青海地区TRMM 3B43降水产品融合降尺度与时空特征分析OA北大核心CSTPCD

Satellite remote sensing precipitation products have been widely used to estimate precipitation,espe-cially in regions with sparse ground observation stations.However,the lower spatial resolution of these satellite products limits their application in localized regions and watersheds.This study proposes a fused downscaling framework based on the area-to-point kriging(ATPOK)algorithm and geographic weighted regression kriging(GWRK)algorithm to downscale TRMM 3B43 data for the Qinghai region of China from 2000 to 2019.The framework incorporates ground observation station data,normalized difference vegetation index(NDVI),digital elevation model,slope,and auxiliary factors for calibration,ultimately obtaining precipitation products at 1 km resolution.The results showed that:(1)The proposed fused downscaling framework can effectively improve the accuracy of TRMM 3B43 products;however,it cannot eliminate the overestimation of TRMM 3B43.(2)Com-pared to the ATPOK algorithm,the GWRK algorithm using data TRMM precipitation data and auxiliary factors can better estimate annual/monthly precipitation data.(3)Based on the study of the relationship between TRMM 3B43 and NDVI,it was found that NDVI responds to precipitation with a delay of 0-2 months.(4)Based on the spatiotemporal variation analysis of downscaled precipitation products,significant increases in monthly precipita-tion were observed in the Qinghai region,with an interannual change rate of 3.33%in dry month(December)and 1.79%in wet month(July).