Estimation of Sea Ice Concentration in the Arctic Using SARAL/AltiKa Data

Sea ice concentration (SIC) is important in determining important climate variables. Together with sea ice thickness, fluxes between air and sea as well as heat transfer through the atmosphere can be determined. We present here an adaptation of the NASA bootstrap (BT) algorithm with significant necessary modifications to estimate daily and monthly SIC in the Arctic for winter and summer seasons separately. Brightness temperature (Tb) data from the dual-frequency microwave radiometer (23.8 and 37 GHz) onboard the SARAL/AltiKa satellite for the period from April 2013 to December 2020 were used to estimate SIC. We compared our daily SIC product quantitatively with the daily National Snow Ice and Data Centre Climate Data Record (NSIDC CDR) SIC, the Ocean and Sea Ice Satellite Application Facility (OSISAF) SIC, and the ARTIST Sea Ice Algorithm SIC (ASI) (3.125 and 6.25 km resolutions). Root-mean-square errors (RMSEs) are 11% (versus NSIDC CDR), 11% (versus OSISAF), and 13% (versus ASI 3.125 and 6.25 km); for winter months about 9% (versus NSIDC CDR), 8% (versus OSISAF), and 11% (versus ASI 3.125 and 6.25 km); and for summer months about 15% (NSIDC CDR), 16% (OSISAF), and 16% (ASI 3.125 and 6.25 km). A bias correction method was performed on our product using NSIDC CDR SIC data, which resulted in an improvement of 6% RMSE. The algorithm was also analyzed for 14 regions in the Arctic to gain confidence in the methodology employed. The results show that our SIC is of good quality and can be used as a basic dataset for SIC records.