Intercalibration of FY-3B/MWRI and GCOM-W1/AMSR-2 brightness temperature over the Arctic

Microwave radiometers have been widely applied in polar region research because of their all-weather and all-time capabilities. Microwave Radiation Imager (MWRI) on FY-3B is the microwave radiometer of China's own research and development and has aroused widespread concern. Long time series of earth observation data records play an important role in the research of earth environment changes and trends. The Arctic region is used as the study area and the data of Advanced Microwave Scanning Radiometer-2 (AMSR-2) on Global Change Observation Mission 1st-Water (GCOM-W1) are considered the standard data in providing the intercalibration result and the basis of retrieving remote sensing parameters in Arctic region in the future. Ascending and descending brightness temperatures at 10 channels in 2015 from FY-3B/MWRI are calibrated against those from GCOM-W1/AMSR-2. Before brightness temperature data analysis and intercalibration, the data are processed in five steps. The first step is reading data, in which the DN value of remote sensing is transferred to brightness temperature value in the research region. The second step is data quality control. If the standard deviation of values in a grid and eight surrounding grids is more than 3 K, then the values in the nine grids should be eliminated. The value that is more than 300 K or less than 10 K should also be eliminated. In the third step, stereographic projection is used to project the brightness temperature value, time, longitude, and latitude into 896×608 grids. In the fourth step, data at the land-sea boundary and Marginal Ice Zone (MIZ) should be eliminated because of the mixed pixel. First, the grid data of 7×7 around the land data are marked as land, and the data marked as land are removed. Then, the ratio of V187 to V365 from AMSR-2 is used to calculate the MIZ. The ratio, which is equal to 0.92, is viewed as the threshold to divide the sea ice and the open water. Thereafter, the 3×3 grid is set as a template. If the template includes the grids that represent sea ice and open water, then the nine grids are eliminated. The fifth and last step is to set the time window as 30 min and the space window as 12.5 km ×12.5 km and convert 2D matched data to 1D data for intercalibration. The intercalibration results of MWRI and AMSR-2 are as follows. First, the brightness temperature data of each channel of MWRI are smaller than those of AMSR-2, and the absolute values of monthly bias of vertical polarization channels are greater than those of horizontal polarization channels at the same frequency. The difference in monthly bias between ascending and descending orbits is small in each channel, which is less than 1 K. Second, the difference in the monthly bias in each channel between the ascending and descending orbits in the sea ice area is less than 1 K, while that in the open water is between 0 and 1.5 K. Third, linear regression analysis shows that most of the correlation coefficients of MWRI and AMSR-2 in each channel are above 0.99, which indicates a good correlation. The slope and intercept of the intercalibration of each channel in the ascending and descending orbits are obtained. Fourth, the brightness temperature of MWRI after calibration is consistent with that of AMSR-2. This consistency indicates that the intercalibration is effective. © 2020, Science Press. All right reserved.