Development and validation of a global thin ice thickness algorithm for SMMR

Abstract

Abstract Thin ice thickness algorithms using satellite passive microwave radiometers are frequently used to quantify sea ice production in coastal polynyas. The objective of this paper is to develop and validate a thin ice thickness algorithm using only low-frequency (LF) data (19 and 37 GHz channels), applicable back to 1978 when the Scanning Multichannel Microwave Radiometer (SMMR) observations started. We compared several recent satellite observations for 22 cases of typical polynya events in the Southern Ocean and developed a new method to discriminate three types of thin ice (active frazil, thin solid ice, and a mixture of the two) using only LF data. Combined with ice type discrimination, the obtained LF ice thickness (20 cm or less) has a bias of less than 1 cm and an RMSD of about 5 cm, as shown by validation data from 24 cases of typical polynya events that occurred prior to 1992 in the Southern Ocean and the Bering Sea. Although the error in ice type discrimination with LF data is relatively large in comparison with previous studies, these results still suggest that the LF ice thickness could be a continuous global data since 1978. We also developed a false thin ice detection algorithm using only LF data and created daily false thin ice masks from 1978 to the present, which are also essential for estimating global sea ice production. The use of our new LF algorithm will allow quantification of the global sea ice production over nearly half a century since 1978.