Sea ice forecasts are operationally produced using physical-based models, but these forecasts are often not accurate enough for maritime operations. In this study, we developed a statistical correction technique (also called calibration) using machine learning in order to improve the skill of short-term (up to 10 days) sea ice concentration forecasts produced by the TOPAZ4 model. This technique allows to reduce the errors from the TOPAZ4 sea ice concentration forecasts by 41 % on average.       

Methods have been developed for calibrating sea ice drift forecasts from an operational prediction system using machine learning algorithms. These algorithms use predictors from sea ice concentration observations during the initialization of the forecasts, sea ice and wind forecasts, and some geographical information. Depending on the calibration method, the mean absolute error is reduced between 3.3 % and 8.0 % for the direction and between 2.5 % and 7.1 % for the speed of sea ice drift.

This study presents new methods to assess how well observations of sea ice expansion are reproduced by results from models. The aim is to provide information about the quality of forecasts for changes in the sea ice extent to operators in or near ice-infested waters. A test using 2 years of model results demonstrates the practical applicability and usefulness of the methods that are presented.

Evaluation of the vertical precipitation rate profiles of CloudSat radar by comparison with two surface-based micro-rain radars (MRR) located at two antarctic stations gives a near-perfect correlation between both datasets, even though climatic and geographic conditions are different for the stations. A better understanding and reassessment of CloudSat uncertainties ranging from −13 % up to +22 % confirms the robustness of the CloudSat retrievals of snowfall over Antarctica.