Advanced method for sea ice concentration retrieval from satellite microwave radiometer measurements at frequencies near 90 GHz
|Zabolotskikh E.V.1, Balashova E.A.1, Chapron Bertrand2
|1 : Russian State Hydrometeorological University, Saint Petersburg, Russia
2 : Institut Français de Recherche pour l’Exploitation de la Mer, Plouzané, France
|Sovremennye problemy distantsionnogo zondirovaniya Zemli iz kosmosaCurrent problems in remote sensing of the Earth from space (2411-0280) (Space Research Institute RAS), 2019 , Vol. 16 , N. 4 , P. 233-243
|Article en russe
|sea ice, sea ice concentration, Arctic, satellite passive microwave radiometers, brightness temperatures, AMSR2, polarization difference, physical modeling
An advanced method for sea ice concentration retrieval from satellite microwave radiometer measurements at frequencies near 90 GHz is presented. The method is based on the new approach for the determination of the tie points ― the polarization differences (PD) of the brightness temperatures (TB) of the ocean-atmosphere system (PDW) and the sea ice-atmosphere system microwave radiation (PDSI). The approach is based on the results of physical modeling of the sea ice – ocean – atmosphere TB and the analysis of the measurements of the Advanced Microwave Scanning Radiometer 2 (AMSR2) in the Arctic region. The TB simulation is carried out for the whole ranges of the Arctic atmospheric conditions and sea ice and ocean parameters. The method of sea ice concentration (SIC) retrieval uses PD in measurements on the vertical and horizontal polarization at the frequency of 89 GHz and the values of tie points over the ice-free sea surface and over the sea ice. The range of PDW and PDSI variability is analyzed basing on the AMSR2 measurement data and the results of TB model calculations. The advancement of the method as compared to those traditionally used is the use of variable PDW values depending on how far from the sea ice edge is the pixel for which C is estimated. The method was tested using the maps of the Norwegian Meteorological Institute (NMI) for the Northeren, Kara and Barents seas. The error of SIC estimation, calculated using the new method, turned out to be 4.2 %, which is almost two times lower than the error of the standard product of the University of Bremen, calculated using the same verification data set.