Search

Revolutionary advances in artificial intelligence (AI) in the past decade have brought transformative innovation across science and engineering disciplines. In the field of Arctic science, we have witnessed an increasing trend in the adoption of AI, especially deep learning, to support the analysis of Arctic big data and facilitate new discoveries. In this paper, we provide a comprehensive review of the applications of deep learning in sea ice remote sensing domains, focusing on problems such as sea ice lead detection, thickness estimation, sea ice concentration and extent forecasting, motion detection, and sea ice type classification. In addition to discussing these applications, we also summarize technological advances that provide customized deep learning solutions, including new loss functions and learning strategies to better understand sea ice dynamics. To promote the growth of this exciting interdisciplinary field, we further explore several research areas where the Arctic sea ice community can benefit from cutting-edge AI technology. These areas include improving multimodal deep learning capabilities, enhancing model accuracy in measuring prediction uncertainty, better leveraging AI foundation models, and deepening integration with physics-based models. We hope that this paper can serve as a cornerstone in the progress of Arctic sea ice research using AI and inspire further advances in this field. [ABSTRACT FROM AUTHOR]

This study explored the interannual characteristics of the out‐of‐phase pattern of summer precipitation over northern China during the past years, as well as the possible underlying mechanisms. The out‐of‐phase pattern is characterized by the positive precipitation anomaly in Northwest China (NWC) and the negative anomaly in North China (NC). Our analyses indicate that the variation of Asian westerly jet (AWJ) is found evidently associated with this out‐of‐phase pattern of summer precipitation. The meridional displacement of AWJ induces opposite trends in water vapour transport and circulation anomalies between NWC and NC, leading to the out‐of‐phase pattern. Further analyses suggest that the anomalies of Arctic sea ice concentration (SIC), North Atlantic and Northwest Pacific sea surface temperature (SST) can impact the variation of AWJ by triggering the Eurasian (EU), polar‐Eurasia (POL) and Pacific‐Japan (PJ) teleconnection patterns, respectively, which in turn modulate the formation of this pattern. The precipitation in NWC is primarily affected by the North Atlantic and Northwest Pacific SST anomalies, whereas the precipitation in NC is notably influenced by both SST anomalies and the Arctic SIC anomaly. However, the role of SST anomalies on precipitation in NWC and NC exhibits an evident contrast, leading to the out‐of‐phase pattern of summer precipitation. [ABSTRACT FROM AUTHOR]

We apply the Canny edge algorithm to imagery from the Utqiaġvik coastal sea ice radar system (CSIRS) to identify regions of open water and sea ice and quantify ice concentration. The radar-derived sea ice concentration (SIC) is compared against the (closest to the radar field of view) 25 km resolution NSIDC Climate Data Record (CDR) and the 1 km merged MODIS-AMSR2 sea ice concentrations within the ∼11 km field of view for the year 2022–2023, when improved image contrast was first implemented. The algorithm was first optimized using sea ice concentration from 14 different images and 10 ice analysts (140 analyses in total) covering a range of ice conditions with landfast ice, drifting ice, and open water. The algorithm is also validated quantitatively against high-resolution MODIS-Terra in the visible range. Results show a correlation coefficient and mean bias error between the optimized algorithm, the CDR and MODIS-AMSR2 daily SIC of 0.18 and 0.54, and ∼−1.0 and 0.7%, respectively, with an averaged inter-analyst error of ±3%. In general, the CDR captures the melt period correctly and overestimates the SIC during the winter and freeze-up period, while the merged MODIS-AMSR2 better captures the punctual break-out events in winter, including those during the freeze-up events (reduction in SIC). Remnant issues with the detection algorithm include the false detection of sea ice in the presence of fog or precipitation (up to 20%), quantified from the summer reconstruction with known open water conditions. The proposed technique allows for the derivation of the SIC from CSIRS data at spatial and temporal scales that coincide with those at which coastal communities members interact with sea ice. Moreover, by measuring the SIC in nearshore waters adjacent to the shoreline, we can quantify the effect of land contamination that detracts from the usefulness of satellite-derived SIC for coastal communities. [ABSTRACT FROM AUTHOR]

A strong cooling event refers to a sharp change in the average temperature over a short period. The rapid change of temperature has important effect on human health and is highly concerned recent years. Based on the observed temperature data set from stations in Southwest China (SWC) from 1979 to 2017, this paper analyses the characteristics of the strong cooling event (SCE). The result shows that SCE occurs with the highest frequency during the time from February to May. Among them, the frequency of SCE in February exhibits an abrupt change before and after 2005 with a significant increase. Further study reveals that the change of SCE frequency in February is associated with the large‐scale background circulation patterns. After 2005, there is a cyclonic circulation anomaly in Northeast Asia and an anticyclonic circulation anomaly in the Tibet Plateau (TP). This pattern provides a favourable condition for the southward movement of cold air mass, thereby increasing the frequency of SCE in SWC. Furthermore, it is revealed that there is a strong correlation between the variation of SCE frequency in February and sea surface temperatures (SST) in the Northwest Pacific in January before 2005. The cold SST anomaly could favour the occurrence of extreme TD events in SWC through vertical circulation. After 2005, the correlation between SCE and Northwest Pacific SST is not significant. The sea ice in the northern Barents Sea and Kara Sea becomes the dominant impact factor. The abnormally low sea ice concentration is conducive to strengthen the meridional circulation over East Asia, leading to an increasing frequency of SCE in SWC. [ABSTRACT FROM AUTHOR]

Objectives: To achieve the accurate monitoring of sea ice conditions in the key area of Arctic passage, high resolution sea ice thickness is essential. However, sea ice thickness with kilometer-scale resolution cannot meet the requirement at present. Methods: Considering the relationship between sea ice thickness and image albedo, this paper tried to establish a regression model between the ICESat-2 ATL10 sea ice freeboard product and the high resolution Sentinel-2 optical image albedo by taking advantage of the intensive altimetry data of ICESat-2 along the orbit to obtain dense sea ice freeboard. Based on high resolution sea ice thickness calculated from dense freeboard with the hydrostatic equilibrium model and combined with high resolution sea ice concentration derived from Sentinel-2 image, we classified ice areas into different navigational categories according to the ice-break capability of ship and compared optimal route design at multiple spatial scales. Results: Regression models established from the ATL10 sea ice freeboard and the Sentinel-2 image albedo have good fitting accuracy. The R² of the regression model is higher than 0.5, the average deviation is less than 0.05 m, and the root mean square error of the accuracy validation is less than 0.2 m. High resolution sea ice parameters derived by the proposed method can describe the distribution of fine leads between floating ice which are difficult to be captured by low resolution sea ice parameters. Conclusions: High resolution sea ice parameters can describe the details of sea ice conditions more accurately. Therefore, the proposed high resolution inversion method of sea ice parameters can improve the capability of navigation planning for ships in Arctic passage, further improving the navigation safety. [ABSTRACT FROM AUTHOR]

The ASI algorithm uses the same sea ice and seawater tie-points when inverting polar sea ice concentration (SIC), but this approach does not fully consider the differences between different polar sea regions and the impact of different sea ice characteristics on SIC results. To make up for this deficiency, the SIC inversion algorithm based on different types of Arctic sea ice is proposed. The proposed algorithm selects pure ice and pure water sample points in different sea regions to derive SIC inversion formulas, and subsequently obtains SIC retrieval results for the entire Arctic. Compare the results of this study with those of traditional ASI algorithm, and perform local validation based on the sea ice distribution obtained from Landsat-8 data. The results show that compared with the traditional ASI algorithm, the proposed algorithm has improved the accuracy of SIC inversion in different sea ice regions by 2%-6%, with an average improvement of 3.3%. Overall, our research has improved the ASI algorithm, which is of great significance for obtaining higher precision polar SIC. [ABSTRACT FROM AUTHOR]

Sea ice concentration (SIC) is an important dimension for characterising the geographical features of the pan-Arctic region. Trends in SIC bring new opportunities for human activities in the Arctic region. In this paper, we propose a deep learning technology-based sea ice concentration prediction model, SICFormer, which can realise end-to-end daily sea ice concentration prediction. Specifically, the model uses a 3D-Swin Transformer as an encoder and designs a decoder to reconstruct the predicted image based on PixelShuffle. This is a new model architecture that we have proposed. Single-day SIC data from the National Snow and Ice Data Center (NSIDC) for the years 2006 to 2022 are utilised. The results of 8-day short-term prediction experiments show that the average Mean Absolute Error (MAE) of the SICFormer model on the test set over the 5 years is 1.89%, the Root Mean Squared Error (RMSE) is 5.99%, the Mean Absolute Percentage Error (MAPE) is 4.32%, and the Nash–Sutcliffe Efficiency (NSE) is 0.98. Furthermore, the current popular deep learning models for spatio-temporal prediction are employed as a point of comparison given their proven efficacy on numerous public datasets. The comparison experiments show that the SICFormer model achieves the best overall performance. [ABSTRACT FROM AUTHOR]

In this study, we explore the interannual variability of Sea Surface Salinity (SSS) in the Dotson–Getz Trough located in West Antarctica, focusing on the month of February. Utilizing the oceanic analysis product EN4, we first validate the EN4 SSS with data from a singular ship-based survey, and then delve into potential factors that may influence SSS, with a particular emphasis on surface freshwater flux, sea ice concentration (SIC), and also the surface stress curl, which will induce upwelling via Ekman transport to affect the SSS. Our findings primarily indicate a link between SSS and sea ice concentration, showcasing a negative correlation where the peak (average) coefficient is around −0.6 (−0.4), further affirming the substantial interannual variability of SSS in this region. [ABSTRACT FROM AUTHOR]

Sea ice concentration (SIC) can be retrieved from thermal infrared (TIR) imagery due to the distinctive thermal properties of ice and water. Nevertheless, existing TIR-based SIC algorithms rely on surface temperature data, which often introduces additional errors. To address this issue, we have developed a new TIR ice concentration algorithm (TIRIA) that directly utilizes TIR brightness temperatures. TIRIA considers factors such as seawater salinity, observation angle and their impacts on seawater brightness temperature. TIRIA and a traditional algorithm, the MODIS potential open water algorithm (MPA), are applied to MODIS TIR imagery. Results are evaluated with near-infrared (NIR) SICs and compared with passive microwave (PM) SICs. Overall, TIRIA outperforms MPA, exhibiting a smaller root mean square error (RMSE) (14.01% compared to 17.63%) and higher correlation coefficient (0.89 compared to 0.81). Both TIRIA and MPA tend to underestimate SIC in high SICs while overestimating it in low SICs. Due to its more accurate identification of water, TIRIA significantly mitigates the overestimation in low SICs. Compared to PM-based SICs, both TIR-based SICs exhibit overall overestimations, with better consistency between TIRIA and PM-based SICs. TIRIA, being independent of surface temperature products and theoretically applicable to any TIR data, showcases great potential for future application.

ABSTRACTLocated at the easternmost sector of the Last Ice Area (LIA) with the oldest and thickest sea ice, the north of Greenland has witnessed some unprecedented polynyas since 2018, receiving considerable attention. Sea ice concentration (SIC) derived from passive microwave data can provide near daily observations of these polynyas. However, these SIC products are limited by the coarse spatial resolution of passive microwave data, with mixed pixels and large uncertainties at the ice–water divide. It is especially difficult in the case of newly formed polynyas in the form of leads with narrow openings, preventing accurate estimation of the regional energy budget and heat fluxes as well as the hindering causal analysis of the polynyas. In this paper, a novel state-of-the-art deep-learning-based super-resolution (SR) method is adopted, and passive microwave SR images, with spatial resolutions of up to 4 times the original, are generated and employed to derive super-resolved SIC (SR-SIC). Experimental results confirm the benefits of SR-SIC for finer monitoring of sea ice at both the polynya and lead scale compared to the original lower-resolution SIC, with an average 7.55% improvement in F1-Score for polynya extent and 31.17% for lead width. Using the SR-SIC, more accurate quantitative parameters of polynyas and leads are extracted to present more detailed spatial distributions at their different development stages. For example, SR-SIC with reduced mixed pixels can distinguish more SIC values below 40%, advantageous for detecting open water pixels within polynya areas and improving the accuracy of heat fluxes estimation and ice production calculation. Moreover, SR-SIC has the capability to identify leads with openings less than 3 km and to locate the finer boundaries of the polynyas. Finally, it is found that the lead development along the 2018 winter polynya is associated with regional wind, air temperature, and ice drift. The prominent summer polynya events in 2018, 2020, and 2021 may be in part caused by the thinning sea ice in the north of Greenland.

This study examines the future spatial distributions and long-term trends of the sea ice concentration (SIC) and the sea ice thickness (SIT) in the Arctic, along with the potential impacts of the Temperature at Surface (TAS), using the outputs of six models under four forcing scenarios of the Coupled Model Intercomparison Project Phase 6 (CMIP6). The results show that, under the four forcing scenarios, the SIC exceeds 50% and the SIT is approximately 1.5 m in most parts of the Arctic in 2030, 2040, and 2050. However, in the regions of East Greenland, the Barents Sea, the Kara Sea, and the Chukchi Sea, the SIC and SIT are comparatively lower and thinner. Both the SIC and the SIT exhibit declining trends during 2015—2050. Specifically, under the high-forcing scenario, the SIC in some regions decreases by over 1% per year. In 2050, under the high-forcing scenario, the SIC in most of the Arctic exceeds 90% during winter and spring months. The SIC reveals a decreasing trend in every month. The SIT exceeds 1 m from winter to summer in most regions, while it is less than or equal to 0.5 m in the fall. From December to May, the SIT exhibits a decreasing trend across the Arctic, with some localized increases observed in other months. Regarding the long-term trend up to 2100, the SIC and SIT in the Central Arctic, the East Greenland Sea, and the Chukchi Sea decrease over time, with the fastest reduction in the Central Arctic, meanwhile the TAS continues to rise in these three regions. Additionally, the spatial distribution and long-term trend between the TAS and the SIC are opposite, implying the possible impact of the TAS on the SIC. These findings provide valuable insights into the future sea ice in the Arctic under different forcing scenarios.

Based on the CICE sea ice model and the PDAF parallel data assimilation framework, this paper uses the local error subspace transform Kalman filter method (LESTKF) to assimilate the sea ice concentration, sea ice thickness and sea ice freeboard data into the model, and designs experiments to study the improvement of multi-parameter assimilation on the simulation of Arctic sea ice concentration and range. The results show that data assimilation has a good improvement effect on the simulation of Arctic sea ice concentration and range. The average deviation, root mean square error and mean absolute error of the assimilation experiment are significantly reduced compared with the control experiment. The assimilation experiment improves the simulation of sea ice concentration and range most obviously in summer. Multi-parameter assimilation can improve the prediction accuracy and reliability of Arctic sea ice change.

Using the ERA‐5 reanalysis datasets, NCEP/NCAR sea ice concentration (SIC) and observed winter surface wind speed (WSWS) in northwest China during 1979–2020, this study examined the regime shift of the leading mode of observed WSWS in northwest China and investigated its possible link with the winter Barents–Kara Sea (BKS) SIC. The leading mode of observed WSWS in northwest China showed negative anomalies in southern and northeastern Xinjiang, Gansu, western and northern Shaanxi Provinces, and positive anomalies in northwestern and middle Xinjiang, Qinghai and Ningxia Provinces. The signs of WSWS in the leading mode is generally opposite to their trends during 1979–2020. Leading principal component (PC1) of observed WSWS in northwest China underwent an obvious decreasing regime shift in 2003/04, as well as the winter BKS SIC. There are significant correlations between PC1 and BKS SIC on both interdecadal and interannual timescales. Decadal decrease of BKS SIC is accompanied with a positive phase of Scandinavia pattern (SP) with positive anomalies of 500‐hPa geopotential height over subtropical North Atlantic, Arctic‐Ural region and southern Asia, and negative anomalies over northern North Atlantic‐West Europe and Siberia regions. BKS SIC and SP significantly influenced the Northern China Wind Index (NCWI) at 200 hPa on interdecadal timescale, which was related to the PC1 on both interdecadal and interannual timescales after removing the signal of Arctic Oscillation. Meridional configuration of Siberian cyclonic anomaly and southern Asian anticyclonic anomaly associated with the winter BKS SIC could enhance the intensity of NCWI, which favoured the increase of observed WSWS in northwest China and thereby may contribute to the regime shift in 2003/04 of the leading mode. However, there were still some stations showing decreasing regime shift, which mainly located at the region of steeping terrain. These decreasing changes of WSWS may be related to circumfluence effect of terrain, urbanization and local microclimate, and these reasons still need further detailed study. [ABSTRACT FROM AUTHOR]

Two coupled climate models that participated in the CMIP6 project (Coupled Model Intercomparison Project Phase 6), the Earth System Model of Chinese Academy of Sciences version 2 (CAS-ESM2-0), and the Nanjing University of Information Science and Technology Earth System Model version 3 (NESM3) were assessed in terms of the impact of four new sea ice parameterization schemes. These four new schemes are related to air–ice heat flux, radiation penetration and absorption, melt ponds, and ice–ocean flux, respectively. To evaluate the effectiveness of these schemes, key sea ice variables with and without these new schemes, such as sea ice concentration (SIC) and sea ice thickness (SIT), were compared against observation and reanalysis products from 1980 to 2014. The simulations followed the design of historical experiments within the CMIP6 framework. The results revealed that both models demonstrated improvements in simulating Arctic SIC and SIT when the new parameterization schemes were implemented. The model bias of SIC in some marginal sea ice zones of the Arctic was reduced, especially during March. The SIT was increased and the transpolar gradient of SIT was reproduced. The changes in spatial patterns of SIC and SIT after adding new schemes bear similarities between the two coupled models. This suggests that the new schemes have the potential for broad application in climate models for simulation and future climate scenario projection, especially for those with underestimated SIT. [ABSTRACT FROM AUTHOR]

Revolutionary advances in artificial intelligence (AI) in the past decade have brought transformative innovation across science and engineering disciplines. In the field of Arctic science, we have witnessed an increasing trend in the adoption of AI, especially deep learning, to support the analysis of Arctic big data and facilitate new discoveries. In this paper, we provide a comprehensive review of the applications of deep learning in sea ice remote sensing domains, focusing on problems such as sea ice lead detection, thickness estimation, sea ice concentration and extent forecasting, motion detection, and sea ice type classification. In addition to discussing these applications, we also summarize technological advances that provide customized deep learning solutions, including new loss functions and learning strategies to better understand sea ice dynamics. To promote the growth of this exciting interdisciplinary field, we further explore several research areas where the Arctic sea ice community can benefit from cutting-edge AI technology. These areas include improving multimodal deep learning capabilities, enhancing model accuracy in measuring prediction uncertainty, better leveraging AI foundation models, and deepening integration with physics-based models. We hope that this paper can serve as a cornerstone in the progress of Arctic sea ice research using AI and inspire further advances in this field.

The features of the spatial distribution of the long-term average ice-free period (IFP) characteristics and their trends in the Laptev Sea are calculated and described quantitatively based on the microwave remote sensing data for 1979–2021. The spatial distribution of the long-term average duration of the IFP mainly follows the spatial distribution of the long-term average IFP start dates. The highest value of the long-term average duration of the IFP is observed in the areas where the Anabar–Lena and Western New Siberian stationary flaw polynyas are formed. It is shown that south of 78° N, the duration of the IFP increased statistically significantly by +17 ± 7 days/10 year on average. The areas of the Laptev Sea north of 80° N are characterized by statistically insignificant changes in IFP characteristics. The duration of the IFP extends due to both the shift of the IFP start dates to the earlier time (an average rate of changes is –10 ± 4 days/10 year) and the shift of the IFP end dates to a later time (an average rate of changes is +8 ± 4 days/10 year). [ABSTRACT FROM AUTHOR]

To effectively extract multi-scale information from observation data and improve computational efficiency, a multi-scale second-order autoregressive recursive filter (MSRF) method is designed. The second-order autoregressive filter used in this study has been attempted to replace the traditional first-order recursive filter used in spatial multi-scale recursive filter (SMRF) method. The experimental results indicate that the MSRF scheme successfully extracts various scale information resolved by observations. Moreover, compared with the SMRF scheme, the MSRF scheme improves computational accuracy and efficiency to some extent. The MSRF scheme can not only propagate to a longer distance without the attenuation of innovation, but also reduce the mean absolute deviation between the reconstructed sea ice concentration results and observations reduced by about 3.2 % compared to the SMRF scheme. On the other hand, compared with traditional first-order recursive filters using in the SMRF scheme that multiple filters are executed, the MSRF scheme only needs to perform two filter processes in one iteration, greatly improving filtering efficiency. In the two-dimensional experiment of sea ice concentration, the calculation time of the MSRF scheme is only 1/7 of that of SMRF scheme. This means that the MSRF scheme can achieve better performance with less computational cost, which is of great significance for further application in real-time ocean or sea ice data assimilation systems in the future. [ABSTRACT FROM AUTHOR]

The strong CO2 sink in Arctic Ocean plays a significant role in the global carbon budget. As a high-latitude oceanic ecosystem, the features of sea surface pCO2 and air-sea CO2 flux are significantly influenced by sea ice melt; however, our understanding of pCO2 evolution during sea ice melt remains limited. In this study, we investigate the dynamics of pCO2 during the progression of sea ice melt in the western Arctic Ocean based on data from two cruises conducted in 2010 and 2012. Our findings reveal substantial spatiotemporal variability in surface pCO2 on the Chukchi Sea shelf and Canada Basin, with a boundary along the shelf breaks at depths of 250-500 m isobaths. On the Chukchi Sea shelf, strong biological consumption dominates pCO2 variability. Moreover, in Canada Basin, the pCO2 dynamics are modulated by various processes. During the active sea ice melt stage before sea ice concentration decreases to 15%, biological production through photosynthetic processes and dilution of icemeltwater lead to a reduction in DIC concentration and subsequent decline in pCO2. Further, these effects are counteracted by the air-sea CO2 exchange at the sea surface which tends to increase seawater DIC and subsequently elevate surface pCO2. Compared to the pCO2 reduction resulting from biological production and dilution effects, the contribution of air-sea CO2 exchange is significantly lower. The combined effects of these factors have a significant impact on reducing pCO2 during this stage. Conversely, during the post sea ice melt stage, an increase in pCO2 resulting from high temperatures and air-sea CO2 exchange outweighs its decrease caused by biological production. Their combined effects result in a prevailing increase in sea surface pCO2. We argue that enhanced air-sea CO2 uptake under high wind speeds also contributes to the high sea surface pCO2 observed in 2012, during both active sea ice melt stage and post sea ice melt stage. The present study reports, for the first time, the carbonate dynamics and pCO2 controlling processes during the active sea ice melt stage. These findings have implications for accurate estimation of air-sea CO2 fluxes and improved modeling simulations within the Arctic Ocean. [ABSTRACT FROM AUTHOR]

To further understand the prediction skill for the interannual variability of the sea ice concentration (SIC) in specific regions of the Arctic, this paper evaluates the NCEP Climate Forecast System version 2 (CFSv2), in predicting the autumn SIC and its interannual variability over the Barents–East Siberian Seas (BES). It is found that CFSv2 presents much better prediction skill for the September SIC over BES than the Arctic as a whole at 1–6-month leads, and high prediction skill for the interannual variability of the SIC over BES is displayed at 1–2-month leads after removing the linear trend. CFSv2 can reasonably reproduce the relationship between the SIC over BES in September and such factors as the surface air temperature (SAT), 200-hPa geopotential height, sea surface temperature (SST), and North Atlantic Oscillation. In addition, it is found that the prescribed SIC initial condition in August as an input to CFSv2 is also essential. Therefore, the above atmospheric and oceanic factors, as well as an accurate initial condition of SIC, all contribute to a high prediction skill for SIC over BES in September. Based on a statistical prediction method, the contributions from individual predictability sources are further identified. The high prediction skill of CFSv2 for the interannual variability of SIC over BES is largely attributable to its accurate predictions of the SAT and SST, as well as a better initial condition of SIC. [ABSTRACT FROM AUTHOR]

We apply the Canny edge algorithm to imagery from the Utqiaġvik coastal sea ice radar system (CSIRS) to identify regions of open water and sea ice and quantify ice concentration. The radar-derived sea ice concentration (SIC) is compared against the (closest to the radar field of view) 25 km resolution NSIDC Climate Data Record (CDR) and the 1 km merged MODIS-AMSR2 sea ice concentrations within the ∼11 km field of view for the year 2022–2023, when improved image contrast was first implemented. The algorithm was first optimized using sea ice concentration from 14 different images and 10 ice analysts (140 analyses in total) covering a range of ice conditions with landfast ice, drifting ice, and open water. The algorithm is also validated quantitatively against high-resolution MODIS-Terra in the visible range. Results show a correlation coefficient and mean bias error between the optimized algorithm, the CDR and MODIS-AMSR2 daily SIC of 0.18 and 0.54, and ∼−1.0 and 0.7%, respectively, with an averaged inter-analyst error of ±3%. In general, the CDR captures the melt period correctly and overestimates the SIC during the winter and freeze-up period, while the merged MODIS-AMSR2 better captures the punctual break-out events in winter, including those during the freeze-up events (reduction in SIC). Remnant issues with the detection algorithm include the false detection of sea ice in the presence of fog or precipitation (up to 20%), quantified from the summer reconstruction with known open water conditions. The proposed technique allows for the derivation of the SIC from CSIRS data at spatial and temporal scales that coincide with those at which coastal communities members interact with sea ice. Moreover, by measuring the SIC in nearshore waters adjacent to the shoreline, we can quantify the effect of land contamination that detracts from the usefulness of satellite-derived SIC for coastal communities.

Sea ice concentration (SIC) is an important dimension for characterising the geographical features of the pan-Arctic region. Trends in SIC bring new opportunities for human activities in the Arctic region. In this paper, we propose a deep learning technology-based sea ice concentration prediction model, SICFormer, which can realise end-to-end daily sea ice concentration prediction. Specifically, the model uses a 3D-Swin Transformer as an encoder and designs a decoder to reconstruct the predicted image based on PixelShuffle. This is a new model architecture that we have proposed. Single-day SIC data from the National Snow and Ice Data Center (NSIDC) for the years 2006 to 2022 are utilised. The results of 8-day short-term prediction experiments show that the average Mean Absolute Error (MAE) of the SICFormer model on the test set over the 5 years is 1.89%, the Root Mean Squared Error (RMSE) is 5.99%, the Mean Absolute Percentage Error (MAPE) is 4.32%, and the Nash–Sutcliffe Efficiency (NSE) is 0.98. Furthermore, the current popular deep learning models for spatio-temporal prediction are employed as a point of comparison given their proven efficacy on numerous public datasets. The comparison experiments show that the SICFormer model achieves the best overall performance.

This research paper explores S M Nazmuz Sakib’s hypothesis of aerosol-sea ice feedback and its implications for climate system dynamics. The hypothesis suggests that changes in aerosol emissions significantly impact sea ice concentration and thickness in the Arctic, which, in turn, affect aerosol transport and deposition over the Tibetan Plateau. The paper presents a comprehensive analysis of the hypothesis, including the underlying facts, a proposed formula for the aerosol-sea ice feedback, and the potential variations of this feedback based on regional and temporal patterns of aerosol emission changes. Furthermore, the paper introduces the Sea Ice-Aerosol-Cloud Feedback (SIACF) Index and its application to historical incidents related to aerosol emission changes. The SIACF Index provides a quantitative measure to evaluate the influence of changes in aerosol emissions on sea ice concentration and thickness. The paper concludes by emphasizing the importance of testing this hypothesis through global aerosol models, observations, and historical incidents to gain a deeper understanding of aerosol-climate interactions and develop effective mitigation strategies.

During the COVID-19 lockdown, large industries in the Russian Arctic region were closed for two months, leading to a significant reduction in anthropogenic aerosol emissions. This decline in aerosols resulted in a decrease in the human-made aerosol cooling effect. The strict lockdown measures implemented after 18 March 2020 caused a drastic downward fall in considered aerosols. Consequently, there was an impact on air temperature, with temperature differences exceeding 3 K in November and February, while clear-sky top net solar flux values were greater than 13 W m−2. Throughout the lockdown period, the Russian Arctic's annual area average sea ice concentration was 7.72% in 2019 and 7.64% in 2020. The decrease in radiation-scattering pollutants contributed to an increase in global warming. Some pollutants, like sulfur, carbon dioxide, and methane, produced a cooling effect by mitigating greenhouse gases, which could affect the speed of sea ice melt. However, due to the short duration of the pandemic, it remains challenging to determine whether the aerosol changes accelerated or decelerated the sea ice melting process. HIGHLIGHTS COVID lockdown resulted in a drastic drop in the amount of anthropogenic aerosol emitted over the Russian Arctic.; Arctic haze was experienced during April 2020.; The decline in aerosol emissions, black carbon, and sulfur dioxide led to reductions in total aerosol cooling.; Net solar radiation flux showed values greater than 13 W m−2 during lockdown.; No significant effect of anthropogenic activities on the sea ice condition was observed.;

The strong CO2 sink in Arctic Ocean plays a significant role in the global carbon budget. As a high-latitude oceanic ecosystem, the features of sea surface pCO2 and air-sea CO2 flux are significantly influenced by sea ice melt; however, our understanding of pCO2 evolution during sea ice melt remains limited. In this study, we investigate the dynamics of pCO2 during the progression of sea ice melt in the western Arctic Ocean based on data from two cruises conducted in 2010 and 2012. Our findings reveal substantial spatiotemporal variability in surface pCO2 on the Chukchi Sea shelf and Canada Basin, with a boundary along the shelf breaks at depths of 250-500 m isobaths. On the Chukchi Sea shelf, strong biological consumption dominates pCO2 variability. Moreover, in Canada Basin, the pCO2 dynamics are modulated by various processes. During the active sea ice melt stage before sea ice concentration decreases to 15%, biological production through photosynthetic processes and dilution of ice melt water lead to a reduction in DIC concentration and subsequent decline in pCO2. Further, these effects are counteracted by the air-sea CO2 exchange at the sea surface which tends to increase seawater DIC and subsequently elevate surface pCO2. Compared to the pCO2 reduction resulting from biological production and dilution effects, the contribution of air-sea CO2 exchange is significantly lower. The combined effects of these factors have a significant impact on reducing pCO2 during this stage. Conversely, during the post sea ice melt stage, an increase in pCO2 resulting from high temperatures and air-sea CO2 exchange outweighs its decrease caused by biological production. Their combined effects result in a prevailing increase in sea surface pCO2. We argue that enhanced air-sea CO2 uptake under high wind speeds also contributes to the high sea surface pCO2 observed in 2012, during both active sea ice melt stage and post sea ice melt stage. The present study reports, for the first time, the carbonate dynamics and pCO2 controlling processes during the active sea ice melt stage. These findings have implications for accurate estimation of air-sea CO2 fluxes and improved modeling simulations within the Arctic Ocean.

Weather filters are frequently used to lessen the interference while using the ARTSIST Sea Ice (ASI) algorithm, which is based on Advanced Microwave Scanning Radiometer for EOS/Advanced Microwave Scanning Radiometer-2 (AMSR-E/AMSR-2) 89 GHz brightness temperature (Tb) data to calculate sea ice concentration (SIC) with high spatial resolution. However, the weather filters may cause low concentration sea ice to be mistaken for sea water, which decreases the accuracy of the SIC inversion. This paper proposed the ASI algorithm combined with the Bootstrap (BT) algorithm for SIC inversion to increase the accuracy of the SIC inversion. First, use the weather filters to screen out an area that has been disrupted by the external factors like clouds and water vapor. Second, the BT algorithm is used in this area to accomplish SIC inversion. Based on the proposed method, the Antarctic SIC results in February 2018 were obtained and compared with the SIC obtained from the ASI algorithm. The results demonstrated that the SIC inversion based on ASI algorithm combined with BT algorithm is practicable. By comparing the SIC results of Landsat-8 data, it was further verified that the SIC inversion based on ASI algorithm combined with BT algorithm has higher accuracy compared to the ASI algorithm. As a result, the accuracy of SIC inversion can be improved to some extent by combining the ASI algorithm and the BT algorithm.

The ongoing decline of sea ice in the Arctic has heightened the need for accurate sea‐ice forecasts to support environmental protection and resource development in the region and beyond. While deep learning has shown promise in seasonal sea‐ice forecasting, most of the existing models overlook the crucial influence of atmospheric factors, thereby limiting their ability to capture the intricate characteristics of the sea‐ice system and improve forecast accuracy. To address this deficiency, we propose an attention convolutional long short‐term memory ensemble network named Atsicn, which integrates atmospheric factors to enhance the precision of multi‐step seasonal sea‐ice concentration forecasts. Our findings reveal that Atsicn outperforms state‐of‐the‐art dynamic and statistical models, and demonstrates remarkable reliability in extreme years. Furthermore, the impact of atmospheric factors on sea‐ice forecasts exhibits significant seasonality, with a relatively minimal impact on forecasts from March to June, a growing impact from July to October, and a persistent yet diminishing impact from November to February. This study provides a practical approach for seasonal sea‐ice forecasts and contributes a new perspective to the understanding of the intricate interplay between sea ice and atmospheric factors. Plain Language Summary: Arctic sea ice is a crucial and sensitive component of the complex Earth system. The ongoing global warming has intensified the instability and unpredictability of the Arctic sea‐ice system, emphasizing the need for accurate sea‐ice forecasts to support environmental protection and resource development in the Arctic and beyond. In this study, we propose a deep learning approach that considers the influence of atmospheric factors on sea ice to improve the accuracy of the seasonal sea ice concentration forecast. Our findings reveal that the proposed model performs well and maintains remarkable reliability, even in extreme years. The impact of atmospheric factors on sea‐ice forecasts exhibits significant seasonality, with a relatively minimal impact on forecasts from March to June, a growing impact from July to October, and a persistent yet diminishing impact from November to February. Key Points: A deep learning ensemble network integrating atmospheric factors named Atsicn is proposed for the seasonal sea‐ice concentration forecastsThe proposed model outperforms state‐of‐the‐art dynamic and statistical models, and demonstrates remarkable reliability in extreme yearsThe results highlight the pronounced seasonality in the influence of atmospheric factors on the sea‐ice concentration forecasts [ABSTRACT FROM AUTHOR]

We conducted a comparison of sea ice concentration (SIC) in the Amundsen Sea using three satellite datasets: Hadley Centre's sea ice and sea surface temperature (HadISST1), Operational Sea Surface Temperature and Ice Analysis (OSTIA), and Advanced Microwave Scanning Radiometer 2 (AMSR2). HadISST1 has the longest time period, while AMSR2 has the shortest. In terms of grid resolution, HadISST1 has the coarsest resolution, while AMSR2 has the finest. The sea ice areas (SIAs) observed in HadISST1, OSTIA, and AMSR2 are similar. We studied the decadal variations in SICs by dividing the study period into four temporal segments. We investigated the differences between HadISST1 and OSTIA for each temporal segment. HadISST1 exhibited a more pronounced positive trend compared to OSTIA between 2005 and 2010. Additionally, we compared the interannual and seasonal variations in SICs between HadISST1 and OSTIA. Lastly, it should be noted that the Amundsen Sea polynya area varies across all three datasets. [ABSTRACT FROM AUTHOR]

The features of the spatial distribution of the long-term average ice-free period (IFP) characteristics and their trends in the Laptev Sea are calculated and described quantitatively based on the microwave remote sensing data for 1979–2021. The spatial distribution of the long-term average duration of the IFP mainly follows the spatial distribution of the long-term average IFP start dates. The highest value of the long-term average duration of the IFP is observed in the areas where the Anabar–Lena and Western New Siberian stationary flaw polynyas are formed. It is shown that south of 78° N, the duration of the IFP increased statistically significantly by +17 ± 7 days/10 year on average. The areas of the Laptev Sea north of 80° N are characterized by statistically insignificant changes in IFP characteristics. The duration of the IFP extends due to both the shift of the IFP start dates to the earlier time (an average rate of changes is –10 ± 4 days/10 year) and the shift of the IFP end dates to a later time (an average rate of changes is +8 ± 4 days/10 year). [ABSTRACT FROM AUTHOR]

Arctic sea ice prediction is of great practical significance in facilitating Arctic route planning, optimizing fisheries management, and advancing the field of sea ice dynamics research. While various deep learning models have been developed for sea ice prediction, they predominantly operate at the seasonal or sub-seasonal scale, often focusing on localized areas, and few cater to full-region daily-scale prediction. This study introduces the use of spatiotemporal sequence data prediction models, namely, the convolutional LSTM (ConvLSTM) and predictive recurrent neural network (PredRNN), for the prediction of sea ice concentration (SIC). Our analysis reveals that, when solely utilizing SIC historical data as the input, the ConvLSTM model outperforms the PredRNN model in SIC prediction. To enhance the models' capacity to capture spatiotemporal relationships between multiple variables, we expanded the range of input data types to form the ConvLSTM-multi and PredRNN-multi models. Experimental findings demonstrate that the prediction accuracy of the four models significantly surpasses the CMIP6 model in three prospective climate scenarios (SSP126, SSP245, and SSP585). Of the four models, the ConvLSTM-multi model excels in assimilating the influence of reanalysis data on sea ice within the sea ice edge region, thus exhibiting superior performance than the PredRNN-multi model in predicting daily Arctic SIC over the subsequent 10 days. Furthermore, sensitivity tests on various model parameters highlight the substantial impact of sea surface temperature and prediction date on the accuracy of daily sea ice prediction, and meteorological and oceanographic parameters primarily affect the prediction accuracy of the thin-ice region at the edge of the sea ice. [ABSTRACT FROM AUTHOR]

Growing interest in finding the optimal route through the arctic ocean, and sea ice concentration is also emerging as a factor to be considered. In this paper, an algorithm to calculate the sea ice concentration was developed based on the images taken during the Arctic voyage of the Korean icebreaker ARAON in July 2019. A sea ice concentration calculation program was developed using the image processing functions in open-source image processing library, called OpenCV. To develop the algorithm, parameter studies were conducted on red, green, blue (RGB) color space and hue, saturation, value (HSV) color space, and k-means clustering. To verify the algorithm for sea ice concentration calculation, it was applied to images taken during Araon's Arctic voyages. Lens curvature and view point were corrected through camera calibration. To improve the accuracy of sea ice concentration calculation, a binarization model based on random forest was proposed. A parameter study for training image numbers and tree numbers was conducted to establish the random forest model. The calculated sea ice concentrations by random forest and k-means clustering were compared and discussed. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Remote Sensing is the property of Editorial Office of Journal of Remote Sensing & Science Publishing Co. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this study, we explore the interannual variability of Sea Surface Salinity (SSS) in the Dotson–Getz Trough located in West Antarctica, focusing on the month of February. Utilizing the oceanic analysis product EN4, we first validate the EN4 SSS with data from a singular ship-based survey, and then delve into potential factors that may influence SSS, with a particular emphasis on surface freshwater flux, sea ice concentration (SIC), and also the surface stress curl, which will induce upwelling via Ekman transport to affect the SSS. Our findings primarily indicate a link between SSS and sea ice concentration, showcasing a negative correlation where the peak (average) coefficient is around −0.6 (−0.4), further affirming the substantial interannual variability of SSS in this region.

The waters near the Antarctic Peninsula are characterized with unique oceanographic conditions and rich krill resources. Based on samples collected around the South Shetland Islands (SSI) in austral summer of 2019/2020, the net- phytoplankton community structure and relevant major biotic and abiotic influencing factors were investigated. Eighty-one taxa were identified by light microscope, and diatoms were the most abundant group. The most abundant species were Chaetoceros atlanticus, C. criophilus, C. dichaeta, Fragilariopsis kerguelensis and Pseudo−nitzschia lineola. The abundance and Shannon-Weaver index of net-phytoplankton ranged from 100 to 2.64×107 cells/m3 and 0.0747 to 4.0176 respectively, with significantly low values detected in the Bransfield Strait (BS) and high values in the west of the SSI. The dissimilarity was mainly caused by the differences in abundance of diatoms (including Thalassiothrix antarctica and the species in genus Rhizosolenia, Chaetoceros, Fragilariophsis). These diatoms and Dictyocha speculum were found in higher abundance in the west of the SSI, while Corethron pennatum and cryptophytes were found in higher abundance in the BS. Combined with acoustic density of krill and environmental data (Sea Surface Temperature and Sea Ice Concentration). The multivariate analysis suggested that phytoplankton community was positively affected by the SST, and the acoustic- derived krill density would be associated with the spatial distribution of pennate diatoms. This study enhances the knowledge about the selective feeding for krill and provides ecological implications for the Antarctic marine ecosystem.

Abstract This study evaluated feasibility statistically and analyzed, during the freezing period, the relationship between brightness temperature (Tb) data of the 37V polarisation and the GR3719 (Gradient Ratio 37V and 19V) obtained by Special Sensor Microwave/Imager from F11 and F13 satellites with sea ice thickness (SIT) data obtained in the Weddell Sea through Antarctic Sea Ice Processes and Climate program. The multiple linear regression (MLR) was applied at 1,520 points, with 70% of these points being randomly separated to generate the MLR and 30% to carry out the validation. To perform the temporal mapping, the MLR was applied only to pixels with sea ice concentration (SIC) ≥ 90%, obtained through the fraction image calculated from the spectral linear mixing model (SLMM) using the Tb in the channels and polarizations 19H, 19V and 37V. The results of the SLMM validation process for estimating the SIC were σ = 10.5%, RMSE = 11.0%, and bias = -2.8%, and the SIT based on the MLR, the results were R² = 0.57, RMSE = 0.268 m, and bias = 0.103 m. In the SIT mapping, we highlight the trend of thickness reduction on the east coast of the Antarctic Peninsula during the period 1992–2009.

Using delay‐Doppler maps of Global Navigation Satellite Systems Reflectometry (GNSS‐R) from the TechDemoSat‐1 satellite and considering sea ice and ocean interaction, an innovative method for retrieval of Arctic sea ice concentration (SIC) based on a deep neural network is proposed. This retrieval method shows the potential of future GNSS‐R applications for Arctic missions. Compared with SIC products from Hamburg University, the root mean square errors (RMSE) of retrieved results in March and June 2016 are 0.0284 and 0.0415, respectively. When the retrieved GNSS‐R SIC data are added into the assimilation as supplementary passive microwave remote‐sensing data, it has a positive influence on improving the accuracy of the Arctic SIC forecast. Especially in some edge regions of sea ice, when compared to only assimilating the remote‐sensing data, the regional RMSE of joint assimilation has a maximum decrease of approximately 17% in the 24‐hr forecast time, and over 5% in 72‐hr. Plain Language Summary: Accurate sea ice forecasting is critical to understanding the risks of Arctic maritime activity and to improving climate forecasting in the mid‐high latitudes of the Northern Hemisphere. Data assimilation of sea ice observations is an effective way to improve the numerical model forecast results, and its effect is related to both the quality and the quantity of observations. As a new remote sensing technology, Global Navigation Satellite Systems Reflectometry (GNSS‐R) shows great potential in sea ice remote sensing. Based on GNSS‐R data from TechDemoSat‐1, we combined Delay‐doppler number eigenvalue method and deep neural network to propose an innovative method for estimating sea ice concentration (SIC) at GNSS‐R subsatellite points. The estimated SIC has reasonable accuracy and provides more information at subsatellite points near the grid points of the passive microwave remote sensing product. When it is joint assimilated with passive microwave remote sensing SIC, it has a positive influence on the forecast effect of SIC in the region with rapid change of sea ice. This is the first time that GNSS‐R data has been applied to the prediction of Arctic SIC, which is of great value in promoting the application and development of GNSS‐R in Arctic sea ice forecast. Key Points: An innovative method for retrieving sea ice concentration (SIC) at Global Navigation Satellite Systems Reflectometry (GNSS‐R) subsatellite points is presented, considering the marine factorsThe joint assimilation of retrieved GNSS‐R SIC and passive microwave remote sensing SIC is a useful means to improve the forecast accuracyMore GNSS‐R SIC data in areas with large SIC gradient will bring more useful information to the sea ice forecast [ABSTRACT FROM AUTHOR]

The Tibetan Plateau (TP) is characterized by heavily local dust activities, however, the mechanism of interannual variations of winter dust frequency over the TP remain poorly understood. Previous studies showed the autumn Arctic sea ice could significantly influence the winter climate over Eurasia. Whether autumn sea ice affects winter dust activity over the TP or not? Here, we used an integrated surface database to investigate possible mechanisms for interannual variability in the frequency of winter dust events above the TP. This variability, which is thought to be mainly caused by local dust emissions, shows significant correlations with sea ice concentration (SIC) in the Barents and Kara Seas during the preceding autumn. Low Barents–Kara SIC is accompanied by reduced snow depth over northern Eurasia between autumn and winter, which can enhance the Eurasian mid-latitude westerly jet stream. This strengthening increases the cyclogenesis and occurrence of strong surface wind speeds in winter, especially over the TP. In addition, a lower SIC is closely associated with reduced precipitation and snow cover in late autumn and winter over the TP, which in turn enhances warming of the land surface and reduces the area of frozen ground. These anomalies in atmospheric circulation patterns and local surface conditions promote dust events above the TP during winter. The ensemble means of Atmospheric Model Intercomparison Project experiments from Phase 6 of the Coupled Model Inter-comparison Project and the Community Atmosphere Model version 4 can generally reproduce the atmospheric circulation anomalies associated with decreased Barents–Kara SIC. This study reveals the crucial effect that SIC anomalies in the Barents and Kara Seas have on winter dust activities over the TP.

Sea ice hinders the navigability of the Arctic, especially in winter and spring. However, three Arc7 ice-class Liquefied Natural Gas carrying vessels safely transited the Northern Sea Route (NSR) without icebreaker assistance in January 2021. More and more Arc7 ice-class vessels are putting into the transit services in the NSR. Therefore, it is necessary to analyze sea-ice conditions and their impact on navigation during wintertime, and the future navigability of Arc7 ice-class vessels along the NSR during winter and spring. Based on sea ice datasets from satellite observations and a model using data assimilation, we explored the sea-ice conditions and their impact during the first three successful commercial voyages through the NSR in winter. In addition, we analyzed the sea ice variation and estimated navigability for Arc7 ice-class vessels in the NSR from January to June of the years 2021–2050 using future projections of the sea-ice cover by the Coupled Model Inter-comparison Project Phase 6 (CMIP6) models under two emission scenarios (SSP2-4.5 and SSP5-8.5). The results reveal lower sea ice thickness and similar sea ice concentration during these three transits relative to the past 42 years (from 1979 to 2020). We found the thickness has a larger impact on the vessels’ speeds than sea ice concentration. Very likely sea ice thickness played a larger role than the sea ice concentration for the successful transit of the NSR in winter 2021. Future projections suggest sea ice thickness will decrease further in most regions of the NSR from January to June under all scenarios enabling increased navigability of the NSR for Arc7 ice-class vessels. Such vessels could transit through the NSR from January to June under all scenarios by 2050, while some areas near the coast of East Siberian Sea remain inaccessible for Arc7 ice-class vessels in spring (April and May). These findings can support the strategic planning of shipping along the NSR in winter and spring.

The sea ice concentration (SIC) over the Weddell Sea displays obvious intraseasonal variability during austral autumn with a dominant frequency of 5–20 days. The intraseasonal SIC variability is manifested as development of sea ice anomalies from the northeastern Antarctic Peninsula toward the central Weddell Sea. Rossby wave train associated with the internal atmospheric variability featuring circumglobal zonal wavenumber 4 pattern induces the development of anomalous surface winds and sea ice drifting, leading to the SIC anomalies directly. The thermodynamical processes associated with the intraseasonal SIC variability include the air temperature, the moisturizing of the lower troposphere, the latent and sensible heating, and the downwelling longwave radiation, which contribute to the variation of the local greenhouse effect over the Weddell Sea, and thus lead to the sea ice variation. The investigation of the intraseasonal SIC variability helps the understanding of sea‐ice‐atmosphere interaction over the Antarctic region. Plain Language Summary: The variation of the Antarctic sea ice can modulate the atmospheric circulation over the Southern Hemisphere and play an important role in the sea‐ice‐atmosphere interaction. The intraseasonal sea ice concentration (SIC) variability over the Antarctic is more evident and with larger extent over the Weddell Sea during austral autumn, featuring a dominant frequency of 5–20 days. The evolution of the intraseasonal SIC variability manifests as SIC anomalies appearing over the northeastern Antarctic Peninsula and then moving eastward to the central Weddell Sea. It is controlled by the propagation of wave train around the Antarctic, which induces surface wind anomalies and sea ice movement. Besides, the air temperature, moisture, the latent and sensible heating, and the doweling longwave radiation can also contribute to the increase or melting of the sea ice through the modulation of the local greenhouse effect. Key Points: The intraseasonal sea ice concentration (SIC) variability over the Weddell Sea manifests an evident frequency of 5–20 days in austral autumnRossby wave train induces anomalous surface winds and sea ice drifting, leading to the intraseasonal SIC anomalies over the Weddell SeaThermodynamical processes including moisture and longwave radiation can modulate local greenhouse effect and intraseasonal SIC anomalies [ABSTRACT FROM AUTHOR]

With the goal of automatic sea ice mapping during the summer sea ice melt cycle, this study involved designing a fully automatic sea ice segmentation method based on a deep learning semantic segmentation network applicable to summer SAR images, which achieved high accuracy and the fully automatic extraction of sea ice segmentation during the summer ice melt cycle by optimizing the process, improving the pixel-level semantic segmentation network, and introducing high-resolution sea ice concentration features. Firstly, a convolution-based, high-resolution sea ice concentration calculation method is proposed and was applied to the deep learning task. Secondly, the proposed DF-UHRNet network was improved upon by designing high- and low-level fusion modules, introducing an attention mechanism, and reducing the number of convolution layers and other operations, and it can effectively fuse high- and low-scale semantic features and global contextual information based on reducing the overall number of network parameters, enabling it to achieve pixel-level classification. The results show that this method meets the needs associated with the automatic mapping and high-precision classification of thin ice, one-year ice, open water, and multi-year ice and effectively reduces the model size. [ABSTRACT FROM AUTHOR]

Arctic sea ice export is important for the redistribution of freshwater and sea ice mass. Here, we use the sea ice thickness, sea ice velocity, and sea ice concentration (SIC) to estimate the exported sea ice volume through the Fram Strait from 2011 to 2018. We further analyse the contributions of the sea ice thickness, velocity and concentration to sea ice volume export. Then, the relationships between atmospheric circulation indices (Arctic Oscillation (AO), North Atlantic Oscillation (NAO), and Arctic Dipole (AD)) and the sea ice volume export are discussed. Finally, we analyse the impact of wind-driven oceanic circulation indices (Ekman transport (ET)) on the sea ice volume export. The sea ice volume export rapidly increases in winter and decreases in spring. The exported sea ice volume in winter is likely to exceed that in spring in the future. Among sea ice thickness, velocity and SIC, the greatest contribution to sea ice export comes from the ice velocity. The exported sea ice volume through the zonal gate of the Fram Strait (which contributes 97% to the total sea ice volume export of the Fram Strait) is much higher than that through the meridional gate (3%) because the sea ice flowing out of the zonal gate has the characteristics of a high thickness (mainly thicker than 1 m), a high velocity (mainly faster than 0.06 m/s) and a high concentration (mainly higher than 80%). The AD and ET explain 53.86% and 38.37% of the variation in sea ice volume export, respectively. [ABSTRACT FROM AUTHOR]

A comprehensive assessment of Antarctic sea ice cover prediction is conducted for twelve CMIP6 models under the scenario of SSP2-4.5, with a comparison to the observed data from the Advanced Microwave Scanning Radiometer 2 (AMSR2) during 2015–2021. In the quantitative evaluation of sea ice extent (SIE) and sea ice area (SIA), most CMIP6 models show reasonable variation and relatively small differences compared to AMSR2. CMCC-CM4-SR5 shows the highest correlation coefficient (0.98 and 0.98) and the lowest RMSD (0.98 × 106 km2 and 1.07 × 106 km2) for SIE and SIA, respectively. In the subregions, the models with the highest correlation coefficient and the lowest RMSD for SIE and SIA are inconsistent. Most models tend to predict smaller SIE and SIA compared to the observational data. GFDL-CM4 and FGOALS-g3 show the smallest mean bias (−4.50 and −1.21 × 105 km2) and the most reasonable interannual agreement of SIE and SIA with AMSR2, respectively. In the assessment of sea ice concentration (SIC), while most models can accurately predict the distribution of large SIC surrounding the Antarctic coastal regions, they tend to underestimate SIC and are unable to replicate the major patterns in the sea ice edge region. GFDL-CM4 and FIO-ESM-2-0 exhibit superior performance, with less bias (less than −5%) and RMSD (less than 23%) for SIC in the Antarctic. GFDL-CM4, FIO-ESM-2-0, and CESM2 exhibit relatively high positive correlation coefficients exceeding 0.60 with the observational data, while few models achieve satisfactory linear trend prediction of SIC. Through the comparison with RMSD, Taylor score (TS) consistently evaluates the Antarctic sea ice cover and proves to be a representative statistical indicator and applicable for its assessment. Based on comprehensive assessments of sea ice cover, CESM2, CMCC-CM4-SR5, FGOALS-g3, FIO-ESM-2-0, and GFDL-CM4 demonstrate more reasonable prediction performance. The assessment findings enhance the understanding of the uncertainties associated with sea ice in the CMIP6 models and highlighting the need for a meticulous selection of the multimodel ensemble. [ABSTRACT FROM AUTHOR]

The safe and efficient navigation of ships traversing the Northern Sea Route demands accurate information regarding sea ice concentration. However, the sea ice concentration forecasts employed to support such navigation are often flawed. To address this challenge, this study advances a statistical interpolation method aimed at reducing errors arising from traditional interpolation approaches. Additionally, this study introduces an autoregressive integrated moving average model, derived from ERA5 reanalysis data, for short-term sea ice concentration forecasts along the Northern Sea Route. The validity of the model has been confirmed through comparison with ensemble experiments from the Coupling Model Intercomparison Project Phase 5, yielding reliable outcomes. The route availability is assessed on the basis of the sea ice concentration forecasts, indicating that the route will be available in the upcoming years. The proposed statistical models are also shown the capacity to facilitate effective management of Arctic shipping along the Northern Sea Route. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Remote Sensing is the property of Editorial Office of Journal of Remote Sensing & Science Publishing Co. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Abstract Using delay‐Doppler maps of Global Navigation Satellite Systems Reflectometry (GNSS‐R) from the TechDemoSat‐1 satellite and considering sea ice and ocean interaction, an innovative method for retrieval of Arctic sea ice concentration (SIC) based on a deep neural network is proposed. This retrieval method shows the potential of future GNSS‐R applications for Arctic missions. Compared with SIC products from Hamburg University, the root mean square errors (RMSE) of retrieved results in March and June 2016 are 0.0284 and 0.0415, respectively. When the retrieved GNSS‐R SIC data are added into the assimilation as supplementary passive microwave remote‐sensing data, it has a positive influence on improving the accuracy of the Arctic SIC forecast. Especially in some edge regions of sea ice, when compared to only assimilating the remote‐sensing data, the regional RMSE of joint assimilation has a maximum decrease of approximately 17% in the 24‐hr forecast time, and over 5% in 72‐hr.

Two coupled climate models that participated in the CMIP6 project (Coupled Model Intercomparison Project Phase 6), the Earth System Model of Chinese Academy of Sciences version 2 (CAS-ESM2-0), and the Nanjing University of Information Science and Technology Earth System Model version 3 (NESM3) were assessed in terms of the impact of four new sea ice parameterization schemes. These four new schemes are related to air–ice heat flux, radiation penetration and absorption, melt ponds, and ice–ocean flux, respectively. To evaluate the effectiveness of these schemes, key sea ice variables with and without these new schemes, such as sea ice concentration (SIC) and sea ice thickness (SIT), were compared against observation and reanalysis products from 1980 to 2014. The simulations followed the design of historical experiments within the CMIP6 framework. The results revealed that both models demonstrated improvements in simulating Arctic SIC and SIT when the new parameterization schemes were implemented. The model bias of SIC in some marginal sea ice zones of the Arctic was reduced, especially during March. The SIT was increased and the transpolar gradient of SIT was reproduced. The changes in spatial patterns of SIC and SIT after adding new schemes bear similarities between the two coupled models. This suggests that the new schemes have the potential for broad application in climate models for simulation and future climate scenario projection, especially for those with underestimated SIT.

Widening and increasing extent of the marginal ice zone (MIZ), a transitional area between the open ocean and the pack ice, underscores the scientific significance of observing the MIZ. In the present study, we employed passive microwave (PM) and synthetic aperture radar (SAR) sea ice concentration (SIC) in the Greenland Sea and Beaufort Sea in November 2021 to detect the MIZ using two different MIZ definitions: SIC threshold-based (MIZ _t ) and SIC anomaly-based (MIZ _σ ). This study is the first to compare the SIC threshold-based with SIC anomaly-based MIZ definition using two different sources of SIC data. Our findings reveal that the SIC anomaly-based definition delineates a spatially extensive MIZ, capturing SIC variation attributed to sea ice growth. We also found that SAR data, compared to PM data, consistently identifies a broader MIZ region and is less sensitive to the threshold for the SIC anomaly standard deviation, underscoring the importance of selecting the appropriate MIZ definition.