Reply on RC2

Polar regions have a profound impact on global climate change, because polar regions are one of the world’s largest sources of cold. Sea ice is the main influencing factor in the polar regions, which has a great influence on the heat transfer, which in turn affects the global climate system. In the past two decades, Arctic sea ice has decreased year by year, while Antarctic sea ice has increased. The use of passive satellite remote sensing data for long-term observation and analysis of polar sea ice will provide an important theoretical basis for studying global climate change and an important data source for polar sea ice research. This study uses the data of the Antarctic and Arctic sea ice density data from 1998 to 2018 to compare and analyze the changes in time and space to find out the changes of the sea ice in the south and Arctic. Using the SSM / I data of the US National Defense Meteorological Satellite and the NASA TEAM algorithm, the sea ice data from 1998 to 2018 was compared and analyzed. The specific process is to use the NASA TEAM algorithm to invert the Arctic and Arctic sea ice, obtain multi-year ice, one-year ice, and overall sea ice density data, and then determine the sea ice area. Through comparative analysis, it is found that the area of Antarctic sea ice is significantly larger than that of the North Pole, and the change trend of the two is similar. The downward trend is greater than that of the South Pole; the total annual ice volume of the South Pole is greater than that of the North Pole, both of which have an upward trend, and the Arctic has a greater increase; the overall sea ice has fallen, and the multi-year ice has decreased while the annual ice has increased. Sea ice has a tendency to transform from multi-year ice to one-year ice; Antarctic multi-year ice is mainly distributed in the Weddell Sea, with stable areas accounting for 60% of the multi-year ice area; Antarctica is dominated by one-year ice, and one-year ice stable area accounts for one 80.1% of the total annual ice area; Arctic multi-year ice mainly exists in the central area of the Arctic Ocean. The Arctic multi-year ice stable area only accounts for 35.06%; the Arctic one-year ice is mainly distributed around the multi-year ice periphery, and the stable existence area accounts for the total annual ice area 48.5%.

The sea ice concentration dataset (covering the period 1992-2008) used in this study is a new dataset based on the Sea Ice Climate Change Initiative (SICCI) algorithm. We investigate whether the SICCI dataset is on a par with other datasets for studying sea ice cover changes in the Southern Ocean. We then examine spatiotemporal variations in sea ice derived from the SICCI dataset over the Southern Ocean, and analyse relationships of sea ice with sea surface temperature (SST). The results indicate that there is no significant difference between the SICCI dataset and the NASA Team dataset, and therefore the former can also be used for studying sea ice changes. Both sea ice extent (SIE) and sea ice area (SIA) derived from the SICCI dataset over the Southern Ocean increased slightly from 1992 to 2008, at rates of (17.75 ± 11.50) × 10^3 and (17.37 ± 9.51) × 10^3 km^2 yr^(–1), respectively. Antarctic sea ice has significant seasonal variations; all seasonally averaged SIE and SIA show an increase, with spring showing the largest positive changing rate. The Weddell Sea, Ross Sea, and Indian Ocean have positive yearly changing rates in SIE and SIA, while the Bellingshausen/Amundsen seas and western Pacific Ocean have negative yearly changing rates. However, overall sea ice over the Southern Ocean has a slight positive trend, which is the same as the sea ice change pattern derived from the NASA Team dataset. This indicates that the contributions to the change in sea ice over the whole Southern Ocean due to the Weddell Sea, Ross Sea, and Indian Ocean dominate over those by the Bellingshausen/Amundsen seas and western Pacific Ocean. Further analysis shows that both SIE and SIA are negatively correlated with SST in the Southern Ocean or each of the 5 longitudinal sectors, and sea ice is more sensitive to SST in spring and autumn.

Abstract. This study investigates the surface albedo of the sea ice areas adjacent to the Antarctic Peninsula during the austral summer. Aircraft measurements of the surface albedo, which were conducted in the sea ice areas of the Weddell and Bellingshausen Seas show significant differences between these two regions. The averaged surface albedo varied between 0.13 and 0.81. The ice cover of the Bellingshausen Sea consisted mainly of first year ice and the sea surface showed an averaged sea ice albedo of αi = 0.64 ± 0.2 (± standard deviation). The mean sea ice albedo of the pack ice area in the western Weddell Sea was αi = 0.75 ± 0.05. In the southern Weddell Sea, where new, young sea ice prevailed, a mean albedo value of αi = 0.38 ± 0.08 was observed. Relatively warm open water and thin, newly formed ice had the lowest albedo values, whereas relatively cold and snow covered pack ice had the highest albedo values. All sea ice areas consisted of a mixture of a large range of different sea ice types. An investigation of commonly used parameterizations of albedo as a function of surface temperature in the Weddell and Bellingshausen Sea ice areas showed that the albedo parameterizations do not work well for areas with new, young ice.

There is low confidence in projections of Antarctic sea ice area (SIA), due to deficiencies in climate model sea ice processes. Ensemble regression techniques can help to reduce this uncertainty. We investigate relationships between SIA climatology and 21st century change in the Coupled Model Intercomparison Project, phase 6 (CMIP6) multi‐model ensemble. In summer, under a strong forcing scenario, each model loses the majority of its sea ice. Therefore, models with greater historical SIA exhibit greater reductions, so the observed climatology of SIA strongly constrains projections. Ensemble spread in historical summer SIA is smaller than in CMIP phase 5 (CMIP5), and CMIP6 gives a more robust constraint on future SIA. In winter, by 2100 under a strong forcing scenario, 40% of SIA disappears on average, and ensemble spread in historical mean SIA explains approximately half the spread in projected change. A greater winter ice loss in CMIP6 than CMIP5 is explained by the higher climate sensitivities of some CMIP6 models.

Most remote sensing sea ice classification methods use single-source remote sensing data, such as synthetic aperture radar (SAR) data and optical remote sensing data. SAR data contain rich sea ice texture information, but the data are relatively single, making it difficult to distinguish detailed sea ice categories. Optical data include abundant spatial-spectral information, but they are often affected by clouds, fog, and severe weather. Hence, given the limitations of single-source data, the remote sensing sea ice classification accuracy cannot be further improved. A sea ice classification method based on deep learning and multisource remote sensing data fusion is proposed utilizing an improved densely connected convolutional neural network (DenseNet) to mine and fuse the multilevel features of sea ice. According to the characteristics of SAR data and optical data, a dual-branch network structure based on an improved DenseNet is employed for feature extraction, and a squeeze-and-excitation attention mechanism is introduced to weight the fusion features to further enhance the feature weights that can effectively distinguish different types of sea ice. The fully connected network is used to perform the deep fusion of features and classify sea ice. To verify the effectiveness of the proposed method, two sets of sea ice data are utilized for sea ice classification. The experimental results show that the proposed method fully excavates and fuses the multilevel characteristics of heterogeneous data using the improved dual-branch network structure, leverages the complementary characteristics of SAR data and optical data, significantly increases the sea ice classification accuracy, and effectively improves the influence of cloud cover on the sea ice classification accuracy. Compared with the typical single-source data classification method and other heterogeneous data fusion classification methods, the proposed method achieves superior overall classification accuracy (98.49% and 98.58%).

Geophysical constraints on the Antarctic sea ice cover

Fully coupled climate models have long shown a wide range of Antarctic sea ice states and evolution over the satellite era. Here, we present a high‐level evaluation of Antarctic sea ice in 40 models from the most recent phase of the Coupled Model Intercomparison Project (CMIP6). Many models capture key characteristics of the mean seasonal cycle of sea ice area (SIA), but some simulate implausible historical mean states compared to satellite observations, leading to large intermodel spread. Summer SIA is consistently biased low across the ensemble. Compared to the previous model generation (CMIP5), the intermodel spread in winter and summer SIA has reduced, and the regional distribution of sea ice concentration has improved. Over 1979–2018, many models simulate strong negative trends in SIA concurrently with stronger‐than‐observed trends in global mean surface temperature (GMST). By the end of the 21st century, models project clear differences in sea ice between forcing scenarios.

Sea ice is critical to map for safe and efficient maritime navigation, to mitigate ship trapping and capsizing. Sea ice is also vital to monitor to assess the state of the changing climate and a critical component in climate and weather models, reflecting sunlight towards space and acting as an insulating material between the ocean surface and the atmosphere.Professional sea ice analysts at national ice services map sea ice based on Synthetic Aperture Radar (SAR) images acquired by satellites, such as the Sentinel-1 (S1) satellite constellation. The ice analysts manually interpret the SAR images using their in-depth knowledge and experience to create sea ice charts with information on the sea ice conditions.A challenge for the S1 5.4 GHz SAR measurements is that the radar wave does not penetrate deep into the sea ice and is scattered/reflected by the surface. Therefore, the SAR images provide information primarily about the sea ice surface, useful for identifying and classify sea ice conditions. The charts describe, among others, the sea ice’s stage of development - the type of sea ice - an indicator of its thickness. The manual charting process apply sea ice classes, defined by the International Ice Charting Working Group (IICWG) on behalf of the World Meteorological Organization (WMO). Considerable uncertainties are associated with the ice classes that can vary from, e.g. 30-200cm or 70-120cm in thickness. Deep-learning models that produce stage-of-development information from S1 radar images exist but has the same inherent limitations of the sea ice charts in the model outputs.Current state-of-the-art sea ice thickness retrieval methods relies on altimeter satellites, such as the CryoSat-2 (CS2) satellite. The distance between the ocean and the sea ice is measured, known as the sea ice freeboard. For a Ku-band radar altimeter like CS2, it is assumed that the radar response penetrates the snow and returns from the sea ice surface. As the true penetration is unknown, and the radar wave propagation is delayed when the signal passes through snow, the measured quantity is known as the radar freeboard.The sea ice thickness can be estimated with an accuracy of 20-40% using the radar freeboard by calculating the sea ice's buoyancy based on snow and ice density estimates, and auxiliary snow depth information. However, CS2 only measures 1600m across the orbit and can thus only monitor sea ice thickness in the Arctic monthly - insufficient for many applications, such as maritime navigation, and leaves data record gaps. S1 SAR on the other hand, cover 400km in Extra Wide mode across the orbit with repeating coverage every week.Here, we present our preliminary results of circumventing the limitations of CS2 and S1 by training supervised deep-learning convolutional neural network (CNN) models to recognise sea ice textures in S1 SAR images and assign sea ice radar freeboard estimates acquired by CS2. This approach transfers information acquired by CS2 to S1, which we call Cryo2S1. A Cryo2S1 dataset is curated, containing several thousand collocated S1 SAR images and along-track CS2 measurements during 2020-2021.

Abstract. The optical properties of snow/sea ice vary with age and by the processes they were formed, giving characteristic types of snow and sea ice. The response of albedo and light penetration depth (e-folding depth) to increasing mass ratio of black carbon is shown to depend on the snow and sea ice type and the thickness of the snow or sea ice. The response of albedo and e-folding depth of three different types of snow (cold polar snow, wind-packed snow and melting snow) and three sea ice (multi-year ice, first-year ice and melting sea ice) to increasing mass ratio of black carbon is calculated using a coupled atmosphere–snow/sea ice radiative-transfer model (TUV-snow), over the optical wavelengths of 300–800 nm. The snow and sea ice types are effectively defined by a scattering cross-section, density and asymmetry parameter. The relative change in albedo and e-folding depth of each of the three snow and three sea ice types with increasing mass ratio of black carbon is considered relative to a base case of 1 ng g−1 of black carbon. The relative response of each snow and sea ice type is intercompared to examine how different types of snow and sea ice respond relative to each other. The relative change in albedo of a melting snowpack is a factor of four more responsive to additions of black carbon compared to cold polar snow over a black carbon increase from 1 to 50 ng g−1, while the relative change in albedo of a melting sea ice is a factor of two more responsive to additions of black carbon compared to multi-year ice for the same increase in mass ratio of black carbon. The response of e-folding depth is effectively not dependent on snow/sea ice type. The albedo of sea ice is more responsive to increasing mass ratios of black carbon than snow.

<p>Ocean reanalyses (ORAs) are used extensively in polar research, hence their realism should be assessed regularly. Here the ORAs performance in the Antarctic region is analyzed with specific emphasis on sea ice concentration and thickness. We used four global ocean-sea ice products: C-GLORSv7, FOAM-GLOSEA5v13, GLORYS2v4, and ORAS5, and their ensemble mean GREP (provided by CMEMS) within the 1993 to 2018 period. All ORAs use the NEMO ocean model in a global eddy-permitting configuration (1/4° horizontal resolution and 75 vertical levels) and are forced by the ECMWF ERA-Interim atmospheric reanalysis.</p><p>Here we examine the ability of ORAs to reproduce sea ice properties in the Southern Ocean taking into account regional characteristics and sea ice types. Seasonal and interannual variability of sea ice concentration (SIC) and sea ice thickness (SIT) is examined in the hemispheric domain and in five sub-regions for three different sea ice classes: pack ice (SIC ≥ 80%), marginal ice zone (MIZ) (15% ≤ SIC < 80%), and sparse ice (0 < SIC <15%).  Modeled sea ice properties are compared to a set of satellite products: NSIDC CDR, Ifremer/CERSAT, and EUMETSAT OSI-SAF for SIC and Envisat and CryoSat-2 for SIT, together with PIOMAS and GIOMAS reanalyses. We revealed shortcomings of reanalysis systems to be improved in the future representation of Antarctic sea ice. Additionally, we focused on the assessment of the GREP ensemble mean product. We found that for certain metrics GREP minimizes the single errors and outperforms individual members. The evidence from this study implies that GREP can be a feasible product for a number of applications.</p>

Despite the overwhelming warming over most oceans under global warming, cooling in the surface of the Southern Ocean between Australia and South America has been previously reported. The surface cooling is due to wind-driven sea-ice transport and its subsequent melting. However, Antarctic sea ice which has a steady increasing trend during several decades, experienced an abrupt decline to a record low in the mid-2010s. Since then, the interannual variability of Antarctic sea ice has been further intensified, showed another record low in 2022 and 2023 consecutively. Accordingly, the Pacific sector of the Southern Ocean has experienced warming during the period of rapid sea ice decline.Nevertheless, we find that a coherent cooling pattern is observed in a very specific area, near the Ross Sea gyre. Here, to investigate the mechanisms associated with the cooling pattern, we examined how the rapid decline in Antarctic sea ice in recent years has contributed to the strengthening of the cooling trend near the Ross Sea gyre, using satellite observations of sea ice, as well as oceanic and atmospheric reanalysis data. Our result shows that the cooling trend has strengthened despite the rapid decrease in Antarctic sea ice in recent years. The significant cooling trend has been attributed to record atmospheric low pressure systems over the Pacific sector of the Southern Ocean. We demonstrate that the cooling near the Ross Sea gyre is likely influenced by the local and remote large-scale atmospheric variabilities that lead to substantial sea ice anomalies. Since the mid-2010s, local anomalous winds and surface heat flux associated with strong events of natural climate oscillations like the Zonal Wavenumber 3, El Niño–Southern Oscillation, and Southern Annular Mode in the Southern Hemisphere that drive the recent change in Antarctic sea ice, induce strong meridional flows led to enhanced sea ice drifts and melting, and consequently to the strong cooling trend.    

Arctic sea ice type information is essential for various operational and scientific applications including the support of marine users and guiding ice thickness retrieval algorithms operating with SMOS and CryoSat-2 data for improved sea ice prediction. A sea ice type analysis system developed at Environment and Climate Change Canada’s (ECCC) generates pan-Arctic ice type analyses at 5 km resolution every 6 hours. The ice type analysis system assimilates ice type information provided by passive microwave (AMSR2, SSMIS) and scatterometer (ASCAT) data, but assimilation of these observations is not reliable in the areas near land and in the narrow channels of the Canadian Arctic Archipelago due to their low spatial resolution of ~20-50 km. Therefore, assimilation of high-resolution ice type observations from synthetic aperture radar (SAR) is highly desired.In this study, we extended our approach for automated detection of winter multi-year ice (MYI) and first-year ice (FYI) at 1.6 km scale from RADARSAT-2 to RCM data. To this end, we collected more than 2,000 RCM images and corresponding image analyses products that were manually generated by the Canadian Ice Service (CIS) ice experts for the period of time between July 1, 2020 and July 31, 2023. From these RCM images we extracted SAR information for more than 30,000 pure MYI and more than 619,000 pure FYI data samples under no melt conditions as identified by the CIS image analyses.We demonstrated that separability measures for MYI and FYI classes in the spaces of the two predictor parameters (HV/HH polarization ratio and standard deviation of HV signal) are consistent with those we previously observed for RADARSAT-2. Furthermore, we found that our RCM-based MYI/FYI detection approach allows us to classify 60% of the winter CIS ice data samples with the accuracy of 99.6%. Preliminary results of assimilating RCM-based MYI/FYI high-resolution retrievals in combination with passive microwave and scatterometer data in the ECCC ice type analysis system will be also presented.

Relationship between Antarctic sea ice and the Southern Annular Mode and tropical climate variability has been widely studied. Focusing on the midlatitude oceanic frontal zones, this study identifies a fingerprint of South Atlantic (SA) variability in the Antarctic summer sea ice. The interaction between spring SA sea surface temperature (SST) and summer storm‐track activity modulates the hemispheric atmospheric circulation, which in turn alters Antarctic summer sea ice concentration (SIC) through thermal and wind‐driven forcing. Specifically, warm SST anomalies in the western SA frontal zone, corresponding to the strengthened SST front, cause the increased SIC in the eastern Ross Sea via anomalous cold air advection and offshore drift of sea ice. The opposite effect (i.e., warm air advection and onshore ice drift) results in the decreased SIC in the northwestern Weddell Sea. The findings also imply a potential impact of the midlatitude SA variability to the South Pole air temperature variability.

This study investigates the response of the Southern Ocean sea surface temperature (SST) and Antarctic sea ice to stratospheric ozone recovery, focusing on the time scale and seasonality of the response. The response is quantified by contrasting two twenty-first-century ensemble simulations conducted with the Goddard Earth Observing System Chemistry–Climate Model: one with decreasing ozone-depleting substances (ODSs) and the other with fixed 2005 levels of ODSs. In our simulations, the response to ozone recovery has large seasonal variations, but it does not show a two-time-scale behavior. Ozone recovery causes Southern Ocean SST warming in austral summer and cooling in other seasons. Ozone recovery mitigates Antarctic sea ice decrease in the twenty-first century in austral spring, fall, and winter. However, the summer Antarctic sea ice extent is not affected by ozone recovery despite strong surface warming, because the warming occurs north of the sea ice edge. The absence of summer sea ice response likely results from the model bias of underestimating summer sea ice climatology. The summer surface warming response is associated with cooling directly below the mixed layer. Temperature tendency budget analysis shows that reduced vertical mixing plays a critical role in driving this vertical dipole temperature response. We also find that the impact of the vertical temperature advection on SST depends not only on changes in upwelling but also on changes in vertical temperature gradient. Significance Statement We study the climate impact of the projected stratospheric ozone recovery in the twenty-first century on Southern Ocean sea surface temperature and Antarctic sea ice using coupled atmosphere–ocean–chemistry model simulations. The model results show that ozone recovery causes weakening of surface winds over the Southern Ocean, which leads to warming of the Southern Ocean surface in summer and cooling in other seasons. Ozone recovery also reduces Antarctic sea ice loss in the twenty-first century. We quantify the relative importance of meridional and vertical advection and vertical mixing in determining the upper Southern Ocean temperature response to stratospheric ozone recovery. These results improve our understanding of how changes in surface winds affect the Southern Ocean temperature, circulation, and Antarctic sea ice.

Global, Arctic, and Antarctic sea ice volume predictions using score-driven threshold climate models