Spatial distribution of 17O-excess in surface snow along a traverse from Zhongshan station to Dome A, East Antarctica

Earth and Planetary Science Letters 414 (2015) 126–133

Contents lists available at ScienceDirect

Earth and Planetary Science Letters www.elsevier.com/locate/epsl

Spatial distribution of 17 O-excess in surface snow along a traverse from Zhongshan station to Dome A, East Antarctica Hongxi Pang a , Shugui Hou a,∗ , Amaelle Landais b , Valérie Masson-Delmotte b , Frederic Prie b , Hans Christian Steen-Larsen b , Camille Risi c , Yuansheng Li d , Jean Jouzel b , Yetang Wang e , Jing He a , Bénédicte Minster b , Sonia Falourd b a

Key Laboratory of Coast and Island development of Ministry of Education, School of Geographic and Oceanographic Sciences, Nanjing University, Nanjing 210093, China b Laboratoire des Sciences du Climat et de l’Environnement, UMR8212, CEA-CNRS-UVSQ/IPSL, Gif-sur-Yvette, France c Laboratoire de Météorologie Dynamique, UMR8539, IPSL/CNRS/UPMC, 4, place Jussieu, 75252 Paris Cedex 05, France d Polar Research Institute of China, Shanghai 200136, China e College of Population, Resources and Environment, Shandong Normal University, Jinan 250014, China

a r t i c l e

i n f o

Article history: Received 5 October 2014 Received in revised form 7 January 2015 Accepted 13 January 2015 Available online 28 January 2015 Editor: J. Lynch-Stieglitz Keywords: water isotopologues 17 O-excess Dome A ice sheet Antarctica

a b s t r a c t The influence of temperature on the triple isotopic composition of oxygen in water is still an open question and limits the interpretation of water isotopic profiles in Antarctic ice cores. The main limitation arises from the lack of 17 O-excess measurements in surface snow and especially for remote regions characterized by low temperature and accumulation rate. In this study, we present new 17 O-excess measurements of surface snow along an East Antarctic traverse, from the coastal Zhongshan station to the highest point of the Antarctic ice sheet at Dome A. The 17 O-excess data significantly decrease inland, with a latitudinal gradient of −1.33 ± 0.41 per meg/degree, an altitudinal gradient of −0.48 ± 0.17 per meg/100 m, and a temperature gradient of 0.35 ± 0.11 per meg/◦ C. Theoretical calculations performed using a Rayleigh model attribute this inland decrease to kinetic isotopic fractionation occurring during condensation from vapor to ice under supersaturation conditions at low temperatures. However, large heterogeneity of 17 O-excess in Antarctic precipitation cannot only be explained by temperature at condensation and/or influences of relative humidity in the moisture source region. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The relative abundance of water isotopologues H2 18 O and HDO (expressed as δ 18 O and δ D) in polar ice cores have long been used as a proxy for past local condensation temperature (Dansgaard et al., 1969; Johnsen et al., 1972; Jouzel et al., 1987; EPICA community members, 2004). The second-order parameter d-excess, defined as d-excess = δ D − 8δ 18 O (Dansgaard, 1964), mostly depends on the sea surface temperature (SST) and relative humidity (RH) at the moisture source region (Merlivat and Jouzel, 1979; Stenni et al., 2001; Uemura et al., 2008; Jouzel et al., 2013; Steen-Larsen et al., 2014b). The d-excess is also modified significantly by the distillation processes along the air mass trajectory (Hendricks et al., 2000; Xiao et al., 2013) and therefore affected by changes in condensation temperature (Uemura et al., 2004;

*

Corresponding author. E-mail address: [email protected] (S. Hou).

http://dx.doi.org/10.1016/j.epsl.2015.01.014 0012-821X/© 2015 Elsevier B.V. All rights reserved.

Masson-Delmotte et al., 2008). Thus, the d-excess records in polar ice cores have been used to estimate past climate conditions (SST and/or RH) in the moisture source region (Jouzel et al., 1982; Vimeux et al., 1999; Stenni et al., 2003; Uemura et al., 2012). However, these reconstructions are limited because there are three unknown parameters (condensation temperature, source SST and RH) and only two observational constraints (δ 18 O and δ D), leading to an under-constrained inverse problem. With the improvement of water fluorination technique (Barkan and Luz, 2005), the ability to measure H2 17 O (δ 17 O) in water with high precision led to the definition of another secondorder parameter: 17 O-excess = 106 × (ln(δ 17 O/1000 + 1) − 0.528 × ln(δ 18 O/1000 + 1)) (Landais et al., 2006; Barkan and Luz, 2007). Unlike the d-excess, the 17 O-excess is mainly controlled by the relative humidity at the moisture source region and is by construction insensitive to source temperature (Landais et al., 2008; Risi et al., 2010; Uemura et al., 2010). First studies with this parameter suggest that precipitation 17 O-excess in polar regions is expected to give information on relative humidity at the evaporation regions

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

127

Fig. 1. Map of Antarctica showing sampling locations of surface snow from Zhongshan station to Dome A (the filled blue circles). The sampling locations from Terra Nova Bay to Dome C (Landais et al., 2008) (the filled green circles) together with other Antarctic ice cores sites (the filled cyan circle: WDC, the WAIS Divide ice core; the unfilled cyan circle: Siple Dome; the unfilled black circle: Taylor Dome; the unfilled red circle: Talos Dome; the filled red circle: Dome C; the filled black circle: Vostok; the filled purple circle: Dome F; the unfilled purple circle: EDML) are also indicated.

(Landais et al., 2008, 2012b; Barkan and Luz, 2007; Winkler et al., 2012). As a consequence, combining the d-excess and 17 O-excess parameters opens the possibility to constrain past changes in evaporation conditions. However, recent studies have suggested that other processes could also impact precipitation 17 O-excess: precipitation re-evaporation (Landais et al., 2010), mixing between vapors from different origins (Risi et al., 2010), kinetic isotopic fractionation associated with condensation of vapor over ice crystals at low temperatures under supersaturation conditions (Angert et al., 2004; Landais et al., 2012a; Risi et al., 2013; Schoeneman et al., 2014), and possible stratospheric vapor intrusions at sites in interior Antarctica where the temperature and snow accumulation rate are very low (Miller, 2008; Winkler et al., 2013). As for d-excess, processes controlling 17 O-excess in polar regions are complex, not fully understood, with only sparsely available data. Up to now, there is only one Antarctic transect from Terra Nova Bay to Dome C (hereafter we call it Dome C transect) where the 17 O-excess in surface snow samples was measured by Landais et al. (2008) and there are only several locations in Antarctica where the 17 O-excess measurements in core cores were performed (Fig. 1 and Table 1). Documenting the spatial variability of 17 O-excess in Antarctica is an important gap to fill to progress toward a quantitative understanding of this new proxy. In this study, we present new 17 O-excess measurements in surface snow along an East Antarctica traverse from Zhongshan station to Dome A (Fig. 1), in an undocumented area which encompasses the highest point of Antarctic ice sheet and the most intense present-day isotopic depletion. Section 2 describes the

sampling and analytical protocols. The results are described in Section 3, and compared with theoretical distillation calculations performed with a Rayleigh-type model introduced in Section 4. The model-data comparison and discussion (Section 5) is followed by our conclusions. 2. Sampling and isotopic measurements Surface snow samples were collected along a route from Zhongshan Station to Dome A (80◦ 22 51 , 77◦ 27 23 , 4093 m a.s.l., the highest point of Antarctic ice sheet) during the 26th Chinese National Antarctic Research Expedition (CHINARE) from December 2009 to January 2010. The CHINARE route is located approximately along the 77◦ E longitude line and is about 1250 km long (Fig. 1). During the journey to Dome A, 42 surface snow samples (named B1, B2, . . . , B42) were collected along the route with an average step of approximately 30 km. For each location, the surface snow layer (10 cm in depth) was sampled, and collected using low density polythene plastic bags. All the samples were kept frozen during transportation and storage. According to field observations by stakes on surface mass balance at 2 km intervals along the Zhongshan station to Dome A conducted in 2005 and 2008 (Ding et al., 2011), the percent of surface 10 cm snow layer to its annual net accumulation ranges from 21% to 38% at the coastal region (elevation below 2000 m a.s.l.), and the average percent 102% over the Dome A region (elevation above 4000 m a.s.l.). Water isotopologues (δ 17 O, δ 18 O and δ D) measurements were performed at the Laboratoire des Sciences du Climat et de l’Environnement

128

Table 1 Observations of

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

17

O-excess in surface snow and ice cores at different sites in Antarctica. Sample type

Time scale

Elevation (m)

Annual mean temperature (◦ C)

δ 18 O (h)

Early Holocene Early Holocene Early Holocene Early Holocene Early Holocene Late Holocene Annual

3502

−55.5

−56

40

3240

−53.5

−50.7

23

2316

−41.0

−36.6

2.6

Winkler et al. (2012)

2892

−43.2

−44.8

8

1766

−28.5

−33.5

29

621

−24.5

−26.2

21

39.70

Ice core Ice core Ice core Ice core Ice core Ice core Snow

3810

−54.3

−58.0

1

158.26

Snow

Annual

2365

−41

−40.2

8.9

Surface snow

Annual

3733–4092

−57.3

−57.4

Risi et al. (2013) (unpublished data) Schoenemann et al. (2014) Schoenemann et al. (2014) Luz and Barkan (2010) Steig et al. (2000); Risi et al. (2013) This study

Location

Latitude

Longitude

Vostok

−78.45

106.85

Dome C

−75.4

123.14

Talos Dome

−72.82

159.18

EDML

−75.0

WDC

−79.47

−112.09

Siple Dome

−81.67

−148.82

Dome F

−77.32

Taylor Dome

−77.28

Dome A regiona

−79.02 to −80.42

0.07

77.02–77.12

17

O-excess (per meg)

21

References

Landais et al. (2008); Winkler et al. (2012) Winkler et al. (2012)

a Here, the Dome A region indicates the area with elevation above 3700 m a.s.l. Within the Dome A region, 3 surface snow samples (B40–B42) were collected and the averages of annual mean temperature, δ 18 O, and 17 O-excess at the three locations are used.

(LSCE), France. Deuterium isotopic ratios (δ D) were analyzed by chromium reduction method with an uncertainty 0.7h. For oxygen isotopes, water samples were converted to oxygen by fluorination with CoF3 , and the produced oxygen was then analyzed by dual inlet mass spectrometer for δ 17 O and δ 18 O. The analytical uncertainty is 0.05h for both δ 17 O and δ 18 O and 5 per meg for 17 Oexcess. Details on the water fluorination procedure and the dual inlet mass spectrometer analysis can be found elsewhere (Landais et al., 2012a, 2012b; Winkler et al., 2012). As for the data calibration, we have used internal standards calibrated vs. V-SMOW and SLAP following the procedure described in Schoenemann et al. (2013) with respective values for δ 18 O and 17 O-excess of 0h and 0 per meg for V-SMOW and −55.5h and 0 per meg for SLAP. At the beginning and at the end of our study, we ran all our internal standards in duplicate on the fluorination line and mass spectrometer to have a faithful relationship between true and measured value. In addition, at least one internal standard was run on the fluorination line each day and then on the mass spectrometer together with the bunch of 5 samples ran in duplicate. The raw δ 18 O and 17 O-excess measured values for the internal standards did not vary by more than 0.1h and 5 per meg during the whole period hence ensuring the validity of our calibration. A quadratic error for d-excess is 0.8h, estimated by the uncertainties of δ 18 O and δ D. 3. Results Fig. 2 displays the spatial distribution of 17 O-excess in surface snow from Zhongshan station to Dome A as a function of latitude, altitude, annual mean surface air temperature, and δ 18 O. Along this route, 17 O-excess varies from 9 per meg to 51 per meg with a mean value of 35 per meg. Despite this large variability, we observe a significant inland decrease, with a latitude gradient of −1.33 ± 0.41 per meg/degree (Fig. 2a), an altitudinal gradient of −0.48 ± 0.17 per meg/100 m (Fig. 2b), and a temperature gradient of 0.35 ± 0.11 per meg/◦ C (Fig. 2c). Because δ 18 O also decreases inland due to distillation caused by air mass cooling, we observe a spatial regression slope of 0.36 ± 0.10 per meg h−1 between 17 Oexcess and δ 18 O (Fig. 2d). It is noted that the spatial distribution of 17 O-excess in surface snow from Zhongshan to Dome A is different with that from Dome C transect. Firstly, we obtain 17 O-excess values with δ 18 O much lower than for the Dome C transect. Analyzing 17 O-excess in

snow over this range of δ 18 O is very important for the interpretation of 17 O-excess in ice cores. Secondly, the 17 O-excess in surface snow along the Dome C transect remains constant from the coast to the continental interior, whereas we observe a significant inland decrease of 17 O-excess. Additionally, we find the observed spatial regression slope between 17 O-excess and δ 18 O (0.36 per meg h−1 ) in surface snow along the route to Dome A is much smaller than the seasonal regression slope between 17 O-excess and δ 18 O in snow precipitation (2.96 per meg h−1 ) observed at the Vostok site (Landais et al., 2012a) (Fig. 2d). 4. Modeling of spatial distribution of 17 O-excess in polar snow 4.1. Mixed cloud isotopic model (MCIM) Theoretical calculations of isotopic fractionation occurring in Antarctica can be performed using Rayleigh-type model. The Mixed Cloud Isotopic Model (MCIM) has been specifically developed for this purpose and has been successful in simulating the spatial distributions of δ 18 O and d-excess in Antarctica (Ciais and Jouzel, 1994; Vimeux et al., 1999; Masson-Delmotte et al., 2004, 2008). This model has recently been equipped with δ 17 O (Landais et al., 2008) and applied for Greenland and Antarctica (Landais et al., 2012a, 2012b; Winkler et al. 2012, 2013). The MCIM accounts for four isotopic fractionation processes: (1) the kinetic fractionation process during evaporation at the original moisture source region, (2) the equilibrium fractionation process during all phase transitions, (3) the kinetic fractionation during the Bergeron–Findesein process at low temperature where vapor, supercooled droplets and ice crystals coexist in the cloud, and (4) the kinetic fractionation process where there is no more liquid droplets in the cloud and the vapor is supersaturated with respect to the ice crystal. The physical frames and parameterizations of these processes can be found elsewhere (Merlivat and Jouzel, 1979; Jouzel and Merlivat, 1984; Ciais and Jouzel, 1994). 4.2. Forcing and tuning of the MCIM The MCIM is run using values for (1) climatic parameters over the moisture source region, such as ocean source temperature (T source ), relative humidity (RH), wind speed and air pressure as

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

129

Fig. 3. Comparison of observed 17 O-excess (upper panel) and d-excess (lower panel) as a function of δ 18 O along the traverse from Zhongshan to Dome A (the filled blue circles) with the MCIM simulated values by varying the supersaturation function S with fixed climatic parameters. The analytical uncertainty associated with the 17 O-excess measurements is indicated for each sample with a vertical error bar. For comparison, the 17 O-excess values in surface snow from the Dome C transect (the filled green circles) and seasonal variations of 17 O-excess in snow precipitation at the Vostok station (the unfilled black squares) are also indicated (Landais et al., 2008, 2012a).

quantifies the proportion of the re-evaporation of liquid, and the experimental supersaturation function of S = p − q × T c . The linear dependence of S with T c is often tuned with p ranging between 1 and 1.02 and q between 0.002 and 0.004 (Ciais and Jouzel, 1994; Masson-Delmotte et al., 2004; Vimeux et al., 2001; Stenni et al., 2010; Uemura et al., 2012; Landais et al., 2012b). We first adjusted different parameters of the MCIM for obtaining the best simulations of δ 18 O, d-excess and 17 O-excess along the route from Zhongshan station to Dome A. We then studied the sensitivity of the spatial distribution of 17 O-excess along this route to changes in the prescribed supersaturation function. Finally, and having in mind the quantitative interpretation of isotopic records from Dome A ice cores, we tested the sensitivity of the isotopic compositions (δ 18 O, d-excess and 17 O-excess) at Dome A to the source and site climatic parameters (T source , RH and T site ). 4.3. Comparison of MCIM simulations with our traverse data Fig. 2. Variations of 17 O-excess in surface snow from Zhongshan station to Dome A with latitude (a), altitude (b), annual mean temperature (c) and δ 18 O (d). The analytical uncertainty associated with the 17 O-excess measurements is indicated for each sample with a vertical error bar. The annual mean temperatures at sample locations from Zhongshan station to Dome A were interpolated by a multiple linear regression model (Wang et al., 2010). The 17 O-excess data from Zhongshan station to Dome A are indicated by the filled blue circles and the 17 O-excess data from Dome C transect are indicated by the filled green circles. The other filled or unfilled color circles are other Antarctic ice cores sites in Fig. 1, which have the same color codes as in Fig. 1. For comparison, seasonal variations of 17 O-excess in snow precipitation at Vostok station with δ 18 O are also indicated in Fig. 2d (the unfilled black squares). The blue lines are regression lines between 17 O-excess and latitude, altitude, annual mean temperature and δ 18 O from Zhongshan station to Dome A. The black line is the regression line between the seasonal 17 O-excess and δ 18 O in snow precipitation at the Vostok station.

well as the isotopic composition of ocean water, and (2) climatic parameters at the site of precipitation, namely the condensation temperature in the cloud (T c ) and surface air pressure. In Antarctica, due to thermal inversion, T c is evaluated at first order following the equation T c = 0.67 × T site − 1.2 (Jouzel and Merlivat, 1984), where T site is the surface temperature at the site of precipitation. There are several adjustable parameters such as the temperature range where liquid and solid water can coexist, a coefficient that

In order to simulate spatial distribution of 17 O-excess and dexcess along the traverse from Zhongshan station to Dome A, we have tested many different sets of the MCIM parameters. For the climatic parameters over the moisture source region, we independently varied T source between 5 ◦ C and 25 ◦ C and RH between 0.65 and 0.95. We set the supersaturation function as S = 1 − q × T c and tuned q between 0.002 and 0.004. The condensation temperature at precipitation site has been adjusted in order to reproduce the observed spatial distribution of d-excess and 17 O-excess. The observations of 17 O-excess and d-excess along the route from Zhongshan station to Dome A can be simulated well when setting T source = 23 ◦ C, RH = 0.78 and S = 1 − 0.0033T c (Fig. 3). The source climatic conditions that enable to reasonable fit the measured isotopic level with the model are representative of subtropical oceans, consistent with the previous results that the primary source of moisture for snow in Antarctica originates from the subtropical oceans (Petit et al., 1991; Ciais et al., 1995; Werner et al., 2001). However, this is not consistent with back-trajectory calculations performed using a Lagrangian moisture source diagnostic for 5 recent years, showing that Dome A moisture mostly comes from the mid-latitude South Indian Ocean (46 ± 4◦ S) (Wang et al., 2013). We stress that the observed 17 O-excess and d-excess along

130

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

the traverse from Zhongshan station to Dome A cannot be simulated by the MCIM model using such climatic conditions (lower source temperature and higher relative humidity) observed today in the mid-latitude South Indian Ocean. Similar mismatches between MCIM results and back-trajectory or water tagging calculations have been identified for other Greenland and Antarctic sites, with a tendency to attribute more tropical sources to the polar sites with the MCIM than in other methods (Armengaud et al., 1998; Masson-Delmotte et al., 2004; Steen-Larsen et al., 2011; Landais et al., 2012b). One reason for such mismatch may arise from the closure assumption used to calculate the initial vapor isotopic composition based on evaporation conditions. Additionally, we are aware that some processes such as mixing of air masses, recycling of water vapor in the atmosphere, and postdeposition processes of snow are not taken into account in the MCIM. Further investigations are needed for clarifying the bias between the MCIM and the Lagrangian moisture source diagnostic, for instance using continuous monitoring of water vapor isotopic composition along a traverse from moisture sources towards Antarctica (Steen-larsen et al., 2013, 2014b), which remains a challenge at very low humidity levels typical of central Antarctic climate. A supersaturation function of S = 1 − 0.004T c to S = 1 − 0.002T c was used in both simple Rayleigh-type fractionation models and complex isotope-enabled GCMs for modeling water isotopologues in polar precipitation (Landais et al., 2012a, 2012b; Winkler et al., 2012; Risi et al., 2013; Schmidt et al., 2005; Lee et al., 2007). Each time, the choice of the supersaturation function has been adjusted to fit the isotopic composition of available data (surface transect, seasonal cycles). The differences in the relationship between supersaturation function and temperature depend on the model that is used and on the available data (the same fitting will not apply for Greenland, coastal Antarctica or central Antarctica). In the case of the transect from the coast to Dome A, our sensitivity tests to different values of q show a reasonable fit with our dataset when using the supersaturation function as S = 1 − 0.0033T c . It is noted that S determined here is the same as the one found for the same MCIM model by Winkler et al. (2012) based on the δ 18 O, 17 O-excess and d-excess on the transect from Terra Nova Bay to Dome C and the mean ice-core isotopic values of early Holocene at Vostok, Dome C and Talos Dome; it is lower than the value tested by Schoenemann et al. (2014). We conclude that a supersaturation function of S = 1 − 0.0033T c seems a good parameterization of the MCIM model applied to the Antarctic plateau since it fits reasonably the isotopic data (δ 18 O, d-excess and 17 O-excess) currently available (Landais et al., 2008; Winkler et al., 2012; this study). Due to the changes in the equilibrium fractionation coefficient of H2 18 O (18 αeq ) with temperature and nonlinearity of 17 O-excess to δ 18 O relationship, 17 O-excess is expected to increase with decreasing δ 18 O by a simple Rayleigh equilibrium distillation in a cooling air mass (Angert et al., 2004; Luz and Barkan, 2010); we call this mechanism the equilibrium distillation effect on 17 Oexcess. The MCIM simulations performed with different supersaturation functions show that this equilibrium distillation effect may be significant for 17 O-excess when the supersaturation effect is low (q = 0.002) (Fig. 3). However, under high supersaturation conditions, this equilibrium distillation effect is counterbalanced by the supersaturation effect (Fig. 3), suggesting that the supersaturation effect is an important factor for controlling spatial distribution of 17 O-excess in polar precipitation. The observed decreasing 17 O-excess in surface snow from Zhongshan station to Dome A with decreasing temperature/δ 18 O (Fig. 2c and 2d) confirms the influence of the supersaturation effect on 17 O-excess at low temperature.

5. Discussion 5.1. Comparison with other sites 5.1.1. Spatial variability We now compare our results with the other available Antarctic surface snow 17 O-excess data (Landais et al., 2008), measured along a transect from Terra Nova Bay to Dome C (Fig. 1). The Dome C transect shows stable 17 O-excess levels from the coast to the inland sites, which contrasts with our inland decreasing trend towards Dome A. The flat pattern of 17 O-excess over the Dome C transect is probably due to the following reasons. Firstly, the Dome C transect covers less than 2◦ latitude in the south– north direction from 74.0◦ S to 75.9◦ S and extends 36◦ longitude in the east–west direction from 124.9◦ E to 160.9◦ E. Due to this configuration, we suspect that the Dome C transect receives moisture from different moisture sources, as suggested by Lagrangian backtrajectories (Sodemann and Stohl, 2009). The supersaturation effect at low temperature is expected to induce a gradual decrease of precipitation 17 O-excess from Terra Nova Bay to Dome C; it could be however masked if the initial vapor formed at the moisture source region of Dome C has a larger 17 O-excess than the vapor formed at the initial source region of Terra Nova Bay. Along this traverse, the variability of temperature, snow accumulation, aerosol deposition and δ 18 O-temperature slopes varying from 0.60h/◦ C to 0.91h/◦ C (Proposito et al., 2002; Masson-Delmotte et al., 2008), support our inference that the Dome C transect may receive moisture from different sources. Secondly, the Dome C transect does not cover the very low δ 18 O range (below −51h) because the annual mean temperature at Dome C (−53.5 ◦ C) (Jouzel et al., 2001) is not low enough. It is noted that the supersaturation effect on 17 Oexcess increases with decreasing temperature. The relatively weak supersaturation effect on 17 O-excess (decreasing 17 O-excess) from the Dome C transect due to its relatively high temperatures probably counteracts the equilibrium distillation effect on 17 O-excess (increasing 17 O-excess). This probably causes a flat evolution of 17 O-excess in surface snow from the Dome C transect. The Dome C traverse is complemented by existing Holocene measurements of 17 O-excess from different Antarctic ice cores (Fig. 1 and Table 1). These data show no evidence for a coastal to inland trend, and a large spatial variability (Fig. 2 and Table 1). For instance, very high and low 17 O-excess values can be found in both interior land (e.g., high 17 O-excess values at Vostok and low value at Dome F) and coastal regions (e.g., high values at nearby Zhongshan station and Terra Nova Bay, and low values at Talos Dome and Taylor Dome). In the central cold Antarctic plateau, we speculate that there is a significant supersaturation effect on 17 Oexcess at Dome C, Vostok, Dome A and Dome F. However, there is no clear relationship between the observed 17 O-excess values and condensation temperature at those sites, probably suggesting that the supersaturation effect on 17 O-excess is different at different sites in the interior Antarctica. The complexity of the observed spatial distribution suggests that the spatial distribution of 17 O-excess in present-day Antarctic surface snow cannot be simply explained by the moisture source location and its relative humidity, nor by a temperature-dependent supersaturation effect only. Our conclusion is not in agreement with more recent study by Schoenemann et al. (2014) who concluded that kinetic isotope effects resulting from supersaturation of water vapor over ice dominate the spatial pattern of modern 17 O-excess in Antarctic precipitation. It is expected that the spatial variability of 17 O-excess in Antarctica should be governed by both the initial 17 O-excess value in vapor over the moisture source region and the supersaturation effect at the precipitation site. For the former, the measurements of 17 O-excess in vapor at the moisture source regions are very sparse to this day, although we can calculate it based on the closure assumption. For

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

the latter, the supersaturation function is often estimated by surface temperature based on a regression slope between condensation temperature and surface temperature. However, the slope may be different at different sites due to variations of the atmospheric inversion layer over Antarctica. As a consequence, more measurements of 17 O-excess in vapor at the moisture source regions and the temperature of the inversion layer at sites in Antarctica are needed. 5.1.2. Comparison with seasonal variability The spatial regression slope between 17 O-excess and δ 18 O (0.36 per meg h−1 , Fig. 2d) found in this study is much smaller than the seasonal regression slope between 17 O-excess and δ 18 O (2.96 per meg h−1 ) observed at the Vostok site (Landais et al., 2012a). This difference is not unexpected since the seasonal variations may not directly be compared to the average values retrieved from the transect at various locations in East Antarctica. Indeed, seasonal variations in the profile of atmospheric temperature are expected. The seasonal variations observed at Vostok also include variations in the types of precipitation (hoar, snow, diamond dust) and the relationship between 17 O-excess and δ 18 O could be influenced by the type of precipitation. Moreover, surface snow 17 O-excess, in opposite to precipitation samples analyzed for the seasonal variations at Vostok, are also affected by post-deposition processes associated with firn diffusion (Steen-Larsen et al., 2014a), which is not yet known. Finally, several authors further suggested that a high 17 O-excess stratospheric water vapor source may provide a contribution to the observed high 17 O-excess level measured in Vostok snow (Miller, 2008; Winkler et al., 2013). Further investigations would also require to estimate the differences in condensation temperature and source humidity between the different sites, which could be achieved thanks to radio-sounding and remote sensing measurements. Finally, our dataset complements existing records of 17 O-excess in Antarctica and highlights a previously undocumented decrease of 17 O-excess under the coldest/driest modern conditions, but is not sufficient to unveil the control mechanisms. It calls for laboratory experiments as well as more measurements of 17 O-excess in Antarctic water vapor, precipitation and surface snow. 5.2. Implications for the interpretation of Dome A ice cores After optimizing the MICM tuning with respect to the traverse data, we can now use this model setup to investigate the sensitivity of water isotopologues (δ 18 O, d-excess and 17 O-excess) at Dome A to the climatic parameters (T source , RH and T site ), having in mind the interpretation of deep ice core records. Following earlier studies applied to other sites (Cuffey and Vimeux, 2001; Vimeux et al., 2002; Masson-Delmotte et al., 2004; Stenni et al., 2010; Landais et al., 2009, 2012a; Risi et al., 2010; Winkler et al., 2012; Uemura et al., 2012), we vary climatic parameters one by one independently. The linear analysis of the results therefore allows to quantify the sensitivity of the Dome A precipitation isotopic composition to each climatic parameter. Below we summarize the outputs of our sensitivity analysis.

δ 18 O = 0.94 T site − 0.52 T source − 0.05RH

(1)

d-excess = 1.6 T source − 1.8 T site − 0.18RH

(2)

17 O-excess = −1.1RH + 0.33 T site

(3) 18

(0.94h/◦ C)

The linear regression slope between δ O and T site is close to what was obtained previously with the MCIM by Landais et al. (2009) (1.15h/◦ C) or Risi et al. (2010) (1.0h/◦ C) at the Vostok site. The simulation δ 18 O-temperature slope is significantly larger than the average observed in Antarctica (0.80h/◦ C), but

131

it is equal to the observed δ 18 O-temperature slope obtained at some sites with the elevation above 3300 m in the central Antarctic plateau, including Dome A (Masson-Delmotte et al., 2008). We therefore expect an enhanced temperature sensitivity of water isotopologues at Dome A. The sensitivity coefficient of d-excess to T source (1.6h/◦ C) is close to previously values obtained with the MCIM by Risi et al. (2010) at Vostok (1.5h/◦ C), Uemura et al. (2012) at Dome F (1.6h/◦ C), and Uemura et al. (2012) at Dome C (1.5h/◦ C). However, the sensitivity of d-excess to T site (−1.8h/◦ C) is significantly larger than previously values obtained with the MCIM by Risi et al. (2010) at Vostok (−1.1h/◦ C), Uemura et al. (2012) at Dome F (−1.3h/◦ C), and Uemura et al. (2012) at Dome C (−1.2h/◦ C). The strong sensitivity of the d-excess to T site at Dome A is due to the dependency of equilibrium fractionation coefficients on temperature, and the decrease of the meteoric water line slope at cold temperatures. We therefore expect a strong equilibrium distillation effect on d-excess at Dome A. As a result, extracting the information on past changes in moisture source surface temperature by ice-core d-excess records from the central Antarctic plateau will critically depend on this correction to site temperature estimates (Uemura et al., 2004, 2012; Stenni et al., 2010). In Eq. (3), Dome A 17 O-excess is simulated to mainly depend on changes in moisture source relative humidity (assuming a constant supersaturation function). The simulated sensitivity of 17 O-excess on RH over the moisture source region (−1.1h/%) is consistent with earlier results (−1.0h/%) obtained with the MCIM model, albeit with different tuning, for the Vostok and Dome C sites (Landais et al., 2009; Risi et al., 2010; Winkler et al., 2012). The simulation sensitivity of 17 O-excess to T site (0.33 per meg/◦ C) is consistent with the observed value (0.35 per meg/◦ C) obtained from Zhongshan station to Dome A (Fig. 2c) and is within the range (−0.2 per meg/◦ C to 0.5 per meg/◦ C) obtained with the MCIM by tuning supersaturation (Winkler et al., 2012). As expected from the definition of 17 O-excess, its sensitivity to T site is lower than the sensitivity of d-excess to T site , implying more limited influence of changes in local temperature on 17 O-excess than for d-excess. In order to deduce humidity information at the oceanic vapor source by the 17 O-excess records in ice cores in the remote Antarctica, the local temperature effect on 17 Oexcess in ice cores (i.e., the supersaturation effect) should be removed. This exercise has already been done in Landais et al. (2012a). In this study the record of 17 O-excess in ice core from Vostok during the last 150,000 yr was corrected by the seasonal tendency (17 O-excess)/(δ 18 O) (2.96 per meg h−1 ) observed in the same site. After correction from this local effect, the remaining 17 O-excess signal does not show any significant increase over the deglaciation, suggesting that the entire 17 O-excess increase observed at this very cold site was only linked to local temperature increase. However, if we use the spatial tendency (17 O-excess)/(δ 18 O) (0.36 per meg h−1 , Fig. 2d) observed in our study to correct the Vostok ice core 17 O-excess record, we remove only 2.4 per meg to the 20 per meg 17 O-excess increase over the deglaciation. This correction is however perhaps underestimated because the spatial relationship (17 O-excess)/(δ 18 O) of 0.36 per meg h−1 is obtained from a regression over a large range of δ 18 O values. A stronger relationship would be obtained if the regression is performed only for δ 18 O between −50 and −60 permil. For very remote regions of Antarctica with very low temperatures, we expect strong distillation and supersaturation effects on d-excess and 17 O-excess. Because the two strong different contributions acting one against the other, it is very important to quantify their influences if one want to obtain any quantitative information on past climate from water isotopologues in these drilling places. Up to now, the best way to parameterize the models to

132

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

infer the contributions of the distillation and supersaturation effects is to fit the model to existing d-excess and 17 O-excess data in these regions. The data presented here are thus very precise and confirm some recent modeling exercises aiming at quantifying the influences of climatic parameters at the site and at the source on water isotopic records in remote regions of Antarctica (Landais et al., 2009; Risi et al., 2010; Winkler et al., 2012; Uemura et al., 2012). 6. Conclusions The 17 O-excess in surface snow along the traverse from Zhongshan station to Dome A shows for the first time a significant decreasing trend from the coast to the central Antarctica, a feature which could not be seen from earlier measurements along the transect from Terra Nova Bay to Dome C, and from Holocene ice core measurements from other East Antarctic ice cores. This observed inland decrease of 17 O-excess is explained theoretically by the impact of kinetic fractionation associated with supersaturation on ice crystal at low temperature. Our dataset expands the documentation of the spatial distribution of 17 O-excess in Antarctic precipitation, showing large variability. This variability may result from differences in moisture sources and differences in supersaturation along air mass trajectories. Our data are used to optimize the parameters of a Rayleigh-type distillation model for the Dome A site, providing a framework for the quantitative interpretation of ice core water isotopologues data expected from the Dome A deep ice core project. Large uncertainties however remain on the understanding of present-day 17 O-excess in Antarctica, calling for an improved documentation of its variability in vapor, precipitation and surface snow. Acknowledgements This work was jointly supported by the National Natural Science Foundation of China (41176165, 41330526, 41171052, 41321062 and 41206175), the Program for Chinese National Antarctic and Arctic Research Expedition (CHINARE2014-02-02), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the Fundamental Research Funds for the Central Universities (1082020904), and the ERC Starting Grant COMBINISO (306045). We are grateful to many scientists, technicians and porters for their hard work in the field. Appendix A. Supplementary material Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.epsl.2015.01.014. References Angert, A., Cappa, C.D., DePaolo, D.J., 2004. Kinetic 17 O effects in the hydrologic cycle: indirect evidence and implications. Geochim. Cosmochim. Acta 68, 3487–3495. Armengaud, A., Koster, R.D., Jouzel, J., Ciais, P., 1998. Deuterium excess in Greenland snow—analysis with simple and complex models. J. Geophys. Res. 103, 8947–8953. Barkan, E., Luz, B., 2005. High precision measurements of 17 O/16 O and 18 O/16 O ratios in H2 O. Rapid Commun. Mass Spectrom. 19, 3737–3742. Barkan, E., Luz, B., 2007. Diffusivity fractionations of H2 16 O/H2 17 O and H2 16 O/H2 18 O in air and their implications for isotope hydrology. Rapid Commun. Mass Spectrom. 21, 2999–3005. Ciais, P., Jouzel, J., 1994. Deuterium and oxygen 18 in precipitation: isotopic model, including mixed cloud processes. J. Geophys. Res. D99, 16793–16803. Ciais, P., White, J.W.C., Jouzel, J., Petit, J.R., 1995. The origin of present-day Antarctic precipitation from surface snow deuterium excess data. J. Geophys. Res. 100, 18917–18927. Cuffey, K.M., Vimeux, F., 2001. Covariation of carbon dioxide and temperature from the Vostok ice core after deuterium-excess correction. Nature 412, 523–527.

Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 16, 436–468. Dansgaard, W., Johnsen, S.J., Moller, J., Langway, C.C., 1969. One thousand centuries of climatic record from Camp Century on the Greenland ice sheet. Science 166, 377–381. Ding, M.H., Xiao, C.D., Li, Y.S., Ren, J.W., Hou, S.G., Jin, B., Sun, B., 2011. Spatial variability of surface mass balance along a traverse route from Zhongshan station to Dome A, Antarctica. J. Glaciol. 57 (204), 658–666. http://dx.doi.org/ 10.3189/002214311797409820. EPICA community members, 2004. Eight glacial cycles from an Antarctic ice core. Nature 429 (6992), 623–628. http://dx.doi.org/10.1038/nature02599. Hendricks, M., DePaolo, D., Cohen, R., 2000. Space and time variations of δ 18 O and δ D: can paleotemperatures be estimated from ice cores? Glob. Biogeochem. Cycles 14, 851–861. Johnsen, S.J., Dansgaard, W., Clausen, H.B., Langway, C.C., 1972. Oxygen isotope profiles through the Antarctic and Greenland ice sheets. Nature 235 (5339), 429–434. Jouzel, J., Merlivat, L., 1984. Deuterium and oxygen 18 in precipitation: modeling of the isotopic effects during snow formation. J. Geophys. Res. D89, 11749–11757. Jouzel, J., Merlivat, L., Lorius, C., 1982. Deuterium excess in an East Antarctic ice core suggests higher relative humidity at the oceanic surface during the last glacial maximum. Nature 299, 688–691. Jouzel, J., Lorius, C., Petit, J.R., Genthon, C., Barkov, N.I., Kotlyakov, V.M., Petrov, V.M., 1987. Vostok ice core: a continuous isotope temperature record over the last climatic cycle (160,000 years). Nature 329, 402–408. Jouzel, J., Masson, V., Cattani, O., Falourd, S., Stievenard, M., Stenni, B., Longinelli, A., Johnsen, S.J., Steffenssen, J.P., Petit, J.R., Schwander, J., Souchez, R., Barkov, N.I., 2001. A new 27 ky high resolution East Antarctic climate record. Geophys. Res. Lett. 28 (16), 3199–3202. http://dx.doi.org/10.1029/2000GL012243. Jouzel, J., Delaygue, G., Landais, A., Masson-Delmotte, V., Risi, C., Vimeux, F., 2013. Water isotopes as tools to document oceanic sources of precipitation. Water Resour. Res. 49, 7469–7486. http://dx.doi.org/10.1002/2013WR013508. Landais, A., Barkan, E., Luz, B., 2006. The triple isotopic composition of oxygen in leaf water. Geochim. Cosmochim. Acta 70, 4105–4115. Landais, A., Barkan, E., Luz, B., 2008. Record of δ 18 O and 17 O-excess in ice from Vostok Antarctica during the last 150,000 years. Geophys. Res. Lett. 35, L02709. http://dx.doi.org/10.1029/2007GL032096. Landais, A., Barkan, E., Vimeux, F., Masson-Delmotte, V., Luz, B., 2009. Combined analysis of water stable isotopes (H2 16 O, H2 17 O, H2 18 O, HD16 O). In: Hondoh, T. (Ed.), PICR “Physics of Ice-Core Records”. Institute of Low Temperature Science, Kokkaido University Press, Sapporo, pp. 315–327. Landais, A., Risi, C., Bony, S., Vimeux, F., Descroix, L., Falourd, S., Bouygues, A., 2010. Combined measurements of 17 O-excess and d-excess in African monsoon precipitation: implications for evaluating convective parameterizations. Earth Planet. Sci. Lett. 298, 104–112. Landais, A., Ekaykin, A., Barkan, E., Winkler, R., Luz, B., 2012a. Seasonal variations of 17 O-excess and d-excess in snow precipitation at Vostok station, East Antarctica. J. Glaciol. 58, 725–733. Landais, A., Steen-Larsen, H.C., Guillevic, M., Masson-Delmotte, V., Vinther, B., Winkler, R., 2012b. Triple isotopic composition of oxygen in surface snow and water vapor at NEEM (Greenland). Geochim. Cosmochim. Acta 77, 304–316. Lee, J.-E., Fung, I., DePaolo, D.J., Henning, C.C., 2007. Analysis of the global distribution of water isotopes using the NCAR atmospheric general circulation model. J. Geophys. Res. 112, D16306. http://dx.doi.org/10.1029/2006JD007657. Luz, B., Barkan, E., 2010. Variations of 17 O/16 O and 18 O/16 O in meteoric waters. Geochim. Cosmochim. Acta 74, 6276–6286. Masson-Delmotte, V., Stenni, B., Jouzel, J., 2004. Common millennial-scale variability of Antarctic and Southern Ocean temperatures during the past 5000 years reconstructed from the EPICA Dome C ice core. Holocene 14 (2), 145–151. Masson-Delmotte, V., et al., 2008. A review of Antarctic surface snow isotopic composition: observations, atmospheric circulation, and isotopic modeling. J. Climate 21, 3359–3387. http://dx.doi.org/10.1175/2007JCLI2139.1. Merlivat, L., Jouzel, J., 1979. Global climatic interpretation of the deuterium–oxygen 18 relationship for precipitation. J. Geophys. Res. 84, 5029–5033. Miller, M.F., 2008. Comments on “Record of δ 18 O and 17 O-excess in ice from Vostok Antarctica during the last 150,000 years” by Amaelle Landais et al. Geophys. Res. Lett. 35, L23709. http://dx.doi.org/10.1029/2008GL034505. Petit, J.R., White, J.W.C., Young, N.W., Jouzel, J., Korotkevich, Y.S., 1991. Deuterium excess in recent Antarctic snow. J. Geophys. Res. 96 (D), 5113–5122. Proposito, M., Becagli, S., Castellano, E., Flora, O., Genoni, L., Gragnani, R., Stenni, B., Traversi, R., Udisti, R., Frezzotti, M., 2002. Chemical and isotopic snow variability along the 1998 ITASE traverse from Terra Nova Bay to Dome C, East Antarctica. Ann. Glaciol. 35 (1), 187–194. Risi, C., Landais, A., Bony, S., Masson-Delmotte, V., Jouzel, J., Vimeux, F., 2010. Understanding the 17 O-excess glacial–interglacial variations in Vostok precipitation. J. Geophys. Res. 115, D10112. http://dx.doi.org/10.1029/2008JD011535. Risi, C., Landais, A., Winkler, R., Vimeux, F., 2013. Can we determine what controls the spatio-temporal distribution of d-excess and 17 O-excess in precipitation using the LMDZ general circulation model? Clim. Past 9, 2173–2193. Schmidt, G.A., Hoffmann, G., Shindell, D.T., Hu, Y., 2005. Modeling atmospheric stable water isotopes and the potential for constraining cloud processes

H. Pang et al. / Earth and Planetary Science Letters 414 (2015) 126–133

and stratosphere–troposphere water exchange. J. Geophys. Res. 110, D21314. http://dx.doi.org/10.1029/2005JD005790. Schoenemann, S.W., Schauer, A.J., Steig, E.J., 2013. Measurement of SLAP2 and GISP δ 17 O and proposed VSMOW-SLAP normalization for δ 17 O and 17 Oexcess . Rapid Commun. Mass Spectrom. 27, 582–590. http://dx.doi.org/10.1002/rcm.6486. Schoenemann, S.W., Steig, E.J., Ding, Q., Markle, B.R., Schauer, A.J., 2014. Triple water-isotopologue record from WAIS Divide, Antarctica: controls on glacial– interglacial changes in 17 O-excess of precipitation. J. Geophys. Res., Atmos. 119. http://dx.doi.org/10.1002/2014JD021770. Sodemann, H., Stohl, A., 2009. Asymmetries in the moisture origin of Antarctic precipitation. Geophys. Res. Lett. 36, L22803. http://dx.doi.org/10.1029/ 2009GL040242. Steen-Larsen, H.C., Masson-Delmotte, V., Sjolte, J., Johnsen, S.J., Winther, B.M., Bréon, F.-M., Clausen, H.B., Dahl-Jensen, D., Falourd, S., Gallée, H., Jouzel, J., Kageyama, M., Lerche, H., Minster, B., Picard, G., Punge, H.J., Risi, C., Salas, D., Schwander, J., Steffen, K., Sveinbjörnsdóttir, A.E., Svensson, A., White, J., 2011. Understanding the climatic signal in the water stable isotope records from the NEEM cores. J. Geophys. Res. 116, D06108. http://dx.doi.org/10.1029/2010JD014311. Steen-Larsen, H.C., Johnsen, S.J., Masson-Delmotte, V., Stenni, B., Risi, C., Sodemann, H., Balslev-Clausen, D., Blunier, T., Dahl-Jensen, D., Ellehoj, M.D., Falourd, S., Simonsen, S.B., Sjolte, J., Steffensen, J.P., Sperlich, P., Sveinbjörnsdóttir, A.E., Vinther, B.M., White, J.W.C., 2013. Continuous monitoring of summer surface water vapor isotopic composition above the Greenland Ice Sheet. Atmos. Chem. Phys. 13, 4815–4828. http://dx.doi.org/10.5194/acp-13-4815-2013. Steen-Larsen, H.C., Masson-Delmotte, V., Hirabayashi, M., Winkler, R., Satow, K., Prié, F., Bayou, N., Brun, E., Cuffey, K.M., Dahl-Jensen, D., Dumont, M., Guillevic, M., Kipfstuhl, S., Landais, A., Popp, T., Risi, C., Steffen, K., Stenni, B., Sveinbjörnsdottir, A.E., 2014a. What controls the isotopic composition of Greenland surface snow? Clim. Past 10, 377–392. http://dx.doi.org/10.5194/cp-10-377-2014. Steen-Larsen, H.C., Sveinbjörnsdottir, A.E., Peters, A.J., Masson-Delmotte, V., Guishard, M.P., Hsiao, G., Jouzel, J., Noone, D., Warren, J.K., White, J.W.C., 2014b. Climatic controls on water vapor deuterium excess in the marine boundary layer of the North Atlantic based on 500 days of in situ, continuous measurements. Atmos. Chem. Phys. 14, 7741–7756. Steig, E.J., Morse, D.L., Waddington, E.D., Stuiver, M., Grootes, P.M., Mayewski, P.A., Twickler, M.S., Whitlow, S.I., 2000. Wisconsinan and Holocene climate history from an ice core at Taylor Dome, western Ross Embayment, Antarctica. Geogr. Ann., Ser. A 82 (2–3), 213–235. Stenni, B., Masson-Delmotte, V., Johnsen, S., Jouzel, J., Longinelli, A., Monnin, E., Röthlisberger, R., Selmo, E., 2001. An oceanic cold reversal during the last deglaciation. Science 293, 2074–2077. Stenni, B., Jouzel, J., Masson-Delmotte, V., Röthlisberger, R., Castellano, E., Cattani, O., Falourd, S., Johnsen, S.J., Longinelli, A., Sachs, J.P., Selmo, E., Souchez, R., Steffensen, J.P., Udisti, R., 2003. A late-glacial high-resolution site and source temperature record derived from the EPICA Dome C isotope records (East Antarctica). Earth Planet. Sci. Lett. 217, 183–195. http://dx.doi.org/10.1016/ S0012-821X(03)00574-0. Stenni, B., Masson-Delmotte, V., Selmo, E., Oerter, H., Meyer, H., Röthlisberger, R., Jouzel, J., Cattani, O., Falourd, S., Fischer, H., 2010. The deuterium excess records

133

of EPICA Dome C and Dronning Maud Land ice cores (East Antarctica). Quat. Sci. Rev. 29, 146–159. http://dx.doi.org/10.1016/j.quascirev.2009.10.009. Uemura, R., Yoshida, N., Kurita, N., Nakawo, M., Watanabe, O., 2004. An observationbased method for reconstructing ocean surface changes using a 340,000-year deuterium excess record from the Dome Fuji ice core, Antarctica. Geophys. Res. Lett. 31, L13216. http://dx.doi.org/10.1029/2004GL019954. Uemura, R., Matsui, Y., Yoshimura, K., Motoyama, H., Yoshida, N., 2008. Evidence of deuterium excess in water vapor as an indicator of ocean surface conditions. J. Geophys. Res. 113, D19114. http://dx.doi.org/10.1029/2008JD010209. Uemura, R., Barkan, E., Abe, O., Luz, B., 2010. Triple isotope composition of oxygen in atmospheric water vapor. Geophys. Res. Lett. 37, L04402. http://dx.doi.org/ 10.1029/2009GL041960. Uemura, R., Masson-Delmotte, V., Jouzel, J., Landais, A., Motoyama, H., Stenni, B., 2012. Ranges of moisture-source temperature estimated from Antarctic ice cores stable isotope records over glacial–interglacial cycles. Clim. Past 8, 1109–1125. Vimeux, F., Masson, V., Jouzel, J., Stievenard, M., Petit, J.R., 1999. Glacial–interglacial changes in ocean surface conditions in the Southern Hemisphere. Nature 398, 410–413. Vimeux, F., Masson, V., Delaygue, G., Jouzel, J., Petit, J.R., Stievenard, M., 2001. A 420,000 year deuterium excess record from East Antarctica: information on past changes in the origin of precipitation at Vostok. J. Geophys. Res. 106, 31863–31873. Vimeux, F., Cuffey, K.M., Jouzel, J., 2002. New insights into Southern Hemisphere temperature changes from Vostok ice cores using deuterium excess correction. Earth Planet. Sci. Lett. 203, 829–843. http://dx.doi.org/10.1016/ S0012-821X(02)00950-0. Wang, Y., Hou, S., Masson-Delmotte, V., Jouzel, J., 2010. A generalized additive model for the spatial distribution of stable isotopic composition in Antarctic surface snow. Chem. Geol. 271, 133–141. http://dx.doi.org/10.1016/ j.chemgeo.2010.01.004. Wang, Y., Sodemann, H., Hou, S., Masson-Delmotte, V., Jouzel, J., Pang, H., 2013. Snow accumulation and its moisture origin over Dome Argus, Antarctica. Clim. Dyn. 40 (3–4), 731. Werner, M., Heimann, M., Hoffmann, G., 2001. Isotopic composition and origin of polar precipitation in present and glacial climate simulations. Tellus B 53, 53–71. Winkler, R., Landais, A., Sodemann, H., Dümbgen, L., Prie, F., Masson-Delmotte, V., Stenni, B., Jouzel, J., 2012. Deglaciation records of 17 O-excess in East Antarctica: reliable reconstruction of oceanic normalized relative humidity from coastal sites. Clim. Past 8, 1–16. Winkler, R., Landais, A., Risi, C., Baroni, M., Ekaykin, A., Jouzel, J., Petit, J.R., Prie, F., Minster, B., Falourd, S., 2013. Interannual variation of water isotopologues at Vostok indicates a contribution from stratospheric water vapor. Proc. Natl. Acad. Sci. USA. http://dx.doi.org/10.1073/pnas.1215209110. Xiao, C., Ding, M., Masson-Delmotte, V., Zhang, R., Jin, B., Ren, J., Li, C., Werner, M., Wang, Y., Cui, X., Wang, X., 2013. Stable isotopes in surface snow along a traverse route from Zhongshan station to Dome A, East Antarctica. Clim. Dyn. 41, 2427–2438.