Integration Study on Oasis Soil Moisture Inversion Using ALOS-2 and Landsat-8

Objectives:Integration of machine learning and multi-source data has become one of the hot topics in soil moisture inversion, where relatively few studies have been performed on L-band SAR imagery. Methods:In this paper, ALOS-2 PALSAR-2 and Landsat-8 imagery of the Ejina Oasis were used to extract the radar and optical characteristic parameters which were then screened according to the importance score. Random forest was adopted to establish different soil moisture inversion models based on radar, optical, and radar-optical integrated parameters. Model accuracies were evaluated and soil moisture content in Ejina Oasis was inversed. Results:(1) Compared with C-band data, L-band SAR data is more sensitive to soil moisture content in arid desert oasis. (2) With regard to radar characteristic parameters, surface and volume scattering components have higher important scores. Dihedral and helix scattering component are less important. As for optical characteristic parameters, vegetation water supply index takes the most important place while enhanced vegetation index is the least important one. (3) The optimal models R² and RMSE of radar characteristic parameter scheme were 0.67 and 2.16% respectively. The accuracy of optical characteristic parameter scheme model is generally low and the accuracy is equivalent, R² and RMSE were about 0.5 and 2.47%. (4) The optimal model R² and RMSE of radar optical parameter integrated inversion were 0.72 and 1.99%, respectively. Compared with a single data source, R² increased by 7.46% and 38.4%, and RMSE decreased by 8.54% and 22.6%, respectively. Conclusions:The research proves that the random forest model based on multi-source data fusion has higher prediction accuracy and better applicability in arid desert oasis area.