Haiyang Kaifa yu guanli (Jun 2024)
Retrieval of Snow Depth in Arctic Multi-year Ice Based on SMRT Model
Abstract
The L-band microwave radiative transfer model can simulate the physical propagation process of microwave radiation signals in the Arctic ice and snow, and serves as an effective tool to invert the snow thickness of the Arctic ice. However, at present, the models available for L-band brightness temperature simulation over Arctic sea ice and snow cover are scarce, which greatly limits the development of snow depth retrieval algorithms based on L-band microwave radiative transfer models. The Snow Microwave Radiative Transfer model (SMRT), developed under the sponsorship of ESA in 2018, has been proved to be suitable for L-band brightness temperature simulation on Arctic sea ice and snow, but there is no research using this model for Arctic snow depth. Therefore, in this paper, a new generation of microwave radiative transfer model of snow, SMRT, developed in 2018, is used to carry out the retrieval research of multi-year ice snow depth in the Arctic based on the L-band radiative transfer model. Based on the SMRT model, the key process of forward simulation is improved to simplify the calculation and improve the accuracy of the simulation. Then the L-band model combined with the snow depth retrieval algorithm of radar altimeter is selected to carry out research, and the snow depth sensitivity of the SMRT model with improved simulation process is analyzed by using this algorithm, and the feasibility of applying the improved SMRT model to the multi-year ice snow depth retrieval is verified. At last, the average monthly snow depth in January to March of each year from November 2012 to April 2021 is inverted. The comparison between the retrieval results and the snow depth dataset shows that the retrieval results can better capture the snow depth distribution on the Arctic ice. The accuracy of the snow depth measured by Operation Ice Bridge (OIB) from 2013 to 2015 is verified. The retrieval results are in good agreement with the snow depth measured by OIB, and the Root Mean Square Error (RMSE) for both is 0.07m, indicating the validity of the multi-year ice snow depth retrieve algorithm based on SMRT model.
