Geographically Weighted Area-to-Point Regression Kriging for Spatial Downscaling in Remote Sensing

摘要: Spatial downscaling of remotely sensed products is one the main ways to obtain earth observations at fine resolution. Area-to-point (ATP) geostatistical techniques, in which regular grids remote sensing are regarded as points, have been applied widely for spatial downscaling. In downscaling, it common use auxiliary information explain some unknown variation target geographic variable. Because ubiquitously heterogeneities, observed variables always exhibit uncontrolled variance. To overcome problems caused by local heterogeneity that cannot meet stationarity requirement ATP regression kriging, this paper proposes a hybrid statistical method incorporates geographically weighted and kriging The proposed (GWATPRK) combines resolution allows non-stationarity model. approach was verified using eight groups four different 25 km-resolution surface soil moisture (SSM) 1 km SSM predictions two experimental regions, conjunction with implementation three benchmark methods. Analyses comparisons downscaled results showed GWATPRK obtained images greater quality an average loss root mean square error value 17.5%. analysis indicated has high potential applications.