Abstract

Observed daily precipitation data were used to investigate the characteristics of precipitation at Antarctic Progress Station and synoptic patterns associated with extreme precipitation events during the period 2003–16. The annual precipitation, annual number of extreme precipitation events, and amount of precipitation during the extreme events have positive trends. The distribution of precipitation at Progress Station is heavily skewed with a long tail of extreme dry days and a high peak of extreme wet days. The synoptic pattern associated with extreme precipitation events is a dipole structure of negative and positive height anomalies to the west and east of Progress Station, respectively, resulting in water vapor advection to the station. For the first time, we apply self-organizing maps (SOMs) to examine thermodynamic and dynamic perspectives of trends in the frequency of occurrence of Antarctic extreme precipitation events. The changes in thermodynamic (noncirculation) processes explain 80% of the trend, followed by the changes in the interaction between thermodynamic and dynamic processes, which account for nearly 25% of the trend. The changes in dynamic processes make a negative (less than 5%) contribution to the trend. The positive trend in total column water vapor over the Southern Ocean explains the change of thermodynamic term.

1. Introduction

As the most common form of precipitation over Antarctica, snowfall is the only remarkable source term in the mass balance of the Antarctic ice sheet, whose variations are important for the global sea level change. Information on snow accumulation and its relationship with atmospheric temperature allows accurate analyses of ice core data and exploration of paleoclimatological changes. Hence, we need to know the formation mechanism, spatial patterns, and temporal variations of Antarctic precipitation. However, the adverse environment with strong winds and low temperatures makes it challenging to measure Antarctic precipitation. Distinguishing between precipitation and blowing snow is one of the challenges. Two types of Antarctic precipitation are identified in terms of formation mechanism: diamond dust and high precipitation (Bromwich 1988). Diamond dust, also called clear-sky precipitation, which occurs in the interior of the ice sheet and shows some seasonal cycle, originates from tenuous and isolated clouds. On the contrary, synoptically induced high precipitation, which occurs episodically in the coastal areas, constitutes a large percentage of the annual precipitation amount (King and Turner 1997; Noone et al. 1999). The high-precipitation events are usually associated with moisture transport from the low-latitude regions (atmospheric rivers) during certain synoptic conditions (Schlosser et al. 2010a; Gorodetskaya et al. 2014). Better understanding of the seasonal occurrence of high-precipitation events and its changes is also important for interpretation of ice core data. During different historical periods high-precipitation events have occurred in different seasons, but to interpret ice core data we have to assume that high precipitation occurs in a certain season. Only with this assumption can the temperature of the season be estimated (Noone and Simmonds 1998; Noone et al. 1999; Noone and Simmonds 2004).

Before reviewing Antarctic high-precipitation events, we first introduce the synoptic environment of the high southern latitudes. The Antarctic continent is surrounded by a circumpolar trough, whose average latitude is located near 66°S (King and Turner 1997). The circumpolar trough is the main activity region of polar cyclones and depressions (Turner et al. 1996). The highest cyclone density occurs near or south of 60°S with the highest values over the southern Indian Ocean and the region south of Australia (Simmonds et al. 2003). The number of cyclones is larger in austral winter than summer, with the exception of the Bellingshausen Sea (Simmonds et al. 2003). The cyclone frequency is positively correlated with the southern annular mode (SAM) index (Thompson and Wallace 2000), especially in austral summer (Grieger et al. 2018). The strength of cyclones is related to the Pacific decadal oscillation (PDO) with more intense cyclones during the positive phase of PDO (Pezza et al. 2007). The increasing cyclone frequency in austral summer is linked with the trend in the summertime SAM index (Grieger et al. 2018). Cyclone activity strongly influences the meridional moisture transport. The major regions of cyclone-related moisture transport into Antarctica include Dronning Maud Land, Wilkes Land, and Marie Byrd Land (Cullather et al. 1998; Grieger et al. 2018).

Because of the importance of high precipitation in Antarctica, we review the studies of the Antarctic high-precipitation events. Previous studies have addressed the features of high-precipitation events and associated synoptic patterns and large-scale circulations using observed and reanalysis precipitation data. Some researchers have investigated the synoptic patterns responsible for the occurrence of high-precipitation events. Noone and Simmonds (1998) noted a zonal dipole structure of the mean sea level pressure (MSLP) anomaly associated with high-precipitation events in Casey and a low pressure anomaly in Vostok (see Fig. 1 for the station locations). The high-precipitation events on the western margin of the Ross Sea are usually related to synoptic-scale low systems and local mesocyclones over the Ross Ice Shelf (Sinclair et al. 2010).

Birnbaum et al. (2006) reviewed three categories of synoptic situations leading to high precipitation at Kohnen Station using observed snowfall data from summers 2001/02 to 2004/05. The first two categories are associated with low pressure systems moving in different directions, whereas the third corresponds to large-scale lifting processes induced by an upper-air low west of Kohnen Station over the plateau of Dronning Maud Land (DML). Using precipitation data over the period 2001–06, Schlosser et al. (2010a) considered five synoptic weather patterns causing high-precipitation events in the DML area. They found that strong cyclone systems only explained 20% of those events and more than half of the events occurred under blocking highs (anticyclones). Welker et al. (2014), however, found that the presence of a cyclone over the Weddell Sea is more important for the occurrence of high-precipitation events in DML than the blocking to its east.

Accompanying these synoptic patterns a large amount of moisture is transported to the high-precipitation sites. The moisture for coastal sites usually has a marine origin (Noone and Simmonds 1998). Most moisture sources for snow falling in DML are located in the southern Atlantic Ocean between 40° and 60°S (Reijmer and Van den Broeke 2001; Gorodetskaya et al. 2014). Local moisture origin in DML also contributed to part of the moisture (Noone et al. 1999). In contrast, Dome Argus in the high plateau mainly receives moisture from regions around 46°S in the Indian Ocean (Wang et al. 2013).

Typical individual high-precipitation events have also been documented. Noone et al. (1999) described two events in DML on 17 and 19 January 1981 and on 5 November 1997. The two events showed a similar synoptic pattern characterized by a dipole structure composed of the Weddell cyclone and a blocking high to the east. Schlosser et al. (2010b) studied an extreme precipitation event in DML on 22–25 February 2003, using data from the Antarctic Mesoscale Prediction System (AMPS), and found a similar synoptic pattern associated with increases in surface air temperature and wind speed. During 22–28 July 1994, 27 July–6 August 1995, and 26 December 2001–13 January 2002, blocking anticyclones in the South Tasman Sea and associated water vapor transportation from 35° to 40°S resulted in high-precipitation events on the East Antarctic Ice Sheet (Massom et al. 2004).

The aforementioned literature demonstrates that previous studies have focused on extreme precipitation events in DML, Wilkes Land, and the Ross Sea regions of the Antarctica. However, in the Prydz Bay region there has been little work done to investigate features of high-precipitation events and the effect of synoptic situation on them. Since high-precipitation events contribute greatly to the annual precipitation amount at other stations, whether the heavily skewed distribution of the precipitation applies to the Prydz Bay region remains unknown. To understand the mass balance of Amery Shelf and to interpret ice core data, we also need to investigate high-precipitation events in the Prydz Bay region and the synoptic situations causing these events. In the past three decades, the sea surface temperature (SST) and air–sea heat fluxes over the Southern Ocean have shown a significant change (Yu et al. 2012, 2017), and meanwhile the SAM has exhibited a significantly increasing trend (Thompson and Solomon 2002). Whether and to what extent these changes influence the precipitation over Prydz Bay region needs to be examined. The aim of this study is to assess precipitation features at Progress Station in Prydz Bay. We focus on high-precipitation events and synoptic circulations responsible for these events, as well as on the variability and trend of precipitation and their causes.

2. Study region, data, and methods

Russian Progress Station (69°22′50″S, 76°23′22″E, elevation 15.5 m above mean sea level) is located at the eastern edge of the Antarctic Larsemann Hills on a rocky/sandy plateau near the shore of Prydz Bay (Fig. 1). For a detailed station map, see Turner and Pendlebury (2004, their Fig. 7.8.2.1.1). The station is located 200 m in from the coast.

Fig. 1.

Location of Russian Progress Station (red asterisk). Pink asterisks denote other stations referred to in the text (Kohnen, Vostok, and Casey).

Fig. 1.

Location of Russian Progress Station (red asterisk). Pink asterisks denote other stations referred to in the text (Kohnen, Vostok, and Casey).

The annual daily mean air temperature at Progress Station is −9.4°C with a maximum value of 0.6°C in January and a minimum value of −16.0°C in July [statistics from the Russian Arctic and Antarctic Research Institute (AARI), http://www.aari.aq]. The mean annual wind is 6.7 m s−1 with a prevailing easterly wind direction. On average, there are 50 days per year when wind speed is more than 15 m s−1 (Turner and Pendlebury 2004). Snow is the major precipitation type at Progress Station, where the number of snow days is nearly 60 per year. The annual precipitation amount is 148.9 mm with the peak value of 18.7 mm in June and the smallest value of 5.2 mm in January (http://www.aari.aq). More information on meteorological conditions at Progress Station is presented in Turner and Pendlebury (2004).

Solid precipitation is measured every 24 h using a 0.19-m-diameter snow gauge housed inside a plastic shield to avoid spurious precipitation readings due to blowing snow. We used observations from the period from February 2003 to December 2016. The high-precipitation events are defined as a daily precipitation exceeding the 95th percentile of the whole period, whose value is 5.9 mm day−1. To compare the synoptic situations using the 95th percentile, we also utilized the precipitation threshold of the 90th percentile with a daily precipitation of 3.9 mm day−1. The two thresholds of high-precipitation events are high enough to ensure that only precipitation events related to synoptic systems are analyzed. During the 2003–16 period, at Progress Station 181 and 89 events were identified according to the 3.9 and 5.9 mm day−1 thresholds, respectively.

To analyze synoptic-scale situations leading to high-precipitation events at Progress Station, we utilized the ERA-Interim reanalysis of the European Centre for Medium-Range Weather Forecasts (ECMWF) (Dee et al. 2011). In comparison with other reanalysis dataset, ERA-Interim has a better performance in Antarctic moisture transport and precipitation (Bromwich et al. 2011; Nicolas and Bromwich 2011). ERA-Interim data have a temporal resolution of 6 h and a spatial resolution of 0.75° latitude × 0.75° longitude for the period from1979 to the present. Atmospheric variables derived from ERA-Interim reanalysis include the 500-hPa geopotential height, 500-hPa vertical velocity, MSLP, surface air temperature, surface wind field, total cloud cover, total column water vapor, and vertically integrated water vapor flux.

In this study we analyze synoptic-scale meteorological patterns related to the occurrence of daily extreme precipitation using three methods: composite analysis, empirical orthogonal functions (EOFs), and self-organizing maps (SOMs; Kohonen 2001). Grotjahn et al. (2016) summarized the three methods and compared their advantages and disadvantages in their Table 2. Here we utilize composite analysis to average the atmospheric variables at each grid point on specific dates when the daily precipitation amount exceeds the thresholds. The EOF analysis of atmospheric variables during the extreme-precipitation period can produce a set of modes composed of a spatial pattern and corresponding time series. The spatial pattern represents the synoptic pattern of extreme precipitation occurrences. Each mode has an eigenvalue that depicts the importance of each synoptic pattern. The SOM technique (Kohonen 2001; Hewitson and Crane 2002; Cassano et al. 2015) employs a neural network algorithm to produce continuous patterns of a dataset. The specific process is to reduce a high-dimensional data into a two-dimensional matrix of nodes. Each node includes a spatial pattern of the data, and all nodes and their distribution can represent the distribution of the data. Unlike EOF analysis, the SOM technique does not require the orthogonality of two spatial patterns. Sheridan and Lee (2011) summarized the application of the SOM technique to the study of extreme events. In this study we carry out SOM analysis of 500-hPa geopotential height field to identify the patterns that co-occur with extreme precipitation events. The three methods (the composite analysis, EOF, and SOM methods) address anomalous values of atmospheric variables. Hence, before applying them the mean annual cycle in 2003–16 is subtracted from all atmospheric variables.

To quantitatively partition the dynamic and thermodynamic contributions to the total trend in precipitation extreme occurrence, we use the method derived from Cassano et al. (2007):

 
formula

where E is the frequency of extreme precipitation occurrence, is the frequency of SOM node i, is the frequency of extreme precipitation occurrence when SOM node i occurs, and K is the number of SOM nodes (). Bars and primes denote the temporal average and deviation from it, respectively.

We differentiate Eq. (1) with respect to time to obtain Eq. (2):

 
formula

The left-hand side shows the total trend in the occurrence of precipitation extremes, and the right-hand side gives the thermodynamic, dynamic, and interaction contributions to the total trend associated with each SOM node i.

For the thermodynamic contribution, we assume that the frequency of each SOM pattern is stationary over the study period and that the trends in precipitation extreme occurrence result from noncirculation factors, such as the changes in surface moisture and temperature. The thermodynamic contribution is determined by the product of the mean occurrence of the SOM pattern and the trend in extreme precipitation occurrence per SOM pattern occurrence. For the dynamic contribution we assume that the extreme precipitation occurrence related to each SOM pattern is stationary and that the trends stem from the changes in each SOM pattern occurrence. The trend from dynamic factors is determined by the product of the trend in SOM pattern occurrences and the mean frequency of extreme event occurrences per SOM pattern occurrence. The third component represents the interaction between the dynamic and thermodynamic factors, such as the land–atmosphere interaction. For a rainfall event, land–atmosphere interaction refers to an anomalous cyclone interacting with anomalous higher water vapor on the land surface. It is obtained from the trend in the product of anomalous SOM pattern occurrence and the anomalous number of precipitation extreme occurrence per SOM pattern occurrence.

3. Results

a. Precipitation characteristics

At Progress Station during the 2003–16 period the mean annual precipitation was 165.5 mm water equivalent and the mean annual number of days with precipitation was 125.4. The annual precipitation shows a large interannual variability and trend (Fig. 2). The SAM is the major atmospheric variability mode in southern middle to high latitudes (Thompson and Wallace 2000). In our dataset, the time series of the annual precipitation is correlated with the annual SAM index (correlation coefficient r = 0.5 at above 90% confidence level, p < 0.1), both having an increasing trend: 7.1 mm yr−1 (p < 0.05) for annual precipitation and 0.04 yr−1 (p = 0.12) for annual SAM index. However, the correlation coefficient between the two detrended time series is only 0.33, which is below the 80% confidence level. Hence, the results do not support any causal relationship between annual precipitation and SAM. The annual precipitation shows no significant relationship with the Niño-3.4 and zonal wave 3 (ZW3) (Raphael 2004) indices. El Niño–Southern Oscillation (ENSO) exerts an insignificant effect on the Antarctic precipitation at East Antarctica (Bromwich et al. 2004). During the positive (negative) phase of the ZW3 index, characterized by three higher (lower) pressure and geopotential height anomalies in the latitude band of 45°–50°S, the easterly (westerly) winds prevail at Progress Station. The zonal winds are unfavorable for the precipitation occurrences there, accounting for insignificant correlation between the ZW3 index and precipitation at Progress Station. Irving and Simmonds (2015) devised a new ZW3 index to allow for longitudinal shifts in the phase of the old ZW3 index. However, according to Fig. 9 of Irving and Simmonds (2015), neither the old nor the new ZW3 index is significantly correlated with the annual precipitation at Progress Station.

Fig. 2.

Time series of (a) annual precipitation amount and (b) the SAM index. The dashed lines denote linear trends. The significant levels of trends in annual precipitation amount and annual SAM index are p < 0.05 and p = 0.12, respectively.

Fig. 2.

Time series of (a) annual precipitation amount and (b) the SAM index. The dashed lines denote linear trends. The significant levels of trends in annual precipitation amount and annual SAM index are p < 0.05 and p = 0.12, respectively.

The annual precipitation is classified by the contributions from various daily precipitation amounts. Figure 3 shows the contribution of seven groups of daily precipitation amount to the annual precipitation and number of days with precipitation. No precipitation was observed in 34.6% of days, as there was no precipitation or it occurred in amounts too small to be observed, often in the form of diamond dust (Bromwich 1988). In 52.9% of days precipitation is less than 0.5 mm, and these days together only contribute to 3.8% of the annual precipitation. In 66.6% of days precipitation is less than 1.0 mm, cumulatively contributing to 10.8% of the annual amount. In contrast, in 10.2% of days precipitation exceeds 3.9 mm, cumulatively contributing to 53.1% of the annual amount. Furthermore, 5.1% of days have more than 5.9 mm of precipitation, which yields 34.5% of the annual amount. It indicates that in the Antarctic coastal regions the rare cases of extreme precipitation play a larger role in the annual precipitation than the common cases of minor precipitation of less than 1 mm, which is different from Antarctic high-elevation stations (Noone et al. 1999).

Fig. 3.

Fraction of number of days (black column) and precipitation amount (gray column) for different amounts of daily precipitation.

Fig. 3.

Fraction of number of days (black column) and precipitation amount (gray column) for different amounts of daily precipitation.

The annual amount of extreme precipitation and the number of days with extreme precipitation have exhibited large interannual variability in the last decades (Fig. 4). The annual amount of extreme precipitation correlates with the annual mean SAM (r = 0.52 for the 90th percentile threshold and r = 0.53 for the 95th percentile), as does also the annual number of days with extreme precipitation (r = 0.53 and 0.58 for 90th and 95th percentile thresholds, respectively; p < 0.05). However, after removing linear trends the correlation coefficients between the annual SAM index and the 90th and 95th percentile thresholds of the annual amount of extreme precipitation are insignificant, and this is the case also for the correlation between the annual SAM index and the 90th and 95th percentile thresholds of the annual number of days with extreme precipitation. The Niño-3.4 index does not have significant correlation with extreme precipitation either.

Fig. 4.

Time series of annual (a) number of days and (b) amount of extreme precipitation, using the 90th (red lines) and 95th (black lines) percentile thresholds. The dashed lines denote linear trends (p < 0.05).

Fig. 4.

Time series of annual (a) number of days and (b) amount of extreme precipitation, using the 90th (red lines) and 95th (black lines) percentile thresholds. The dashed lines denote linear trends (p < 0.05).

In addition, during the last decade the annual amount and number of days of extreme precipitation exhibit an increasing trend. The trends in the annual amount of extreme precipitation are 5.2 and 3.8 mm yr−1 (p < 0.05) for the thresholds of the 90th and 95th percentiles, respectively. The trends in the number of days with extreme precipitation are 0.7 and 0.4 day yr−1 (p < 0.05), respectively. However, the weak relationship between the detrended annual SAM index and extreme precipitation does not support any causal relationship between the positive trends in extreme precipitation and the annual SAM index (Fig. 2b).

The annual amount and numbers of days of extreme precipitation also show a seasonal cycle (Fig. 5). Overall the amount of extreme precipitation is larger and such events occur more frequently in austral winter (June–August) than in other seasons. This is consistent with the winter maximum in the occurrence of strong polar mesoscale cyclones in the southern Indian Ocean (Pezza et al. 2016). However, certain months of austral spring and autumn show a high frequency of occurrence and a large amount of extreme precipitation. For example, May has the highest number of days of extreme precipitation (22 days in 14 years) using the 90th percentile threshold and the second highest for the 95th percentile threshold. For the total amount of extreme precipitation the values in May and September exceeded those in June and August for the 95th percentile threshold.

Fig. 5.

The 14-yr averaged (a) number of days and (b) amount (mm) of extreme precipitation using the 90th (black column) and 95th (gray column) percentile threshold for the 2003–16 period.

Fig. 5.

The 14-yr averaged (a) number of days and (b) amount (mm) of extreme precipitation using the 90th (black column) and 95th (gray column) percentile threshold for the 2003–16 period.

b. Synoptic patterns of extreme precipitation

As shown above, more than half of the annual precipitation falls in extreme events, which are associated with cyclones. To better understand the mechanism underlying these extreme precipitation events, we characterize the main synoptic patterns resulting in these events using three methods: composite analysis, EOF, and SOM. We also intercompare these patterns.

1) Composite analysis

Figure 6 shows the composites of 500-hPa geopotential height anomaly, MSLP anomaly, and vertically integrated water vapor flux in the region of 40°–80°S, 40°–130°E. The synoptic patterns show similar features for different strengths of extreme precipitation events. The spatial patterns of extreme events for the 95th percentile threshold are more pronounced than those for the 90th percentile threshold. The composite of 500-hPa geopotential height is characterized by a dipole pattern of negative (positive) anomalies west (east) of Progress Station. The MSLP composite map shows an almost identical spatial pattern. The circulation system has a slight westward (baroclinic) tilt. The locations of anomalous MSLP centers shift eastward compared to 500-hPa height. The anomalous northerly winds from the southern midlatitudes transport moisture to Progress Station, which favors extreme precipitation there. Such water vapor transportation patterns also can be seen in high-accumulation events in May 2009 and February 2011 (Gorodetskaya et al. 2014).

Fig. 6.

Composited (top) 500-hPa geopotential height anomaly (gpm), (middle) MSLP (hPa), and (bottom) vertically integrated water vapor flux (kg m−1 s−1) for the (left) 90th and (right) 95th percentiles of extreme precipitation events. The red asterisk indicates the location of Progress Station.

Fig. 6.

Composited (top) 500-hPa geopotential height anomaly (gpm), (middle) MSLP (hPa), and (bottom) vertically integrated water vapor flux (kg m−1 s−1) for the (left) 90th and (right) 95th percentiles of extreme precipitation events. The red asterisk indicates the location of Progress Station.

To assess the consistency among the members of the composites, we calculated the sign counts at each grid point, following Grotjahn (2011) and Gao et al. (2014). The sign counts show the number of individual members whose anomalies have the same sign as the composite. A negative sign has been placed in front of the sign counts to indicate that the composite anomaly is negative over this region. The larger the absolute value of the sign count is, the more consistent sign exists among the members. If at a grid point all the members have the same sign as the composite, the sign count for the grid point is the number of extreme precipitation events (181 or 89). The similarity of the spatial patterns of the sign count and the composites of 500-hPa geopotential height demonstrates a high consistency of the composite members (Fig. 7). Spatial patterns of sign counts for anomalous MSLPs resemble those of 500-hPa height (not shown).

Fig. 7.

The fraction of sign counts of the composite members for 500-hPa geopotential height anomaly with the (a) 90th and (b) 95th percentiles of extreme precipitation. The dotted region indicates areas above the 95% confidence level. A negative sign has been placed in front of the sign counts to indicate regions where the composite anomaly is negative.

Fig. 7.

The fraction of sign counts of the composite members for 500-hPa geopotential height anomaly with the (a) 90th and (b) 95th percentiles of extreme precipitation. The dotted region indicates areas above the 95% confidence level. A negative sign has been placed in front of the sign counts to indicate regions where the composite anomaly is negative.

2) EOF

An EOF analysis is applied to 500-hPa geopotential height (40°–80°S, 40°–130°E) of 181 extreme precipitation events using the 90th percentile threshold and 89 extreme precipitation events using the 95th percentile threshold. The spatial patterns of the first three modes are shown in Fig. 8. The first three modes explain more than 60% of the variance. These modes resemble the spatial patterns of the different extreme precipitation events. Using the rule of North et al. (1982), we note that the first three modes are significantly different from each other. The first modes explain 24.1% and 26.3% of variances for extreme events using the 90th and 95th percentile thresholds, respectively. The northwest–southeast orientation dipole has elongated positive height anomalies over the Antarctic continent and coast, indicating a well-developed blocking pattern over Wilkes Land (Fig. 8). Because of the similarity of spatial patterns for the 90th and 95th percentile thresholds, we only show MSLP and water vapor flux for the 90th percentile threshold (Fig. 9). The anomalous MSLP map for mode 1 shows a dipole structure of south–north orientation, which is somewhat different from the anomalous 500-hPa pattern. It indicates that extreme events above the 90th percentile are driven by baroclinic energy conversion with a westward tilt with height. The anomalous northeasterly winds north of the strongest positive MSLP anomalies transport a large amount of moisture to Progress Station, resulting in the occurrence of high-precipitation events (Fig. 9). Mode 2 (20%) for the 90th percentile of extreme precipitation and mode 3 (16%) for the 95th percentile of extreme precipitation show a positive height anomaly over the ocean north of Wilkes Land and a negative height anomaly over the ocean north of DML, which produces a northeasterly advection of warm and moist air to Progress Station (Fig. 9). Mode 2 for the 90th percentile and mode 3 for the 95th percentile of precipitation resemble the composite analysis patterns. The spatial pattern of the second mode means a blocking high with a northwesterly flow toward the station. An anomalously strong cyclone occurs over the north of Progress Station in mode 3 (16%) for the 90th percentile of extreme precipitation and mode 2 (20%) for the 95th percentile of extreme precipitation. With such a circulation pattern, Schlosser et al. (2010a,b) reported that extreme precipitation at Progress Station is associated with the frontal system of a cyclone (Fig. 9).

Fig. 8.

Spatial patterns of the first three modes for 500-hPa geopotential height anomaly (gpm) (40°–80°S, 40°–130°E) of (left) 181 extreme precipitation events using the 90th percentile threshold and (right) 89 extreme precipitation events using the 95th percentile threshold. The number in the top-right corner indicates the explained variance.

Fig. 8.

Spatial patterns of the first three modes for 500-hPa geopotential height anomaly (gpm) (40°–80°S, 40°–130°E) of (left) 181 extreme precipitation events using the 90th percentile threshold and (right) 89 extreme precipitation events using the 95th percentile threshold. The number in the top-right corner indicates the explained variance.

Fig. 9.

The (top) first, (middle) second, and (bottom) third modes of (left) the MSLP anomaly (hPa) and (right) the corresponding vertically integrated water vapor flux (kg m−1 s−1) of 181 extreme precipitation events. The red asterisk indicates the location of Progress Station.

Fig. 9.

The (top) first, (middle) second, and (bottom) third modes of (left) the MSLP anomaly (hPa) and (right) the corresponding vertically integrated water vapor flux (kg m−1 s−1) of 181 extreme precipitation events. The red asterisk indicates the location of Progress Station.

3) SOM

We applied the SOM technique to anomalous daily 500-hPa geopotential heights from 2003 to 2016 for the domain 40°–130°E, 40°–80°S, which encompasses the synoptic circulation influencing extreme precipitation at Progress Station. A key issue to employ the SOM is the number of nodes. More (fewer) nodes indicate more (less) detail and less (more) generalization of each synoptic circulation pattern (Cassano et al. 2015). To find a compromise between them we tried different sizes ranging from 4 to 48. Analogous to root-mean-square error (RMSE), quantization error is a metric measuring how well the SOM nodes match the data and is equal to mean-square error multiplied by the number of grid points (Cassano et al. 2015). Figure 10 shows the change in the quantization error with increase in node count. As the node count increases from 4 to 20, quantization error decreases from 3269 to 2855 geopotential meters (gpm) and then gradually levels off. A 20-node SOM was chosen to produce the synoptic circulation patterns shown in Fig. 11. Among the 20 patterns, nodes 3 and 8 show a dipole of eastern negative and western positive anomalies; an opposite phase pattern occurs in nodes 13 and 18. Nodes 7 and 15 exhibit positive and negative phases of the SAM index. Other nodes represent transition states among the zonal dipole and the SAM patterns.

Fig. 10.

Quantization error vs number of SOM nodes.

Fig. 10.

Quantization error vs number of SOM nodes.

Fig. 11.

The 500-hPa geopotential height (gpm) anomaly patterns of 5 × 4 SOM nodes.

Fig. 11.

The 500-hPa geopotential height (gpm) anomaly patterns of 5 × 4 SOM nodes.

The different patterns have a different frequency of occurrence (Fig. 12). Only nodes 1, 3, 5, 16, and 17 occur more than 6% of time, among which nodes 1 and 5 have the highest frequency of occurrence. The spatial patterns of 20 nodes show seasonal difference (Fig. 13). During winter nodes 1, 5, 16, and 20 occur more than 7% of time, representing the northwest–southeast and northeast–southwest orientation of dipole structures. The dipole structure for nodes 1 and 5 is favorable for the enhancement of katabatic wind at Progress Station, while the dipole for nodes 16 and 20 helps the moisture transport to Progress Station, increasing the probability of extreme precipitation there. In contrast, in summer the nodes show a smaller difference in the frequency of occurrence. In autumn and spring the frequency distribution of nodes shows transition features between summer and winter. In autumn, apart from nodes 1 and 5, the west–east structure dipole (node 3) exhibits more than 7% of occurrence frequency, while the most common node (mode 1) occurs nearly 10% of the time in spring.

Fig. 12.

Frequency of occurrence of each node pattern for the 2003–16 period.

Fig. 12.

Frequency of occurrence of each node pattern for the 2003–16 period.

Fig. 13.

Frequency of occurrence of each node pattern in austral (a) autumn (March–May), (b) winter (June–August), (c) spring (September–November), and (d) summer (December–February).

Fig. 13.

Frequency of occurrence of each node pattern in austral (a) autumn (March–May), (b) winter (June–August), (c) spring (September–November), and (d) summer (December–February).

Compared with the seasonal difference of these patterns we are more concerned about which patterns are conducive for the occurrence of extreme precipitation at Progress Station. Figure 14 shows the occurrence of patterns during extreme precipitation events in terms of two thresholds (the 90th and 95th percentiles). With the exception of node 14, the other 19 synoptic patterns can produce extreme precipitation at Progress Station, although most of the patterns have less than 5% occurrence frequency in extreme precipitation events. Synoptic patterns of more than 7% frequency are those of nodes 16, 17, 19, and 20, and the sum of their frequencies is 48.1% and 51.7% for the 90th and 95th percentiles, respectively. The spatial patterns of these four nodes are the main synoptic patterns of extreme precipitation events at Progress Station. We present the anomalous MSLP and water vapor flux for the four nodes (Fig. 15). Spatial patterns of anomalous MSLPs for the four nodes are similar to those of 500-hPa heights, indicating that those synoptic systems influencing extreme precipitation at Progress Station are nearly barotropic. The spatial pattern of node 20 is similar to that of the first EOF mode, which produces the northeasterly airflow transporting moisture to Progress Station. The spatial patterns of nodes 16 and 17 lie in the transition between EOF modes 2 and 3. The wind direction over Progress Station is northeasterly for nodes 16, 17, 19, and 20. For nodes 16, 17, and 19, synoptic systems favoring extreme precipitation show a dipole structure with an anomalous low in its west and an anomalous high in its east, although the shape and location of the spatial patterns vary. These patterns also represent strong cyclones and blocking highs with northeasterly winds passing Progress Station in the study region. For node 20, an anomalous cyclone over the ocean and an anomalous anticyclone over the land produce anomalous northeasterly winds toward the station.

Fig. 14.

Frequency of occurrence of each node pattern for the extreme precipitation events with the threshold of the (a) 90th and (b) 95th percentiles.

Fig. 14.

Frequency of occurrence of each node pattern for the extreme precipitation events with the threshold of the (a) 90th and (b) 95th percentiles.

Fig. 15.

(left) MSLP anomaly (hPa) and (right) the corresponding vertically integrated water vapor flux (kg m−1 s−1) for nodes 16, 17, 19, and 20.

Fig. 15.

(left) MSLP anomaly (hPa) and (right) the corresponding vertically integrated water vapor flux (kg m−1 s−1) for nodes 16, 17, 19, and 20.

We further analyze other meteorological variables during the occurrences of nodes 16, 17, 19, and 20 (Figs. 1618). For node 16, Progress Station is located in a region between an anomalous surface anticyclone to its northeast and a small anomalous surface cyclone to its northwest. Anomalous northeasterly surface winds transport warm and moisture air to Progress, providing a favorable condition of extreme precipitation events. For nodes 17, 19, and 20, the anomalous surface cyclone influences Progress more strongly than in the case of node 16. Stronger anomalous northeasterly winds over Progress Station advect warm and moist air from the Southern Ocean to Progress and the Antarctic continent. Positive anomalies in the surface air temperature, cloud cover, and total column water vapor clearly demonstrate the advection process, which is favorable for the occurrence of extreme precipitation events. In addition to these moist thermodynamic conditions, the occurrence of extreme precipitation events also requires favorable dynamic conditions. For the four nodes, a distinct upward motion occurs at the 500-hPa level over Progress Station (Fig. 18). The ascent can be largely associated with the pressure pattern and orographic lifting.

Fig. 16.

As in Fig. 15, but for the (left) surface air temperature anomaly (°C) and (right) surface wind field (m s−1).

Fig. 16.

As in Fig. 15, but for the (left) surface air temperature anomaly (°C) and (right) surface wind field (m s−1).

Fig. 17.

As in Fig. 15, but for anomalies of the (left) total cloud cover and (right) total column water vapor (kg m−2).

Fig. 17.

As in Fig. 15, but for anomalies of the (left) total cloud cover and (right) total column water vapor (kg m−2).

Fig. 18.

As in Fig. 15, but for the 500-hPa vertical velocity anomaly (Pa s−1). Positive (negative) vertical velocity indicates descending (ascending) airflow.

Fig. 18.

As in Fig. 15, but for the 500-hPa vertical velocity anomaly (Pa s−1). Positive (negative) vertical velocity indicates descending (ascending) airflow.

Next, we analyze the contributions of 20 nodes to the trends in extreme precipitation frequency at Progress Station from the dynamic and thermodynamic perspectives. Among the three components (thermodynamic, dynamic, and their interaction), thermodynamic factors make the largest contribution, explaining 77.0% and 83.3% of the total trend in frequency of extreme precipitation occurrence, for the 90th and 95th thresholds, respectively. The second-largest contribution comes from the interaction, which accounts for 26.6% and 21.1%, whereas the dynamic contributions are only −3.6% and −4.4% and slightly offset the contribution from the other two components. Figures 19 and 20 show the total trend and its three components for each SOM node using the 90th and 95th percentile thresholds. As shown in Fig. 19a, nodes 1, 2, and 19 have the three largest contributions to the total trend, and their thermodynamic components explain the majority of the trends. The fourth largest contribution is from node 16 but, unlike the above three nodes, its dynamic and interaction components account for most of its trend. Among the 20 nodes, node 15 has the largest thermodynamic contribution (Fig. 19b), but negative trends in other two components offset the contribution (Figs. 19c,d). Nodes 20 and 16 are the first two largest magnitudes of the dynamic component, but their signs are opposite, corresponding to the opposite trends in their occurrences (0.75 day yr−1 for node 16 and 20.79 day yr−1 for node 20). The interaction term means the changes in node pattern frequency act on the changes in extreme precipitation occurrence for each node. Node 20 has the largest interaction component among the 20 nodes, which accounts for more than 0.1 day yr−1. Because of the large negative magnitude of its dynamic component, node 20 only has a small contribution to the total trend.

Fig. 19.

Trends in the frequency of precipitation extreme occurrences using the 90th percentile threshold of the (a) total, (b) thermodynamic contribution, (c) dynamic contribution, and (d) interaction contribution (day yr−1) at Progress Station for each SOM node for the 2003–16 period.

Fig. 19.

Trends in the frequency of precipitation extreme occurrences using the 90th percentile threshold of the (a) total, (b) thermodynamic contribution, (c) dynamic contribution, and (d) interaction contribution (day yr−1) at Progress Station for each SOM node for the 2003–16 period.

Fig. 20.

As in Fig. 19, but for the 95th percentile threshold.

Fig. 20.

As in Fig. 19, but for the 95th percentile threshold.

Similarly, we also examined the contributions for each node using the 95th percentile threshold (Fig. 20). Unlike the 90th percentile threshold, the first four largest contributions to the total trend are nodes 16, 19, 2, and 12, which mainly stem from their large thermodynamic components. In contrast, node 17 has the most negative trend, which similarly results from the most negative thermodynamic component. Like the 90th percentile threshold, nodes 16 and 20 have the largest dynamic and interaction components, respectively; the largest interaction component is node 20, followed by node 16.

Among the three terms the thermodynamic term plays the greatest role in the trend in the occurrence of extreme precipitation events. We investigate the trend in total column water vapor to explain the reason for the thermodynamic term. The water vapor of extreme precipitation events at Progress Station originates from the Southern Ocean to its northeast (Fig. 16). The daily total column water vapor shows a significantly increasing trend in the source region (Fig. 21), which provides ample water vapor for the increasing frequencies of extreme precipitation events accompanying the favorable atmospheric circulation (northeasterly winds).

Fig. 21.

The trend in daily total column water vapor for each year over the 2003–16 period. Regions with significant (p < 0.05) trends are dotted.

Fig. 21.

The trend in daily total column water vapor for each year over the 2003–16 period. Regions with significant (p < 0.05) trends are dotted.

It should be noted that our daily precipitation data are derived from gauge observations at Progress Station. Blowing snow may distort the snowfall measurements in Antarctica, leading to errors of precipitation data. The spatial patterns of node 2 and 3 are favorable for the development of katabatic winds at Progress Station, which may cause blowing snow. However, the presented synoptic patterns can explain well the occurrence of extreme precipitation. To determine the effect of blowing snow, we examined the wind speed on days of extreme precipitation occurrences (Fig. 22). There is no significant positive correlation between wind speed and snow accumulation during extreme precipitation events, which is in agreement with Pomeroy and Gray (1995). King and Turner (1997) used a wind speed of 10 m s−1 as a blowing snow threshold. In our dataset, the wind speed exceeds 10 m s−1 in only 30% of days, indicating fewer occurrences of blowing snow during the period of extreme precipitation. According to Palm et al. (2011, their Fig. 11), during winter months (April–October) blowing snow occurs at Progress Station on less than 10% of days. Hence the precipitation measurement error caused by blowing snow does not have a large influence on our conclusions. In addition to blowing snow, wind affects snowfall observations via the wind-induced undercatch of gauges (Aleksandrov et al. 2005). The effect itself is often large; however, as similar gauges have been used in the Progress Station throughout our study period, the effect on interannual variations and trends should be minor.

Fig. 22.

The relationship between daily extreme precipitation (≥3.9 mm) and 10-m wind speed.

Fig. 22.

The relationship between daily extreme precipitation (≥3.9 mm) and 10-m wind speed.

4. Conclusions and discussion

Based on the daily observed precipitation data for the 2003–16 period, we have investigated the characteristics of extreme precipitation at Progress Station and linked the occurrence of extreme precipitation to synoptic patterns using three methods.

Like other Antarctic stations, the precipitation at Progress Station has a heavily skewed distribution with a long tail of extreme dry days and a high peak of extreme wet days. The high precipitation occurs mostly in austral winter (May–September). The annual precipitation at Progress Station has a significantly increasing trend. The increasing trend can also be seen in time series of the annual amount and number of days with extreme precipitation. From thermodynamic and dynamic perspectives, we investigated the trend in the frequency of extreme precipitation occurrence. The thermodynamic components explained 80% of the trend, which is related to a positive trend in total column water vapor over the Southern Ocean. Dynamic components had a negative and less than 5% contribution. The changes in the interaction of thermodynamic and dynamic processes explained nearly one-quarter of the total trend. Overall, the increasing trends in annual precipitation and extreme precipitation are not related to the positive trend in annual SAM index but are associated with the increasing column water vapor over the southern Indian Ocean.

We find no significant correlation between the detrended annual SAM index and annual precipitation and annual extreme precipitation at Progress Station, although the positive (negative) phase of SAM is related to more (fewer) cyclones around Antarctica (80°–60°S) in the southern Indian Ocean (Pezza et al. 2008, 2012; Reboita et al. 2015), producing likely more precipitation over East Antarctica. Noone et al. (1999) found a moderate modulation of ENSO on annual precipitation at 75°S, 0°E, and Divine et al. (2009) also noted an unstable relationship between ENSO and stable isotope records in Antarctic ice cores from Dronning Maud Land. During 2003–16 there was no significant relationship between precipitation at Progress Station and ENSO and ZW3. Longer-term precipitation data need to be used to examine the relationships between precipitation and large-scale oceanic and atmospheric circulations such as ENSO, SAM, and ZW3.

Three methods were utilized to investigate the synoptic patterns associated with the occurrence of extreme precipitation events at Progress Station. Among the methods, SOMs captured the most diverse synoptic patterns of the extreme precipitation events and gave the frequency of occurrences of these synoptic patterns. Although extreme precipitation at Progress Station occurred during nearly all of the SOM patterns, the main four synoptic patterns accounted for most of the extreme precipitation occurrences. Their common features demonstrate that a dipole structure of negative (positive) height anomalies in its west (east) with northeasterly flows over Progress Station is a major weather pattern producing extreme precipitation there. The ascending motion over Progress Station also provides a favorable dynamic condition for the occurrence of extreme precipitation. In previous studies, analogous dipole patterns with strong cyclones and blocking highs have been associated with extreme precipitation in Dronning Maud Land (Noone and Simmonds 1998; Birnbaum et al. 2006; Schlosser et al. 2010a; Sinclair et al. 2010; Welker et al. 2014) with a preceding upper-level atmospheric pattern and a preceding Rossby wave train in the Southern Ocean (Welker et al. 2014).

Water vapor flux induced by the dipole structure is also of importance for the extreme precipitation events. Welker et al. (2014) emphasized the importance of meridional moisture flux to the occurrence of extreme precipitation events in Dronning Maud Land. Further study needs to be done to quantify the role of meridional moisture flux in extreme precipitation events at Progress Station.

Although our study may be influenced by errors in precipitation measurements, the results should provide some suggestions for the interpretation of ice core data and the investigation of the surface mass balance in Amery Ice shelf. Our results also highlight that changes in atmospheric variables over the Southern Ocean strongly affect the changes in the Antarctic precipitation.

Acknowledgments

We are deeply grateful for the late Victor E. Lagun from AARI for his insightful comments and advice on the study. We thank the European Centre for Medium-Range Weather Forecasts (ECMWF) for the ERA-Interim reanalysis data. This study is financially supported by the National Key R&D Program of China (2018YFA0605701 and 2018YFA0605901), the National Natural Science Foundation of China (41376005), Shanghai Pujiang Program (17PJ1409800), the AARI/Russian Antarctic Expedition’s Subprogram “Study and Research of the Antarctic” of the Federal Target Program “World Ocean,” and the Academy of Finland (Contract 304345). In addition we thank the Helsinki University of Technology’s Laboratory of Computer and Information Science (http://www.cis.hut.fi/research/som-research/) for information on SOM and software for producing it.

REFERENCES

REFERENCES
Aleksandrov
,
Y. I.
,
N. N.
Bryazgin
,
E. J.
Førland
,
V. F.
Radionov
, and
P. N.
Svyashchennikov
,
2005
:
Seasonal, interannual and long-term variability of precipitation and snow depth in the region of the Barents and Kara Seas
.
Polar Res.
,
24
,
69
85
, https://doi.org/10.3402/polar.v24i1.6254.
Birnbaum
,
G.
,
R.
Brauner
, and
H.
Ries
,
2006
:
Synoptic situation causing high-precipitation rates on the Antarctic Plateau: Observations from Kohnen Station, Dronning Maud Land
.
Antarct. Sci.
,
18
,
279
288
, https://doi.org/10.1017/S0954102006000320.
Bromwich
,
D. H.
,
1988
:
Snowfall in high southern latitudes
.
Rev. Geophys.
,
26
,
149
168
, https://doi.org/10.1029/RG026i001p00149.
Bromwich
,
D. H.
,
A. J.
Monaghan
, and
Z.
Guo
,
2004
:
Modeling the ENSO modulation of Antarctic climate in the late 1990s with the polar MM5
.
J. Climate
,
17
,
109
132
, https://doi.org/10.1175/1520-0442(2004)017<0109:MTEMOA>2.0.CO;2.
Bromwich
,
D. H.
,
J. P.
Nicolas
, and
A. J.
Monaghan
,
2011
:
An assessment of precipitation changes over Antarctica and the Southern Ocean since 1989 in contemporary global reanalyses
.
J. Climate
,
24
,
4189
4209
, https://doi.org/10.1175/2011JCLI4074.1.
Cassano
,
E. N.
,
J. M.
Glisan
,
J. J.
Cassano
,
W. J.
Gutowski
, and
M. W.
Seefeldt
,
2015
:
Self-organizing map analysis of widespread temperature extremes in Alaska and Canada
.
Climate Res.
,
62
,
199
218
, https://doi.org/10.3354/cr01274.
Cassano
,
J. J.
,
P.
Uotila
,
A. H.
Lynch
, and
E. N.
Cassano
,
2007
:
Predicted changes in synoptic forcing of net precipitation in large Arctic river basins during the 21st century
.
J. Geophys. Res.
,
112
,
G04S49
, https://doi.org/10.1029/2006JG000332.
Cullather
,
R. I.
,
D. H.
Bromwich
, and
M. L.
Van Woert
,
1998
:
Spatial and temporal variability of Antarctic precipitation from atmospheric methods
.
J. Climate
,
11
,
334
367
, https://doi.org/10.1175/1520-0442(1998)011<0334:SATVOA>2.0.CO;2.
Dee
,
D. P.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, https://doi.org/10.1002/qj.828.
Divine
,
D. V.
, and Coauthors
,
2009
:
Tropical Pacific-high latitude south Atlantic teleconnections as seen in δ18O variability, in Antarctic coastal ice cores
.
J. Geophys. Res.
,
114
,
D11112
, https://doi.org/10.1029/2008JD010475.
Gao
,
X.
,
A.
Schlosser
,
P.
Xie
,
E.
Monier
, and
D.
Entekhabi
,
2014
:
An analogue approach to identify heavy precipitation events: Evaluation and application to CMIP5 climate models in the United States
.
J. Climate
,
27
,
5941
5963
, https://doi.org/10.1175/JCLI-D-13-00598.1.
Gorodetskaya
,
I. V.
,
M.
Tsukernik
,
K.
Claes
,
M. F.
Ralph
,
W. D.
Neff
, and
N. P. M.
Van Lipzig
,
2014
:
The role of atmospheric rivers in anomalous snow accumulation in East Antarctica
.
Geophys. Res. Lett.
,
41
,
6199
6206
, https://doi.org/10.1002/2014GL060881.
Grieger
,
J.
,
L. G. C.
Leckebusch
,
C.
Raible
,
I.
Rudeva
, and
I.
Simmonds
,
2018
:
Subantarctic cyclones identified by 14 tracking methods, and their role for moisture transports into the continent
.
Tellus
,
70A
,
1454808
, https://doi.org/10.1080/16000870.2018.1454808.
Grotjahn
,
R.
,
2011
:
Identifying extreme hottest days from large scale upper air data: A pilot scheme to find California Central Valley summertime maximum surface temperature
.
Climate Dyn.
,
37
,
587
604
, https://doi.org/10.1007/s00382-011-0999-z.
Grotjahn
,
R.
, and Coauthors
,
2016
:
North American extreme temperature events and related large scale meteorological patterns: A review of statistical methods, dynamics, modeling, and trends
.
Climate Dyn.
,
46
,
1151
1184
, https://doi.org/10.1007/s00382-015-2638-6.
Hewitson
,
B. C.
, and
R. G.
Crane
,
2002
:
Self-organizing maps: Applications to synoptic climatology
.
Climate Res.
,
22
,
13
26
, https://doi.org/10.3354/cr022013.
Irving
,
D.
, and
I.
Simmonds
,
2015
:
A novel approach to diagnosing Southern Hemisphere planetary wave activity and its influence on regional climate variability
.
J. Climate
,
28
,
9041
9057
, https://doi.org/10.1175/JCLI-D-15-0287.1.
King
,
J. C.
, and
J.
Turner
,
1997
: Antarctic Meteorology and Climatology. Cambridge University Press, 422 pp.
Kohonen
,
T.
,
2001
: Self-Organizing Maps. 3rd ed. Springer, 501 pp.
Massom
,
R. A.
,
M. J.
Pook
,
J. C.
Comiso
,
N.
Adams
,
J.
Turner
,
T.
Lachlan-Cope
, and
T. T.
Gibson
,
2004
:
Precipitation over the interior East Antarctic ice sheet related to midlatitude blocking-high activity
.
J. Climate
,
17
,
1914
1928
, https://doi.org/10.1175/1520-0442(2004)017<1914:POTIEA>2.0.CO;2.
Nicolas
,
J. P.
, and
D. H.
Bromwich
,
2011
:
Precipitation changes in high southern latitudes from global reanalyses: A cautionary tale
.
Surv. Geophys.
,
32
,
475
494
, https://doi.org/10.1007/s10712-011-9114-6.
Noone
,
D.
, and
I.
Simmonds
,
1998
:
Implications for the interpretation of ice-core isotope data from analysis of modelled Antarctic precipitation
.
Ann. Glaciol.
,
27
,
398
402
, https://doi.org/10.3189/1998AoG27-1-398-402.
Noone
,
D.
, and
I.
Simmonds
,
2004
:
Sea ice control of water isotope transport to Antarctica and implications for ice core interpretation
.
J. Geophys. Res.
,
109
,
D07105
, https://doi.org/10.1029/2003JD004228.
Noone
,
D.
,
J.
Turner
, and
R.
Mulvaney
,
1999
:
Atmospheric signals and characteristics of accumulation in Dronning Maud Land, Antarctica
.
J. Geophys. Res.
,
104
,
19 191
19 211
, https://doi.org/10.1029/1999JD900376.
North
,
G. R.
,
T. L.
Bell
,
R. F.
Cahalan
, and
F. J.
Moeng
,
1982
:
Sampling errors in the estimation of empirical orthogonal functions
.
Mon. Wea. Rev.
,
110
,
699
706
, https://doi.org/10.1175/1520-0493(1982)110<0699:SEITEO>2.0.CO;2.
Palm
,
S. P.
,
Y.
Yang
,
J. D.
Spinhirne
, and
A.
Marshak
,
2011
:
Satellite remote sensing of blowing snow properties over Antarctica
.
J. Geophys. Res.
,
116
,
D16123
, https://doi.org/10.1029/2011JD015828.
Pezza
,
A. B.
,
I.
Simmonds
, and
J. A.
Renwick
,
2007
:
Southern Hemisphere cyclones and anticyclones: Recent trends and links with decadal variability in the Pacific Ocean
.
Int. J. Climatol.
,
27
,
1403
1419
, https://doi.org/10.1002/joc.1477.
Pezza
,
A. B.
,
T.
Durrant
,
I.
Simmonds
, and
I.
Smith
,
2008
:
Southern Hemisphere synoptic behavior in extreme phases of SAM, ENSO, sea ice extent, and southern Australia rainfall
.
J. Climate
,
21
,
5566
5584
, https://doi.org/10.1175/2008JCLI2128.1.
Pezza
,
A. B.
,
H. A.
Rashid
, and
I.
Simmonds
,
2012
:
Climate links and recent extremes in Antarctic sea ice, high-latitude cyclones, southern annular mode and ENSO
.
Climate Dyn.
,
38
,
57
73
, https://doi.org/10.1007/s00382-011-1044-y.
Pezza
,
A. B.
,
K.
Sadler
,
P.
Uotila
,
T.
Vihma
,
M. D. S.
Mesquita
, and
P.
Reid
,
2016
:
Southern Hemisphere strong polar mesoscale cyclones in high-resolution datasets
.
Climate Dyn.
,
47
,
1647
1660
, https://doi.org/10.1007/s00382-015-2925-2.
Pomeroy
,
J. W.
, and
D. M.
Gray
,
1995
: Snowcover accumulation, relocation and measurement. National Hydrology Research Institute Sci. Rep. 7, 144 pp.
Raphael
,
M. N.
,
2004
:
A zonal wave 3 index for the Southern Hemisphere
.
Geophys. Res. Lett.
,
31
,
L23212
, https://doi.org/10.1029/2004GL020365.
Reboita
,
M. S.
,
R. P.
da Rocha
,
T.
Ambrizzi
, and
C. D.
Gouveia
,
2015
:
Trend and teleconnection patterns in the climatology of extratropical cyclones over the Southern Hemisphere
.
Climate Dyn.
,
45
,
1929
1944
, https://doi.org/10.1007/s00382-014-2447-3.
Reijmer
,
C. H.
, and
M. R.
Van den Broeke
,
2001
:
Moisture source of precipitation in Western Dronning Maud Land, Antarctica
.
Antarct. Sci.
,
13
,
210
220
, https://doi.org/10.1017/S0954102001000293.
Schlosser
,
E.
,
K. W.
Manning
,
J. G.
Powers
,
M. G.
Duda
,
G.
Birnbaum
, and
K.
Fujita
,
2010a
:
Characteristics of high-precipitation events in Dronning Maud Land, Antarctica
.
J. Geophys. Res.
,
115
,
D14107
, https://doi.org/10.1029/2009JD013410.
Schlosser
,
E.
,
J. G.
Powers
,
M. G.
Duda
,
K. W.
Manning
,
C. H.
Reijmer
, and
M. R.
van den Broeke
,
2010b
:
An extreme precipitation event in Dronning Maud Land, Antarctica: A case study with the Antarctic Mesoscale Prediction System
.
Polar Res.
,
29
,
330
344
, https://doi.org/10.3402/polar.v29i3.6072.
Sheridan
,
S. C.
, and
C. C.
Lee
,
2011
:
The self-organizing map in synoptic climatological research
.
Prog. Phys. Geogr.
,
35
,
109
119
, https://doi.org/10.1177/0309133310397582.
Simmonds
,
I.
,
K.
Keay
, and
E.-P.
Lim
,
2003
:
Synoptic activity in the seas around Antarctica
.
Mon. Wea. Rev.
,
131
,
272
288
, https://doi.org/10.1175/1520-0493(2003)131<0272:SAITSA>2.0.CO;2.
Sinclair
,
K. E.
,
N. A. N.
Bertler
, and
W. J.
Trompetter
,
2010
:
Synoptic controls on precipitation pathways and snow delivery to high-accumulation ice core sites in the Ross Sea region, Antarctica
.
J. Geophys. Res.
,
115
,
D22112
, https://doi.org/10.1029/2010JD014383.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
2000
:
Annual modes in the extratropical circulation. Part I: Month-to-month variability
.
J. Climate
,
13
,
1000
1016
, https://doi.org/10.1175/1520-0442(2000)013,1000:AMITEC.2.0.CO;2.
Thompson
,
D. W. J.
, and
S.
Solomon
,
2002
:
Interpretation of recent Southern Hemisphere climate change
.
Science
,
296
,
895
899
, https://doi.org/10.1126/science.1069270.
Turner
,
J.
, and Coauthors
,
1996
:
The Antarctic First Regional Observing Study of the Troposphere (FROST) project
.
Bull. Amer. Meteor. Soc.
,
77
,
2007
2032
, https://doi.org/10.1175/1520-0477(1996)077<2007:TAFROS>2.0.CO;2.
Turner
,
J.
, and
S. F.
Pendlebury
,
2004
: The International Antarctic Weather Forecasting Handbook. British Antarctic Survey, 663 pp.
Wang
,
Y.
,
H.
Sodemann
,
S.
Hou
,
V.
Masson-Delmotte
,
J.
Jouzel
, and
H.
Pang
,
2013
:
Snow accumulation and its moisture origin over Dome Argus, Antarctica
.
Climate Dyn.
,
40
,
731
742
, https://doi.org/10.1007/s00382-012-1398-9.
Welker
,
C.
,
O.
Martius
,
P.
Froidevaux
,
C. H.
Reijmer
, and
N.
Fischer
,
2014
:
A climatological analysis of high-precipitation events in Dronning Maud Land, Antarctica, and associated large-scale atmospheric conditions
.
J. Geophys. Res.
,
119
,
11932
11954
, https://doi.org/10.1002/2014JD022259.
Yu
,
L.
, and Coauthors
,
2012
:
Trends in latent and sensible heat fluxes over the Southern Ocean
.
Atmos. Climate Sci.
,
2
,
159
173
, https://doi.org/10.4236/acs.2012.22017.
Yu
,
L.
,
S.
Zhong
,
J. A.
Winkler
,
M.
Zhou
,
D. H.
Lenschow
,
B.
Li
,
X.
Wang
, and
Q.
Yang
,
2017
:
Possible connections of the opposite trends in Arctic and Antarctic, sea-ice cover
.
Sci. Rep.
,
7
,
45804
, https://doi.org/10.1038/srep45804.

Footnotes

© 2018 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).