Atmospheric precursors to the Antarctic sea ice record low in February 2022

Antarctic sea ice expansion and recession are asymmetric in nature, with regional and temporal variations. The decade-long overall increase in the Antarctic sea ice extent (SIE) until 2015 showed a decrease in recent years since satellite records were available. The present study focused on determining the atmospheric forcing and climate fluctuations responsible for the lowest SIE record in February 2022. Here, the lowest SIE record was assumed to result from the sea ice recession that began in September 2021. The SIE reached a record low of 2.16 × 106 km2 in February 2022, which was 43% lower than the mean extent of the previous February months since the satellite era. However, the second-lowest SIE was recorded from November 2021 to January 2022. The Weddell Sea, Ross Sea, and Bellingshausen/Amundsen Seas (ABS) sectors experienced the maximum sea ice change on a regional scale. The record-low SIE occurred when the Amundsen Sea Low (ASL) pressure center was intensified, with the Southern Annular Mode (SAM) at its positive phase. Together, these two climate fluctuations played a role in modifying the pressure and wind patterns in Antarctica. The warm northerly winds largely contributed to decreased SIE. Further, the study investigated the Polar Cap Height (PCH), which demonstrates a strengthening of the stratospheric polar vortex and positive polarity of the SAM.


Introduction
Polar sea ice fluctuation is an essential component of the complex global climate system, as it can reflect and affect other climatic components of the system (Eayrs et al 2019(Eayrs et al , 2021. The advent of satellite remote sensing in the late 1970s enabled the continuous observation of sea ice distribution in remote and harsh polar environments with higher spatiotemporal coverage . Since 1979, the Antarctic sea ice extent (SIE) has shown a weak but significant positive trend until 2015, when it unexpectedly declined by ∼50%, with no discernible trend in subsequent years (Kusahara et al 2018, Parkinson 2019. The Antarctic annual mean SIE decreased by 2.03 × 10 6 km 2 between 2015-2017 (Simmonds and Li 2021). However, Antarctic sea ice variability is regional and seasonal in nature (Zwally et al 2002, Meehl et al 2019, as sea ice cover changes rapidly, and approximately 15 million km 2 of ice forms and melts each year (Eayrs et al 2019). Since the satellite era, the overall decade-long weak positive Antarctic SIE trend has been explained by several mechanisms Overland 2009, Holland et al 2017). These mechanisms include freshening and cooling of the Southern Ocean surface due to decreased upwelling of warmer water (Bintanja et al 2013, Armour et al 2016, Purich et al 2018, strengthening of the westerly wind belt due to the Southern Annular Mode (SAM) (Pezza et al 2012, Comiso et al 2017, Doddridge and Marshall 2017, and changes in the Amundsen Sea Low (ASL) pressure center in Antarctica, resulting in variations in sea ice concentration (SIC) and transport anomalies (Turner et al 2015, Holland et al 2018. The ASL is a low-pressure system in the Southern Ocean that traverses across the Ross Sea and Bellingshausen-Amundsen Sea sectors, resulting in warmer conditions and increased northerly wind anomalies (Donat-Magnin et al 2020). The Antarctic SIE is influenced by atmospheric circulation through the process of sea ice drift and atmospheric heat transport, which are induced by climatic modes. In addition, the surface climatic variations are linked to the stratosphere-troposphere coupling processes. The earlier studies demonstrated a linkage between the stratospheric polar vortex and the low springtime SIE record in 2016 (Wang et al 2019. Since the availability of satellite records, the yearly average of the total Antarctic SIE had a maximum extent of 12.8 × 10 6 km 2 in 2014 and a minimum of 10.7 × 10 6 km 2 in 2017 (Parkinson 2019). Over the first three decades, the Antarctic SIE showed a gradual increase, followed by a sharp decrease from 2015 onward (Cerrone et al 2017, Eayrs et al 2021. The Antarctic SIE trend from 1979-2015 was significantly positive from September to February at 95% and 99% significance levels (table S1). The trend analysis shows a positive trend from September to February in the Indian Ocean sector, western Pacific Ocean sector, and Ross sea, while a negative trend persisted in the ABS and the Weddell Sea sector. Since 2016, the maximum SIE loss has been recorded from September to February (figure 1). In addition, the second-lowest monthly SIE was recorded in 2017 (2.29 × 10 6 km 2 ) after the maximum sea ice recession in the spring of 2016 , Wang et al 2019. A negative trend was observed from December to February for 2015-2021/22 (figures 1(c), (d)). The annual cycle of monthly Antarctic SIE always shows an increase in September and a decrease in February (Parkinson 2019). During February 2022, Antarctic sea ice recorded an unprecedented record low SIE (2.16 × 10 6 km 2 ) (figures 2(a)-(c)). Notably, the difference between the maximum February Antarctic SIE of 2008 (3.9 × 10 6 km 2 ) and the minimum Antarctic SIE of 2022 was only 1.74 × 10 6 km 2 (Turner et al 2022).
Similarly, from October 2021 to January 2022, the Antarctic SIE was close to the minimum sea ice record since 1979 (figures 2(b), (c), and 3). As shown in figure 2(c), the SIE anomaly for January and February 2022 is very close to the lowest record set in 2017 (figure 2(c)). Based on the monthly SIE trend from 1979-2021/22, November has the highest negative trend, followed by January ( figure 2(d)). In addition, the austral summer SIE, the highest minimum, was recorded in 2017 (4.7 × 10 6 km 2 ), followed by the second lowest record in 2022 (5 × 10 6 km 2 ) (figure 3(a)). The recent low sea ice records in the Antarctic may indicate the beginning of a protracted sea ice recession in the coming years, as observed in the Arctic, or it may reverse after these short-term variations (Fogt et al 2022, Raphael and. Here, we investigated the temporal and spatial evolution of the low Antarctic sea ice record in February 2022 since the satellite record began (figure 1(f)). This study examines the monthly (September to February) and seasonal (spring and summer) sea ice variability in response to climate changes and polar vortex variability. We studied atmospheric variables to determine their fidelity to the sea ice. Furthermore, we discussed the prevailing atmospheric conditions from September 2021 that may have contributed to the new record low SIE of February 2022.

Materials and method
This study analyzed sea ice data derived from multichannel passive microwave satellites since the late 1970s. We examined daily and monthly sea ice data from the United States National Snow and Ice Data Centre (https:// nsidc.org/) generated using the NASA Team algorithm (Cavalieri et al 1996, Cavalieri and. Once the data from each sensor were received, they were mapped onto rectangular grids superimposed on polar stereographic projections with grid squares (or pixels) of approximately 25 km × 25 km (NSIDC 2004). Because sea ice data were available only on alternate days prior to 1987, we interpolated the missing days using data from the preceding and following days using the nearest neighbor interpolation technique. Since no daily data were available from 3 December 1987 to 12 January 1988, interpolation was not performed for these missing days. NASA's near-real-time data were used for the years 2021-2022. SIC is the percentage of the area occupied by ice compared to the reference area. Furthermore, SIE was determined as the sum of the areas of the grid squares with at least 15% ice concentration (Parkinson et al 1999, Zwally et al 2002, Serreze et al 2003. We evaluated the SIE variability for the entire Antarctic and its five sectors (Cavalieri et al 2003, Parkinson 2019. Antarctica is divided into five sectors: the Indian Ocean, Western Pacific Ocean, Ross Sea, Amundsen-Bellingshausen Sea (ABS), and Weddell Sea (figure 2(a)).
We investigated the sensitivity of ocean-atmospheric forcing on sea ice using data from the Fifth Generation European Centre for Medium-range Weather Forecasts (ECMWF) reanalysis data (ERA5; https://www.ecmwf. int/). ERA5 provides higher spatiotemporal resolution data for a large number of oceanic, atmospheric, and land climate variables and is the most reliable dataset for understanding recent Antarctic climate fluctuations (Tetzner et al 2019, Hersbach et al 2020. The present study used real-time monthly data updated monthly after quality verification. The datasets obtained at the surface level from ERA5 for the period 1979-2022 include SIC, mean sea level pressure (MSLP), and 10 m zonal and meridional wind components with a spatial resolution of 0.25°× 0.25°. Similarly, the ERA5 pressure level data were retrieved for Geopotential height (GPH) and temperature from the equator to the 90°S latitude. The GPH at 50 hPa and 10 hPa was used to determine the Polar Cap Height (PCH), which is averaged over the south of 65°S (Siegmund 2005). The PCH anomaly at each pressure level was normalized by its standard deviation and is unitless.
Two dominant cycles, SAM and ASL, which regulate climatic fluctuations and sea ice in the Southern Hemisphere, were studied. Both indices were obtained from https://climatedataguide.ucar.edu/ for the period 1979-2022. The station-based SAM index is the zonal pressure difference between the Southern Hemisphere latitudes (40°and 65°) (Marshall 2003). However, the ASL is a climatological low-pressure center located towards the coastal flank of western Antarctica and over the southern Pacific Ocean. This study used the Actual Central Pressure (ACP) of the ASL index, defined as the pressure at the ASL location (Bromwich et al 2012. Hosking et al (2013) found a significant correlation (p < 0.01) between ACP and SAM across all seasons. The ACP is strongly associated with and influenced by large-scale variability, including both SAM and ENSO. Therefore, we have also analyzed the Relative Central Pressure (RCP), which provides a more precise influence of ASL at a regional scale (Simmonds and Wu 1993). In the present study, anomaly calculations were performed with reference to the 42-year period (1979-2020) for all of the aforementioned ocean-atmospheric variables.
From September to February, the monthly SIE trend for the total Southern Ocean remained positive over the 42-year (1979-2020) period, except in November, when the trend was near zero. However, the monthly trend for the 43/44-year (1979-2021/22) period was consistently lower than the 42-year SIE trend. The SIE trend (1979 was 70% lower in December 2021 than the trend for the period . In most parts of Antarctica, January marks the beginning of the summer sea ice retreat. The second-lowest extent (3.86 × 10 6 km 2 ) was recorded in January 2022, with a difference of 0.08 × 10 6 km 2 from the previous lowest record in 2017 ( figure 2(b)).
The total Antarctic SIE reached its fifth-lowest (17 × 10 6 km 2 ) record in spring 2021, a difference of 0.4 × 10 6 km 2 from the first lowest record in 2016 (16.6 × 10 6 km 2 ) (figure 3(a)). However, the average summer SIE for 2021/22 was the third-lowest (6.1 × 10 6 km 2 ) and was very close to the lowest values recorded in the year 1998. Overall, the summer 2022 SIE was 18.6% lower than the mean SIE over the period 1979 to 2020. Seasonally, the Antarctic SIE trend was positive in both spring (4.5 ± 5.0 × 10 3 km 2 y −1 ) and summer (3.7 ± 9.1 × 10 3 km 2 y −1 ) over the period 1979-2021/22. In spring 2021, the rate of change in the Antarctic SIE trend was 30% lower than the 42-year average, and in the summer of 2022, it was 54% lower. Sea ice advance and retreat occur on a distinct timescale and at different rates in each of Antarctica's five sectors. The ABS sector experienced maximum sea ice loss in September, with a 60% (0.8 ± 3.4 × 10 3 km 2 y −1 ) reduction compared to the 42-year trend (1.3 ± 3.6 × 10 3 km 2 y −1 ). Sea ice loss in this sector continues until December, with maximum loss in November (table 1). The Weddell Sea and Western Pacific Ocean sectors exhibited a negative trend from September to February. In addition, the Weddell Sea had the highest monthly SIE change compared to other sectors, while the Indian Ocean sector had the lowest monthly SIE change. In January and February 2022, when the Ross Sea and the Weddell Sea experienced the highest sea ice loss, the ABS sector recorded a positive SIE trend. Seasonally, during spring, the ABS and Western Pacific Ocean sectors registered maximum sea ice changes of −90% and −76%, respectively, relative to the 42-year trend. However, in summer, the maximum sea ice retreat was recorded in the Ross Sea (−83%) and Weddell Sea (−32%), while growth occurred in the ABS and the Indian Ocean sectors (table 1).
Over the period 1979-2015, the total Antarctic minimum monthly SIE in February was consistently between 3 × 10 6 km 2 and 4 × 10 6 km 2 , with the exception of 1993 and 1997 (2.4 × 10 6 km 2 ) ( figure 1(f)). However, the SIE was well below 3 × 10 6 km 2 after 2015. Recently, the total Antarctic SIE reached a new record minimum of 2.16 × 10 6 km 2 in February 2022, with a record low of 1.9 × 10 6 km 2 on 25 February (figures 1(f) and 3(b)). It is worth mentioning that the sea ice retreat began in 2021, with the largest mean total SIE change occurring in December (−7.12 × 10 6 km 2 ) (table S2). During 1979-2020, the SIE change was consistently negative, with the exception of September, when the SIE increased by 0.26 × 10 6 km 2 . In contrast, in September 2021 (−0.59 × 10 6 km 2 ), sea ice loss was three times greater than the 42-year mean loss. The monthly difference shows that February 2022 had the largest change of 43% compared to the previous year.
Compared to other sectors, the Weddell Sea sector experienced the greatest sea ice loss in September and October and growth from January to February 2022 (table S2). This sector exhibits positive seasonal trends in summer and autumn but negative seasonal trends in winter and Spring (Kumar et al 2021). In October, a positive yearly rate of change was recorded only in the Ross Sea (0.49 × 10 6 km 2 ). However, the magnitude at which sea ice increased in the Ross Sea experienced a rapid loss in the following months compared to the previous year's mean. The western Pacific Ocean showed a negative rate of SIE change from October to February, with a maximum difference in October. The Weddell Sea experienced the highest rate of SIE change during spring (−2.18 × 10 6 km 2 ), followed by the ABS (−0.86 × 10 6 km 2 ) and the Ross Sea sectors (−0.83 × 10 6 km 2 ). In contrast, the Weddell Sea and ABS both recorded positive changes in summer relative to the previous year's mean loss. The Ross Sea sector recorded a maximum sea ice retreat of −3.21 × 10 6 km 2 in summer (table S2).

Sea ice loss at a record low is linked to physical forcings 3.2.1. Monthly variability
The lowest total Antarctic SIE in February was determined by comparing the monthly change in SIC, meridional and zonal wind components, and MSLP from September 2021 to February 2022 to the long-term mean of 1979-2020 ( figure 5). Similarly, to understand the monthly variability in the GPH and temperature at pressure level from the equator to 90°S has been shown in figure 6. The regions with maximum negative SIC anomalies include the Weddell Sea, the western Pacific Ocean, and the Ross Sea. In contrast, positive anomalies were observed over the Indian Ocean and western ABS sector from September to November 2021 (figures 3(a), (b), and 7(a)). The MSLP anomaly is negative across the circumpolar trough with a low-pressure center in the ABS sector ( figure 7(b)). The driver of anomalous low pressure is the Amundsen Sea Low (ASL), the deepest of the three climatological low pressures formed around Antarctica (Holland et al 2018). It is suggested that maximum inter-annual pressure variability occurs within the ABS sector (Fogt et al 2012). Further, the low-pressure center allows the intrusion of warm and moist air over the ABS sector by developing the atmospheric rivers (Djoumna and Holland 2021). The strength and position of the wind moving across the Southern Ocean are largely modified by the ASL (Turner et al 2013). The ACP and RCP indices are used here to determine the ASL index. The RCP is the difference between the central pressure and the climatological pressure of the cyclone location at a particular time of year (Lim and Simmonds 2002). The large-scale variability can be understood from the ACP, whereas the RCP provides accurate measurements of the regional pressure difference in West Antarctica (Hosking et al 2013, Raphael et al 2019). In this study, a monthly analysis of the RCP shows the maximum pressure difference in October (figure 3(c)). Similarly, from September to November 2021, the ACP recorded the lowest minimum pressure values in October (959 hPa) ( figure 3(d)).
The ASL central pressure variations determine the temperature, surface wind patterns, downward longwave radiation, and cloud cover in the Western Antarctic (Yeung et al 2019). The downward longwave radiation in Antarctica regulates the sea ice variation in relation to the air and skin temperature. Since more downward longwave radiation causes higher skin temperatures and less sea ice, less radiation causes lower skin temperatures and more ice (Lee et al 2017, Sato and Simmonds 2021, Zhang et al 2021. ASL is also associated with the dominant mode of climatic fluctuations, that is, SAM. According to previous studies, SAM has been more positive over the past 30-year (Arblaster and Meehl 2006, Abram et al 2014, Simmonds 2015. Similarly, in the recent year 2021-2022, SAM was found to be in a positive phase, suggesting that it may intensify the westerlies and deepen the ASL. The positive and negative SAM perturbations correspond to the contraction of the westerlies towards the pole and equator, respectively Wallace 2000, Stuecker et al 2017). The phase change of the SAM was the cause of the pressure differences between higher and mid-latitudes (figures 7(b), (c)). Additionally, the stratospheric polar vortex significantly influences the ASL and SAM (England et al 2016, Screen et al 2018. Here, we have investigated the extent to which the stratospheric polar vortex modulates the climatic indices. The positive (negative) PCH anomaly corresponds to the weakening (strengthening) of the stratospheric polar vortex. The relation between the PCH50 and PCH10 with the ASL (SAM) during spring and summer shows a significantly positive (negative) correlation (table S3). Thus, indicating that a strong polar vortex is inducing positive SAM. The standardized PCH at 50 hPa and 10 hPa during spring and summer shows large inter-annual fluctuations (figure 4). During spring 2021 and summer 2022, the standardized PCH shows nearly the lowest values signifying the strengthening of the polar vortex. In 2021, the PCH10 recorded the second and third lowest values in spring and summer. In contrast, the PCH50 showed the third lowest negative value during spring 2021 and summer 2022. Consequently, the positive SAM is associated with strong westerlies, as demonstrated in figure 5.
On the other hand, the deepening of the ASL is identified with the strong northerly winds across the ABS and the Weddell Sea sectors, resulting in sea ice retreat (figures 3 and 7(d)). Meanwhile, wind-driven positive sea ice anomalies were observed in sectors with increasing southerly winds. The ABS and the Weddell Sea sectors recorded the maximum sea ice retreat from September to November ( figure 7(a)). However, November 2021 had the lowest negative SIE trend, which may be the result of sea ice retreats in the western Antarctic and western Pacific Oceans. The sea ice retreat in western Antarctica is affected by shifting the low-pressure center from the western to the eastern flank of the ABS sector. In the Ross Sea, a positive SIC anomaly was observed during November and December, which was partially explained by the deepening of the ASL and strong southerly winds off the Ross Ice Shelf (figures 5(b), (c)). Similarly, warmer northwesterly winds pushed sea ice in the Indian Ocean and western Pacific sectors, resulting in positive (ice compaction towards the coast) and negative SIC anomalies off the coast. Further, the zonal mean temperature and GPH anomaly are illustrated at different pressure levels in figure 6 using the vertical-latitude section. It is evident from the figure that a positive temperature anomaly persisted below 500hPa in the region extending from the equator towards the pole. In the months of September and October, a warm air mass from the upper stratosphere is advancing towards the surface and influencing the negative GPH anomaly center towards the pole (figures 6(a), (b)). There is a positive temperature anomaly on the   surface and a negative anomaly above 850hPa over the Antarctic region (60°S-90°S) in November. Therefore, it may be suggested that prolonged atmospheric warming on the surface beginning in September led to the maximum sea ice loss in November.
From December 2021 to February 2022, a negative SIC anomaly persisted across most Antarctic sectors (figures 3(d)-(f) and 7(a)). ASL set the lowest record (968 hPa) in December and the second-lowest record in February ( figure 3(d)). Maximum growth and retreat of sea ice were observed in the Indian Ocean and Weddell Sea sectors, respectively. The intensification of westerlies drives cold surface waters off the Antarctic mainland and further sea ice advection northwards by Ekman transport (Hall and Visbeck 2002). This enhanced transport results in more northwards sea ice movement toward the northern Indian Ocean sector and the coastal Weddell Sea. Meanwhile, the increased poleward warmer wind flow regions corresponded to low sea ice, which was evident in the Ross Sea (figures 5(d), (e)). In January, the MSLP anomaly increased around the circumpolar trough from 60°S to 70°S, particularly in the western AP and ABS regions (figures 3(e) and 7(b)). This was reinforced by an increase in the magnitude of the zonal wave three (ZW3) pattern. It is an asymmetric pattern composed of three ridges close to 20°E, 90°E, and 150°W that affect climatic conditions and wind flow around the Antarctic coastal flank (Connolley 2002, Raphael 2004. The strong ZW3 pattern around Antarctica resembles a strong meridional flow with an alternate pattern of warm northerly, and cold southerly winds (Raphael 2007). Similarly, the earlier study found an association between the enhanced ZW3 states and the sea ice anomalies over the ABS sector (Irving and Simmonds 2015). The persistence of ZW3 contributes to sea ice retreat in the Ross Sea, western Weddell Sea, and the Indian Ocean, as well as sea ice advancement in the Bellingshausen Sea ( figure 5(e)). The atmospheric temperature and GPH anomaly are also in agreement with the above findings (figures 6(d), (e)). In February, a positive temperature anomaly was observed extending from the surface to the upper atmosphere in the Antarctic region. Together, these mechanisms mentioned above may have set the ground for a new record of sea ice loss in February 2022. Unprecedented sea ice reductions are evident in figure 5(f), as there is little to no sea ice in most Antarctic regions. The positive SIC anomaly prevailed in the ABS sector in January, which decreased in the area once again, possibly due to the wind flow and warm air intrusion from the mid-latitudes in the previous month.

Seasonal variability
During spring 2021, a positive SIC anomaly persisted over the western ABS sector, Ross Sea, and the western Indian Ocean, while the rest of the sectors showed a negative SIC anomaly ( figure 5(g)). The SIC anomaly pattern is consistent with the fifth-lowest Antarctic SIE in spring 2021 ( figure 3(a)). The low sea ice in spring may be explained by a robust low-pressure center over the ABS sector and the deepening of the ASL ( figure 5(g)). The anomalous westerlies towards the pole indicate a positive SAM, leading to SIC anomalies (Li et al 2018). In the Ross Sea and the western Indian Ocean, the prevailing wind flow pattern results in Ekman transport and hence the advection of sea ice northwards. However, a strong northerly wind results in sea ice retreat in the Weddell Sea. The temperature variation at different pressure levels in spring shows a weak positive anomaly in the higher latitudes of the lower tropospheric region (figure 6(g)). In the Antarctic region, the GPH anomaly was negative from 300hPa to the upper stratosphere.
The second lowest Antarctic SIE was recorded in summer, close to the record low set in 2017 since satellite records began ( figure 3(a)). The SIC anomaly is negative in the Ross Sea, western ABS, and the Weddell Sea. The equilibrium response of sea ice to climatic fluctuations in the previous season (month) is clear from the low summer sea ice record in summer (lowest in February) (figure 5). The positive SIC anomaly patch in the Indian Ocean, ABS sector, and the Weddell Sea was caused by the northwards advection of sea ice ( figure 5(h)). In summer, a negative GPH anomaly center is observed in the Antarctic, which is bounded by the two positive temperature anomaly centers in the upper and lower atmosphere ( figure 6(h)). At the same time, the positive surface temperature anomaly accelerates the sea ice loss towards higher latitudes.

Summary and conclusions
In this study, we analyzed monthly and seasonal SIE trends and the rate of change in sea ice across the entire Southern Ocean and its five sectors. These analyses were conducted across two distinct periods, 42-year (1979-2020) and 43/44 years (1979-2021/2022), to ascertain the influence of short-term monthly variability over longer periods. The Antarctic sea ice retreat from September 2021 was close to the minimum SIE records since 1979, with the lowest in February 2022. Seasonally, Antarctic SIE records reached the fifth lowest in spring 2021 and the second lowest during summer 2022. The anomalous sea ice retreat was in good agreement with the series of records observed for the climatic indices. The intensification of the low-pressure center in the Amundsen Sea and the positive SAM from September 2021 influence the wind patterns across the entire Southern Ocean. In addition, the PCH at 10 hPa and 50 hPa showed a strengthening of the stratospheric polar vortex with the lowest negative values. Thus, accelerating westerlies and inducing positive polarity of SAM. On the other hand, the ASL also deepens in association with the positive SAM index. The increased meridional flow of northerly winds contributed to the maximum spring sea ice retreat in the Bellingshausen Sea and Weddell Sea sectors. However, negative sea ice anomalies in the western Pacific Ocean are due to westerly flow. In October 2021, remarkable sea ice variation occurred in spring, possibly owing to the lowest ACP recorded since 1979. Positive sea ice anomalies towards the coastal area of the Antarctic mainland are related to northward sea ice advection due to Ekman transport. In December 2021, the fourth-highest positive SAM index and the lowest ACP resulted in an anomalous increase in westerlies across Antarctica and maximum negative sea ice anomalies. In contrast, a decrease in the MSLP anomaly around the circumpolar trough in January led to the ZW3 pattern. The existing pattern of MSLP anomalies contributed to the maximum sea ice recession in the Ross Sea, western Weddell Sea, and the Indian Ocean while advancing in the Bellingshausen Sea. Overall, the lowest Antarctic sea ice record in February 2022 resulted from the long months of sea ice retreat induced by the cumulative response of the positive SAM and the deepening of the ASL. Consistent with these observations, the atmospheric temperature and GPH anomaly also revealed evidence of warming at the surface, extending from the upper to the lower atmosphere and from the equator to the higher latitudes.
The findings of this study have implications for the future understanding of the role of atmospheric circulation and climate variability in driving specific short-term events of low Antarctic sea ice. The drivers of changes in the stratospheric polar vortex, SAM, ASL, wind, and pressure patterns over Antarctica are complex to understand. Due to the regional nature of Antarctic sea ice variability, small changes in atmospheric circulation in any sector may impact the SIE for the entire Southern Ocean. These findings indicate that ASL and SAM play a greater role in controlling sea ice variability in East Antarctica and the Weddell Sea during the ice retreat phase. The influence of the former on sea ice is determined by the amplitude of wind circulation and pressure patterns, as well as their direction and strength. This may provide insight into how Antarctic sea ice evolved before the satellite era and how it may vary in the future.

Data availability statement
Daily and seasonal sea ice extent data for the total Antarctic and its five sectors are obtained from U.S. National Snow and Ice Data Center based on Bootstrap and NASA algorithm (available at https://daacdata.apps.nsidc. org/pub/DATASETS /nsidc0192_seaice_trends_climo_v3/total-ice-area-extent/). We acknowledge the European Centre for Medium-range Weather Forecasts for providing the higher spatiotemporal resolution fifthgeneration reanalysis monthly data at single and pressure levels for sea ice concentration, mean sea level pressure, 10 m zonal and meridional wind components, geopotential height, and temperature (available at https:// cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels-monthly-means?tab=overview and https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-pressure-levels-monthly-means?tab=overview). The monthly data of the Amundsen Sea Low Actual Central Pressure index (https://scotthosking. com/asl_index) and Southern Annular Mode index (https://legacy.bas.ac.uk/met/gjma/sam.html) were obtained from National Center for Atmospheric Research.