist in precisely reconstructing dietaryvegetation using the measured δ13Cshell. It is considered that the main findings of this study will facilitatea better understanding of stable isotopes in land snail shell carbonates to enable environmental interpreta-tions in monsoonal regions.

2. Materials and Methods2.1. Sample Collection

We collected living land snails from 69 sites in China between the years 2013 and 2017 (Figure 1). Fewer sam-pling sites were used in western China because of the lack of land snails found during sample collection; thiswas considered to be due to the low amount of annual rainfall (<200 mm) in the region. During the samplingperiod, Bradybaena was collected from a total of 64 sites and Cathaica from 27 sites. The Bradybaena landsnail species included Bradybaena (Acusta) ravida ravida, Bradybaena (B.) similaris similaris, Bradybaena(B.) fuchsia, and Bradybaena (Acusta) ravida sieboldiana, and the Cathaica snail species included Cathaicacunlunensis, C. pulveratricula, C. (Pliocathaica) pulveratrix, Cathaica (Xerocathaica) pekinensis, Cathaica(C.) fasciola fasciola, C. gansuica, and Cathaica sp. Photos of some of the selected species are shown inSupporting Information Figure S1. Information on the land snail species at each study site are listed inTable S1. The land snails were collected from greenlands and parks in rural area or on the outskirts of a citiesfar from obvious human disturbances. The vegetation at the sampling sites was either trees or shrubs withgrass species as the understory. Trees were mostly conifer, willow, and poplar, and shrubs were predomi-nantly Loropetalum chinense var. rubrum, Symplocaceae Pittosporum, Nanshan Berberidaceae bamboo,Schefflera heptaphylla, and Photinia serrulata. Most of plants were classified as being C3 plants. Data from16 sites reported by Liu et al. (2006), Liu, Gu, et al. (2007) were also incorporated, and in this respect, 1and 16 extra data were included for Bradybaena and Cathaica, respectively. Collectively, 23 sites containedboth Bradybaena and Cathaica land snails.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3358

With respect to physioecological and/or growth‐seasonal differences, the two land snail species havedifferent geographic distribution ranges, for example, Bradybaena is found in a broad area from southto north China, whereas Cathaica is mainly confined to north China. The study sites follow a similar geo-graphical distribution in that they incorporate both inland and coastal areas. Table S1 summarizes thegeographical, topographical, and environmental data of these sites. Bradybaena were sampled over anarea distributed between 18.26–45.80°N (latitude) and 87.62–126.53°E (longitude) at an altitude between3 and 1,679 m, with a MAT of 1.27–25.15 °C, mean annual precipitation (MAP) of 88.40–2128.8 mm,and mean annual relative humidity (MARH) of 41.42–82.17%. In contrast, Cathaica were sampled froma narrower geographical range (31.89–41.78°N and 103.82–124.33°E), which has smaller ranges of MAT(6.35–15.97 °C), MAP (155–1000.80 mm), and MARH (41.42–74.83%) but larger topographicalvariations (3–1982m).

Figure 1. (a) Schematic map shows locations of sampling sites for two modern land snail genera in this study.Sampling sites for isotopic data cited from Liu et al. (2006), Liu, Gu, et al. (2007) are also shown (include two Cathaicasites of Xi'an and Zhengzhou); (b) monthly distribution of δ18OP and environmental variables inWulumuqi, Puer, Beijing,and Sanya.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3359

2.2. Chemical Pretreatments

The shell was first cleaned in distilled water using a toothbrush to remove large particles of soil and debrisand then further cleaned using an ultrasonic utility to remove surficial sediments and organic fluid stuckto the shell surface. The cleaned shell was dried at 60 °C overnight and ground into fine powder with a pestleand mortar. Some of the selected fine powder was tested for carbonate polymorphs by X‐ray diffraction usinga PANAlytical diffractometer with Ni‐filtered Cu‐Kα radiation (40 kV and 40 mA) and aragonite was identi-fied (Figure S2). The fine aragonite powder was reacted with 3% H2O2 at room temperature for 24 hr tofurther remove organic substances occluded in the shell carbonate, and it was then centrifuged and rinsedthree times. The residue was finally dried overnight at 60 °C.

The shell‐bound organic matter was extracted following the method developed by Yanes et al. (2018). Briefly,an aliquot of homogenized shell powder from multiple shells (i.e., from 4–10 shells, depending on availabil-ity) from each study site was placed into a beaker with 2M HCl solution at room temperature. After the shellcarbonate had fully dissolved following a reaction time of 12 hr, the released organic matter was centrifugedat 4,000 rpm for 10 min and then rinsed with deionized water. Following centrifugation three times and rin-sing, the organic matter was subsequently recovered as a pellet, and the pellet residue was oven dried at 40 °C overnight.

2.3. Carbon and Oxygen Isotopic Analyses

The δ13C and δ18O of pretreated shell carbonate powders were measured using a GasBench II linked to DeltaV plus isotope ratio mass spectrometer (Thermo Fisher). Sample powders weighing approximately 100 μgwere reacted with 100% H3PO4 at 72 °C for 1 hr. The resultant CO2 was purified by passing through twoNafion™ water traps and a PoraPlot Q chromatograph column (maintained at 45 °C) and it was then intro-duced to a mass spectrometer for isotopic measurements. The δ13C of shell‐bound organic matter (δ13CSBOM)was determined using an elemental analyzer (Flash EA 1112) coupled with a MAT 253 isotope ratio massspectrometer (Thermo Fisher). Stable isotope results were reported relative to the Vienna Pee DeeBelemnite (VPDB) standards. As the stable isotope standards (GBW4416: δ18OVPDB = ‐ 11.59‰; andNBS19: δ18OVPDB = ‐2.20‰) employed for calibration are calcite, which has a different oxygen isotope acidfractionation factor from our snail shell aragonite (Kim et al., 2007), a correction factor of 0.38‰was appliedto generate δ18Oshell data. External analytical precisions of 0.06‰ and 0.10‰ were obtained for δ13C andδ18O of carbonate, respectively, and better than 0.2‰ for δ13CSBOM.

2.4. Examining Representativeness of Multiple‐Shell Samples

Previous studies have shown apparent variabilities in δ13C and δ18O between individual shells collected fromthe same site. It was thus decided to conduct isotopic analyses of multiple shells from each site to obtain therepresentative average condition for the population (e.g., Goodfriend & Ellis, 2002; Liu et al., 2006). In mostcases, 10 shells from each location were combined to make one sample (referred to as multiple shell) to com-pile a representative sample and to approximate the average conditions under which the land snails existed.As it was difficult to find snails at some sites, a minimum number of four shells were employed. A total of 38multiple‐shell samples of Bradybaena were analyzed and 24 of Cathaica.

To examine the representativeness of our multiple‐shell samples, we also randomly selected 10 extra indivi-dual shells at two sites (Sanmenxia and Dalian) for Cathaica and four sites (Sanmenxia, Dalian, Yinchuan,and Ningbo) for Bradybaena (although only five shells of Bradybaena were taken from Dalian) to conductδ13C and δ18O analyses. The mean δ13C and δ18O for those individual shells from each of the selected siteswere then compared with the isotopic data of the multiple‐shell sample from the same site.

2.5. Shell Size Measurements

To compare the shell sizes of Bradybaena and Cathaica, we measured the shell width and height, the relativeaperture area, and the number of whorls on each shell of the two species at the same site. The shell width wasmeasured as the maximum length of the base body whorl (“1” in Figure S3a) and the shell height as the dis-tance from the top to the base of the shell (“2” in Figure S3a). The relative aperture area was the area of theshell mouth, which was calculated approximately based on an ellipse shape (S = πab), where a or b refers tothe aperture width or aperture height, respectively (“3” or “4” in Figure S3a). Furthermore, the number of

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3360

whorls relates to the number of shell rings that extend along the central columellae (e.g., the snail in FigureS3b has five whorls).

2.6. Data Source for Environmental Variables and Oxygen Isotope of Precipitation (δ18OP)

Instrumental data of environmental variables for each sampling site were obtained from the ChinaMeteorological Data Service Center; these data represented environmental variables monitored daily at localmeteorological observation stations. We used averaged annual or monthly data over the past 30 years (from1981 to 2010).

For δ18OP, we used actual measured rainfall isotope data for available sites collected by the Global Networkof Isotopes in Precipitation (https://gnip.com/) but used predicted data for other sites by employing theonline isotope in precipitation calculator (at wateriso.utah.edu/waterisotopes/) provided by Bowen (2008).

3. Results3.1. Comparison Between Isotopic Composition of Individual Shells and Multiple‐Shell FromSame Sites

As shown in Figure S4, the intrapopulation variation was between 1.5‰ and 3.5‰ for both δ13Cshell andδ18Oshell at the same site, except for Cathaica snails from Sanmenxia, which had variability of up to 5.5‰for δ18Oshell. Although there were relatively large variations in δ13C and δ18O among individual shells fromeach site, the mean δ13C and δ18O of those shells were almost the same as those of the multiple‐shell samplefrom the same site (the largest difference was <0.3‰). These results show that the δ13C and δ18O values ofour multiple‐shell sample can be used to represent the average condition gained from discrete measurementson several individual shells. Therefore, all the δ13Cshell and δ18Oshell data discussed in this study wereobtained from multiple‐shell samples.

3.2. Carbon Isotopes of Land Snail Shell

All measured data are shown in Table S1. The ranges of δ13Cshell values for Bradybaena and Cathaica are ‐14.16‰ to ‐9.52‰ (amplitude: 4.64‰) and ‐13.03‰ to ‐5.21‰ (amplitude: 7.82‰), respectively, and themean δ13Cshell value for Bradybaena is ‐11.31‰while that for Cathaica is ‐10.42‰. When the δ13Cshell valuesof both species from the same sites are put together in a plot, 14 out of 22 δ13Cshell values of Cathaica arehigher than those of Bradybaena, with a mean offset of 0.43‰ (Figures 2a and 2b). The t test for theδ13Cshell of the two genera was conducted using SPSS software, and the δ

13Cshell values of Cathaica are statis-tically indistinguishable from those of Bradybaena (Table S2).

Spatially, the δ13Cshell values of Bradybaena at sites east of 112°E exhibit a V‐shaped changing pattern by lati-tude (Figure 3a), for example, δ13Cshell becomes more positive from the southeast to the northwest. However,in regions north of 40°N, such as Shenyang (41.78°N, 123.40°E), δ13Cshell values are more negative than thosein the 31°–40°N region. In contrast, this trend is not obvious for δ13Cshell values of Cathaica (Figure 3b);instead, δ13Cshell values of Cathaica show an increasingly positive trend from coastal areas to inland areas.However, when we plot δ13Cshell against latitude, the V‐shaped changing pattern is more evident for bothBradybaena and Cathaica. It should be noted that rainfall amount (MAP) displays the same latitudinalchange (Figure S5).

A correlation analysis was conducted to examine the relationship of δ13Cshell with certain environmental orgeographic variables. Results show that δ13Cshell of Bradybaena and Cathaica both correlate with MAP,MAT, and altitude (Alt; Figure 4). In addition, δ13Cshell values for both Bradybaena and Cathaica show a gen-erally decreasing trend with increases in MAP and MAT (Figures 4a, 4b, 4c, and 4d), but the correlations arerelatively weak between δ13Cshell and MAP for Cathaica and between δ13Cshell and MAT for Bradybaena(Figures 4b and 4c). In addition, δ13Cshell values of both land snail species display a positive correlation withaltitude, although the correlation is a little stronger for Cathaica (Figures 4e and 5f). Furthermore, δ13Cshell

of Bradybaena also negatively correlates with MARH whereas that of Cathaica does not (Table S3).

δ13CSBOM from four sites varies from ‐24.64‰ to ‐28.09‰ for Bradybaena and from ‐24.00‰ to ‐26.78‰ forCathaica (Table S4). Meanwhile, a strong positive correlation is observed between δ13CSBOM and its corre-sponding δ13Cshell for both species (Figure S6a): the slopes of both regression lines are close to 1 (i.e., 1.048

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3361

for Bradybaena and 0.991 for Cathaica). There is a mean 13C enrichment factor of +15.18 ± 0.26‰ forBradybaena δ13Cshell with respect to its δ13CSBOM, but Cathaica has a relatively low 13C enrichment factorof +13.89 ± 0.15‰.

3.3. Oxygen Isotope of Land Snail Shell

The δ18Oshell values for Bradybaena and Cathaica range from ‐7.02‰ to 0.45‰ (amplitude = 7.47‰) andfrom ‐7.65‰ to ‐1.98‰ (amplitude = 5.67‰), respectively. The mean δ18Oshell value of Bradybaena is ‐

2.99‰, which is less negative than that of Cathaica (‐4.55‰). For snails collected from the same sites, 16out of 23 Bradybaena show higher δ18Oshell values than Cathaica, with a mean offset of 0.97‰(Figure 2b). A t test for the δ18Oshell of both species was conducted using SPSS software, and results demon-strate that δ18Oshell values of Cathaica are statistically more negative than those of Bradybaena (Table S2).Geographically, the δ18Oshell for Bradybaena presents a more complicated changing pattern: the values aremore negative in the southwest, which is the area controlled by the Indian Monsoon; less negative valuesare found in the northwest where the westerlies dominates; and the shift becomes more negative from south-east to northwest in the East Asian monsoon region (Figure 5a). In comparison, δ18Oshell values for Cathaicabecome more negative towards inland regions (from east to west) but are less negative furtherinland (Figure 5b).

With respect to the relationship with environmental variables, an overall positive correlation can be foundbetween temperature (MAT) and δ18Oshell of Bradybaena in the East Asian Monsoon region (R2 = 0.26, P

Figure 2. Comparison of stable isotopes between Cathaica and Bradybaena at the same sampling sites. (a) δ13Cshell com-parison between the two genera; (b) δ18Oshell comparison between the two genera.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3362

< 0.001, and n= 56), and the correlation is obviously improved when five deviated data from the big lake area(triangle data) or coastal area (square data) are eliminated (Figure 6a; R2 = 0.47, P < 0.001, and n = 51).Moreover, in coastal areas, there is a strong and convincing negative correlation between precipitation andδ18Oshell of Bradybaena (Figure 6b; R2 = 0.67). In the westerlies and Indian monsoon regions, the δ18Oshell

values of Bradybaena exhibit a significantly negative correlation with MAP (Figure 6c, R2 = 0.96; andFigure 6d, R2 = 0.88, respectively). In contrast, the correlation between Cathaica δ18Oshell and either MATor MAP has no statistical meaning, despite δ18Oshell of Cathaica showing a slight correlation with δ18OP forall the study sites (Table S3). In addition, the positive correlation between Bradybaena δ18Oshell andtemperature is more robust in northeast China, where temperatures are usually low throughout the year,for example, MAT is as low as 5.0 °C in Harbin, which is situated in the northernmost (R2 = 0.84; FigureS7a). In addition, the δ18Oshell values forCathaica also show a significantly positive correlation with tempera-ture in this region (R2 = 0.74; Figure S7b).

4. Discussion4.1. Contrasting Ecophysiological Habits of the Two Land Snail Species

The lifestyle and habits of the Bradybaena genus have been extensively studied through field observationsand culturing experiments, and our previous study has thoroughly summarized the findings in this respect

Figure 3. Contour maps showing spatial changing pattern of (a) δ13Cshell of Bradybaena and (b) δ13Cshell of Cathaica.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3363

(Wang et al., 2016). We thus briefly introduce the ecophysiological habits of the land snails. Bradybaenagenerally lives for up to 2 years, and its optimal environmental conditions are a temperature of 10–30 °Cand RH higher than 66% (the most optimal is ~90%). In this respect, the growing season for Bradybaena

Figure 4. Plots showing (a) changes in δ13Cshell of Bradybaena with mean annual precipitation (MAP), (b) changes inδ13Cshell of Cathaica with MAP, (c) changes in δ13Cshell of Bradybaena with mean annual temperature, (d) changes inδ13Cshell of Cathaicawith mean annual temperature, (e) changes in δ13Cshell of Bradybaenawith altitude, and (f) changesin δ13Cshell of Cathaica with altitude. Red dash line shows the trend of decrease in δ13Cshell of Bradybaena when MAPis less than 800 mm whereas green dash line indicates the slope of δ13Cshell decrease when MAP is higher than 800 mm.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3364

land snails is from April to October in many parts of China, and they may hibernate in July and August whentemperature becomes too high.

In contrast, studies of the lifestyle of Cathaica are rather scarce. Nevertheless, an appropriate temperature forthis genus is 10–30 °C, and Cathaica land snails usually hibernate in winter time (from mid October to earlyMarch), but they seldom hibernate in summer. This may suggest that Cathaica is tolerant to relatively hightemperatures. Cathaica live in groups on the trees, and they crouch on the trunk or branches to ingest food(Yang & Ren, 2012). In addition, our previous studies showed that the clumped isotope temperatures ofCathaica were 3–5 °C higher than those of Bradybaena at the same study sites (Wang et al., 2016; Zhaiet al., 2019). This is consistent with Cathaica's tolerance to relatively high temperature and may also revealthat the two genera have different ecophysiological habits. For example, Cathaicamay be more active (thanBradybaena) in the summer and therefore register a greater summer signal.

Shell form and sizes can offer clues about the ecophysiological habits of land snails. Previous studies haveshown that larger snails are often associated with moister conditions, and the relative aperture area tends

Figure 5. Contour maps showing spatial changing pattern of (a) the δ18Oshell of Bradybaena and (b) the δ18Oshell ofCathaica.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3365

to be smaller under drier conditions (Goodfriend, 1986 and references therein). The larger snail sizeassociated with a moister environment is commonly explained by an increased food intake and highergrowth rate under such conditions (Gould, 1984; Sacchi, 1965). Moreover, several studies have relatedlarger snail shells with warmer conditions, although this has not been universally confirmed (Bengtson

Figure 6. Plots showing changes in δ18Oshell of two genera along with temperature and precipitation. (a) Changes in δ18Oshell of Bradybaena with mean annualtemperature (MAT) in the East Asia monsoon region except for some coastal sites (square data) and near lake sites (triangle data). (b) Changes in δ18Oshell ofBradybaenawith mean annual precipitation (MAP) in coastal area of the East Asia monsoon region. (c) Changes in δ18Oshell of Bradybaenawith MAP in westerliesregion. (d) Changes in δ18Oshell of Bradybaena with MAP in India monsoon region.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3366

et al., 1979; Goodfriend, 1986); therefore, we measured the shell sizes of Cathaica and Bradybaena at thesame sites in this study.

Our results showed that Bradybaena is overwhelmingly larger than Cathaica, with respect to shell height,width, and relative aperture area (Figures S8a, S8b, and S8c), but Cathaica tends to have more number ofwhorls than Bradybaena (Figure S8d). The observed smaller size of Cathaica than that of Bradybaena couldbe attributed to certain factors, such as population density, insolation, and ecophysiological traits. For exam-ple, previous studies have observed a decreasing shell size trend with an increase in population density forsome European land snail species (table 1 in Goodfriend, 1986). During snail collection in the field, wenoticed that Cathaica were usually more abundant than Bradybaena at the same site, and Cathaica tendsto be gregarious land snails that usually live in groups, whereas Bradybaena snails are sporadically distribu-ted throughout their habitat. In the summer months of some years, thickly dotted Cathaica snails have beenseen crawling on the trunks of fruit trees and eating leaves in north China, which damages fruit trees andgreatly reduces fruit production (Yang & Ren, 2012). In addition, a previous study found that smaller snailshad better survival rates than larger snails under unshaded conditions (Knights, 1979). It is also consideredthat the relatively large whorl number of Cathaicamay manifest its high growth rate. This needs to be inves-tigated in further study.

From a perspective of stable isotopes, there are no statistical differences between δ13Cshell values of Cathaicaand Bradybaena at the same sites (Figure 2a), which suggests that the two have the same food source.However, Cathaica have generally lower δ18Oshell than Bradybaena at the same study sites in north China(Figure 2b and Table S2). As the two genera were collected from the same sites, they would experience thesame climatic and environmental conditions; therefore, they may use the same water sources, such as pre-cipitation, dew water, and plant water, if so, the oxygen isotopes in their ingested waters would be similar.Therefore, the different δ18Oshell values of the two genera may indicate differing ecophysiological traits.The relative small aperture area of Cathaica shells (Figure S8c) may effectively reduce body water loss andkeep the oxygen isotope from shifting to more positive during the evaporative process. In our previous study,it was found that the clumped isotope temperatures of Cathaica were 3–5 °C higher than those ofBradybaena at the same study sites, which means that Cathaica snails precipitate their shells at higher tem-peratures, and this drives δ18Oshell towards more negative values due to a reduction in oxygen isotope frac-tionation between shell carbonate and body water at elevated temperatures. This discovery deserves furtherinvestigation in future studies.

Because of the large variability in climatic seasonality in China, the length of the growing season for landsnails differs in different regions. For example, the length of snail activity differs by approximately 40days (i.e., from 202 to 162 days) between the eastern and western parts of the Chinese Loess Plateau(Huang et al., 2012). Therefore, the exact growing season for land snail at each study site still needsfurther investigation. In this study, we compared δ13Cshell and δ18Oshell with annual instrumental data.This may incur some uncertainties in calculation but provide baseline relations of stable isotopes toenvironmental variables. Moreover, it allows for direct comparisons with many previous studies, whichused annual averages of environmental variables to build relationships with stable isotopes in land snailshells and plants.

4.2. Carbon Isotopes in Shell Carbonate of Land Snails4.2.1. Spatial Changes in δ13Cshell of Land Snails in China and Their Linkages to VegetationAs aforementioned, both culturing and field studies have shown that the carbon isotope composition of theshell carbonate of land snails is mainly related to the assimilation of plant‐based food resources(Balakrishnan & Yapp, 2004; Francey, 1983; Goodfriend & Ellis, 2002; McConnaughey et al., 1997;McConnaughey & Gillikin, 2008; Metref et al., 2003; Stott, 2002). Based on the photosynthetic pathway,plants can bemainly classified as C3 and C4, and the two groups have radically different carbon isotopic com-positions: with the δ13C values of C3 plants ranging from ‐33‰ to ‐21‰ and those of C4 plants are typicallyfrom ‐17‰ to ‐9‰ (Cerling & Quade, 1993; Jacobs et al., 1999). When land snails digest and utilize carbonfrom differing plant types, the δ13Cshell registers different isotopic signals. In this respect, δ13Cshell can beemployed to reflect the composition of the local vegetation (i.e., C3 vs. C4), and this approach has been widelyapplied to reconstruct past vegetation types in a variety environments throughout the world (Leone et al.,2000; Goodfriend & Ellis, 2000, 2002; Balakrishnan et al., 2005; Baldini et al., 2007; Yanes et al., 2008, 2009).

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3367

To retrieve knowledge of the dietary vegetation consumed by our studied Bradybaena and Cathaica, it isnecessary to know the CIF of δ13Cshell relative to δ13C within plants ingested by the land snails. As snailshell‐bound organic matter is considered to mainly be derived from ingested plant material (Burleigh &Kernney, 1982; DeNiro & Epstein, 1978), it is possible to estimate the CIF based on the difference betweenδ13Cshell and δ13CSBOM (namely the 13C enrichment factor in the shell carbonate relative to shell‐boundorganic matter). The 13C enrichment factors obtained in this study are +15.18 ± 0.26‰ and +13.89 ±0.15‰ for Bradybaena and Cathaica, respectively (Table S4). The value of +13.89‰ for our Cathaica isvery close to the previously published enrichment factor of +14.2‰ in δ13Cshell relative to δ13C of snailbody (δ13Cbody) for living land snails (mainly Cathaica family) in north China (Liu, Gu, et al., 2007).This means that the shell organic matter and the snail body have almost the same CIF with respect toshell carbonate. Yanes et al. (2018) also observed little difference in offsets of δ13Cshell ‐ δ

13CSBOM andδ13Cshell ‐ δ

13Cbody for the snail, Neohelix, in North Carolina (USA), and their CIF values were +14.1‰and +14.5‰, respectively. A similar positive offset of 14.5‰ was determined for δ13Cshell relative toδ13Cbody of Helix melanostoma in the eastern Mediterranean (Prendergast et al., 2017). In addition, theregression line between δ13Cshell and δ13CSBOM/δ

13Cbody for Cathaica in this study and that of Liu,Gu, et al. (2007) are indistinguishable from that for Helix aspersa developed in a culturing study (Stott,2002). In contrast, however, our regression line for Bradybaena shows an apparent deviation and has ahigher intercept (Figure S6b); this is possibly attributed to the larger CIF at a lower temperature becauseour previous study exhibited that the calcification temperature of Bradybaena was 3 ~ 5 °C lower thanthat of Cathaica (Wang et al., 2016; Zhai et al., 2019). Similar changes in shell CIF with temperature wereobserved in a recent land snail culturing experiment (Zhang, Yamada, et al., 2014). Based on the newlyobtained CIFs for the two genera determined here, the calculated δ13C of snail organic tissues is‐28.29‰ to ‐24.74‰ for Bradybaena snails and ‐26.82‰ to ‐19.10‰ for Cathaica snails. The inferredδ13C of snail organic tissues for the two species indicate that these land snails generally have C3 plantsas their main food source. However, some Cathaica snails, which have more positive inferred dietaryδ13C values (of around ‐19‰) may also ingest a certain amount of C4 plants.

As shown in Figures 3 and S5, δ13Cshell values of Bradybaena at sites east of 112°E in China show a latitudinalV‐shaped changing pattern, with values that are more positive located between 31°–40°N. Many studies havedetermined that the region of 31°–40°N in China is very suitable for the growth of C4 plants and provides amaximum C4 biomass (e.g., Rao et al., 2008; Wang et al., 2000). As a result, this region has the least negativeδ13C values for soil organic matters, while the other regions (i.e., to the north or south) have δ13C values thatare more negative. However, since the inferred dietary δ13C values for Bradybaena, which fall in the range ofcarbon isotope composition of C3 plants, show the same spatial pattern as the rainfall amount, it could beinferred that the V‐shaped change in the δ13Cshell values of Bradybaena may not reflect the spatial changeof C3/C4 vegetation but instead reflect the effect of water stress on C3 vegetation along a latitudinal transect.An investigation of MAP data shows that the region of 31°–40°N is quite arid, and high aridity causes plantsto close their stomata to conserve water for photosynthesis, which in turn reduces the intercellular partialpressures of CO2 and increases plant δ13C values accordingly (Liu et al., 2005). However, at the sites westof 112°E and with latitude between 24° and 30°N, the δ13Cshell values of Bradybaena have rather positivevalues due to the high elevations (see the altitude effect referred to in the following section). In comparison,the spatial changes in the δ13Cshell values of Cathaica possibly results from both water stress and latitudinaldistribution of C3/C4 vegetation. Previous studies have recognized that δ13Cshell can be used to infer the rela-tive abundance of C3/C4 plants in the regions with C3‐C4 mixed vegetation (i.e., Balakrishnan et al., 2005;Baldini et al., 2007; Goodfriend & Ellis, 2000, 2002; Leng & Lewis, 2016; Yanes et al., 2008; Yanes et al.,2009) or to estimate plants in C3‐crassulacean acid metabolism mixed environments (Yanes et al., 2013).However, our study implies that the ecophysiological traits of land snails (e.g., Cathaica registers a greatersummer signal than Bradybaena) may play important roles in determining the composition of their dietaryvegetation in a C3‐C4mixed region. This fact therefore requires consideration when using δ13Cshell to retrievethe C3/C4 biomass of the geological past.4.2.2. Potential Environmental Factors Affecting δ13Cshell of Land SnailsOur results show that the δ13Cshell values of both Bradybaena and Cathaica generally decrease with anincrease in MAP (Figures 4a and 4b). As discussed above, this relation could reflect the negative responseof the carbon isotope composition of C3 plants to precipitation because the main dietary intake of the two

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3368

genera is C3 plants. This assumption also agrees with the observed increase in the δ13C values of C3 plantswith a decrease in MAP in north China, which is associated with carbon isotope discrimination relating tostomata closure under water stress (Liu et al., 2005; Wang et al., 2003). We also noted that the decrease ratein δ13Cshell of Bradybaena changed at 800‐mm MAP, which is a critical precipitation isoline point used todivide humid and semihumid zones in monsoonal regions (Figure 4a). The δ13Cshell of Bradybaena reducedat ‐0.33‰/100 mm when MAP was less than 800 mm (red dash line in Figure 4a) and was close to ‐0.50‰/100 mm for δ13Cshell of Cathaica at most sampling sites in north China (Figure 4b). These two rates are com-parable with the sensitivity of ‐0.4‰/100 mm for C3 plants in response to precipitation in north China (Wanget al., 2008). WhenMAPwas higher than 800mm, the δ13Cshell of Bradybaena diminished at ‐0.12‰/100mm(green dashed line in Figure 4a). This means that the carbon isotope composition of C3 plants has a less sen-sitive response to precipitation changes in the humid climatic zone (precipitation > 800 mm/year). The shal-lower slope at higher rainfall levels is in accordance with the changing pattern of C3 plant δ13C withprecipitation on a global scale (Kohn, 2010). The inference that water stress (water use efficiency) in plantsis recorded in snails (determined using snail δ13Cshell) has previously been confined to only C3‐plant domi-nated landscapes (e.g., Colonese et al., 2007; Prendergast et al., 2017; Yanes, 2015). We extended it to differ-ent landscapes in this study.

We also observed a negative correlation between δ13Cshell and temperature for both Bradybaena andCathaica (Figures 4c and 4d). However, a recent study has shown a strong positive relationship betweenδ13C of C3 plants and MAT in north China with a coefficient of 0.104‰/°C (Wang et al., 2013), where theprecipitation effect was eliminated because all samples were collected along the 400‐mmisoline of annualprecipitation. It is thus considered that the observed negative correlation with temperature is actually areflection of the relationship between δ13Cshell and precipitation because temperature is positively correlatedwith precipitation in monsoonal regions. To explore the clear relationship between δ13Cshell and tempera-ture, we narrowed down the study sites by determining those with MAP in the range of 450–550 mm. A posi-tive correlation was then found between δ13Cshell of Bradybaena and MAT for these selected sites, with aslope of approximately 0.13‰/°C (Figure S9a), which is comparable to the results obtained by Wang et al.(2013) and thus indicates that carbon isotopes have a similar sensitivity to temperature. However, therewas no positive relationship between δ13Cshell of Cathaica andMAT, which is possibly because of the smallertemperature gradient (about 5 °C) at the study sites (Figure S9b).

For both Bradybaena and Cathaica, the δ13Cshell values exhibited an overall increasing trend at a higheraltitude. Moreover, this positive correlation between the δ13Cshell and altitude was stronger for Cathaica(R2 = 0.51; Figure 4f) than Bradybaena (R2 = 0.32; Figure 4e). This could be attributed to the larger eleva-tion gradient at sites where Cathaica land snails were sampled. The δ13Cshell is a reflection of the carbonisotopic composition of a land snail's diet (e.g., Balakrishnan & Yapp, 2004; Francey, 1983; Goodfriend &Ellis, 2002); therefore, the above changing pattern of the δ13Cshell infers that the δ13C of the digestedplants shifts to a less negative values when the altitude increases, which is consistent with modern obser-vations of the relationship between δ13C of C3 plants and altitude on a global scale (Körner et al., 1988;Körner et al., 1991). This occurs because carbon isotope discrimination during photosynthesis of terrestrialC3 plants decreases with altitude, and it is associated with a greater carboxylation efficiency under lowertemperatures and atmospheric CO2/O2 partial pressures at high elevations (Körner et al., 1991). However,controversial changing patterns of δ13C in C3 plants with altitude have been observed on local scales indifferent mountainous areas of China (Pan et al., 2015; Wang et al., 2010). For example, leaf δ13C showeda positive correlation with altitude in a subtropical monsoon forest at the Mt. Tianmu Reserve in easternChina (Pan et al., 2015), while no consistent variation was observed in the δ13C values of C3 species alongan altitudinal gradient at Mt. Donglingshan in northern China under a temperate semihumid climate(Wang et al., 2010). This discrepancy could be attributed to differing amounts of soil water availabilityalong with altitude on a local scale. Nevertheless, over a large geographical range, we observe a positiveshift in δ13Cshell at a higher elevation in this study, which suggests a positive δ13C response at altitude inC3 plants.

As many environmental variables are not independent, we performed multivariate statistical analyseson δ13Cshell, altitude (Alt), MAT, and MAP. Multiple linear regressions revealed the strongestpredictive model for Bradybaena δ13Cshell using Alt and MAP as predictor variables (R2 = 0.39, P < 0.001;

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3369

Table S5), and a significant multiple linear regression conducted using SPSS software yielded thefollowing relationship:

δ13Cshell;Bradybaena ¼ 0:001 Altð Þ−0:001 MAPð Þ−11:15: (1)

In contrast, Cathaica δ13Cshell showed only a relationship with altitude (R2 = 0.51, P < 0.001; Table S6), asshown by the following equation:

δ13Cshell;Cathaica ¼ 0:002 Altð Þ−11:49: (2)

4.3. Oxygen Isotopes in Shell Carbonate of Land Snails4.3.1. Spatial Changes in δ18Oshell of Land Snails in China and Their Relationship With OxygenIsotopes in PrecipitationIn the East Asian monsoon region, there is a gradually decreasing trend in δ18Oshell for Bradybaena from thesoutheast coastal area to inland areas (Figure 5a). In northwest China (westerlies influenced region),δ18Oshell is more positive with respect to the reduced precipitation and increased evaporation. In southwestChina (Indian monsoon region), δ18Oshell is more negative, and this result is generally consistent with thenegative excursion of oxygen isotopes of precipitation (δ18OP) towards inland areas, as observed in the mon-soonal regions of China (Liu, 2008; Liu et al., 2009). However, the spatial pattern of δ18Oshell for Cathaica isnot as clear as that for Bradybaena, as shown by the slightly negative shift in δ18Oshell from coastal to inlandareas (Figure 5b). The reason for the less clear spatial pattern of δ18Oshell for Cathaica could be attributed tothe relatively small gradient of the environmental variables (i.e., MAT and MAP), as the sampled Cathaicaland snails are confined to a narrow geographical range.

To examine the relationship between δ18Oshell and δ18OP, we conducted a statistical analysis of the two datasets (https://gnip.com/). The result showed that δ18Oshell of both Bradybaena and Cathaica are moderatelycorrelatedwith annual δ18OP (Figures S10a and S10b). The relatively weak correlations are attributed to influ-ence of shell‐forming temperature that determines oxygen isotopic fractionation between δ18Oshell and δ

18OP.A similar relationship has also been observed in a variety of regions with different climatic conditions, such asthe oxygen isotopic compositions of land snails in central Europe (Lécolle, 1985), Italy (Zanchetta et al., 2005),the southern Great Plains of the USA (Goodfriend & Ellis, 2002), and Mediterranean and North Africanregions (Prendergast et al., 2015), which are all primarily a function of δ18OP.4.3.2. Potential Environmental Factors Affecting δ18Oshell of Land SnailsOur results show that the isotopic signature of δ18Oshell of Bradybaena differs among regions controlled bydifferent climatic systems (Figure 5a), and this is attributed to the different precipitation sources. We there-fore discuss the potential environmental factors affecting δ18Oshell with respect to each region. In the EastAsian monsoon (namely, the southeastern monsoon) region (Group 1 with 56 sites), which covers a vast areaof mainland China, rainfall mainly originates from the southeastern Pacific Ocean, and a typical monsoonalclimate prevails with clearly distinct wet/dry seasons and rainfall mostly concentrated during summer (i.e.,Beijing; Figure 1b), although in some of the southernmost coastal areas (such as Sanya; Figure 1b), rainfalloccurs nearly all year round. In the westerlies region (Group 2 with only three sites), which lies in the wes-tern inland part of the country, rainfall is mostly derived from the North Atlantic Ocean and/orMediterranean; rainfall is comparatively low, and the amount of evaporation may be higher than precipita-tion (i.e., Wulumuqi; Figure 1b). Rainfall in the Indian monsoon region (Group 3 with six sites), which islocated in southwest China, is derived from the Indian Ocean, and the area receives a relatively large amountof precipitation (>1,000 mm) and has a moderate temperature (13–20°C; i.e., Puer; Figure 1b).

For Group 1, there is a generally positive correlation between δ18Oshell of Bradybaena and MAT (R2 = 0.26, P< 0.001, n = 56; Figure 6a). However, we noticed the apparent deviation of some data points (square and tri-angle data) from the regression line in relation to sites located either in the southernmost coastal area withhigh precipitation (triangles; namely, Sanya [2,128 mm] and Guangzhou [1,801 mm]) or near the large lakes(squares; namely, Huoshan County, Nanchang, and Ganzhou). Huoshan County is close to the Chao Lake,and Nanchang and Ganzhou are situated around the Poyang Lake. In coastal areas, the Bradybaena δ18Oshell

values are more negative as they are influenced by the amount of rainfall (the so‐called rainfall amounteffect): δ18OP values become more negative as the amount of rainfall increases (Dansgaard, 1964; Liu,

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3370

2008; Rozanski et al., 1993; Yapp, 1982). Indeed, the rainfall amount effect has been proven to exist insouthern China, and it tends to be more profound in coastal areas (e.g., Liu, 2008). To further confirm this,we made a plot of Bradybaena δ18Oshell against MAP for the study sites in coastal areas. The statistical resultsexhibit a strong negative correlation between δ18Oshell and MAP (R2 = 0.67; Figure 6b), which indicates thatthe amount of rainfall has a prominent effect on δ18Oshell. In large lake areas, water vapor recycled from thelakes may contribute to local precipitation or dew water, which are eventually ingested by land snails anddrive δ18Oshell to become more negative. However, we also observed a slightly positive correlation betweenδ18Oshell and MAP for the study sites except the deviated 5 points (square and triangle data) in Group 1(Table S3), which contradicts the aforementioned amount effect on δ18Oshell, and we thus consider that thisis a “fake amount effect” relating to a reflection of the temperature effect, as MAT and MAP are highly cor-related in the East Asian monsoon region of China (Figure S11).

After removing those data points, the positive correlation between Bradybaena δ18Oshell and MAT becomesstronger in the East Asian monsoon region (R2 = 0.47, P < 0.001, n = 51; Figure 6a). This is in accordancewith the expected relationship that is derived from the following two facts: one, annual mean δ18O of preci-pitation correlates positively with annual mean air temperature (the so‐called “temperature effect”; e.g.,Dansgaard, 1964); two, δ18Oshell links with δ18O of local rainfall, as shown by the moderate δ18Oshell ‐

δ18OP correlation (Group 1 in Table S3) as well as the results presented in previous studies (Goodfriend &Ellis, 2002; Lécolle, 1985; Zanchetta et al., 2005). Many studies conducted in diverse environments have alsoproven a correlation between δ18Oshell and MAT (Balakrishnan & Yapp, 2004; Magaritz et al., 1981; Yaneset al., 2009). To further support these results, the temperature effect is more pronounced in northeast China(i.e., northernmost part of the East Asian monsoon region), where temperatures are usually low throughoutthe year (e.g., the MAT ranges from 5.0 °C to 11.5 °C from Harbin to Dalian), as shown by the high correla-tion coefficient between MAT and δ18Oshell for both Bradybaena (Figure S5a) and Cathaica (Figure S5b).Furthermore, Bradybaena δ18Oshell in Group 1 have a positive relationship withMARH and δ18OP (Table S3).

For Group 2, there is a negative correlation between δ18Oshell andMAP (R2 = 0.96, P> 0.05, n= 3; Figure 6c),which infers a rainfall amount effect on δ18Oshell because a large amount of rainfall effectively reduces theimpact of evaporation that drives oxygen isotope of precipitation towards a more positive value. However,the P value for this correlation is >0.05, which is possibly due to the limited number of sites sampled inthis region.

A statistical analysis conducted on Group 3 shows an obvious negative correlation between Bradybaenaδ18Oshell and MAP (R2 = 0.88, P < 0.05, n = 6; Figure 6c), which is a reflection of the rainfall amount effecton δ18OP.

If we assume that shell carbonate occurs at isotopic equilibrium, we can obtain the theoretical oxygenisotopic composition of shell (δ18Oeq) that precipitates at isotopic equilibrium with rainfall, with respect tothe mean temperature of the sampling site, and then compare it with measured δ18Oshell values. Such acomparison assists in better understanding the processes involved in snail shell formation and associatedcontrolling factors (Zanchetta et al., 2005). Following the method of Zanchetta et al. (2005), whoconsidered that the following equation provided the most reasonable δ18O‐temperature relationship foraragonite (Zanchetta et al., 2005 and reference therein), we calculated δ18Oeq using the equation ofPatterson et al. (1993).

1; 000 lnα ¼ 18:56 103T−1� �

−33:49; (3)

where T is in degree Kelvin and α is the isotopic fractionation factor between shell carbonate and water. Therelation between α and δ18O of shell carbonate and water is presented in the following equation:

α ¼ 1; 000þ δ18Oshell� �

= 1; 000þ δ18OP� �

; (4)

where δ18OP is considered to be the δ18O of the water from which the shell aragonite precipitated.

The mean monthly temperature and δ18OP in April to October (the assumed growing season) were respec-tively employed in the above calculation for both Bradybaena and Cathaica. Although the land snail

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3371

growth season may vary from southern to northern China, we assume that the general growing season forland snails is the warm season (April–October) when most rainfall occurs. Although it is acknowledged thatthis may incur some uncertainty in the calculated δ18Oeq, this provides a baseline for oxygen isotope fractio-nation in the water of land snail bodies. As can be seen, δ18Oshell is uniformly higher than δ18Oeq, and thedifference between them (Δδ18Oshell − eq) varies from ‐ 0.10 to +8.13‰ (mean: +3.97‰) for Bradybaenaand from ‐ 1.21 to +5.63‰ (mean: +1.95‰) forCathaica (Figure S12). Many previous studies have also foundthat δ18Oshell is more positive than expected if it is in equilibrium with rainfall (Baldini et al., 2007;Goodfriend et al., 1989; Goodfriend & Ellis, 2002; Lécolle, 1985; Prendergast et al., 2015; Yanes et al.,2009; Zanchetta et al., 2005). As the snail's shell carbonate is commonly precipitated from a body fluid atan isotopic steady state (Balakrishnan & Yapp, 2004), the overall positiveΔδ18Oshell − eq suggests a more posi-tive δ18O in snail body water than that in precipitation, which is due to evaporation, and this result is in fairlygood agreement with the previously observed 18O‐enriched snail body fluid relative to rainwater (Goodfriendet al., 1989; Prendergast et al., 2015). Moreover, the Δδ18Oshell − eq for Bradybaena positively correlates withtemperature and RH in growing season (growing season temperature [GST] and growing season relativehumidity, respectively, in Figures S12a and S12b). In contrast, the Δδ18Oshell − eq for Cathaica exhibits nocorrelation with GST and growing season relative humidity (Figure S12c and S12d). The positive correlationbetween Bradybaena Δδ18Oshell − eq and GST indicates that temperature plays an important role in the eva-poration process of body water during the snail's active seasons and infers that higher rates of evaporationoccur with higher temperatures. However, the snail flux balance model proposed by Balakrishnan andYapp (2004) contradicts the observed positive relation between Bradybaena Δδ18Oshell − eq and RH andshows that RH has a negative effect on the δ18O of snail body fluids. The reason for the positive effect ofRH on Bradybaena Δδ18Oshell − eq observed in this study is currently unclear. It possibly relates to expandingof land snail growing season (to relatively dry seasons) at sites with high RH. The less effective correlationbetween Cathaica Δδ18Oshell − eq and temperature possibly results from the relatively small temperaturerange at Cathaica study sites.

We also determined the RH required for Bradybaena andCathaica that cohabitated at same sites by using thesnail evaporative steady‐state flux‐balance mixing model developed by Balakrishnan and Yapp (2004). Thismodel effectively integrates the δ18O of water vapor, the flux and δ18O of water imbibed from local precipita-tion, temperature, and RH as the common controlling factors used to predict δ18O of snail body water andeventually δ18Oshell. In this study, a total of 23 sites were co‐occupied by both Bradybaena and Cathaica(Table S1 and Figure 1a). The average air temperature and δ18OP from April to October (the assumed grow-ing season) for these sites are 19.77 °C and ‐ 6.09‰ Standard Mean Ocean Water (SMOW), and the meanδ18Oshell values over the sites are ‐ 3.59 ± 1.22‰ and ‐ 4.56 ± 1.43‰ for Bradybaena and Cathaica, respec-tively. By projecting the mean δ18Oshell onto the lines, we obtained an RH of 93.06% for Bradybaena and95.58% for Cathaica (Figure 7). This suggests that both Bradybaena and Cathaica grow their shells duringthe condition of high RH (RH > 90%), which is consistent with the fact that land snails are often active afterrain or during the evening.

For Cathaica snails, we found no reliable relationship between δ18Oshell and environmental factors, such asaltitude, MAT, MAP, and MARH (Table S3), and Cathaica δ18Oshell only shows a robust correlation withMAT on a local scale (i.e., in northeast China; Figure S7b).

To avoid problems of multicolinearity in our data set, we also conducted multivariate statistical analyses onBradybaena δ18Oshell and multiple environmental variables (Table S3). However, we could not find a reliablepredictive model that could be used as a multiple regression model, and further analysis was not possible.Furthermore, there was also no appropriate predictive modeling equation for Cathaica δ18Oshell due to itsunclear relationship with environmental variables.

Our results show a moderate correlation between δ18Oshell of both Bradybaena and Cathaica and δ18OP. Ofall the environmental variables, temperature acts as a major control on δ18Oshell of Bradybaena in the EastAsian monsoonal region, which differs from results of precipitation or RH control gained in the ChineseLoess Plateau (Liu et al., 2006) and in other parts of the world (Bonadonna et al., 1999; Goodfriend &Ellis, 2002; Lécolle, 1985; Yanes et al., 2008; Yapp, 1979; Zanchetta et al., 2005). As discussed above, the pro-minent temperature control on δ18Oshell for Bradybaena could be attributed to the effective temperatureinfluences on δ18OP and on evaporation processes that drive δ18O of snail body water in growth seasons.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3372

However, in the westerlies and Indian monsoon regions, rainfall amount plays a key role in determining theδ18Oshell of Bradybaena.

Under these circumstances, a certain degree of uncertainty would be introduced if paleoclimatic reconstruc-tion was conducted using Bradybaena δ18Oshell because the determined correlation coefficient (R2) is nothigh enough, and this uncertainty may stem from the complex ecophysiological response of land snails toenvironmental changes. For example, δ18Oshell should reflect conditions within relatively narrow ranges ofRH (>70%) and moderate temperatures (10–27 °C; Balakrishnan & Yapp, 2004; Balakrishnan et al., 2005).Furthermore, metabolic activity may also affect δ18Oshell through the changing isotopic composition of snailbody water (Goodfriend et al., 1989; Lécolle, 1985).

5. Conclusions

This paper presents a systematic study of stable carbon and oxygen isotopes in shell carbonates (δ13Cshell andδ18Oshell) of two land snail genera (Bradybaena and Cathaica) within China, with the aim of determiningtheir potential usefulness in reconstructing paleoclimate and paleovegetation. The conclusions are madeas follows:

1. Stable isotope data and shell size comparison further confirm that Bradybaena and Cathaica have differ-ent ecophysiological habits.

2. δ13Cshell reflects the carbon composition of dietary vegetation. Bradybaena δ13Cshell documents the isoto-pic signal of C3 plants, even in C3/C4 mixing regions, and shows the effects of water stress, temperature,and altitude on δ13C of C3 plants (where the values decrease with rainfall and increase with temperatureor altitude). In contrast, Cathaica δ13Cshell registers a C3/C4 mixing signal in the C4 maximum zone andexhibits a strong altitude effect and a moderate water stress influence.

3. δ18Oshell of both Bradybaena and Cathaica show a moderate correlation with δ18OP. The temperatureeffect on δ18Oshell is prominent in the East Asian monsoonal region. Although there is a robust positivecorrelation between Bradybaena δ18Oshell and MAT, which confirms its usefulness in paleotemperaturereconstruction, it should be emphasized that a certain degree of uncertainty would be introduced if con-ducting a paleotemperature reconstruction using Bradybaena δ18Oshell, as the correlation coefficient (R2)determined in this study was not high enough. In contrast, the rainfall amount is the dominant factoraffecting Bradybaena δ18Oshell in the westerlies and the Indian monsoon regions; therefore, Bradybaenaδ18Oshell in these regions has potential of being used to retrieve paleorainfall.

Figure 7. Predicted curves of δ18Oshell as a function of relative humidity using the evaporative steady‐state flux‐balancemixing model by Balakrishnan and Yapp (2004). The curves were calculated for temperature and rainfall δ18O valueof 19.8°C and 6.1‰. Arrows illustrate that, on average, Bradybaena and Cathaica shells grew at relative humidity of 93.1%and 95.6%, respectively.

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3373

ReferencesBalakrishnan, M., & Yapp, C. J. (2004). Flux balance models for the oxygen and carbon isotope compositions of land snail shells.Geochim. et

Cosmochim. Acta, 68(9), 2007–2024. https://doi.org/10.1016/j.gca.2003.10.027Balakrishnan, M., Yapp, C. J., Theler, J. L., Carter, B. J., & Wyckoff, D. G. (2005). Environmental significance of 13C/12C and 18O/16O ratios

of modern land‐snail shells from the southern Great Plains of North America. Quaternary Research, 63(1), 15–30. https://doi.org/10.1016/j.yqres.2004.09.009

Baldini, L. M., Walker, S. E., Railsback, L. B., Baldini, J. U., & Crowe, D. E. (2007). Isotope ecology of the modern land snails Cerion, SanSalvador, Bahamas: Preliminary advances toward establishing a low‐latitude island palaeoenvironmental proxy. Palaios, 22(2), 174–187.https://doi.org/10.2110/palo.2005.p05.091r

Bengtson, S. A., Nilsson, A., Nordström, S., & Rundgren, S. (1979). Selection for adult shell size in natural populations of the land snailCepaea hortensis (Müll.). Ann. Zool. Fennici., 16, 187–194.

Bonadonna, F. P., Leone, G., & Zanchetta, G. (1999). Stable isotope analyses on the last 30ka molluscan fauna from Pampa grassland,Bonaerense region, Argentina. Palaeogeogr., Palaeoclimatol. Palaeoecol., 153(1‐4), 289–308. https://doi.org/10.1016/S0031‐0182(99)00063‐2

Bowen, G. J. (2008). The online isotopes in precipitation calculator, Version 2.2, The University of Utah, Salt Lake City, Utah, USA.(Available at http://www.waterisotopes.org.)

Burleigh, R., & Kernney, M. P. (1982). Some chronological implications of a fossil moluscan assemblage from a Neolithic site at Brook, Kent,England. Journal of Archeological Sciences, 9(1), 29–38. https://doi.org/10.1016/0305‐4403(82)90004‐8

Cerling, T. E., & Quade, J. (1993). Stable carbon and oxygen isotopes in soil carbonates. In P. K. Swart et al. (Eds.), Climate change incontinental isotopic records, Geophys. Monogr. Ser. (Vol. 78, pp. 217–231). Washington, D. C: AGU.

Colonese, A. C., Zanchetta, G., Fallick, A. E., Manganelli, G., Lo Cascio, P., Hausmann, N., et al. (2014). Oxygen and carbon isotopiccomposition of modern terrestrial gastropod shells from Lipari Island, Aeolian Archipelago (Sicily). Palaeogeogr., PalaeoclimatolPalaeoecol., 394, 119–127. https://doi.org/10.1016/j.palaeo.2013.12.003

Colonese, A. C., Zanchetta, G., Fallick, A. E., Martini, F., Manganelli, G., & Vetro, D. L. (2007). Stable isotope composition of Late Glacialland snail shells from Grotta del Romito (Southern Italy): Palaeoclimatic implications. Palaeogeogr., Palaeoclimatol Palaeoecol., 254(3‐4),550–560. https://doi.org/10.1016/j.palaeo.2007.07.005

Dansgaard, W. (1964). Stable isotopes in precipitation. Tellus, 16, 436–468.DeNiro, M. J., & Epstein, S. (1978). Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta,

42(5), 495–506. https://doi.org/10.1016/0016‐7037(78)90199‐0Francey, R. J. (1983). A comment on 13C/12C in land snail shells. Earth and Planetary Science Letters, 63(1), 142–143. https://doi.org/

10.1016/0012‐821X(83)90031‐6Goodfriend, G. A. (1986). Variation in land‐snail shell form and size and its causes: A review. Syst. Zool., 35(2), 204–223. https://doi.org/

10.2307/2413431Goodfriend, G. A., & Ellis, G. L. (2000). Stable carbon isotope record of middle to late Holocene climate changes from land snail shells at

Hinds Cave, Texas. Quaternary International, 67(1), 47–60. https://doi.org/10.1016/S1040‐6182(00)00008‐2Goodfriend, G. A., & Ellis, G. L. (2002). Stable carbon and oxygen isotopic variations in modern Rabdotus land snail shells in the southern

Great Plains, USA, and their relation to environment. Geochim. et Cosmochim. Acta, 66(11), 1987–2002. https://doi.org/10.1016/S0016‐7037(02)00824‐4

Goodfriend, G. A., & Hood, D. G. (1983). Carbon isotope analysis of land snail shells: Implications for carbon sources and radiocarbondating. Radiocarbon, 25(3), 810–830. https://doi.org/10.1017/S0033822200006226

Goodfriend, G. A., & Magaritz, M. (1987). Carbon and oxygen isotope composition of shell carbonate of desert land snails. Earth andPlanetary Science Letters, 86(2‐4), 377–388. https://doi.org/10.1016/0012‐821X(87)90234‐2

Goodfriend, G. A., Magaritz, M., & Gat, J. R. (1989). Stable isotope composition of land snail body water and its relation to environmentalwaters and shell carbonate. Geochim. et Cosmochim. Acta, 53(12), 3215–3221. https://doi.org/10.1016/0016‐7037(89)90102‐6

Gould, S. J. (1984). Covariance sets and ordered geographic variation in Cerion from Aruba, Bonaire and Curacao: A way of studyingnonadaptation. Syst. Zool., 33(2), 217–237. https://doi.org/10.2307/2413022

Hassan, K. M. (2015). Stable isotopic signatures of the modern land snail Eremina desertorum form a low‐latitude (hot) dry desert—A studyfrom the Petrified Forest, New Cairo, Egypt. Chemie der Erde‐Geochemistry, 75(1), 65–72. https://doi.org/10.1016/j.chemer.2014.09.002

Huang, L. P., Wu, N. Q., Gu, Z. Y., & Chen, X. Y. (2012). Variability of snail growing season at the Chinese Loess Plateau during the last 75ka. Chinese Science Bulletin, 57(9), 1036–1045. https://doi.org/10.1007/s11434‐011‐4931‐z

Jacobs, B. F., Kingston, J. D., & Jacobs, L. L. (1999). The origin of grass‐dominated ecosystems. Annals of the Missouri Botanical Gardens,86(2), 590–643. https://doi.org/10.2307/2666186

Kim, S. T., Mucci, A., & Taylor, B. E. (2007). Phosphoric acid fractionation factors for calcite and aragonite between 25 and 75°C: Revisited.Chemical Geology, 246(3‐4), 135–146. https://doi.org/10.1016/j.chemgeo.2007.08.005

Knights, R. W. (1979). Experimental evidence for selection on shell size in Cepaea hortensis (Müll). Genetica, 50(1), 51–60. https://doi.org/10.1007/BF00122678

Kohn, M. J. (2010). Carbon isotope compostions of terrestrial C3 plants as indicators of (paleo)ecology and (paleo)climate. Proceedings of theNational Academy of Sciences of the United States of America, 107(46), 19,691–19,695. https://doi.org/10.1073/pnas.1004933107

Körner, C., Farquhar, G. D., & Roksandic, Z. (1988). A global survey of carbon isotope discrimination in plants from high altitude.Oecologia, 74(4), 623–632. https://doi.org/10.1007/BF00380063

Körner, C., Farquhar, G. D., & Wong, S. C. (1991). Carbon isotope discrimination by plants follows latitudinal and altitudinal trends.Oecologia, 88(1), 30–40. https://doi.org/10.1007/BF00328400

Laurin, B., & Rousseau, D. D. (1985). Analyse multivariée des associations malacologiques d'Achenheim. Implications climatiques etenvironnementales. Bulletin de l'Association française pour l'étude du quaternaire, 22(1), 21–30. https://doi.org/10.3406/quate.1985.1523

Lécolle, P. (1985). The oxygen isotope composition of landsnail shells as climatic indicator: application to hydrogeology and paleoclima-tology. Chem. Geol.: Isotope Geoscience section, 58(1‐2), 157–181. https://doi.org/10.1016/0168‐9622(85)90036‐3

Leng, M. J., Heaton, T. H. E., Lamb, H. F., & Naggs, F. (1998). Carbon and oxygen isotope variations within the shell of an African land snail(Limicolaria kambeul chudeaui Germain): A high‐resolution record of climate seasonality? Holocene, 8(4), 407–412. https://doi.org/10.1191/095968398669296159

Leng, M. J., & Lewis, J. P. (2016). Oxygen isotopes in Molluscan shell: Applications in environmental archaeology. EnvironmentalArchaeology, 21(3), 295–306. https://doi.org/10.1179/1749631414Y.0000000048

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL.

Acknowledgments

This study was supported by theStrategic Priority Research Program ofthe Chinese Academy of Sciences(Grant XDA19050104), the NationalNatural Science Foundation of China(Grant 41572163), and the StateScholarship Fund for Study Abroad byChina Scholarship Council (CSC). Wethank Professor Deniu Chen for assis-tance in identifying land snail species.The meteorological data of study siteswere provided by the ChinaMeteorological Data Service Center,and data for producing Figures 1 andFigures S1–S12 are available from thedata share website of https://figshare.com/articles/Data_for_2019JG005243/8120825.

3374

Leone, G., Bonadonna, F., & Zanchetta, G. (2000). Stable isotope record in mollusca and pedogenic carbonate from Late Pliocene soilsof Central Italy. Palaeogeogr., Palaeoclimatol. Palaeoecol., 163(3‐4), 115–131. https://doi.org/10.1016/S0031‐0182(00)00148‐6

Liu, J. R. (2008). A study of deuterium and oxygen‐18 in precipitation in China (Master's thesis), Retrieved from (Dissertation knowledge dis-covery system), (LW036361). Beijing: University of Chinese Academy of Science. (In Chinese)

Liu, W. G., Huang, Y. S., An, Z. S., Clemens, S. C., Li, L., Prell, W. L., & Ning, Y. F. (2005). Summer monsoon intensity controls C4/C3 plantabundance during the last 35ka in the Chinese Loess Plateau: Carbon isotope evidence from bulk organic matter and individual leafwaxes. Palaeogeogr, Palaeoclimatol, Palaeoccol., 220(3‐4), 243–254. https://doi.org/10.1016/j.palaeo.2005.01.001

Liu, Z. F., Tian, L. D., Yao, T. D., & Chai, X. R. (2009). Spatial distribution of δ18O in precipitation over China. Chinese Science Bulletin, 54,804–811. (In Chinese)

Liu, Z. X., Gu, Z. Y., Wu, N. Q., & Xu, B. (2007). Diet control on carbon isotopic compostion of land snail shell carbonate. Chinese ScienceBulletin, 52(3), 388–394. https://doi.org/10.1007/s11434‐007‐0045‐z

Liu, Z. X., Gu, Z. Y., Xu, B., & Wu, N. Q. (2006). Monsoon precipitation effect on oxygen isotope composition of land snail shell carbonatefrom the Loess Plateau. Quaternary Sciences, 26, 643–648. (In Chinese)

Magaritz, M., & Heller, J. (1980). A desert migration indicator—Oxygen isotopic composition of land snail shells. Palaeogeogr.,Palaeoclimatol. Palaeoecol., 32, 153–162. https://doi.org/10.1016/0031‐0182(80)90038‐3

Magaritz, M., Heller, J., & Volokita, M. (1981). Land–air boundary environment as recorded by the 18O/16O and 13C/12C isotope ratios in theshells of land snails. Earth and Planetary Science Letters, 52(1), 101–106. https://doi.org/10.1016/0012‐821X(81)90212‐0

McConnaughey, T. A., Burdett, J., Whelan, J. F., & Paull, C. K. (1997). Carbon isotopes in biological carbonates: Respiration and photo-synthesis. Geochim. et Cosmochim. Acta, 61(3), 611–622. https://doi.org/10.1016/S0016‐7037(96)00361‐4

McConnaughey, T. A., & Gillikin, D. P. (2008). Carbon isotopes in mollusk shell carbonates. Geo‐Marine Letters, 28(5‐6), 287–299. https://doi.org/10.1007/s00367‐008‐0116‐4

Metref, S., Rousseau, D.‐D., Bentaleb, I., Labonne, M., & Vianey‐Liaud, M. (2003). Study of the diet effect on δ13C of shell carbonate of theland snail Helix aspersa in experimental conditions. Earth and Planetary Science Letters, 211(3‐4), 381–393. https://doi.org/10.1016/S0012‐821X(03)00224‐3

Milano, S., Demeter, F., Hublin, J.‐J., Duringer, P., Patole‐Edoumba, E., Ponche, J.‐L., et al. (2018). Environmental conditions framing thefirst evidence of modern humans at Tam Pà Ling, Laos: A stable isotope record from terrestrial gastropod carbonates. Palaeogeogr.Palaeoclimatol. Palaeoecol., 511, 352–363. https://doi.org/10.1016/j.palaeo.2018.08.020

Moine, O., Rousseau, D. D., Jolly, D., & Vianey‐Liaud, M. (2002). Paleoclimatic reconstruction using mutual climatic range on terrestrialmollusks. Quaternary Research, 57(1), 162–172. https://doi.org/10.1006/qres.2001.2286

Pan, S., Zhang, W. P., Zhao, M. S., Li, Y., Xu, S. S., &Wang, G. X. (2015). Altitude pattern of leaf carbon isotope composition in a subtropicalmonsoon forest. Polish Journal of Ecology, 63(4), 512–522. https://doi.org/10.3161/15052249PJE2015.63.4.004

Patterson, W. P., Smith, G. R., & Lohmann, K. C. (1993). Continental paleothermometry and seasonality using the isotopic composition ofaragonite otoliths of freshwater fishes. In P. K. Swart et al. (Eds.), Climatic Change in Continental Isotopic Records,Geophys. Monogr. (Vol.78, pp. 191–202). Washington DC: American Geophysical Union.

Pigati, J. S., Rech, J. A., & Nekola, J. C. (2010). Radiocarbon dating of small terrestrial gastropod shells in North America. QuaternaryGeochronology, 5(5), 519–532. https://doi.org/10.1016/j.quageo.2010.01.001

Prendergast, A. L., Stevens, R. E., Barker, G., & O'Connell, T. C. (2015). Oxygen isotope signatures from land snail (Helix melanostoma)shells and body fluid: Proxies for reconstructing Mediterranean and North African rainfall. Chem. Geol., 409, 87–98. https://doi.org/10.1016/j.chemgeo.2015.05.014

Prendergast, A. L., Stevens, R. E., Hill, E. A., Hunt, C., O'Connell, T. C., & Barker, G. W. (2017). Carbon isotope signatures from land snailshells: Implications for palaeovegetation reconstruction in the eastern Mediterranean. Quaternary International, 432, 48–57. https://doi.org/10.1016/j.quaint.2014.12.053

Rangarajan, R., Ghosh, P., & Naggs, F. (2013). Seasonal variability of rainfall recorded in growth bands of the giant African land snailLissachatina fulica (Bowdich) from India. Chem. Geol., 357, 223–230. https://doi.org/10.1016/j.chemgeo.2013.08.015

Rao, Z. G., Jia, G. D., Zhu, Z. Y., Wu, Y., & Zhang, J. W. (2008). Comparison of the carbon isotope composition of total organic carbon andlong‐chain n‐alkanes from surface soils in eastern China and their significance. Chinese Science Bulletin, 53, 3921–3927.

Rozanski, K., Araguás‐Araguás, L., & Gonfiantini, R. (1993). Isotopic patterns in modern global precipitation. In Climate change inContinental isotopic records, geophysical monograph series (pp. 1–36). Washington, D. C: American Geophysical Union.

Sacchi, C. F. (1965). Ecological and historical bases for a study of the Iberian terrestrial Mollusca. Proc. 1st Europ. Malacol. Congr., 1962,243‐257.

Stott, L. D. (2002). The influence of diet on the δ13C of shell carbon in the pulmonate snail Helix aspersa. Earth and Planetary Science Letters,195(3‐4), 249–259. https://doi.org/10.1016/S0012‐821X(01)00585‐4

Wang, G. A., Feng, X., Han, J., Zhou, L., Tan, W., & Su, F. (2008). Paleovegetation reconstruction using δ13C of soil organic matter.Biogeosciences, 5(5), 1325–1337. https://doi.org/10.5194/bg‐5‐1325‐2008

Wang, G. A., Han, J. M., & Liu, T. S. (2003). The carbon isotopic composition of C3 herbaceous plants in loess area of North China. Science inChina (Series D), 46(10), 1069–1076. https://doi.org/10.1007/BF02959402

Wang, G. A., Li, J. Z., Liu, X. Z., & Li, X. Y. (2013). Variations in carbon isotope ratios of plants across a temperature gradient along the 400mm isoline of mean annual precipitation in north China and their relevance to paleovegetation reconstruction. Quaternary ScienceReviews, 63, 83–90. https://doi.org/10.1016/j.quascirev.2012.12.004

Wang, G. A., Zhou, L. P., Liu, M., Han, J. M., Guo, J. H., Faiia, A., & Su, F. (2010). Altitudinal trends of leaf δ13C of follow different patternsacross a mountainous terrain in north China characterized by a temperate semi‐humid climate. Rapid Communications in MassSpectrometry, 24(11), 1557–1564. https://doi.org/10.1002/rcm.4543

Wang, X., Cui, L. L., Zhai, J. X., & Ding, Z. L. (2016). Stable and clumped isotopes in shell carbonates of land snails Cathaica sp. andBradybaena sp. in north China and implications for ecophysiological characteristics and paleoclimate studies. Geochemistry, Geophysics,Geosystems, 17, 219–231. https://doi.org/10.1002/2015GC006182

Wang, Y. J., Lv, H. Y., Wang, G. A., Yang, H., & Li, Z. (2000). Carbon isotope analysis of silicates in C3, C4 plants and modern soils. ChineseScience Bulletin, 45, 978–982. (In Chinese)

Wu, B., & Wu, N. Q. (2008). Paleoenvironmental study of S5 paleosol formation: Based on mollusk evidence. Quaternary Sciences, 28(5),901–908. (In Chinese)

Wu, B., & Wu, N. Q. (2011). Terrestrial mollusc records from Xifeng and Luochuan L9 loess strata and their implications for paleoclimaticevolution in the Chinese Loess Plateau during marine oxygen isotope stages 24‐22. Climate of the Past, 7(2), 349–359. https://doi.org/10.5194/cp‐7‐349‐2011

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3375

Wu, N. Q., Liu, T. S., Liu, X. P., & Gu, Z. Y. (2002). Mollusk record of millennial climate variability in the Loess Plateau during the LastGlacial Maximum. Boreas, 31(1), 20–27. https://doi.org/10.1080/03009480210648

Wu, N. Q., Pei, Y. P., Lv, H. Y., Guo, Z. T., Li, F. J., & Liu, T. S. (2006). Marked ecological shifts during 6.2‐2.4 Ma revealed by a terrestrialmolluscan record from the Chinese Red Clay Formation and implication for palaeoclimatic evolution. Palaeogeogr., Palaeoclimatol.Palaeoecol., 233(3‐4), 287–299. https://doi.org/10.1016/j.palaeo.2005.10.006

Wu, N. Q., Rousseau, D. D., & Liu, T. S. (1996). Land mollusk records from the Luochuan loess sequence and their paleoenvironmentalsignificance. Science in China (Series D), 39, 494–502.

Xu, B., Gu, Z. Y., Han, J. T., Liu, Z. X., Pei, Y. P., Lu, Y. W., et al. (2010). Radiocarbon and stable carbon isotope analyses of land snails fromthe Chinese loess plateau: environmental and chronological implications. Radiocarbon, 52(1), 149–156. https://doi.org/10.1017/S0033822200045094

Yanes, Y. (2015). Stable isotope ecology of land snails from a high latitude site near Fairbanks, interior Alaska, USA. Quat. Res., 83(3),588–595. https://doi.org/10.1016/j.yqres.2015.03.004

Yanes, Y., Asta, M. P., Ibáñez, M., Alonso, M. R., & Romanek, C. S. (2013). Paleoenvironmental implications of carbon stable isotopecomposition of land snail tissues. Quaternary Research, 80(3), 596–605. https://doi.org/10.1016/j.yqres.2013.08.010

Yanes, Y., Delgado, A., Castillo, C., Alonso, M. R., Ibáñez, M., De la Nuez, J., & Kowalewski, M. (2008). Stable isotope (δ18O, δ13C, and δD)signatures of recent terrestrial communities from a low‐latitude, oceanic setting: endemic land snails, plants, rain, and carbonate sedi-ments from the eastern Canary Islands. Chem. Geol., 249(3‐4), 377–392. https://doi.org/10.1016/j.chemgeo.2008.01.008

Yanes, Y., Graves, G. R., & Romanek, C. S. (2018). Stable isotope ecology (δ18O, δ13C, δ15N) of modern land snails along an altitudinalgradient in southern Appalachian Mountains, USA. Palaeogeogr., Palaeoclimatol. Palaeoecol., 492, 92–103. https://doi.org/10.1016/j.palaeo.2017.12.015

Yanes, Y., Romanek, C. S., Delgado, A., Brant, H. A., Noakes, J. E., Alonso, M. R., & Ibáñez, M. (2009). Oxygen and carbon stable isotopes ofmodern land snail shells as environmental indicators from a low‐latitude oceanic island. Geochim. Cosmochim. Acta, 73(14), 4077–4099.https://doi.org/10.1016/j.gca.2009.04.021

Yang, J. Z., & Ren, H. T. (2012). Preliminary report on the damage of Cathaica (Cathaica) orestias on fruit trees and the prevention mea-sures. Chin. Horticulture Abstr., 9, 157–158. (In Chinese)

Yapp, C. J. (1979). Oxygen and carbon isotope measurements of land snail shell carbonate. Geochim. et Cosmochim Acta, 43(4), 629–635.https://doi.org/10.1016/0016‐7037(79)90170‐4

Yapp, C. J. (1982). A model for the relationship between precipitation D/H ratios and precipitation intensity. Journal of GeophysicalResearch, 87(C12), 9614–9620. https://doi.org/10.1029/JC087iC12p09614

Zaarur, S., Olack, G., & Affek, H. P. (2011). Paleo‐environmental implication of clumped isotopes in land snail shells. Geochim. etCosmochim. Acta, 75(22), 6859–6869. https://doi.org/10.1016/j.gca.2011.08.044

Zanchetta, G., Leone, G., Fallick, A. E., & Bonadonna, F. P. (2005). Oxygen isotope composition of living land shell shells: data from Italy.Palaeogeogr., Palaeoclimatol. Palaeoecol., 223(1‐2), 20–33. https://doi.org/10.1016/j.palaeo.2005.03.024

Zhai, J. X., Wang, X., Qin, B., Cui, L. L., Zhang, S. H., & Ding, Z. L. (2019). Clumped isotopes in land snail shells over China: Towardsestablishing a biogenic carbonate paleothermometer.Geochim. et Cosmochim. Acta, 257, 68–79. https://doi.org/10.1016/j.gca.2019.04.028

Zhang, N., Yamada, K., Suzuki, N., & Yoshida, N. (2014). Factors controlling shell carbon isotopic composition of land snail Acusta despectasieboldiana estimated from laboratory culturing experiment. Biogeosciences, 11(19), 5335–5348. https://doi.org/10.5194/bg‐11‐5335‐2014

Reference From Supporting InformationLi, F. (1997). Carbon and oxygen isotopic compositions of snail shells in the Loess Plateau and their relationships with major climatic indicators,

(Doctoral dissertation). Retrieved from (Dissertation knowledge discovery system), (N82678). Beijing: Institute of Geology andGeophysics, Chinese Academy of Sciences. In Chinese

10.1029/2019JG005255Journal of Geophysical Research: Biogeosciences

WANG ET AL. 3376