Antarctic Peninsula Regional Circulation and Its Impact on the Surface Melt of Larsen C Ice Shelf

Chongran Zhang aNorth Carolina A&T State University, Greensboro, North Carolina

Search for other papers by Chongran Zhang in
Current site
Google Scholar
PubMed
Close
,
Jing Zhang aNorth Carolina A&T State University, Greensboro, North Carolina

Search for other papers by Jing Zhang in
Current site
Google Scholar
PubMed
Close
, and
Qigang Wu bFudan University, Shanghai, China

Search for other papers by Qigang Wu in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Enhanced surface melt over the ice shelves of the Antarctic Peninsula (AP) is one of the precursors to their collapse, which can be proceeded by accelerated ground glacier flow and increased contribution to sea level rise. With the collapse of Larsen A and B and the major 2017 calving event from Larsen C, whether Larsen C is bound for a similar fate has received increasing attention. Here, the interannual variation of regional circulation over the AP region is studied using the empirical orthogonal function (EOF)/principal component (PC) analysis on the sea level pressure of ERA5. The EOF modes capture the variations of depth, location, and extent of Amundsen Sea low and Weddell Sea low in each season. Statistically significant positive correlations exist between Larsen C surface temperature and the PC time series of EOF mode 1 in winter and spring through northerly/northwesterly wind anomalies west of the AP. The PC time series of EOF mode 2 is negatively correlated with Larsen C surface temperature in autumn and summer and surface melt in summer, all due to southerly wind anomalies east of the AP. Surface energy budget analysis associated with EOF mode 2 shows that downwelling longwave radiation over Larsen C has negative statistically significant correlations with EOF mode 2 and is the major atmospheric forcing regulating the variation of Larsen C surface melt. Positively enhanced EOF mode 2 since 2004 is responsible for the recent cooling and decline of surface melt over Larsen C.

SIGNIFICANCE STATEMENT

With the collapse of Larsen A and B Ice Shelves and the splitting up of Larsen C Ice Shelf (LCIS) in 2017, whether LCIS of the Antarctic Peninsula (AP) is bound for a similar fate has drawn great attention in recent years. We examine the impact of AP regional circulation on the surface melt of LCIS. Variations in the western extent of the Weddell Sea low and the intensity of high pressure along AP account for a considerable amount of the interannual variation of summer surface melt of LCIS. These variations cause anomalous air movement, which influences the LCIS melt mainly through longwave radiation. These results implicate the influence of atmospheric forcing on the surface melt of LCIS.

© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Chongran Zhang, chongran.zhang@gmail.com

Abstract

Enhanced surface melt over the ice shelves of the Antarctic Peninsula (AP) is one of the precursors to their collapse, which can be proceeded by accelerated ground glacier flow and increased contribution to sea level rise. With the collapse of Larsen A and B and the major 2017 calving event from Larsen C, whether Larsen C is bound for a similar fate has received increasing attention. Here, the interannual variation of regional circulation over the AP region is studied using the empirical orthogonal function (EOF)/principal component (PC) analysis on the sea level pressure of ERA5. The EOF modes capture the variations of depth, location, and extent of Amundsen Sea low and Weddell Sea low in each season. Statistically significant positive correlations exist between Larsen C surface temperature and the PC time series of EOF mode 1 in winter and spring through northerly/northwesterly wind anomalies west of the AP. The PC time series of EOF mode 2 is negatively correlated with Larsen C surface temperature in autumn and summer and surface melt in summer, all due to southerly wind anomalies east of the AP. Surface energy budget analysis associated with EOF mode 2 shows that downwelling longwave radiation over Larsen C has negative statistically significant correlations with EOF mode 2 and is the major atmospheric forcing regulating the variation of Larsen C surface melt. Positively enhanced EOF mode 2 since 2004 is responsible for the recent cooling and decline of surface melt over Larsen C.

SIGNIFICANCE STATEMENT

With the collapse of Larsen A and B Ice Shelves and the splitting up of Larsen C Ice Shelf (LCIS) in 2017, whether LCIS of the Antarctic Peninsula (AP) is bound for a similar fate has drawn great attention in recent years. We examine the impact of AP regional circulation on the surface melt of LCIS. Variations in the western extent of the Weddell Sea low and the intensity of high pressure along AP account for a considerable amount of the interannual variation of summer surface melt of LCIS. These variations cause anomalous air movement, which influences the LCIS melt mainly through longwave radiation. These results implicate the influence of atmospheric forcing on the surface melt of LCIS.

© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Chongran Zhang, chongran.zhang@gmail.com

1. Introduction

Abnormally prolonged or intense levels of surface melt over the Larsen Ice Shelves of the Antarctic Peninsula (AP) can contribute to ice shelf collapse (Rack and Rott 2004; van den Broeke 2005; Bell et al. 2018; Lai et al. 2020). The flow of grounded glaciers is known to accelerate in the absence of buttressing provided by ice shelves, thereby increasing the rate of global sea level rise (Rignot et al. 2004; Rott et al. 2002; Royston and Gudmundsson 2016; Tuckett et al. 2019). Major events in the Larsen Ice Shelves include the collapse of Larsen A in 1995, Larsen B in 2002, and the calving of a large iceberg from Larsen C in 2017 (Hogg and Gudmundsson 2017). Modeling studies that span the order of days indicate that foehn winds can coincide with temperature increases and enhanced surface melt of the Larsen Ice Shelves (Elvidge et al. 2015; Grosvenor et al. 2014; Luckman et al. 2014; Zhang and Zhang 2018). Climatology studies of foehn winds and surface melt over the Larsen Ice Shelves have shown that foehn winds influence annual melt over Larsen B; however, no significant correlations exist over most of Larsen C (Cape et al. 2015). Foehn-induced melt ranges from about 2% to 18% annually over the Larsen Ice Shelves with peak values of 18% in certain sections of the ice shelf adjacent to the base of the AP mountain range (Laffin et al. 2021). In spite of this, foehn wind–related melting has been demonstrated to exist during winter and can influence the presence of liquid water deep in the snowpack (Kuipers Munneke et al. 2018; Datta et al. 2019). Relationships have also been discovered between annual foehn wind occurrence and Larsen C annual temperature (Wiesenekker et al. 2018); however, the annual mean temperature anomalies in the AP are known to be dominated by the changes in winter temperature (Turner et al. 2020), thereby obscuring the signal from summer conditions. It has been suggested that circulation patterns which influence melt remains a topic for further research (Elvidge et al. 2020).

The major circulation patterns that impact the AP region include the Amundsen Sea low (ASL) (Bozkurt et al. 2020; Bromwich et al. 2013; Cape et al. 2015; Hosking et al. 2016; Hosking et al. 2013; Jones et al. 2019; Raphael et al. 2016; Turner et al. 2020) and the Weddell Sea low (WSL) (Bozkurt et al. 2020; Turner et al. 2016). With the self-organizing maps (SOM) and cluster analysis algorithms, Ambrožová et al. (2020) and Gonzalez et al. (2018) have also identified these synoptic patterns and others such as the low over the Drake Passage and the high pressure ridge along the AP to name a few. The influence of the ASL’s location and intensity on the climate of West Antarctica has been studied by Hosking et al. (2013); however, statistically robust results (significance above the 95% level) are only obtained for regions west of the AP and show that warmer temperatures are present when the ASL deepens or is located more eastward. By analyzing the relation between various synoptic patterns and station observation at the northern edge of the AP, Gonzalez et al. (2018) found that the ASL transports warm and moist air mass and often slight precipitation to the area, whereas the WSL is associated with dry and cold air. Ambrožová et al. (2020) also found warm anomalies at stations in northeastern AP when the ASL synoptic pattern is deep and cold anomalies connected to the WSL and pressure ridge along the AP. It has also been demonstrated using the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis that autumn has a warming trend on the western side of the AP caused by strengthened ASL for the period of 1991–2015 (Bozkurt et al. 2020). On the other hand, the negative annual temperature trends in the AP since the late 1990s (Oliva et al. 2017; Turner et al. 2016) and negative summer temperature trends in Larsen Ice Shelves from 1991 to 2015 (Bozkurt et al. 2020) are found to correspond to a deepening WSL. The causes of temperature and melt interannual variations over the AP and Larsen C for each season have not been fully addressed.

Earlier studies have examined the individual impact of the ASL or WSL; however, there is a lack of understanding regarding how the ASL and the WSL combine to impact the interannual variations of surface temperature and melt in the AP region. The interannual variation of surface melt over Larsen C has been shown to be considerable over the 18-yr period from 1999 to 2017 with melt index values in some years exceeding 3 times that of other years (Bevan et al. 2018). Our paper uses empirical orthogonal function (EOF)/principal component (PC) analysis on the regional sea level pressure (SLP) field to study the spatial and temporal variations of regional atmospheric circulation over the AP and surrounding Amundsen, Bellingshausen, and Weddell Seas. The variations in circulation patterns captured by EOF modes are analyzed to determine their impacts on surface temperature and wind in the AP region including the Larsen C Ice Shelf for each season. Furthermore, the impacts of atmospheric circulation on the interannual variations of surface melt and surface energy budget over the Larsen C Ice Shelf are examined for the summer season. The data and the analysis methods used for this study are described in section 2. Section 3 includes a climatology of SLP, temperature, and winds, a description of the EOF modes, and how they impact surface temperature and melt over Larsen C. Then the summary, discussion, and conclusions are given in section 4.

2. Data and methods

This study uses monthly SLP, 10-m wind, 2-m and 700-hPa temperature, downwelling and net longwave radiation, downwelling shortwave radiation, latent and sensible heat fluxes, total-column cloud liquid and ice and water vapor, and 6-hourly skin temperature from ECMWF reanalysis data (ERA5) at a resolution of 0.25° × 0.25° for the 40-yr period from 1979 to 2018 (Hersbach et al. 2020). Note that all radiation variables are for the surface. Upwelling longwave radiation is not available within ERA5 and is calculated by subtracting net longwave from downwelling longwave radiation. The paper will evaluate clouds using the sum of total-column cloud liquid and ice (a quantity that will be referred to throughout the paper as “total cloud”). The monthly climatology of SLP surrounding the AP area includes two low pressure centers: the ASL to the west of the AP and the WSL to the east of the AP. The seasons are grouped by combining the months with similar patterns in temperature and wind circulation. In particular, the four seasons in this study are defined as March–April for autumn, May–August for winter, September–October for spring, and November–February for summer.

The EOF/PC analysis is performed on the area-weighted SLP field over the 40-yr study period for the domain that spans the longitudes of 180°–0°W and latitudes of 50°–80°S. Area weighting is performed by multiplying by the square root of the cosine of latitude before computing the covariance matrix. This domain will be referred to as domain 1 (Fig. 1). To evaluate the relation between the leading EOF modes and surface climatology over the AP area, regressions of surface temperature and wind are performed onto the PC time series of the EOF modes over the 40-yr study period for the domain that spans the longitudes of 95°–35°W and latitudes of 75°–60°S. This domain will be referred to as domain 2 (Fig. 1). In addition, domain 3 inside of domain 2 represents the Larsen C region used in the analysis of regional circulation impact on the Larsen C surface melt.

Fig. 1.
Fig. 1.

Map of study domains with the outermost boundary representing the area used to perform the EOF/PC analysis (domain 1; same region as in Figs. 2 and 3), the dashed box indicating the AP region for the regression analysis (domain 2; same region as in Figs. 5 and 7), and the inner box showing the region of the Larsen C Ice Shelf (domain 3; same region as in Fig. 6).

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

Note that ERA5 includes 6-hourly products and the 6-hourly skin temperature is used to calculate the melt count and melt day for Larsen C. The 6-hourly melt count within a season is the number of times that 6-hourly skin temperature exceeds or equals 0°C in that season. The number of melt days within a season is calculated by counting the number of days that contain one or more 6-hourly skin temperature above 0°C. The melt days and melt count are used to quantify the summer surface melt climatology over Larsen C. The boundaries of Larsen C are determined with the Bedmap2 ice shelf coverage (Fretwell et al. 2013) and the ERA5 land–sea mask.

Using the method above, the annual melt days (from August to July) are calculated over Larsen C and spatially averaged for each year from 1999 to 2017 to compare with the results obtained using satellite remote sensing methods from Bevan et al. (2018). It is found that the average melt days from 1999–2017 range from 10 to 60 days when skin temperature is used. In particular, the spatial distribution indicates a melt gradient increasing from south to north and with peak values in the western and northeastern regions. When the 2-m air temperature is used, the melt days range from 10 to 100 days. When compared to the values from Bevan et al. (2018), which range from 10 to 100 days, the results using 2-m temperature match closely while the melt days with skin temperature are smaller. The correlation of the time series of melt day from 1999 to 2017 over Larsen C with the melt index from Bevan et al. (2018) is 0.62 when skin temperature is used and 0.69 when 2-m surface temperature is used, both with significance above the 95% level. The scatterplots that correspond to these correlations are included in Fig. S3 in the online supplemental material. Since skin temperature is more reflective of surface conditions, it was used to calculate surface melt day and 6-hourly melt counts for the remainder of this study. In addition, a p value of 0.05 is used as the threshold of statistical significance throughout the paper for regressions, correlations, and t tests.

3. Results

a. AP regional climate and circulation variation

The climatology of the SLP indicates the presence of low pressure centers, called the ASL to the west of the AP and the WSL to the east of the AP, in all seasons (Figs. 2a–d). The location of the ASL center shows slight variations from season to season with the most drastic change in summer. In particular, the ASL center is located at coordinates of about 70°S, 145°W; 75°S, 150°W; 75°S, 140°W; and 70°S, 100°W in autumn, winter, spring, and summer, respectively. The intensity of the ASL center also varies with seasons and have values of about 980, 980, 974, and 984 hPa in autumn, winter, spring, and summer, respectively. It should be noted that the northwesterly wind patterns on the western side of the AP follow a similar seasonal cycle to the ASL, with stronger northwesterly winds in autumn and spring with deeper ASL and weaker winds in winter and summer when the ASL weakens (Figs. 2a–d). This is known as the semiannual oscillation, which results from differences in heat storage between land and ocean (van den Broeke 1998). Also note that the ASL center is located significantly closer to the AP in the summer compared with the other seasons. As a result, the northwesterly winds west of the AP in summer are comparable in strength to those in autumn despite summer having a weaker ASL intensity. The WSL east of the AP also follows a seasonal cycle with stronger intensity in autumn and spring compared to winter and summer. The southerly wind patterns on the eastern side of the AP follow a similar cycle to the WSL, with stronger winds in autumn and spring.

Fig. 2.
Fig. 2.

Climatology of (a)–(d) seasonal sea level pressure (hPa) and 10-m wind vectors (m s−1) and (e)–(h) 2-m surface temperature (°C) for domain 1.

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

The temperature climatology indicates a significant seasonal cycle (Figs. 2e–h). The temperature west of the AP is warmer than the east in all seasons. The difference in temperature between the regions to the east and to the west of the AP follows a seasonal cycle with the largest difference in autumn and the smallest difference in summer. The isotherms indicate the existence of temperature gradients, which also vary according to season and location. Over the west of the AP, the seasonal cycle of temperature gradients indicates that the strongest gradients are in winter and spring; they are slightly weaker in autumn, and the weakest gradients occur in summer. Over the east of the AP, the temperature gradient is strongest in autumn, slightly weaker in winter and spring, and the weakest in summer.

The EOF/PC analysis is used to study the spatial and temporal variations in regional circulation over domain 1. The spatial distributions of the EOF mode 1 in autumn, winter, spring, and summer have anomalies with maximum centers on the western side of the AP at longitudes of about 90°, 110°, 110°, and 115°W, respectively (Figs. 3a–d). These anomalies represent the spatial distributions of variance associated with changes in (i) the depth and location of the ASL, including its center, and (ii) the spatial extent of the ASL. For example, during the positive phase of the EOF mode 1, there is a deeper ASL with larger spatial extent compared to the negative phase of EOF mode 1 as shown in the SLP composites for the positive and negative phases (Fig. S1 in the online supplemental material). Also note that the ASL center is located farther east (west) during the positive (negative) phase in autumn, winter, and spring. The eastern boundary of the anomaly extends past the eastern side of the AP and into the Weddell Sea, and the anomaly spans longitudes of about 140°–40°W in autumn, 160°–45°W in winter and spring, and about 160°–30°W in summer (Figs. 3a–d). This suggests that during the positive phase of EOF mode 1 when the ASL becomes deeper, covers a larger spatial extent, and has a center at a more eastern location, the low pressure system that is part of the WSL extends farther west (Fig. S1). Although variations of depth, spatial extent, and location of the ASL center are the primary features captured by EOF mode 1, the corresponding variations of the WSL are also captured. The total percentage of SLP variance explained by EOF mode 1 is 32%, 47%, 36%, and 56% for autumn, winter, spring, and summer, respectively.

Fig. 3.
Fig. 3.

The spatial pattern of EOF (a)–(d) mode 1 and (e)–(h) mode 2 in each season for domain 1.

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

The major spatial features of EOF mode 2 are anomalies of opposite signs or dipole-like anomalies that generally span longitudes of about 180°–30°W as shown in Figs. 3e–h. The center of the western anomaly is located along the longitudes of 120°, 145°, 110°, and 130°W, and the center of the eastern anomaly along 60°, 70°, 30°, and 50°W for autumn, winter, spring, and summer, respectively. By examining Fig. S2, the western anomaly center captures the changes in the northern boundary or the meridional extent of the ASL as evidenced by the largest change in the northern boundary or meridional extent of the ASL occurring at the same longitudes as the western anomaly centers of EOF mode 2 mentioned above. In particular, during the positive (negative) phase, the meridional extent of the ASL at these longitudinal locations is smaller (larger). For example, the 994-hPa isobar changes position by around 2°–4° in latitude for all seasons and reflect changes in the meridional extent of the ASL.

A notable circulation feature that corresponds to the eastern anomaly center of EOF mode 2 is a high pressure ridge along the AP, which is more prominent in the negative phase (Fig. S2). The longitudinal location of the ridge closely matches the longitudinal location of the eastern anomaly center of EOF mode 2 for all seasons (Fig. S2 and Figs. 3e–h). The change in the intensity of the ridge is also related to other circulation features such as the WSL. In particular, in the positive phase of EOF mode 2 when the ridge is less prominent, the western part of the WSL extends farther west toward the eastern AP, which causes stronger southerly winds near Larsen C Ice Shelf. The opposite is true for the negative phase. The total variance explained by EOF mode 2 is 25%, 23%, 23%, and 14% for autumn, winter, spring, and summer, respectively. Figures 4a–d and 4e–h show the PC time series of EOF modes 1 and 2, respectively. These time series summarize the amplitude variations (with a range around −2 to 2) in circulation patterns associated with each EOF mode. Note that no statistically significant trends exist in the PC time series associated with EOF modes 1 or 2 for any season.

Fig. 4.
Fig. 4.

PC time series of EOF (a)–(d) mode 1 and (e)–(h) mode 2 in each season in solid lines. The normalized surface temperature over Larsen C is included as dashed lines whenever it has a correlation with the PC time series that is significant at the 95% level.

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

b. Impacts of regional circulation on the AP surface wind and temperature

The pattern of surface temperature and wind vector anomalies which relate to EOF modes 1 and 2 described above is represented in Fig. 5 over domain 2, which focuses on the AP area as shown in Fig. 1. In particular, the surface temperature and wind vector anomalies are represented by correlations with the PC time series of each EOF mode with the respective variables at each spatial ERA5 point. As discussed earlier, EOF mode 1 mainly captures variations in the ASL in terms of its depth, location, and spatial extent. The pattern of wind vector anomalies associated with this mode shows northerly/northwesterly anomalies in regions with longitudes of about 80°–40°W in the positive phase, when the ASL is more intense, spatially larger, and located more eastward for all seasons (Figs. 5a–d). The direct impact of these anomalies when combined with temperature gradients is enhanced warm advection anomalies. It should be noted that temperatures in the region are warmer to the north, thus creating a gradient that is favorable for advection in the presence of meridional winds. Figure 5a–d also shows positive surface temperature anomalies over the AP region associated with the EOF mode 1 and the longitudinal span matches that of the northerly/northwesterly wind anomalies (i.e., 80°–40°W). This confirms the contribution of advection to the temperature anomalies in the area. Furthermore, large temperature anomalies with significance level above the 95% level are more spatially pervasive in winter and spring, suggesting the more widespread impact from the variations of ASL captured by the EOF mode 1 in these seasons compared to autumn and summer. In particular, the spatial region in winter and spring with temperature anomalies above the 95% significance level covers the entire AP and the adjacent Larsen C Ice Shelf as well as parts of the Bellingshausen and Weddell Seas. In autumn and summer the positive temperature anomalies with significance level above 95% are mainly over the northern AP.

Fig. 5.
Fig. 5.

Surface temperature and wind vector anomalies associated with EOF (a)–(d) mode 1 and (e)–(h) mode 2 in each season for domain 2. Both temperature and wind anomaly distributions are calculated using the correlation coefficients between PC time series of EOF modes and surface temperature and wind vector of domain 2. The dotted regions indicate correlations between the PC time series of the EOF mode and temperature are significant at the 95% level.

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

Surface winds associated with EOF mode 2 show mainly southerly (including southeasterly and southwesterly) anomalies in the positive phase in regions with longitudes that span about 95°–55°W for all seasons (Figs. 5e–h), which are caused by a more westward-extended WSL and diminished high pressure ridge along the AP during this phase. The corresponding surface temperature anomalies under these southerly winds are negative for all seasons, confirming enhanced cold advection or diminished warm advection. In addition, along the eastern boundary of domain 2, northerly wind anomalies are present for all seasons except spring with warm anomalies in autumn and winter. It is worth noting that the temperature anomalies with significance level above 95% are pervasive over the AP and the adjacent Larsen C Ice Shelf and Bellingshausen and Weddell Seas only in autumn and summer (Figs. 5e,h), implying a more significant impact on the interannual variation of temperature in these regions due to the variations in circulation patterns captured by EOF mode 2. In winter and spring, negative temperature anomalies with significance level above 95% rarely occur east of the AP and are mainly present west of the AP (Figs. 5f,g).

c. Correlation between regional circulation and Larsen C surface temperature and melt

The relationship between the variations in circulation patterns captured by the EOF/PC analysis and the interannual variation of surface temperature over the Larsen C Ice Shelf are examined for all seasons as shown in Table 1. Summer is the only season for which the relationship between regional circulation and surface melt is analyzed because greater than 80% of surface melt occurs during this season using calculations indicated in the methods section. There is a correlation of 0.17, 0.64, 0.65, and 0.11 (−0.38, −0.12, −0.14, and −0.44) between Larsen C surface temperature and the PC time series of EOF mode 1 (mode 2) in autumn, winter, spring, and summer, respectively. Note that EOF mode 1 mainly impacts the northwesterly wind anomalies northwest of the AP and EOF mode 2 impacts the southerly wind anomalies east of the AP near Larsen C, which explains the systematic positive/negative correlations above. Furthermore, the winter and spring correlations for EOF mode 1 with values of 0.64 and 0.65 and the autumn and summer correlations for EOF mode 2 with values of −0.38 and −0.44 all have significance exceeding the 95% level. This is consistent with the spatial significance distribution of temperature anomalies associated with the EOF modes shown in Fig. 5. The time series of the surface temperature over Larsen C is displayed alongside the PC time series of EOF mode whenever there is a statistically significant correlation (Figs. 4b,c,e,h). Temperature variations during winter and spring are positively correlated with the PC time series of EOF mode 1 (Figs. 4b,c) while they negatively correlated with the PC time series of EOF mode 2 during autumn and summer (Figs. 4e,h).

Table 1.

Correlation coefficients between surface temperature of Larsen C and the PC time series of EOF modes 1 and 2 for each season. Correlations with significance level above 95% are in boldface.

Table 1.

As analyzed in section 3a, the anomaly center on the western side of the AP in the EOF mode 1 captures the variations in the depth and location of the ASL including its center and the spatial extent (Figs. 2a–d and Fig. S1). The variation in this circulation pattern regulates the surface temperature over Larsen C during winter and spring. To be specific, when the ASL is deeper, located farther east, and has a larger spatial extent, stronger northwesterly winds prevail on the windward side of the AP and enhanced warm advection over Larsen C. The eastern anomaly center over the AP in the EOF mode 2 captures variations in the magnitude of the high pressure ridge along the AP as well as the western extent of the WSL (Figs. 2e–h and Fig. S2), which significantly influences the interannual variation of Larsen C surface temperature during autumn and summer. In this case, when the WSL is farther west and the intensity of the AP ridge is diminished, there are stronger southerly winds near Larsen C and enhanced cold advection.

Similar to summer surface temperature, summer surface melt over Larsen C is negatively correlated with the PC time series of EOF mode 2 with coefficients of −0.41 for melt counts and −0.37 for melt days, all with significance levels exceeding 95% as shown in Table 2, in which correlations with surface energy budget terms and total cloud that are included will be discussed in section 3d below. The correlations of the summer surface melt indices with EOF mode 1 are not statistically significant. Again, variation in surface winds associated with the eastern anomaly center of summer EOF mode 2 explains the causes of this negative correlation. When stronger southerly winds are present during the positive phase of EOF mode 2, there will be stronger cold-air advection to the Larsen C, which is unfavorable for surface melt.

Table 2.

Correlation coefficients between the PC time series of summer EOF mode 2 and Larsen C summer surface melt counts, melt days, surface downwelling and upwelling longwave radiation (LW), surface downwelling shortwave radiation (SW), and total cloud. Correlations with significance level above 95% are in boldface.

Table 2.

d. Larsen C summer surface melt and energy budget

To better understand the physical processes of how the regional circulation influences the variation of Larsen C surface melt, we analyze the Larsen C surface energy budget associated with the EOF mode 2. The budget terms analyzed include downwelling and upwelling longwave radiation, downwelling shortwave radiation, and latent and sensible heat fluxes, which have mean values of 257.30, 297.33, 275.50, −6.06, and 2.97 W m−2, respectively (positive for downward for the sensible and latent heat fluxes). As discussed above, EOF mode 2 captures variations in the western extent of the WSL and the high pressure ridge along the AP, which influence southerly winds near Larsen C. Correlations between EOF mode 2 and each surface energy budget term mentioned above show that only downwelling and upwelling longwave radiation, with correlations of −0.50 and −0.45, respectively, have a significance above the 95% level (Table 2). Downwelling shortwave and total cloud correlations with EOF mode 2 are 0.25 and −0.23, respectively, as shown in Table 2, and are not statistically significant. Correlations with the latent and sensible heat fluxes are extremely weak with values of less than 0.1 and are not included in Table 2.

Positive and negative composites of downwelling and upwelling longwave radiation, downwelling shortwave radiation along with surface temperature, melt days, and total cloud over Larsen C are calculated by averaging over the years when the PC time series of summer EOF mode 2 is larger than positive 1 standard deviation and smaller than −1 standard deviation, respectively (Fig. 6 and/or Table 3). The spatial distribution of temperature and melt days over Larsen C generally show an overall gradient for both phases with increasing values from south to north (Figs. 6a–d). In particular, the spatial distribution of temperature shows a narrow warm belt along the western edge of Larsen C, which is likely to be caused by the foehn wind warming (Figs. 6a,b). The spatial distribution of melt days has two peaks, one located in the southwest just east of Joerg Peninsula (68°S, 64°W) and another located in the northeast. The location of the southwest peak in surface melt shown in Figs. 6c and 6d matches the location of the peak in surface downwelling longwave radiation shown in Figs. 6e and 6f. This might be connected with the AP topography, shown by the shaded gray in Fig. 6f, which has a minimum immediately northwest of the southwest peak. This topographic minimum can cause more warm air to be advected over the mountain barrier and explains the presence of the southwest peak in surface downwelling longwave radiation. The surface downwelling shortwave radiation shows a gradient with increasing values from west to east and peak values along the eastern edge of Larsen C for both phases (Figs. 6g,h). This is apparently caused by the cloud distribution over Larsen C, which demonstrates a similar spatial pattern (Figs. 6i,j). Peak cloud values adjacent to the mountain ranges of eastern AP indicate the influence of topography lifting on the cloud formation. This topography footprint should explain why EOF mode 2 has relatively weak correlations with the downwelling shortwave radiation and total cloud (Table 2).

Fig. 6.
Fig. 6.

Composites for (a),(b) surface temperature (°C), (c),(d) surface melt (melt days per season), (e),(f) surface downwelling longwave radiation (W m−2), (g),(h) surface downwelling shortwave radiation (W m−2), and (i),(j) total cloud (kg m−2) in summer. Rows show the (top) positive and (bottom) negative phases of EOF mode 2. Topography (m) is included with darker shades indicating higher elevations.

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

Table 3.

Mean Larsen C surface temperature, melt days, downwelling and upwelling surface longwave radiation (LW), downwelling shortwave radiation (SW), and total cloud during the positive and negative phases of summer EOF mode 2 and the difference between them. Note that the positive (negative) phase is defined as the years where the PC time series of EOF mode 2 is above 1 (below −1) standard deviation from the mean. Numbers in boldface indicate that the difference has significance above the 95% level.

Table 3.

The composite differences for surface temperature, melt days, downwelling and upwelling longwave radiation, downwelling shortwave radiation, and total cloud over Larsen C for summer EOF mode 2 are −0.91°C, −12.48 days, −7.07 W m−2, −3.86 W m−2, 5.43 W m−2, and −0.59 × 10−2 kg m−2 respectively as shown in Table 3. Among these differences, surface temperature and melt and downwelling and upwelling longwave radiation have significance above the 95% level. The decline in cloud of −0.59 × 10−2 kg m−2 between the phases suggests some impact of EOF mode 2, which is discussed in more detail later. Considering that the difference in absorbed downwelling shortwave radiation between the positive and negative phases of summer EOF mode 2 is much smaller (~5.43 W m−2 × 20% = 1.09 W m−2) due to large values of ice shelf surface albedo (80% assumed in the calculation above), a larger difference in downwelling longwave radiation (7.07 W m−2) compared to that of upwelling longwave radiation (3.86 W m−2) suggests that additional downwelling longwave radiation might be used for enhanced surface melt (12.48 more melt days) during the negative phase of EOF mode 2, when southerly winds near Larsen C are weaker.

That strong influence of downwelling longwave radiation on surface melt can be also manifested as strong correlations of 0.65 and 0.67 with melt days and counts, respectively, both with significance above the 95% level (Table 4). On the other hand, the impacts of downwelling shortwave radiation and cloud on the surface melt are relatively weak with smaller correlations as shown in Table 4. In addition, air temperature and water vapor are closely related as indicated by a strong and statistically significant correlation of 0.92 between 700-hPa temperature and water vapor over Larsen C. As a result, when stronger southerly winds associated with the positive phase of EOF mode 2 are present near Larsen C, cold and dry air will be advected into the region which also diminishes cloud. Strong correlations of downwelling longwave radiation with both total water vapor (0.82) and total cloud (0.80) over Larsen C further confirm that summer EOF mode 2 exert an influence on the surface melt of Larsen C mainly through altering the downwelling longwave radiation forcing at surface.

Table 4.

Correlation coefficients for Larsen C surface melt days and counts with surface downwelling longwave (LW), shortwave (SW) radiation, and total cloud in summer. All correlations have significance above the 95% level.

Table 4.

Advection of cold dry or warm moist air over the surface can influence near-surface vertical temperature/humidity gradient and therefore the surface sensible/latent heat flux. However, competing effects from winds, cloud, and radiation forcing can complicate the calculation of sensible (latent) heat flux, which has a mean value of 2.97 (−6.06) W m−2 over Larsen C during summer, much smaller than the radiation terms of the surface energy budget. The complicated processes involved within the sensible/latent heat flux also weaken the connection to EOF mode 2; as a result, our analysis did not detect considerable impacts on these budget components from the variations in circulation patterns captured by the summer EOF mode 2.

e. Recent decline of Larsen C summer surface melt

During the 40-yr study period, no significant trends are detected in the surface temperature or melt over Larsen C or in the time series of the EOF modes in summer. However, when using the time series with a 5-yr running mean, a negative trend in surface melt and temperature and positive trend in the index for EOF mode 2 occurs between 2004 and 2016 (Fig. 7a). The positive trend of the PC time series for EOF mode 2 and the negative trend in surface melt and temperature from 2004 to 2018 all have significance above the 95% level. As discussed above, the positively enhanced EOF mode 2 generates stronger southerly winds near Larsen C. It is therefore responsible for its declining surface melt. The spatial distribution of linear trends of surface wind, temperature, and melt for the period of 1979–2003 clearly indicate enhanced northwesterly winds and warming over the entire AP region as well as enhanced melt over Larsen C (Fig. 7b). From 2004 to 2018 there are opposite trends of enhanced southeasterlies and cooling, and declining melt (Fig. 7c). The regions where the temperature trends have significance above the 95% level are dotted and indicate that positive trends near most of the southern AP region and parts of the northern AP region from 1979–2003 have significance above the 95% level (Fig. 7b). Also, negative trends are pervasive in the northern AP region and the Larsen Ice Shelves with significance exceeding the 95% level (Fig. 7c). The negative trends in surface temperature and melt match results from Cape et al. (2015) that show a negative trend in summer station temperature from 2004 to 2010. Bevan et al. (2018) have a negative trend in annual melt between 1999 and 2017. The cooling trend is similar with the results of Turner et al. (2016) and is caused by cold advection associated with enhanced cyclonic circulation in the Weddell Sea.

Fig. 7.
Fig. 7.

(a) Time series of 5-yr running mean of normalized PC time series for EOF mode 2 (solid line), Larsen C surface melt count (dotted line), and temperature (dashed line) for the summers of 1981–2016. Also shown are spatial trends of surface temperature (°C decade−1; shaded), sea level pressure (hPa decade−1; dashed line), and surface wind (m s−1 decade−1; vectors) for the summers of (b) 1979–2003 and (c) 2004–18. The melt count trends over Larsen C are embedded in the top right of (b) and (c) for their respective time periods. Dotted regions indicate the trend of surface temperature or melt is significant above the 95% level.

Citation: Journal of Climate 34, 17; 10.1175/JCLI-D-20-1002.1

4. Summary and discussion

In this study, the spatial and temporal variations of regional atmospheric circulation over the AP and the surrounding Amundsen, Bellingshausen, and Weddell Seas are analyzed in each season using EOF/PC analysis over the 40-yr study period from 1979 to 2018. The regional circulation patterns are persistent throughout the year and mainly include the ASL to the west of the AP, the WSL to the east of the AP, and the high pressure ridge along the AP. The wind patterns associated with these circulation patterns include northwesterly winds which impinge on the western AP coast and are influenced by the ASL and southerly winds along the eastern AP coast and the Larsen ice shelves influenced by the WSL. The ASL and WSL follows a seasonal cycle with stronger lows in autumn and spring and weaker lows in winter and summer, known as the semiannual oscillation (van den Broeke 1998). The major variations of these regional circulation patterns captured by the leading EOF modes include (i) the depth, location, and extent of the ASL captured by EOF mode 1; (ii) the meridional extent of the ASL captured by EOF mode 2; and (iii) the intensity of the pressure ridge along the AP and the western boundary of the WSL captured by EOF mode 2.

The variations in circulation patterns captured by EOF modes 1 and 2 influence the surface temperature over the AP region in all seasons over the 40-yr period. Furthermore, the spatial extent of this influence varies from season to season as demonstrated by statistically robust results. In winter and spring, it is shown that variations in northerly/northwesterly wind anomalies associated with the ASL variations captured by EOF mode 1 influence the surface temperature over the AP and the adjacent regions including the Larsen C Ice Shelf and the Bellingshausen and Weddell Seas. By contrast, in summer and autumn analogous results show that the influence of variations in circulation patterns captured by EOF mode 1 on surface temperature is limited to the northern AP. These results categorize the spatial influence of EOF mode 1 into two groups, (i) winter and spring and (ii) summer and autumn, demonstrating that there is seasonal variability in the spatial extent of the influence on the AP regional temperature. It should also be noted that conditions where northerly/northwesterly wind anomalies exist are favorable for the occurrence of foehn winds but can also be responsible for warm-air advection due to the presence of temperature gradients from the northwest to the southeast of the AP.

A similar grouping has also been found for EOF mode 2. For example, statistically significant results over the 40-yr period indicate that variations in circulation patterns captured by EOF mode 2 influence the temperature west of the AP in winter and spring while in summer and autumn the areas impacted cover both west and east of the AP, including the Larsen Ice Shelves. The influence of circulations associated with EOF mode 2 on surface temperature of Larsen Ice Shelves in summer and autumn is modulated by the variability of southerly winds (cold-air advection) east of the AP. The variations in circulation patterns captured by the EOF modes and their impacts on temperature are consistent with the impacts of some synoptic patterns detected by the SOM and cluster analysis (Ambrožová et al. 2020; Gonzalez et al. 2018). For example, warm and cold anomalies occur over the AP when the ASL is deeper (i.e., positive EOF mode 1) and the WSL is stronger (i.e., positive EOF mode 2), respectively.

Since summer surface melt over Larsen C accounts for over 80% of annual surface melt, an in depth analysis is performed on the relationship between the variations in circulation patterns captured by the EOF modes and Larsen C surface melt for this season. Statistically robust results indicate that EOF mode 2 influences the interannual variation of summer surface melt over the 40-yr period. This is due to variations in the advection of cold air caused by the southerly wind anomalies east of the AP. Statistically robust results indicate that the summer EOF mode 2 has a considerable impact on the interannual variation of surface downwelling longwave radiation over the 40-yr period, which is one of the largest components of the surface energy budget. Since diminished water vapor and reduced cloud accompany the cold-air advection, the influence of these factors when combined with diminished temperature helps to physically explain the correlation between downwelling longwave radiation and summer EOF mode 2. Through its influence on the surface downwelling longwave radiation, the melt energy is altered by the influence of cold-air advection. Considering that the change in absorbed shortwave radiation is relatively small due to high ice shelf albedo and that sensible and latent heat fluxes are exceptionally weak, we can expect the difference in melt energy due to thermal advection to have a similar order of magnitude as the difference of net longwave radiation between the positive and negative phases of EOF mode 2, which is around 3.21 W m−2. This is the same as the difference between downwelling and upwelling longwave radiation (i.e., 7.07–3.86 W m−2 from Table 3). This extra melt energy owing to the variation of southerly winds associated with the summer EOF mode 2 should be responsible for the difference (~12 melt days) of surface melt in Larsen C. Analysis of the 5-yr running-mean time series indicates that positively enhanced EOF mode 2 from 2004 to 2018 is responsible for the cooling and declining surface melt over Larsen C via stronger southerly winds in the region.

In light of the collapse of Larsen A and B, the fate of Larsen C becomes an increasing concern. The major circulation features in the AP region include the ASL and WSL, which exist to the west and east of the AP, respectively, and their integrated impacts on the area surface temperature and melt are thoroughly investigated in this study. Surface melt plays a role in ice shelf collapse (Rack and Rott 2004; van den Broeke 2005), which reduces the buttressing of grounded glaciers, causing enhanced discharge and increasing the rate of global sea level rise (Rott et al. 2002, Rignot et al. 2004). Our findings indicate that the atmospheric circulation patterns associated with variations in the intensity of the pressure ridge along the AP and the western boundary of the WSL account for a considerable amount of surface melt in the austral summer months of November–February mainly through varied downwelling longwave radiation. In addition, variations in downwelling longwave radiation associated with thermal advection contributes to the interannual variability of melt over Larsen C. These results have important implications in the prediction of surface melt and the attribution of surface melt to the atmospheric forcing.

These findings imply that future studies of foehn wind warming over Larsen C should also consider the influence of regional circulation, such as the EOF modes presented in this study. The impact of large-scale circulation on the magnitude and spatial extent of northwesterly winds over the AP area can in turn influence the occurrence and intensity of foehn wind warming on the leeward side of the AP which is a subject for further study. It should be noted that the ASL and WSL which occur on large spatial scales, can influence phenomena on much smaller spatial scales. In particular, the ASL impacts northwesterly winds northwest of the AP which in turn influences foehn winds and advection. Foehn winds and advection can both produce warming over Larsen C. On the other hand, the WSL impacts southerly barrier winds east of the AP, which can cause cold advection that is capable of removing any warming that has accumulated as a result of foehn winds and warm advection. Case studies can be used to study the occurrence of this series of processes in further detail. In terms of the mesoscale feature of foehn wind warming, it should be noted that ERA5 has a resolution of 0.25°, which is capable of capturing the foehn wind warming to a certain degree. However high-resolution data can yield stronger warming, which is reflected in the positive melt trends from 1999 to 2017 in the inlets northwest of Larsen C captured by high-resolution satellite retrievals (Bevan et al. 2018), but not in ERA5. However both datasets capture the overall negative melt trend over most of Larsen C.

Acknowledgments

This study was sponsored by the NSF Grants OPP-1649713 and OPP-1543445. Computational resources were mainly provided by the Research Computing Systems (RCS) at the University of Alaska Fairbanks. The authors acknowledge the two anonymous reviewers whose suggestions helped to improve the manuscript.

Data availability statement

ERA5 data used in this study are available at https://cds.climate.copernicus.eu/#!/home. Bedmap2 data used are available at https://www.bas.ac.uk/project/bedmap-2/. Analysis data presented in this study have been elaborated in section 2.

REFERENCES

  • Ambrožová, K., K. Laska, and J. Kavan, 2020: Multi-year assessment of atmospheric circulation and impacts on air temperature variation on James Ross Island, Antarctic Peninsula. Int. J. Climatol., 40, 15261541, https://doi.org/10.1002/joc.6285.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, R. E., A. F. Banwell, L. D. Trusel, and J. Kingslake, 2018: Antarctic surface hydrology and impacts on ice-sheet mass balance. Nat. Climate Change, 8, 10441052, https://doi.org/10.1038/s41558-018-0326-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bevan, S. L., A. J. Luckman, P. Kuipers Munneke, B. Hubbard, B. Kulessa, and D. W. Ashmore, 2018: Decline in surface melt duration on Larsen C ice shelf revealed by the Advanced Scatterometer (ASCAT). Earth Space Sci., 5, 578591, https://doi.org/10.1029/2018EA000421.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bozkurt, D., D. H. Bromwich, J. Carrasco, K. M. Hines, J. C. Maureira, and R. Rondanelli, 2020: Recent near-surface temperature trends in the Antarctic Peninsula from observed, reanalysis and regional climate model data. Adv. Atmos. Sci., 37, 477493, https://doi.org/10.1007/s00376-020-9183-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bromwich, D. H., J. P. Nicolas, A. J. Monaghan, M. A. Lazzara, L. M. Keller, G. A. Weidner, and A. B. Wilson, 2013: Central West Antarctica among the most rapidly warming regions on Earth. Nat. Geosci., 6, 139145, https://doi.org/10.1038/ngeo1671.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cape, M. R., M. Vernet, P. Skvarca, S. Marinsek, T. Scambos, and E. Domack, 2015: Foehn winds link climate-driven warming to ice shelf evolution in Antarctica. J. Geophys. Res., 120, 11 03711 057, https://doi.org/10.1002/2015JD023465.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Datta, R. T., M. Tedesco, X. Fettweis, C. Agosta, S. Lhermitte, J. T. M. Lenaerts, and N. Wever, 2019: The effect of foehn-induced surface melt on firn evolution over the northeast Antarctic Peninsula. Geophys. Res. Lett., 46, 38223831, https://doi.org/10.1029/2018GL080845.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elvidge, A. D., I. A. Renfrew, J. C. King, A. Orr, T. A. Lachlan-Cope, M. Weeks, and S. L. Gray, 2015: Foehn jets over the Larsen C Ice Shelf, Antarctica. Quart. J. Roy. Meteor. Soc., 141, 698713, https://doi.org/10.1002/qj.2382.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elvidge, A. D., P. Kuipers Munneke, J. C. King, I. A. Renfrew, and E. Gilbert, 2020: Atmospheric drivers of melt on Larsen C ice shelf: Surface energy budget regimes and the impact of foehn. J. Geophys. Res. Atmos., 125, e2020JD032463. https://doi.org/10.1029/2020JD032463.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fretwell, P., and Coauthors, 2013: Bedmap2: Improved ice bed, surface and thickness datasets for Antarctica. Cryosphere, 7, 375393, https://doi.org/10.5194/tc-7-375-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gonzalez, S., F. Vasallo, C. Recio-Blitz, J. A. Guijarro, and J. Riesco, 2018: Atmospheric patterns over the Antarctic Peninsula. J. Climate, 31, 35973608, https://doi.org/10.1175/JCLI-D-17-0598.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grosvenor, D. P., J. C. King, T. W. Choularton, and T. Lachlan-Cope, 2014: Downslope föhn winds over the Antarctic Peninsula and their effect on the Larsen ice shelves. Atmos. Chem. Phys., 14, 94819509, https://doi.org/10.5194/acp-14-9481-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 19992049, https://doi.org/10.1002/qj.3803.

  • Hogg, A. E., and G. H. Gudmundsson, 2017: Impacts of the Larsen-C Ice Shelf calving event. Nat. Climate Change, 7, 540542, https://doi.org/10.1038/nclimate3359.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hosking, J. S., A. Orr, G. J. Marshall, J. Turner, and T. Phillips, 2013: The influence of the Amundsen-Bellingshausen Seas low on the climate of West Antarctica and its representation in coupled climate model simulations. J. Climate, 26, 66336648, https://doi.org/10.1175/JCLI-D-12-00813.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hosking, J. S., A. Orr, T. J. Bracegirdle, and J. Turner, 2016: Future circulation changes off West Antarctica: Sensitivity of the Amundsen Sea low to projected anthropogenic forcing. Geophys. Res. Lett., 43, 367376, https://doi.org/10.1002/2015GL067143.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jones, M. E., D. H. Bromvvich, J. P. Nicolas, J. Carrasco, E. Plavcova, X. Zou, and S. H. Wang, 2019: Sixty years of widespread warming in the southern middle and high latitudes (1957–2016). J. Climate, 32, 68756898, https://doi.org/10.1175/JCLI-D-18-0565.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kuipers Munneke, P., and Coauthors, 2018: Intense winter surface melt on an Antarctic ice shelf. Geophys. Res. Lett., 45, 76157623, https://doi.org/10.1029/2018GL077899.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Laffin, M. K., C. S. Zender, S. Singh, J. M. Van Wessem, C. Smeets, and C. H. Reijmer, 2021: Climatology and evolution of the Antarctic Peninsula föhn wind-induced melt regime from 1979–2018. J. Geophys. Res., 126, 19, https://doi.org/10.1029/2020JD033682.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lai, C. Y., and Coauthors, 2020: Vulnerability of Antarctica’s ice shelves to meltwater-driven fracture. Nature, 584, 574578, https://doi.org/10.1038/s41586-020-2627-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Luckman, A., A. Elvidge, D. Jansen, B. Kulessa, P. Kuipers Munneke, J. King, and N. E. Barrand, 2014: Surface melt and ponding on Larsen C Ice Shelf and the impact of föhn winds. Antarct. Sci., 26, 625635, https://doi.org/10.1017/S0954102014000339.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Oliva, M., and Coauthors, 2017: Recent regional climate cooling on the Antarctic Peninsula and associated impacts on the cryosphere. Sci. Total Environ., 580, 210223, https://doi.org/10.1016/j.scitotenv.2016.12.030.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rack, W., and H. Rott, 2004: Pattern of retreat and disintegration of the Larsen B ice shelf, Antarctic Peninsula. Ann. Glaciology, 39, 505510, https://doi.org/10.3189/172756404781814005.

    • Search Google Scholar
    • Export Citation
  • Raphael, M. N., and Coauthors, 2016: The Amundsen sea low variability, change, and impact on Antarctic climate. Bull. Amer. Meteor. Soc., 97, 111121, https://doi.org/10.1175/BAMS-D-14-00018.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rignot, E., G. Casassa, P. Gogineni, W. Krabill, A. Rivera, and R. Thomas, 2004: Accelerated ice discharge from the Antarctic Peninsula following the collapse of Larsen B ice shelf. Geophys. Res. Lett., 31, L18401, https://doi.org/10.1029/2004GL020697.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rott, H., W. Rack, P. Skvarca, and H. De Angelis, 2002: Northern Larsen Ice Shelf, Antarctica: Further retreat after collapse. Ann. Glaciol., 34, 277282, https://doi.org/10.3189/172756402781817716..

    • Search Google Scholar
    • Export Citation
  • Royston, S., and G. H. Gudmundsson, 2016: Changes in ice-shelf buttressing following the collapse of Larsen A Ice Shelf, Antarctica, and the resulting impact on tributaries. J. Glaciol., 62, 905911, https://doi.org/10.1017/jog.2016.77.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tuckett, P. A., J. C. Ely, A. J. Sole, S. J. Livingstone, B. J. Davison, J. M. van Wessem, and J. Howard, 2019: Rapid accelerations of Antarctic Peninsula outlet glaciers driven by surface melt. Nat. Commun., 10, 10, https://doi.org/10.1038/s41467-019-12039-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Turner, J., and Coauthors, 2016: Absence of 21st century warming on Antarctic Peninsula consistent with natural variability. Nature, 535, 411415, https://doi.org/10.1038/nature18645.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Turner, J., G. J. Marshall, K. Clem, S. Colwell, T. Phillips, and H. Lu, 2020: Antarctic temperature variability and change from station data. Int. J. Climatol., 40, 29863007, https://doi.org/10.1002/joc.6378.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van den Broeke, M., 1998: The semi-annual oscillation and Antarctic climate. Part 1: Influence on near surface temperatures (1957–79). Antarct. Sci., 10, 175183, https://doi.org/10.1017/S0954102098000248.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van den Broeke, M., 2005: Strong surface melting preceded collapse of Antarctic Peninsula ice shelf. Geophys. Res. Lett., 32, 32, https://doi.org/10.1029/2005GL023247.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wiesenekker, J. M., P. Kuipers Munneke, M. R. van den Broeke, and C. Smeets, 2018: A multidecadal analysis of föhn winds over Larsen C Ice Shelf from a combination of observations and modeling. Atmosphere, 9, 9, https://doi.org/10.3390/atmos9050172.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, C. R., and J. Zhang, 2018: Modeling study of foehn wind events in Antarctic Peninsula with WRF forced by CCSM. J. Meteor. Res., 32, 909922, https://doi.org/10.1007/s13351-018-8067-9.

    • Crossref
    • Search Google Scholar
    • Export Citation

Supplementary Materials

Save
  • Ambrožová, K., K. Laska, and J. Kavan, 2020: Multi-year assessment of atmospheric circulation and impacts on air temperature variation on James Ross Island, Antarctic Peninsula. Int. J. Climatol., 40, 15261541, https://doi.org/10.1002/joc.6285.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, R. E., A. F. Banwell, L. D. Trusel, and J. Kingslake, 2018: Antarctic surface hydrology and impacts on ice-sheet mass balance. Nat. Climate Change, 8, 10441052, https://doi.org/10.1038/s41558-018-0326-3.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bevan, S. L., A. J. Luckman, P. Kuipers Munneke, B. Hubbard, B. Kulessa, and D. W. Ashmore, 2018: Decline in surface melt duration on Larsen C ice shelf revealed by the Advanced Scatterometer (ASCAT). Earth Space Sci., 5, 578591, https://doi.org/10.1029/2018EA000421.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bozkurt, D., D. H. Bromwich, J. Carrasco, K. M. Hines, J. C. Maureira, and R. Rondanelli, 2020: Recent near-surface temperature trends in the Antarctic Peninsula from observed, reanalysis and regional climate model data. Adv. Atmos. Sci., 37, 477493, https://doi.org/10.1007/s00376-020-9183-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bromwich, D. H., J. P. Nicolas, A. J. Monaghan, M. A. Lazzara, L. M. Keller, G. A. Weidner, and A. B. Wilson, 2013: Central West Antarctica among the most rapidly warming regions on Earth. Nat. Geosci., 6, 139145, https://doi.org/10.1038/ngeo1671.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cape, M. R., M. Vernet, P. Skvarca, S. Marinsek, T. Scambos, and E. Domack, 2015: Foehn winds link climate-driven warming to ice shelf evolution in Antarctica. J. Geophys. Res., 120, 11 03711 057, https://doi.org/10.1002/2015JD023465.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Datta, R. T., M. Tedesco, X. Fettweis, C. Agosta, S. Lhermitte, J. T. M. Lenaerts, and N. Wever, 2019: The effect of foehn-induced surface melt on firn evolution over the northeast Antarctic Peninsula. Geophys. Res. Lett., 46, 38223831, https://doi.org/10.1029/2018GL080845.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elvidge, A. D., I. A. Renfrew, J. C. King, A. Orr, T. A. Lachlan-Cope, M. Weeks, and S. L. Gray, 2015: Foehn jets over the Larsen C Ice Shelf, Antarctica. Quart. J. Roy. Meteor. Soc., 141, 698713, https://doi.org/10.1002/qj.2382.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Elvidge, A. D., P. Kuipers Munneke, J. C. King, I. A. Renfrew, and E. Gilbert, 2020: Atmospheric drivers of melt on Larsen C ice shelf: Surface energy budget regimes and the impact of foehn. J. Geophys. Res. Atmos., 125, e2020JD032463. https://doi.org/10.1029/2020JD032463.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fretwell, P., and Coauthors, 2013: Bedmap2: Improved ice bed, surface and thickness datasets for Antarctica. Cryosphere, 7, 375393, https://doi.org/10.5194/tc-7-375-2013.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gonzalez, S., F. Vasallo, C. Recio-Blitz, J. A. Guijarro, and J. Riesco, 2018: Atmospheric patterns over the Antarctic Peninsula. J. Climate, 31, 35973608, https://doi.org/10.1175/JCLI-D-17-0598.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Grosvenor, D. P., J. C. King, T. W. Choularton, and T. Lachlan-Cope, 2014: Downslope föhn winds over the Antarctic Peninsula and their effect on the Larsen ice shelves. Atmos. Chem. Phys., 14, 94819509, https://doi.org/10.5194/acp-14-9481-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 19992049, https://doi.org/10.1002/qj.3803.

  • Hogg, A. E., and G. H. Gudmundsson, 2017: Impacts of the Larsen-C Ice Shelf calving event. Nat. Climate Change, 7, 540542, https://doi.org/10.1038/nclimate3359.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hosking, J. S., A. Orr, G. J. Marshall, J. Turner, and T. Phillips, 2013: The influence of the Amundsen-Bellingshausen Seas low on the climate of West Antarctica and its representation in coupled climate model simulations. J. Climate, 26, 66336648, https://doi.org/10.1175/JCLI-D-12-00813.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hosking, J. S., A. Orr, T. J. Bracegirdle, and J. Turner, 2016: Future circulation changes off West Antarctica: Sensitivity of the Amundsen Sea low to projected anthropogenic forcing. Geophys. Res. Lett., 43, 367376, https://doi.org/10.1002/2015GL067143.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jones, M. E., D. H. Bromvvich, J. P. Nicolas, J. Carrasco, E. Plavcova, X. Zou, and S. H. Wang, 2019: Sixty years of widespread warming in the southern middle and high latitudes (1957–2016). J. Climate, 32, 68756898, https://doi.org/10.1175/JCLI-D-18-0565.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kuipers Munneke, P., and Coauthors, 2018: Intense winter surface melt on an Antarctic ice shelf. Geophys. Res. Lett., 45, 76157623, https://doi.org/10.1029/2018GL077899.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Laffin, M. K., C. S. Zender, S. Singh, J. M. Van Wessem, C. Smeets, and C. H. Reijmer, 2021: Climatology and evolution of the Antarctic Peninsula föhn wind-induced melt regime from 1979–2018. J. Geophys. Res., 126, 19, https://doi.org/10.1029/2020JD033682.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lai, C. Y., and Coauthors, 2020: Vulnerability of Antarctica’s ice shelves to meltwater-driven fracture. Nature, 584, 574578, https://doi.org/10.1038/s41586-020-2627-8.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Luckman, A., A. Elvidge, D. Jansen, B. Kulessa, P. Kuipers Munneke, J. King, and N. E. Barrand, 2014: Surface melt and ponding on Larsen C Ice Shelf and the impact of föhn winds. Antarct. Sci., 26, 625635, https://doi.org/10.1017/S0954102014000339.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Oliva, M., and Coauthors, 2017: Recent regional climate cooling on the Antarctic Peninsula and associated impacts on the cryosphere. Sci. Total Environ., 580, 210223, https://doi.org/10.1016/j.scitotenv.2016.12.030.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rack, W., and H. Rott, 2004: Pattern of retreat and disintegration of the Larsen B ice shelf, Antarctic Peninsula. Ann. Glaciology, 39, 505510, https://doi.org/10.3189/172756404781814005.

    • Search Google Scholar
    • Export Citation
  • Raphael, M. N., and Coauthors, 2016: The Amundsen sea low variability, change, and impact on Antarctic climate. Bull. Amer. Meteor. Soc., 97, 111121, https://doi.org/10.1175/BAMS-D-14-00018.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rignot, E., G. Casassa, P. Gogineni, W. Krabill, A. Rivera, and R. Thomas, 2004: Accelerated ice discharge from the Antarctic Peninsula following the collapse of Larsen B ice shelf. Geophys. Res. Lett., 31, L18401, https://doi.org/10.1029/2004GL020697.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rott, H., W. Rack, P. Skvarca, and H. De Angelis, 2002: Northern Larsen Ice Shelf, Antarctica: Further retreat after collapse. Ann. Glaciol., 34, 277282, https://doi.org/10.3189/172756402781817716..

    • Search Google Scholar
    • Export Citation
  • Royston, S., and G. H. Gudmundsson, 2016: Changes in ice-shelf buttressing following the collapse of Larsen A Ice Shelf, Antarctica, and the resulting impact on tributaries. J. Glaciol., 62, 905911, https://doi.org/10.1017/jog.2016.77.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tuckett, P. A., J. C. Ely, A. J. Sole, S. J. Livingstone, B. J. Davison, J. M. van Wessem, and J. Howard, 2019: Rapid accelerations of Antarctic Peninsula outlet glaciers driven by surface melt. Nat. Commun., 10, 10, https://doi.org/10.1038/s41467-019-12039-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Turner, J., and Coauthors, 2016: Absence of 21st century warming on Antarctic Peninsula consistent with natural variability. Nature, 535, 411415, https://doi.org/10.1038/nature18645.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Turner, J., G. J. Marshall, K. Clem, S. Colwell, T. Phillips, and H. Lu, 2020: Antarctic temperature variability and change from station data. Int. J. Climatol., 40, 29863007, https://doi.org/10.1002/joc.6378.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van den Broeke, M., 1998: The semi-annual oscillation and Antarctic climate. Part 1: Influence on near surface temperatures (1957–79). Antarct. Sci., 10, 175183, https://doi.org/10.1017/S0954102098000248.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van den Broeke, M., 2005: Strong surface melting preceded collapse of Antarctic Peninsula ice shelf. Geophys. Res. Lett., 32, 32, https://doi.org/10.1029/2005GL023247.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wiesenekker, J. M., P. Kuipers Munneke, M. R. van den Broeke, and C. Smeets, 2018: A multidecadal analysis of föhn winds over Larsen C Ice Shelf from a combination of observations and modeling. Atmosphere, 9, 9, https://doi.org/10.3390/atmos9050172.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhang, C. R., and J. Zhang, 2018: Modeling study of foehn wind events in Antarctic Peninsula with WRF forced by CCSM. J. Meteor. Res., 32, 909922, https://doi.org/10.1007/s13351-018-8067-9.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Map of study domains with the outermost boundary representing the area used to perform the EOF/PC analysis (domain 1; same region as in Figs. 2 and 3), the dashed box indicating the AP region for the regression analysis (domain 2; same region as in Figs. 5 and 7), and the inner box showing the region of the Larsen C Ice Shelf (domain 3; same region as in Fig. 6).

  • Fig. 2.

    Climatology of (a)–(d) seasonal sea level pressure (hPa) and 10-m wind vectors (m s−1) and (e)–(h) 2-m surface temperature (°C) for domain 1.

  • Fig. 3.

    The spatial pattern of EOF (a)–(d) mode 1 and (e)–(h) mode 2 in each season for domain 1.

  • Fig. 4.

    PC time series of EOF (a)–(d) mode 1 and (e)–(h) mode 2 in each season in solid lines. The normalized surface temperature over Larsen C is included as dashed lines whenever it has a correlation with the PC time series that is significant at the 95% level.

  • Fig. 5.

    Surface temperature and wind vector anomalies associated with EOF (a)–(d) mode 1 and (e)–(h) mode 2 in each season for domain 2. Both temperature and wind anomaly distributions are calculated using the correlation coefficients between PC time series of EOF modes and surface temperature and wind vector of domain 2. The dotted regions indicate correlations between the PC time series of the EOF mode and temperature are significant at the 95% level.

  • Fig. 6.

    Composites for (a),(b) surface temperature (°C), (c),(d) surface melt (melt days per season), (e),(f) surface downwelling longwave radiation (W m−2), (g),(h) surface downwelling shortwave radiation (W m−2), and (i),(j) total cloud (kg m−2) in summer. Rows show the (top) positive and (bottom) negative phases of EOF mode 2. Topography (m) is included with darker shades indicating higher elevations.

  • Fig. 7.

    (a) Time series of 5-yr running mean of normalized PC time series for EOF mode 2 (solid line), Larsen C surface melt count (dotted line), and temperature (dashed line) for the summers of 1981–2016. Also shown are spatial trends of surface temperature (°C decade−1; shaded), sea level pressure (hPa decade−1; dashed line), and surface wind (m s−1 decade−1; vectors) for the summers of (b) 1979–2003 and (c) 2004–18. The melt count trends over Larsen C are embedded in the top right of (b) and (c) for their respective time periods. Dotted regions indicate the trend of surface temperature or melt is significant above the 95% level.

All Time Past Year Past 30 Days
Abstract Views 391 0 0
Full Text Views 344 128 6
PDF Downloads 334 82 8