Search

An operational synoptic-scale sea ice forecasting system for the Southern Ocean, namely Southern Ocean Ice Prediction System (SOIPS), has been developed to support ship navigation in the Antarctic sea ice zone. Practical application of the SOIPS forecasts had been implemented for the 38th Chinese National Antarctic Research Expedition for the first time. The SOIPS is configured on an Antarctic regional sea-ice‒ocean‒ice-shelf coupled model and an ensemble-based Localized Error Subspace Transform Kalman Filter data assimilation model. Daily near-real-time satellite sea ice concentration observations are assimilated into the SOIPS to update sea ice concentration and thickness in the 12 ensemble members of model state. By evaluating the SOIPS performance on forecasting sea ice metrics in a complete melt-freeze cycle from October 1, 2021 to September 30, 2022, this study shows that the SOIPS can provide reliable Antarctic sea ice forecasts. In comparison with the OSISAF data, annual mean root mean square errors of the sea ice concentration forecasts at leading time of up to 168-hour are lower than 0.19, and the integrated ice-edge errors of sea ice forecasts in most freezing months at leading times of 24-hour and 72-hour maintain around 0.5 × 106 km2 and below 1.0 × 106 km2, respectively. With respect to the scarce ICESat-2 observations, the mean absolute errors of the sea ice thickness forecasts at leading time of 24-hour are lower than 0.3 m, which is in range of the ICESat-2 uncertainties. Specifically, the SOIPS has a promised capacity in forecasting sea ice drift, both in magnitude and direction. The derived sea ice convergence rate forecasts have a high potential in supporting ship navigation on local fine scale. [ABSTRACT FROM AUTHOR]

Due to global warming, the Arctic sea ice extent (SIE) is rapidly decreasing each year. According to the Intergovernmental Panel on Climate Change (IPCC) climate model projections, the summer Arctic will be nearly sea-ice-free in the 2050s of the 21st century, which will have a great impact on global climate change. As a result, accurate predictions of Arctic sea ice are of significant interest. In most current studies, the majority of deep-learning-based SIE prediction models focus on one-step prediction, and they not only have short lead times but also limited prediction skill. Moreover, these models often lack interpretability. In this study, we construct the Ice temporal fusion transformer (IceTFT) model, which mainly consists of the variable selection network (VSN), the long short-term memory (LSTM) encoder, and a multi-headed attention mechanism. We select 11 predictors for the IceTFT model, including SIE, atmospheric variables, and oceanic variables, according to the physical mechanisms affecting sea ice development. The IceTFT model can provide 12-month SIE directly, according to the inputs of the last 12 months. We evaluate the IceTFT model from the hindcasting experiments for 2019–2021 and prediction for 2022. For the hindcasting of 2019–2021, the average monthly prediction errors are less than 0.21 ×106 km 2 , and the September prediction errors are less than 0.1 ×106 km 2 , which is superior to the models from Sea Ice Outlook (SIO). For the prediction of September 2022, we submitted the prediction to the SIO in June 2022, and IceTFT still has higher prediction skill. Furthermore, the VSN in IceTFT can automatically adjust the weights of predictors and filter spuriously correlated variables. Based on this, we analyze the sensitivity of the selected predictors for the prediction of SIE. This confirms that the IceTFT model has a physical interpretability. [ABSTRACT FROM AUTHOR]

The satellite observations unveiled that the July sea ice extent of the Arctic shrank to the lowest value, since 1979, in 2020 with a major ice retreat in the Eurasian shelf seas including Kara, Laptev, and East Siberian seas. Based on the ERA-5 reanalysis products, we explored the impacts of warm and moist air-mass transport on this extreme event. The results revealed that anomalously high energy and moisture converged into these regions in the spring months (April to June) of 2020, leading to a burst of high moisture content and warming within the atmospheric column. The convergence is accompanied by local enhanced downward longwave surface radiation and turbulent fluxes, which is favorable for initiating an early melt onset in the region with severe ice loss. Once the melt begins, solar radiation plays a decisive role in leading to further sea ice depletion due to ice–albedo positive feedback. The typical trajectories of the synoptic cyclones that occurred on the Eurasian side in spring 2020 agree well with the path of atmospheric flow. Assessments suggest that variations in characteristics of the spring cyclones are conducive to the severe melt of sea ice. We argue that large-scale atmospheric circulation and synoptic cyclones acted in concert to trigger the exceptional poleward transport of total energy and moisture from April to June to cause this record minimum of sea ice extent in the following July. [ABSTRACT FROM AUTHOR]

The Arctic regional coupled sea-ice–ocean–atmosphere model (ArcIOAM) has been developed to provide reliable Arctic sea ice prediction on seasonal timescales. The description and implementation of ArcIOAM and its preliminary results for the year of 2012 are presented in this paper. In the ArcIOAM configuration, the Community Coupler 2 (C-Coupler2) is used to couple the Arctic sea-ice–oceanic configuration of the MITgcm (Massachusetts Institute of Technology general circulation model) with the Arctic atmospheric configuration of the Polar WRF (Weather Research and Forecasting) model. A scalability test is performed to investigate the parallelization of the coupled model. As the first step toward reliable Arctic seasonal sea ice prediction, ArcIOAM, implemented with two-way coupling strategy along with one-way coupling strategy, is evaluated with respect to available observational data and reanalysis products for the year of 2012. A stand-alone MITgcm run with prescribed atmospheric forcing is performed for reference. From the comparison, all the experiments simulate reasonable evolution of sea ice and ocean states in the Arctic region over a 1-year simulation period. The two-way coupling has better performance in terms of sea ice extent, concentration, thickness and sea surface temperature (SST), especially in summer. This result indicates that sea-ice–ocean–atmosphere interaction plays a crucial role in controlling Arctic summertime sea ice distribution. [ABSTRACT FROM AUTHOR]

The Pacific sector of the Arctic Ocean (PA, hereafter) is a region sensitive to climate change. Given the alarming changes in sea ice cover during recent years, knowledge of sea ice loss with respect to ice advection and melting processes has become critical. With satellite-derived products from the National Snow and Ice Center (NSIDC), a 38-year record (1979–2016) of the loss in sea ice area in summer within the Pacific-Arctic (PA) sector due to the two processes is obtained. The average sea ice outflow from the PA to the Atlantic-Arctic (AA) Ocean during the summer season (June–September) reaches 0.173×106 km 2 , which corresponds to approximately 34 % of the mean annual export (October to September). Over the investigated period, a positive trend of 0.004×106 km 2 yr -1 is also observed for the outflow field in summer. The mean estimate of sea ice retreat within the PA associated with summer melting is 1.66×106 km 2 , with a positive trend of 0.053×106 km 2 yr -1. As a result, the increasing trends of ice retreat caused by outflow and melting together contribute to a stronger decrease in sea ice coverage within the PA (0.057×106 km 2 yr -1) in summer. In percentage terms, the melting process accounts for 90.4 % of the sea ice retreat in the PA in summer, whereas the remaining 9.6 % is explained by the outflow process, on average. Moreover, our analysis suggests that the connections are relatively strong (R=0.63), moderate (R=-0.46), and weak (R=-0.24) between retreat of sea ice and the winds associated with the dipole anomaly (DA), North Atlantic Oscillation (NAO), and Arctic Oscillation (AO), respectively. The DA participates by impacting both the advection (R=0.74) and melting (R=0.55) processes, whereas the NAO affects the melting process (R=-0.46). [ABSTRACT FROM AUTHOR]