Comparison of soil moisture inversion of three types of CYGNSS Land reflectivity in Guizhou
Considering the unclear performance of the three commonly used Cyclones Global Navigation Satellite System(CYGNSS)surface reflectivity in the soil moisture(SM)retrieval,this paper takes Guizhou Province as the study area and uses LightGBM as the modeling tool to experimentally compare and analyze their respective SM inversion accuracy.The results indicate that the SM models,which incorporate three types of CYGNSS surface reflectivity as explanatory variables,achieved higher modeling accuracy and SM inversion accuracy.However,the SM model based on signal-to-noise ratio(SNR)reflectivity was more susceptible to vegetation cover,resulting in poorer generalization ability and lower accuracy in SM inversion results.The SM model,based on peak power reflectivity,exhibited stronger generalization ability and higher accuracy in SM inversion.The root mean square error,mean absolute error,and Pearson correlation coefficient for the SM inversion results were 0.053 6,0.042 3 cm3·cm-3,and 0.767 0 respectively.The research findings can provide reference for the selection of CYGNSS observations in soil moisture inversion.
NETL
NSTL
万方数据
