Twentieth-century sea ice variability in the Weddell Sea and its effect on moisture transport: Evidence from a coastal East Antarctic ice core record

Abstract

High-resolution study of deuterium excess (d-excess), sea salt sodium (ss-Na⁺), and methane sulfonic acid (MSA) in an ice core from coastal Dronning Maud Land (cDML), East Antarctica, revealed the history of moisture transport and sea ice extent (SIE) during the last century. Backward wind trajectories show that air parcels were mainly derived from the Weddell Sea region. The d-excess profile shows a dramatic shift from an average value of 8‰ during 1905–1920 to −1‰ during ~1940 and thereafter positive excursion during 1940–1980. The dramatic shift during 1920–1940 has been attributed to the reduced moisture supply from low/mid-latitude to Antarctica associated with shifting of Southern Annular Mode (SAM) from positive to negative mode. The ss-Na⁺ flux profile shows systematic positive excursion during 1940–1980 which coincide with that of the d-excess profile. The MSA flux shows a negative excursion during 1950–1965, overlapping with the period of positive excursions in ss-Na⁺ and d-excess profiles. The concomitant increase of ss-Na⁺ and d-excess values and positive excursions during 1940–1980 indicate higher SIE. Based on significant correlation between the Na⁺ flux and satellite-derived winter SIE record of the Weddell Sea, ~10% increase in SIE is estimated compared to its average value of the last century. The power spectrum analysis of d-excess and ss-Na⁺ flux shows a significant periodicity at ~3.5 years, exactly matching with that of the winter SIE in the Weddell Sea. Wavelet analysis of SAM index and Southern Oscillation Index (SOI) shows the highest common power in 4 to 8 year band, overlapping with the periods of higher SIE and opposite phase in 10 to 16 year band, overlapping with the periods of higher d-excess. This study highlights the role of SAM and its teleconnection to El Niño Southern Oscillations (ENSO) in controlling sea ice and moisture source variability in annual to decadal scale in the coastal regions of Antarctica.

Keywords

Antarctica d-excess El Niño Southern Oscillations ice core sea ice extent sea salt Southern Annular Mode

Introduction

Climate change and its impact on atmospheric and hydrological conditions in Antarctica during the last century are poorly understood because of paucity of long-term observational data beyond the instrumental records. Antarctic ice cores provide high-resolution proxy records of regional changes in key climate variables such as precipitation, temperature, and sea ice cover, as well as changes in climate forcing (Abram et al., 2007; Curran et al., 2003; Masson-Delmotte et al., 2008; Mulvaney et al., 2012; Schneider et al., 2004). To address the issues related to 20th-century climate change in Antarctica and predict for the future climate, it is crucial to have knowledge of the two important climate parameters, that is, moisture source variability and sea ice conditions in the past. These parameters are very much sensitive to climate change particularly to atmospheric circulation and hydrological conditions (Cavalieri et al., 1997; Ciais et al., 1995; Masson-Delmotte et al., 2005; Wolff et al., 2003; Zwally et al., 2002). Chemical and isotope proxy records from the coastal ice cores are sensitive enough to reflect changes in moisture source variability and sea ice conditions (Noone and Simmonds, 2004). Within Antarctica, coastal regions show higher snow accumulation compared with inland and thus offer important sites for high-resolution climate records. However, complexities arise because of variable and nonlinear relation between the proxy records and climate parameters in coastal sites (Abram et al., 2013; and reference therein).

Large scale changes in atmospheric circulation are sensitive to hydrological conditions in and around Antarctica, which can be understood through the study of moisture source, transport, and its precipitation history. In order to reconstruct past hydrological conditions in Antarctic regions, deuterium excess (d-excess = δD − 8*δ¹⁸O; hereafter d-excess) records from ice cores (Dansgaard, 1964) together with isotope models have been used extensively (Ciais et al., 1995; Masson-Delmotte et al., 2005). The d-excess records provide information about the processes affected by kinetic fractionation and therefore, it has been used as a tracer for evaporation conditions such as surface air temperature and relative humidity (RH) over moisture source regions (Jouzel et al., 1982; Merlivat and Jouzel, 1979; Pfahl and Sodemann, 2014). Furthermore, it was also used as a proxy for tracing variations in moisture source, especially their origin and transport pathways of precipitation (Ichiyanagi et al., 2002; Vimeux et al., 2002).

Sea ice in Antarctica is another important component of polar climate system that plays a crucial role in global albedo, ocean–atmosphere heat exchange, and global carbon cycle through CO₂ exchange between ocean and atmosphere (Stephens and Keeling, 2000; Wolff et al., 2003). In addition, sea ice also plays an important role in deep water formation, and thereby global ocean circulation (Dieckmann and Hellmer, 2010). Therefore, changes in sea ice conditions in Antarctica affect the climate and hydrological systems on both regional and global scales. In several studies, it has been highlighted that sea ice is highly sensitive to regional climate and has undergone dramatic changes during the last few decades (Cavalieri et al., 1997; Parkinson et al., 1999; Zwally et al., 2002). Interannual changes in sea ice extent (SIE) have been quantified using satellite data, which are available since the late 1970s (Cavalieri et al., 2003). However, intrinsic variability and sensitivity of SIE to climate change are not well understood due to limited SIE records beyond the satellite era. In order to resolve these issues, we need to have long-term records of sea ice conditions. Several chemical and isotope proxies have been used to reconstruct past sea ice conditions in Antarctica (Curran et al., 2003; Röthlisberger et al., 2010; Wolff et al., 2003). However, large uncertainty associated with the quantitative estimates of SIE in Antarctica raises several questions related to their reliability (Abram et al., 2013; Levine et al., 2014). Traditionally, sea salt sodium (ss-Na⁺) has been used as a proxy for SIE in Antarctica (Iizuka et al., 2008; Wolff et al., 2003); however, it could be significantly influenced by meteorological factors other than sea ice such as wind strength at the source regions, wind directions, and residence time of sea salts. Alternative to ss-Na⁺, records of methane sulfonic acid (MSA) were proposed as a robust proxy particularly in high precipitation regions (Abram et al., 2008; Curran et al., 2003). However, comparisons between MSA records and satellite data on SIE have shown variable relations in different oceanic sectors surrounding Antarctica (Abram et al., 2008; Curran et al., 2003; Pasteur et al., 1995; Sun et al., 2002). In view of these complexities, multiple proxy approach would be appropriate to reconstruct SIE more accurately and to resolve the issues related to their application as a sea ice proxy.

In this study, high-resolution d-excess record together with ss-Na⁺ and MSA records in an ice core from cDML have been examined to evaluate their potential as a proxy for moisture source and SIE. Finally, these proxy records have been employed to extract useful information about moisture source variability, transport and evaporation conditions at source regions (e.g. temperature and RH), and reconstruction of SIE in oceanic sectors around Antarctica, particularly in the Weddell Sea. Furthermore, we have also made an attempt to understand the possible influence of Southern Annular Mode (SAM) and El Niño Southern Oscillations (ENSO) on moisture source variability and sea ice conditions in the coastal sectors of Antarctica during the last century, as these climate modes have been recognized as primary drivers of climate variability in Southern Hemisphere at inter annual to decadal scale (Bertler et al., 2006; Ciasto and Thompson, 2008; Gregory and Noone, 2008; Turner, 2004).

Materials and methods

A 65-m-long ice core (IND-25/B5), recovered from the cDML area in Antarctica within the South Atlantic sector, is used for this study (Figure 1). The distance between the calving line and the core site is ~170 km with an elevation of ~1300 m above the mean sea level. The coastal areas with relatively high accumulation rate (discussed in section ‘Temporal variability of snow accumulation’) can be dated more reliably and therefore provide an excellent opportunity to work with annually dated records comparable to instrumental records (Helsen et al., 2006, Naik et al., 2010). The sampling details have already been discussed elsewhere (Naik et al., 2010). Chronology for this core was established by Naik et al. (2010) using multiple methods such as annual layer determination based on summer maxima in δ¹⁸O values, nssSO₄²⁻ (nonsea-salt sulfate) markers of volcanic eruptions, and Tritium markers. Based on the chronological controls, the core was dated back to the year 1905 with an uncertainty of ±2 years. This core represents the depositional history of the past century (1905–2005).

Figure 1.

Location map of ice core IND-25/B5 in coastal Dronning Maud Land, East Antarctica, using Quantarctica software.

Major ions (Na⁺ and Ca²⁺ and MSA) were measured in melted ice samples using an ion exchange chromatograph (IC; Dionex™ Model ICS-2500 equipped with EG50 Eluent generator and CD25 detector). Details of the analytical methods are discussed elsewhere (Laluraj et al., 2011; Thamban et al., 2010). The detection limits of the RF-IC system for Na⁺ and Ca²⁺ measurements were up to 0.25 µg l⁻¹ and for MSA up to 0.5 µg l⁻¹. Quality of the analysis was checked by measuring several replicates of chromatographic standards and samples. The analytical precision based on several replicate measurements was better than 5% for MSA and Na⁺, and better than 10% for Ca⁺².

Oxygen and hydrogen isotope ratios were measured using a dual inlet Isoprime Mass Spectrometer (GV instruments, UK), following the methods discussed in Naik et al. (2010). The external precision (1σ) obtained using in-house laboratory standard (CDML1) on oxygen and hydrogen was ±0.05‰ and ±0.77‰, respectively. Replicate analysis performed on samples yielded reproducibility of ±0.05‰ and ±0.8‰ (1σ) on oxygen and hydrogen isotope ratios, respectively. The estimated precision of d-excess values based on the precision of oxygen and hydrogen isotope measurements was ±0.87‰.

Backward trajectory of 10-day air parcels was analyzed to identify the moisture source using HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) Model access via NOAA Air Resources Laboratory (ARL) READY (n.d.) (http://ready.arl.noaa.gov/HYSPLIT.php). This model was fed with the Reanalysis archives from NOAA/NCEP to perform atmospheric data assimilation from 1948 to present. Single trajectories were later grouped into major clusters for the core site following statistical treatments. The details of the backward wind trajectory and cluster analysis are discussed below (section ‘Pathways of moisture and sea salt transport to the study region’).

Results

Temporal variability of snow accumulation

Annual snow accumulation rates estimated based on δ¹⁸O seasonality show high temporal variability during the entire record (1905–2005). It ranges from 110 to 528 kg m⁻² a⁻¹ with an average of 280 ± 93 (1 σ) kg m⁻² a⁻¹ (Figure 2a), which is similar to that of the previously reported value of 200–250 kg m⁻² a⁻¹ from the Dronning Maud Land (DML) region based on ice core records (Giovinetto and Zwally, 2000; Vaughan et al., 1999). Furthermore, the snow accumulation rates in this site are also consistent with the report based on surface mass balance studies of this region (Altnau et al., 2015; Arthern et al., 2006; Lenaerts et al., 2014; Van de Berg et al., 2006). Large temporal variations in accumulation rates at this core site could be attributed to the significant variations in precipitation owing to cyclonic activities along the coastal regions of DML (Naik et al., 2010; Schlösser and Oerter, 2002).

Figure 2.

(a) Annual snow accumulation rate plotted with age. (b) The δ¹⁸O profile. (c) The d-excess profile shows a dramatic reduction during 1920–1940 and systematic increase during 1940–1960 and 1995–2005. (d) Na⁺ concentration profile shows systematically higher values with positive excursion during 1940–1980 and 1995–2005. (e) MSA concentration shows negative excursion during 1955–1965 which overlaps with the period of positive excursion in Na⁺ concentration profile.

Stable isotopes and ionic concentrations

The oxygen and hydrogen isotope records from this core have already been reported elsewhere (Naik et al., 2010). The relationship between δ¹⁸O and δD follows the meteoric waterline (Craig, 1961) defined by δD = (8.18 ± 0.01) * δ¹⁸O + (9.81 ± 1.59) (r = 0.97, p < 0.01). The d-excess record ranges from −7‰ to 27‰ with an average value of 4.4‰ ± 3.8‰ (Figure 2). The d-excess record shows a dramatic shift from an average value of ~8‰ during 1905–1920 and reduced up to about −1‰ during ~1940 (Figure 2c). Similar magnitude of d-excess shift has been reported from the western Ross Sea sector (Sinclair et al., 2014). Furthermore, the d-excess record reveals a systematic increase from an average value of about −1‰ at around 1940 to systematic rise up to ~5‰ around 1960; upward curvature during 1940–1980. Subsequently, the d-excess profile shows large variability and a sharp rise during 1995–2005 (Figure 2c). However, the magnitudes of the systematic excursions during 1940–1980 and 1995–2005 are smaller compared with that of the shift during 1920–1940.

Sodium (Na⁺) concentration, a major sea salt species, shows large variability; ranges from 0.3 to 962 µg l⁻¹ with a mean value of 81 ± 95 µg l⁻¹ (Figure 2d). The Na⁺ concentration profile shows systematic positive excursion during 1940–1980 and 1995–2005, which overlaps with that of the d-excess profile (Figure 2c and d). The MSA concentration exhibits large variability, ranging from 0.5 to 44.9 µg l⁻¹ with a mean value of 8.3 ± 6.9 µg l⁻¹ (Figure 2e), which is close to that of the reported values of ~4.9 µg l⁻¹ from the DML region (Abram et al., 2008) and ~6.5 µg l⁻¹ from Law Dome (Curran et al., 2003). The MSA profile shows a systematic decrease (negative excursion) during 1955–1965, overlapping with the period of systematic increase (positive excursion) in Na⁺ concentration profile (Figure 2d and e). Since dust could act as an additional source of Na⁺ ions in Antarctic ice/snow (Benassai et al., 2005; Röthlisberger et al., 2002), possible nonsea-salt Na⁺ was subtracted from the total Na⁺. Sea-salt Na⁺ = Na⁺– (nssCa²⁺/Rt), where Ca⁺ and Na⁺ are the measured concentrations of these ions in snow. Rt is the mean ratios of Ca²⁺/Na⁺ in the marine aerosols (~0.038) and in average crust (~1.78), respectively (Bowen, 1979).

In order to account for the dilution effect, Na⁺ and MSA concentrations were converted into annual fluxes by multiplying annual average concentrations with annual accumulation rates. The estimated annual flux of ss-Na⁺ exhibits large variability; ranges between 2.7 and 64.7 mg m^–2 a^–1 with an average value of 18 ± 13 mg m^–2 a^–1 (Figure 3d). Similarly, the estimated MSA flux ranges between 0.3 and 7.4 mg m^–2 a^–1 with an average value of 2.3 ± 1.5 mg m^–2 a^–1 (Figure 3d).

Figure 3.

(a) Annual accumulation rate and annually averaged values of (b) δ¹⁸O, (c) d-excess, and (d) ss-Na⁺ and MSA fluxes plotted with age. Decadal trend derived based on 10-point running average of these parameters are also shown (red line). These data are equally spaced in time. These profiles are compared with 10-point running average of (e) Southern Oscillation Index (SOI) and (f) Southern Annular Mode (SAM) profiles. The principal SAM index (red line) was taken from Marshall (2003) and the modeled SAM index (blue line) taken from Jones et al. (2009). Both profiles show good agreement.

Discussion

In the present study, we discuss new results of ss-Na⁺ and MSA together with the published data of stable isotopes (Naik et al., 2010) to reconstruct the history of the moisture transport in the coastal Antarctica and sea ice variability in the Weddell Sea region.

Influence of snow accumulation and meteorological parameters on proxy records

Stable isotopes and ionic signals in polar firn/ice core are often obscured by post-depositional changes particularly in low accumulation sites because of the snow exposure, ventilation, and vapor condensation–sublimation processes (Hoshina et al., 2014; Neumann and Waddington, 2004). Therefore, it is important to assess the role of accumulation affecting the ionic and isotopic signals before employing them as a proxy. Annual average of snow accumulation rates were compared with the records of stable isotope (δ¹⁸O and d-excess) and sea salts (ss-Na⁺ and MSA fluxes) (Figure 3). Interestingly, the profiles of annually averaged ss-Na⁺ and MSA fluxes are similar to their concentration profiles (Figures 2 and 3), suggesting limited influence of accumulation on these glaciochemical records. To smooth the short-term fluctuations and derive the long-term decadal trend from the proxy records, 10-point running mean were plotted (Figure 3). The ss-Na⁺ flux profile exhibits a major positive excursion during 1940–1980, which overlaps with the period of negative excursion in MSA flux profile (Figure 3d). Decadal trends based on 10-point moving average of the parameters show that accumulation rates vary synchronously with the d-excess, ss-Na⁺, and MSA fluxes during 1905–1930 (Figure 3a, c and d). However, these parameters do not show any significant correlations when compared with accumulation rates for the entire records of 1905–2005 (Table 1). Furthermore, the accumulation trend is almost steady during 1940–1980 when ss-Na⁺ and MSA flux profiles show maximum excursion. These evidences indicate that the effect of snow accumulation on isotopic and chemical records of this core is negligible. In general, the effect of post-depositional modification is expected in low snow accumulation sites. Therefore, higher accumulation observed throughout this core (annual average: 280 ± 93 kg m⁻² a⁻¹) compared with inland ice cores from Antarctica also supports minimum effect of post-depositional processes on the chemical and isotope records. In view of high accumulation at the core site, the glaciochemical records are expected to be primarily controlled by the wet deposition.

Table 1.

Pearson correlation coefficients (n = 45) between the annual average of meteorological parameters from the Novo station available since 1961and measured parameters from IND-25 core.

Variables	P_atm on Met site	P_atm on SL	Surface wind	Surface air temp	RH	Accumulation	ss-Na	MSA	δ¹⁸O	d-excess
P_atm on Met site		0.83	−0.31	−0.26	0.10	−0.28	0.32	0.05	−0.26	−0.14
P_atm on SL	0.83		−0.23	0.05	0.15	−0.29	0.14	0.23	−0.20	−0.23
Surface wind	−0.31	−0.23		0.39	0.08	0.29	−0.24	−0.11	0.05	−0.03
Surface air temp	−0.26	0.05	0.39		0.10	0.21	−0.40	0.10	0.13	−0.35
RH	0.10	0.15	0.08	0.10		0.23	−0.25	0.12	−0.02	−0.49
Accumulation	−0.28	−0.29	0.29	0.21	0.23		−0.10	0.05	−0.06	−0.06
ss-Na	0.32	0.14	−0.24	−0.40	−0.25	−0.10		0.22	−0.05	0.37
MSA	0.05	0.23	−0.11	0.10	0.12	0.05	0.22		0.25	0.00
δ¹⁸O	−0.26	−0.20	0.05	0.13	−0.02	−0.06	−0.05	0.25		0.17
d-excess	−0.14	−0.23	−0.03	−0.35	−0.49	−0.06	0.37	0.00	0.17

P_atm: Atmoshpheric pressure; RH: Relative Humidity; Met: Metrological; MSA: Methane Sulfonic Acid.

Coefficients highlighted in bold indicate significant correlations (df = 43, p < 0.01).

To interpret variability in chemical (ss-Na⁺ and MSA) and isotopic (d-excess) proxy records, it is important to have knowledge on the meteorological parameters from the region as well as the information about the sources of moisture and aerosols being transported to the study area. Comparison of these proxy records with the meteorological parameters can assess the sensitivity and uncertainty associated with them. Therefore, these proxy records were compared with meteorological parameters (atmospheric pressure, surface wind speed, temperature, and RH) available since 1960 from the station ‘Novolazarevskaya’ (Novo). The Novo meteorological records have been found to be broadly representing the environmental records of the IND-25/B5 core site (Naik et al., 2010). The results of the correlations are presented in Table 1. The d-excess record, proxy for moisture source variability, and conditions over source regions show significant negative correlation (r = –0.49, df = 42, p < 0.01) with RH. The ss-Na⁺ flux, proxy for SIE, exhibits significant negative correlation with surface air temperature (r = –0.40, df = 43, p < 0.01). However, it is noteworthy to observe that the MSA flux, another proxy for SIE, does not show significant correlation with any of the meteorological parameters. Significant correlation between ss-Na⁺ and air temperature could be due to the colder/extended winter which might have favored more sea ice formation in the Weddell Sea region (detailed discussion in section ‘Changes in moisture source, transport and climatic conditions during the last century’). It is also observed that ss-Na⁺ and d-excess show significant positive correlation (r = 0.37, df = 43, p < 0.01). This indicates that common oceanographic and meteorological parameters were responsible for d-excess and ss-Na⁺ variability at this core site.

Pathways of moisture and sea salt transport to the study region

Analysis of monthly mean precipitation data since 1961 from the Novo station reveals that the highest mean precipitation occurs in austral winter and lowest in austral summer. In order to investigate the dominant sources of moisture and aerosols that arrive at the core site, backward wind trajectory analysis was carried out for the months of July–August (peak winter) and January–February (peak summer) of representative time intervals in the past when major excursions were observed in the proxy records around 1960 and 1980. The air parcel circulation model HYSPLIT (http://ready.arl.noaa.gov/HYSPLIT.php) was used to reconstruct a 10-day backward trajectory of every precipitation event registered at core location arriving at 1500 m a.s.l. (~800 mb). Considering the availability of 20th-century Reanalysis data since 1948 and the major excursions observed during 1960 and 1980, back trajectory analyses were carried out for the winter and summer months of these particular years. These two years represent maxima and minima of the common excursions observed in d-excess and ss-Na⁺ profiles. A total of 62 trajectories for 2 months of each season (summer and winter) were clustered (Figure 4). Backward wind trajectories show a cyclonic curvature reflecting the substantial influence of cyclones on the air parcel paths over Antarctica. After running the cluster analysis, optimal 3–4 mean clusters were recognized (Figure 4). The most common origin of the air parcels were found over the Weddell Sea region (40–47%), followed by the Bellingshausen–Amundsen (15–34%). The Ross Sea contribution was significant only during the 1960 winter (42%). The mean clusters during summer months do not show significant differences whereas discernable differences were observed between two winters. This could be due to strong wind events and cyclonic activity during winter (Turner et al., 2009). Based on the distribution of clustered components, Weddell Sea seems to be the dominant source of moisture and aerosols to the core site during these periods. Similar observations have also been reported from western DML ice core sites (Noone et al., 1999; Reijmer et al., 1999; Russell et al., 2004).

Figure 4.

The cluster analysis of backward wind trajectory for (a, c) the peak winter months (July and August) and (b, d) peak summer months (Jan and Feb) of 1960 and 1980. A total of 62 trajectories were clustered separately for winter and summer of these two years which shows optimal three clusters (component) except for the 1980 winter. The green colored cluster represents highest fraction of air parcel reaching to core site.

Changes in moisture source, transport, and climatic conditions during the last century

The d-excess in Antarctic snow/ice is primarily controlled by various meteorological parameters over moisture source regions such as RH, sea surface temperature (SST), wind speed, as well as geographical factors, that is, elevation of the sites and distance from the coast (Masson-Delmotte et al., 2008; Stenni et al., 2001). However, d-excess records of ice core could be affected by the post-depositional processes like sublimation and metamorphism in low accumulation sites (Neumann and Waddington, 2004; Satake and Kawada, 1997). As discussed above (section ‘Influence of snow accumulation and meteorological parameters on proxy records’), impact of snow accumulation and post-depositional processes on d-excess record at the core site is insignificant. Furthermore, slope and intercept of the linear relation between δD and δ¹⁸O values in this core are similar to that of the global meteoric waterline (GMWL) which indicates that the moisture mass precipitating at the core site went through similar degree of kinetic fractionation. Therefore, d-excess records of IND-25/B5 could be used to extract information about the moisture source, transport, and climatic conditions which prevailed over the source regions in the past.

In general, higher d-excess values in polar ice cores reflect a more distant moisture origin, while the lower d-excess indicate evaporation at higher latitudes (Vimeux et al., 1999). Measurements and modeling studies have shown that d-excess in Antarctic ice correlates negatively with RH and positively with SST (Merlivat and Jouzel, 1979; Uemura et al., 2008). However, influence of advection height on the final value of d-excess in precipitation can obscure the classical interpretation of parameters in terms of temperature and RH in the moisture source region (Helsen et al., 2006). As discussed earlier, the significant negative correlation observed between RH and d-excess records indicates that the d-excess variability at the core site is apparently associated with the variability in moisture source conditions and its transport to the core site (Masson-Delmotte et al., 2008; Vimeux et al., 1999). Several possibilities have been explored here to explain the dramatic decrease in d-excess profile from 8‰ to −1‰ during 1920–1940 and a relatively less pronounced, but systematic increase during 1940–1965 and 1995–2005 (Figures 2c and 3c). The dramatic decrease in d-excess profile might be associated with the negative shift of SAM index (Figure 3c and f), probably linked to large scale change and/or reorganization of moisture source and its transport pathways. In general, the higher values of d-excess are associated with higher meridional transport carrying warm moist air when the low-pressure center is positioned over Antarctica (Ciais et al., 1995; Masson-Delmotte et al., 2010). Therefore, such a dramatic reduction in d-excess values during 1920–1940 suggests a major shift in moisture source contributing to this region. This invokes a large scale change in atmospheric circulation in the Southern Hemisphere coupled with local factors such as higher RH and lower temperature in the Weddell Sea. The lowest values in the δ¹⁸O profile around this interval indicating the lowest temperature (Figure 3b) also corroborate such a scenario of shifting moisture source regions.

The magnitude of the excursions in d-excess profile during 1940–1980 and 1995–2005 are much lower compared with the large shift during 1920–1940. Therefore, these two excursions with lower magnitude are considered as a second-order change associated with the local factors like sea ice conditions, probably the SIE. Noone and Simmonds (2004) suggested that variations in SIE could have a considerable effect on atmospheric transport and ultimately on the d-excess records. An increase in distance between moisture source and core site because of increase in SIE could be responsible for the systematic increase in d-excess. Interestingly, the timings of the d-excess increase coincide with that of higher ss-Na⁺ supporting higher sea ice production in the Weddell Sea (discussed below). Furthermore, significant positive correlation between ss-Na⁺ and d-excess (Table 1) supports a change in distance between the source regions and core site due to SIE variations. This suggests that large scale changes in moisture source and transport are associated with SAM whereas relatively smaller scale changes are associated with SIE.

Sea ice conditions in the Weddell Sea sector during the last century

The ss-Na⁺ concentration in the coastal Antarctic atmosphere is directly linked to brine and frost flower formation and therefore, has been used as a proxy for SIE in Antarctica (Iizuka et al., 2008; Wolff et al., 2003). Seasonal cycles in MSA over Antarctica are directly linked to phytoplankton blooms that occur during spring, following the breakup of sea ice (Legrand et al., 1992). Therefore, MSA is also considered as an alternate proxy for sea ice, particularly in high precipitation regions (Curran et al., 2003; Welch et al., 1993). However, these proxies could be significantly affected by factors other than the sea ice, influencing their production, transport, and preservation in Antarctic ice sheet. One of the major influencing factors is wind (strength and direction). Furthermore, these proxy records show variable relation with the SIE in different sectors of Antarctica (Abram et al., 2013); a significant positive co-relation with ss-Na⁺ (Iizuka et al., 2008) and inverse relation with MSA (Abram et al., 2008) have been reported for the ice cores which come under the influence of the Weddell Sea sector. However, the correlation between these two proxies and SIE could explain maximum up to ~70% of the total SIE variability in the Weddell Sea sector (Abram et al., 2008, 2010; Iizuka et al., 2008). Remaining unexplained variability in SIE has been attributed to the sensitivity of the proxies and noise associated with the meteorological parameters (Abram et al., 2013).

To identify complexities associated with the proxies like sensitivity to SIE change and the factors contributing to the uncertainty associated with quantitative estimates, the ss-Na⁺ and MSA flux records of IND-25/B5 were compared with the meteorological parameters from the Novo station and satellite based SIE data during 1979–2005 from http://nsidc.org/data/seaice_index/ (Figure 5). The ss-Na⁺ and MSA flux records do not show any significant correlations with the surface wind speed (Table 1). This suggests that the wind speed might not have played any significant role in controlling ss-Na⁺ and MSA variability in this core site. The significant negative correlation of ss-Na⁺ flux with the surface air temperature (discussed in section ‘Influence of snow accumulation and meteorological parameters on proxy records’) indicates a role of temperature in influencing sea ice production in Antarctica. Hence, ss-Na⁺ flux variations could be attributed to sea ice production and thereby SIE in this part of Antarctica. Furthermore, significant positive correlation (r = 0.37, p < 0.01) between ss-Na⁺ flux and d-excess invokes a common causative factor, that is, sea ice production and/or extent. Back trajectory reconstruction at the core site indicates that the moisture and aerosol were derived mainly from the Weddell Sea. Therefore, ss-Na⁺ and MSA flux records from the core are expected to reflect sea ice conditions in the Weddell Sea. Maximum sea salt and aerosol concentrations in Antarctic snow/ice are reported during winter because of the formation of extended sea ice surfaces during winter which act as an effective source for the sea salt aerosols over coastal and inland Antarctica (Jourdain et al., 2008; Rankin et al., 2000).

Figure 5.

(a, b) Annual average records of ss-Na⁺ and MSA fluxes are compared with annual and winter (June–July–August) SIE records of the Weddell Sea. (c) Moving correlation coefficients between the ss-Na flux and winter SIE (at 5-year window) are plotted with years. The dashed lines represent 95% confidence levels at 0.878 and −0.878. (d) Cross plot between ss-Na⁺ flux and winter SIE of the Weddell Sea. Averaged winter SIE in the Weddell Sea shows significant correlation (r = 0.355, p < 0.1, n = 24) with ss-Na⁺ flux. The period of 1979–1982 (four data points) was not considered in this co-relation.

Annual average fluxes of ss-Na⁺ and MSA were compared with the satellite records of annual and winter (June–July–August) average SIE in the Weddell Sea sector for the period of 1978–2005. The annual average of ss-Na⁺ and MSA fluxes do not show any significant correlation with the annual average of SIE. However, average winter SIE shows significant positive correlation (r = 0.355, n = 24, p < 0.1) with ss-Na⁺ flux excluding 3 years from 1979–1982 (Figure 5c and d). Since the ss-Na⁺ and winter SIE data show significant periodicity at ~3.5 and ~4.5 years (discussed in section ‘SAM–ENSO teleconnection in modulating moisture source and sea ice variability’), correlation coefficients between them were plotted in a 5-year window. This shows overall positive correlations except for the period 1979–1982 (Figure 5c). This indicates that total ss-Na⁺ flux variability in this core site is mainly influenced by the winter sea ice production. Although MSA flux does not show any significant relation with SIE, it is noteworthy to observe that negative excursion of MSA flux during 1940–1965 overlaps with the period of positive excursion in ss-Na⁺ profile. Abram et al. (2007) have reported a significant inverse correlation between MSA and SIE in the Weddell Sea sector which is consistent with this above observation. This further supports the explanation of SIE increase during 1940–1980. The concomitant increase of d-excess during 1940–1980 indicates a shifting of the moisture source away from the previous position because of increase of SIE. The above lines of evidence based on multiple proxies indicate higher SIE during 1940–1980 in the Weddell Sea sector which is an unprecedented increase within the last century. Considering a proportional response of ss-Na⁺ flux signal to the change in winter SIE, it is proposed that about 10% increase in SIE with respect to its average value is expected during 1940–1980 based on the linear regression equation (Figure 5d). However, uncertainty associated with this estimate needs to be taken into consideration as other factors, that is, wind speed at the source and residence time of sea salt aerosol, could affect sea salt production and transport to coastal ice sheet (Abram et al., 2013).

Sea ice proxies (ss-Na⁺ and MSA records) in this study also reveal variable relations within the last century; co-variations during 1905–1940 and systematic opposite trend during 1940–1965 (Figure 3d). To find the time scale at which these proxy records responded to SIE variability, spectral analyses of d-excess, ss-Na⁺, and MSA were performed (Figure 6). The d-excess and ss-Na⁺ records show significant (90% χ² level) periodicities at ~3.5 years (Figure 6a and b). The significant common periodicity indicates a common factor, possibly SIE, might have influenced d-excess and ss-Na⁺ at ~3.5-year period. MSA data show significant (90% χ² level) periodicities at ~7 and 2.5 years (Figure 6c), which are consistent with the observation from northern Weddell Sea (Abram et al., 2007). However, we refrain from drawing any major inferences based on MSA record because it does not show any significant co-relation with the satellite-derived SIE data.

Figure 6.

(a, b, c) Spectral analysis of annually averaged d-excess, ss-Na⁺, and MSA fluxes using REDFIT software (Schulz and Mudelsee, 2002) shows significant periodicities; d-excess at ~3.5 years, ss-Na at ~3.5 and ~4.5 years, and MSA at ~2.5 and ~7 years. Analyses of the winter (d) SAM index data (from Marshal, 2003) and (e) SIE in the Weddell Sea also reveal significant periodicity at ~3.5 years which is exactly similar to that of d-excess, ss-Na⁺.

SAM–ENSO teleconnection in modulating moisture source and sea ice variability

SAM, ENSO, and their linkages have been recognized as primary drivers for the large scale change in atmospheric circulations and sea ice variations in the Southern Hemisphere (Marshall, 2003; Marshall et al., 2010; Thompson and Solomon, 2002; Thompson et al., 2000). Sea ice variations in Southern Hemisphere on annual to decadal time scales are often linked to the variations of SAM (Fogt and Bromwich, 2006; Hall and Visbeck, 2002; Lefebvre et al., 2004). Furthermore, SAM–ENSO teleconnection is considered as an integrated mechanism influencing sea ice conditions in Antarctica. In general, the high-latitude cryosphere–atmosphere response to ENSO is strongest when negative SAM coincides with El Niño and positive SAM with La Niña (Stammerjohn et al., 2008).

In order to investigate the link between climate parameters (moisture and SIE variability) and climate modes (SAM and ENSO), annually averaged profiles of SAM index and Southern Oscillation Index (SOI) were compared with the chemical (ss-Na⁺ and MSA) and isotope (δ¹⁸O and d-excess) proxy records (Figure 3e and f). The SAM index reconstructions used for this comparison were taken from Marshall (2003) and Jones et al. (2009). Reconstruction of the SAM indices before 1979 based on various methods is known to have large uncertainty because of sparse data sets in the high-latitude Southern Hemisphere (Bromwich and Fogt, 2004; Marshall, 2003). Therefore, in any comparison with SAM indices prior to 1979, the associated uncertainty needs to be considered.

Comparison of SAM indices with our chemical and isotope proxy records revealed systematic changes in d-excess, ss-Na⁺, and MSA profiles during 1940–1965, indicating higher SIE coincides with the shifting of SAM index from positive to negative mode (Figure 3f). The dramatic shift in d-excess from 8‰ to −1‰ during 1920–1940 overlaps with the period when SAM shifted to negative mode (Figure 3), which apparently resulted in reduction of moisture transport from low and mid-latitude to Antarctica. To determine the time scale at which SAM influenced the moisture transport and SIE variability in this region, spectral analysis of winter SIE records of the Weddell Sea (1979–2005) and winter SAM index (1957–2005) was performed. This shows significant periodicity (95% χ² level) at ~3.5 years (Figure 6d and e) which exactly matches with that of d-excess and ss-Na⁺ (Figure 6a and b). This supports the role of SAM influencing winter sea ice variability in the Weddell Sea on an annual to sub-decadal scale.

To investigate SAM–ENSO teleconnection and their phase relation influencing SIE and moisture source variability, wavelet analyses of SAM and SOI time series data were performed for the entire period of 1905–2005 (Figure 7), following the methods of Torrence and Compo (1998) and Grinsted et al. (2004). SOI data are taken from the Australian Bureau of Meteorology (http://www.bom.gov.au/climate/current/soihtm1.shtml). Based on the cross wavelet analysis, significant highest common power is observed in 4- to 8-year band during 1945–1965 and 1990–2000 (Figure 7a). These periods overlap with the periods of higher SIE, as reflected in the ss-Na flux and d-excess profiles (Figure 7c). Furthermore, the wavelet coherency plot shows SOI overriding SAM in 4- to 8-year band during the same periods. It is interesting to mention here that the fast-ice variability with periodicities of 3–7 years in the Weddell Sea was found to be associated with nonstationary relationships between ENSO and SAM (Murphy et al., 2014). The periods of highest common power in SAM–ENSO at 4- to 8-year band and the time when SOI overrides SAM (phase lag ~90°) coincide with the higher periods of SIE during 1940–1965 and 1990–2000. This suggests a possible role of SAM and ENSO and their teleconnection influencing sea ice condition in the Weddell Sea sector during the last century.

Figure 7.

Cross wavelet transform and wavelet coherence of SAM and SOI time series. The 5% significance level against red noise is shown as a thick contour. The thin solid line indicates cone of influence. The relative phase relationship is shown as arrows (with in-phase pointing right, anti-phase pointing left, and SAM leading SOI by 90° pointing straight down and vice versa). (a) Significant highest common power in 4- to 8-year band is observed during 1945–1965 and 1990–2000. (b) The wavelet coherency plot shows SOI overriding SAM in 4- to 8-year band during the same period. SAM–ENSO shows anti-phase in 10- to 16-year band during 1905–1930 and 1990–2005. (c) The timings of the major excursions in ss-Na⁺ and MSA profiles based on 10-point running average are compared with that of dominant periods and coherency of SAM–ENSO index. Higher values of d-excess during 1905–1930 and 1990–2005 coincide with anti-phase of SAM–ENSO in 10- to 16-year band. The maximum positive excursion in ss-Na flux profile during 1945–1965 coincides with significant highest common power in 4- to 8-year band.

SAM–ENSO shows anti-phase with the highest common power at 10- to 16-year band during 1905–1930 and 1990–2005 (Figure 7a and b), coinciding with the periods of higher d-excess (Figure 7c). The higher values of d-excess could be resulting from intensified long-range meridional moisture transport from the mid-latitude. On the contrary, SAM–ENSO teleconnection coupled with shifting of SAM from positive to negative mode during 1940–1980 might have weakened the meridional transport which resulted in reduced moisture supply from mid to higher latitudes.

Conclusion

Combined multi-proxy (d-excess, ss-Na⁺, and MSA) records from a cDML ice core are employed in this study to reconstruct the history of SIE in the Weddell Sea and moisture source variability in the coastal regions of Antarctica. Comparison of the proxy records with the instrumental records of climate parameters highlights sensitivity of the individual proxy and uncertainty associated with their quantitative estimates. The dramatic shift of d-excess from 8‰ to −1‰ during 1920–1940 is attributed to the reduced moisture supply from low to mid-latitudes associated with shifting of SAM from positive to negative mode. Furthermore, concomitant increase of ss-Na⁺ and d-excess profiles and their positive excursions during 1940–1980 indicate higher SIE in the Weddell Sea. The ss-Na⁺ flux shows significant correlation with satellite record of winter SIE in the Weddell Sea. Based on linear response of the proxies to SIE variability, the excursion corresponds to ~10% increase in SIE during 1940–1980. The SAM and SOI time series show that the highest common power in 4- to 8-year band during 1945–1965 and 1990–2000 overlaps with the period of higher SIE. This suggests that SAM–ENSO teleconnection possibly influenced the sea ice conditions at 4- to 8-year band in the Weddell Sea sector. Furthermore, SAM–SOI shows opposite phase in 10- to 16-year band during 1905–1930 and 1990–2005, coinciding with the periods of higher d-excess supporting a strong SAM–ENSO teleconnection that allowed more meridional transport of warm and moist air from remote sources in the lower latitudes. Our study suggests that moisture source and sea ice variability in annual–decadal scale in Antarctica seems to be largely influenced by SAM and its teleconnection to ENSO.

Footnotes

Acknowledgements

We thank Shri Arun Chaturvedi (Geological Survey of India) and Dr A Rajakumar (IIT Kharagpur) for their field efforts in ice core recovery and BL Redkar (NCAOR) for laboratory support. We also acknowledge NOAA Air Resources Laboratory (ARL) for the HYSPLIT, C Torrence, GP Compo, A Grinsted, JC Moore, and S Jevrejeva, for Wavelet software and Norwegian Polar Institute for Quantarctica software package. We thank Atle Nesje, Associate Editor and two anonymous referees for their useful comments and suggestions. This is NCAOR contribution No. 28/ 2015.

Funding

We sincerely acknowledge Ministry of Earth Sciences and Director, NCAOR, for funding and support.

References

Abram

Wolff

Curran

(2013) A review of sea ice proxy information from polar ice cores. Quaternary Science Reviews 79: 168–183.

Abram

Curran

MAJ

Mulvaney

. (2008) The preservation of methanesulphonic acid in frozen ice-core samples. Journal of Glaciology 54: 680–684.

Abram

Mulvaney

Wolff

. (2007) Ice core records as sea ice proxies: An evaluation from the Weddell Sea region of Antarctica. Journal of Geophysical Research: Atmospheres 112: D15101.

Abram

Thomas

McConnell

. (2010) Ice core evidence for a 20th century decline of sea ice in the Bellingshausen Sea, Antarctica. Journal of Geophysical Research: Atmospheres 115: D23101.

Altnau

Schlosser

Isaksson

. (2015) Climatic signals from 76 shallow firn cores in Dronning Maud Land, East Antarctica. The Cryosphere 9: 925–944.

Arthern

Winebrenner

Vaughan

(2006) Antarctic snow accumulation mapped using polarization of 4.3-cm wavelength microwave emission. Journal of Geophysical Research: Atmospheres 111(D6): D06107.

Benassai

Becagli

Gragnani

. (2005) Sea-spray deposition in Antarctic coastal and plateau areas from ITASE traverses. Annals of Glaciology 41: 32–40.

Bertler

NAN

Naish

Mayewski

. (2006) Opposing oceanic and atmospheric ENSO influences on the Ross Sea Region, Antarctica. Advance Geosciences 6: 83–86.

Bowen

HJM

(1979) Environmental Chemistry of the Elements. London: Academic Press, 333 pp.

10.

Bromwich

Fogt

(2004) Strong trends in the skill of the ERA-40 and NCEP-NCAR reanalyses in the high and midlatitudes of the Southern Hemisphere, 1958–2001*. Journal of Climate 17: 4603–4619.

11.

Cavalieri

Parkinson

Vinnikov

(2003) 30-year satellite record reveals contrasting Arctic and Antarctic decadal sea ice variability. Geophysical Research Letters 30: 1970.

12.

Cavalieri

Gloersen

Parkinson

. (1997) Observed hemispheric asymmetry in global sea ice changes. Science 278: 1104–1106.

13.

Ciais

White

JWC

Jouzel

. (1995) The origin of present day Antarctic precipitation from surface deuterium excess data. Journal of Geophysical Research: Atmospheres 100:18917–18927.

14.

Ciasto

Thompson

DWJ

(2008) Observations of large-scale ocean–atmosphere interaction in the Southern Hemisphere. Journal of Climate 21: 1244–1259.

15.

Craig

(1961) Isotopic variations in meteoric waters. Science 133: 1702–1703.

16.

Curran

MAJ

van Ommen

Morgan

. (2003) Ice core evidence for Antarctic sea ice decline since the 1950s. Science 302: 1203–1206.

17.

Dansgaard

(1964) Stable isotopes in precipitation. Tellus 16: 436–468.

18.

Dieckmann

Hellmer

(2010) The importance of sea ice: An overview. In: Thomas

Dieckmann

(eds) Sea Ice. 2nd Edition. Oxford: Wiley-Blackwell, pp. 1–22.

19.

Fogt

Bromwich

(2006) Decadal variability of the ENSO teleconnection to the high latitude South Pacific governed by coupling with the Southern Annular Mode. Journal of Climate 19: 979–997.

20.

Giovinetto

Zwally

(2000) Spatial distribution of net surface accumulation on the Antarctic ice sheet. Annals of Glaciology 31: 171–178.

21.

Gregory

Noone

(2008) Variability in the teleconnection between the El Niño–Southern Oscillation and West Antarctic climate deduced from West Antarctic ice core isotope records. Journal of Geophysical Research: Atmospheres 113: D17.

22.

Grinsted

Moore

Jevrejeva

(2004) Application of the cross wavelet transform and wavelet coherence to geophysical time series. Nonlinear Processes in Geophysics 11: 561–566.

23.

Hall

Visbeck

(2002) Synchronous variability in the Southern Hemisphere atmosphere, sea ice, and ocean resulting from the Annular Mode. Journal of Climate 15: 3043–3057.

24.

Helsen

van

Wal

RSW

van den Broecke

. (2006) Modeling the isotopic composition of Antarctic snow using backward trajectories: Simulation of snow pit records. Journal of Geophysical Research: Atmospheres 111: D15109.

25.

Hoshina

Fujita

Nakazawa

. (2014) Effect of accumulation rate on water stable isotopes of near-surface snow in inland Antarctica. Journal of Geophysical Research: Atmospheres 119: 274–283.

26.

HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) Model access via NOAA ARL READY (n.d.) Silver Spring, MD: NOAA Air Resources Laboratory. Available at: http://www.arl.noaa.gov/ready/hysplit4.html.

27.

Ichiyanagi

Numaguti

Kato

(2002) Interannual variation of stable isotopes in Antarctic precipitation in response to El Nino-Southern Oscillation. Geophysical Research Letters 29: 1001.

28.

Iizuka

Hondoh

Fujii

(2008) Antarctic sea ice extent during the Holocene reconstructed from inland ice core evidence. Journal of Geophysical Research: Atmospheres 113: D15114.

29.

Jones

Fogt

Widmann

. (2009) Historical SAM variability, Part I: Century-length seasonal reconstructions. Journal of Climate 22: 5319–5345.

30.

Jourdain

Preunkert

Cerri

. (2008) Year-round record of size-segregated aerosol composition in central Antarctica (Concordia station): Implications for the degree of fractionation of sea-salt particles. Journal of Geophysical Research: Atmospheres 113: D14308.

31.

Jouzel

Merlivat

Lorius

(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.

32.

Laluraj

Thamban

Naik

. (2011) Nitrate records of a shallow ice core from East Antarctica: Atmospheric processes, preservation and climatic implications. The Holocene 21: 351–356.

33.

Lefebvre

Goose

Timmermann

. (2004) Influence of the Southern Annular Mode on the sea ice-ocean system. Journal of Geophysical Research: Oceans 109: C09005.

34.

Legrand

Fenietsaigne

Saltzman

. (1992) Spatial and temporal variations of methanesulfonic-acid and non-sea salt sulfate in Antarctic ice. Journal of Atmospheric Chemistry 14: 245–260.

35.

Lenaerts

JTM

Brown

Van Den Broeke

. (2014) High variability of climate and surface mass balance induced by Antarctic ice rises. Journal of Glaciology 60: 1101–1110.

36.

Levine

Yang

Jones

. (2014) Sea salt as an ice core proxy for past sea ice extent: A process-based model study. Journal of Geophysical Research: Atmospheres 119: 2013JD020925.

37.

Marshall

(2003) Trends in the Southern annular mode from observations and reanalysis. Journal of Climate 16: 4134–4143.

38.

Marshall

Di Battista

Naik

. (2010) Analysis of a regional change in the sign of the SAM-temperature relationship in Antarctica. Climate Dynamics 36(1–2): 277–287.

39.

Masson-Delmotte

Hou

Ekaykin

. (2008) A review of Antarctic surface snow isotopic composition: Observations, atmospheric circulation, and isotopic modeling. Journal of Climate 21: 3359–3387.

40.

Masson-Delmotte

Jouzel

Landais

. (2005) Deuterium excess reveals millennial and orbital scale fluctuations of Greenland moisture origin. Science 309: 118–121.

41.

Masson-Delmotte

Stenni

Blunier

. (2010) Abrupt change of Antarctic moisture origin at the end of Termination II. Proceedings of the National Academy of Sciences 107: 12091–12094.

42.

Merlivat

Jouzel

(1979) Global climatic interpretation of the deuterium-oxygen 18 relationship for precipitation. Journal of Geophysical Research: Oceans 84: 5029–5033.

43.

Mulvaney

Abram

Hindmarsh

RCA

. (2012) Recent Antarctic peninsula warming relative to Holocene climate and ice-shelf history. Nature 489: 141–144.

44.

Murphy

Clarke

Abram

. (2014) Variability of sea-ice in the northern Weddell Sea during the 20th century. Journal of Geophysical Research: Oceans 119: 4549–4572.

45.

Naik

Thamban

Laluraj

. (2010) A century of climate variability in the central Dronning Maud Land, East Antarctica and its relation to Southern Annular Mode and El Niño Southern Oscillation. Journal of Geophysical Research: Atmospheres 115: D16102.

46.

Neumann

Waddington

(2004) Effects of firn ventilation on isotopic exchange. Journal of Glaciology 169: 183–194.

47.

Noone

Simmonds

(2004) Sea ice control of water isotope transport to Antarctica and implications for ice core interpretation. Journal of Geophysical Research: Atmospheres 109: D07105.

48.

Noone

Turner

Mulvaney

(1999) Atmospheric signals and characteristics of accumulation in Dronning Maud Land, Antarctica. Journal Geophysical Research 104: 19191–19211.

49.

Parkinson

Cavalieri

Gloersen

. (1999) Arctic sea ice extents, areas, and trends, 1978–1996. Journal of Geophysical Research: Oceans 104: 20837–20856.

50.

Pasteur

Mulvaney

Peel

. (1995) A 340 year record of biogenic sulphur from the Weddell Sea area, Antarctica. Annals of Glaciology 21: 169–174.

51.

Pfahl

Sodemann

(2014) What controls deuterium excess in global precipitation? Climate of the Past 10: 771–781.

52.

Rankin

Wolff

Auld

(2000) Frost flowers as a source of fractionated sea salt aerosol in the polar regions. Journal of Geophysical Research 27: 3469–3472.

53.

Reijmer

Greuell

Oerlemans

(1999) The annual cycle of meteorological variables and the surface energy balance on Berkner Island, Antarctica. Annals of Glaciology 29: 49–54.

54.

Röthlisberger

Crosta

Abram

. (2010) Potential and limitations of marine and ice core sea ice proxies: An example from the Indian Ocean sector. Quaternary Science Reviews 29, 296–302.

55.

Röthlisberger

Mulvaney

Wolff

. (2002) Dust and sea-salt variability in central East Antarctica (Dome C) over the last 45 k years and its implications for southern high-latitude climate. Journal of Geophysical Research 29: 20, 24-1–24-4.

56.

Russell

McGregor

Marshall

(2004) An examination of the precipitation delivery mechanisms for Dolleman Island, eastern Antarctic Peninsula. Tellus 56A: 501–513.

57.

Satake

Kawada

(1997) The quantitative evaluation of sublimation and the estimation of original hydrogen and oxygen of a firn core at East Queen Maud Land, Antarctica. Bulletin of Glacier Research 15: 93–97.

58.

Schlösser

Oerter

(2002) Shallow firn cores from Neumayer, Ekstromisen, Antarctica: A comparison of accumulation rates and stable-isotope ratios. Annals of Glaciology 35: 91–96.

59.

Schneider

Steig

Comiso

(2004) Recent climate variability in Antarctica from satellite-derived temperature data. Journal of Climate 17: 1569–1583.

60.

Schulz

Mudelsee

(2002) REDFIT: Estimating red-noise spectra directly from unevenly spaced paleoclimatic time series. Computers and Geosciences 28, 421–426.

61.

Sinclair

Bertler

NAN

Bowen

. (2014) 20th Century sea ice trends in the Ross Sea from a high resolution, coastal ice core record. Geophysical Research Letters 41: 3510–3516.

62.

Stammerjohn

Martinson

Smith

. (2008) Trends in Antarctic annual sea ice retreat and advance and their relation to El Niño-Southern Oscillation and Southern annular mode variability. Journal of Geophysical Research: Oceans 113: C3.

63.

Stenni

Masson-Delmotte

Johnsen

. (2001) An oceanic cold reversal during the last deglaciation. Science 293: 2074–2077.

64.

Stephens

Keeling

(2000) The influence of Antarctic sea ice on glacial-interglacial CO2 variations. Nature 404: 171–174.

65.

Sun

Ren

Qin

(2002) 60 years record of biogenic sulfur from Lambert Glacier basin firn core, East Antarctica. Annals of Glaciology 35: 362–367.

66.

Thamban

Laluraj

Mahalinganathan

. (2010) Glaciochemistry of surface snow from the Ingrid Christensen Coast, East Antarctica and its environmental implications. Antarctic Science 22: 435–441.

67.

Thompson

DWJ

Solomon

(2002) Interpretation of recent Southern Hemisphere climate change. Science 296: 895–899.

68.

Thompson

DWJ

Wallace

Hegerl

(2000) Annular modes in the extratropical circulation. Part II: Trends. Journal of Climate 13: 1018–1036.

69.

Torrence

Compo

(1998) A practical guide to wavelet analysis. Bulletin of American Meteorological Society 79: 61–78.

70.

Turner

(2004) Review: The El Niño–Southern oscillation and Antarctica. International Journal of Climatology 24: 1–31.

71.

Turner

Comiso

Marshall

. (2009) Non-annular atmospheric circulation change induced by stratospheric ozone depletion and its role in the recent increase of Antarctic sea ice extent. Geophysical Research Letters 36: L08502.

72.

Uemura

Matsui

Yoshimura

. (2008) Evidence of deuterium excess in water vapor as an indicator of ocean surface conditions. Journal of Geophysical Research: Atmospheres 113: D19.

73.

Van de Berg

van den Broeke

Reijmer

. (2006) Reassessment of the Antarctic surface mass balance using calibrated output of a regional atmospheric climate model. Journal of Geophysical Research: Atmospheres 111(D11): D11104.

74.

Vaughan

Bamber

Giovinetto

. (1999) Reassessment of net surface mass balance in Antarctica. Journal of Climate 12: 933–946.

75.

Vimeux

Cuffey

Jouzel

(2002) New insights into Southern Hemisphere temperature changes from Vostok ice cores using deuterium excess correction. Earth and Planetary Science Letter 203: 829–843.

76.

Vimeux

Masson

Jouzel

. (1999) Glacial-interglacial changes in ocean surface conditions in the Southern Hemisphere. Nature 398: 410–413.

77.

Welch

Mayewski

Whitlow

(1993) Methanesulphonic acid in coastal Antarctic snow related to sea-ice extent. Geophysical Research Letters 20: 443–446.

78.

Wolff

Rankin

Röthlisberger

(2003) An ice core indicator of Antarctic sea ice production? Geophysical Research Letters 30: 2158.

79.

Zwally

Comiso

Parkinson

. (2002) Variability of Antarctic sea ice 1979–1998. Journal of Geophysical Research: Oceans 107: 9-1–9-19.