Ground validation of Fengyun-4A and Global Precipitation Measurement satellite observations over an alpine and canyon basin of the southeastern Tibetan Plateau

Satellite-based precipitation observations with high spatiotemporal resolution are essential for studying rainfall-induced natural hazards,especially in alpine and canyon areas of the southeastern Tibetan Plateau,which are prone to such hazards yet sparsely gauged.Here,we evaluated precipitation estimated from the Chinese Fengyun-4A meteorological satellite(FY-4A AGRI)versus the Integrated Multi-satellitE Retrievals for GPM(IMERG),by using rain gauge data collected in the Parlung Zangbo Basin from May through September in both 2018 and 2019.Our results showed that(1)FY-4A AGRI generated smaller values of RMSE(root mean square error)on hourly to daily scales,and larger correlation coefficients(R-values)and smaller RMSE values for both moderate and heavy rain,indicating its greater accuracy at rainfall estimation,which is most likely due to the denser rain gauge network at a finer temporal scale used when calibrating FY-4A AGRI;(2)Both satellite products underestimated the volume of moderate and heavy rain,with the larger degree of underestimation by FY-4A AGRI,which could lower their performance in flood monitoring and forecasting;(3)Worse performance and greater inconsistency between the two products were observed in high-elevation areas,perhaps because of orographic cloud effects in these mountainous areas;and(4)Both products revealed that the Gangrigabu Range blocked incoming water vapor from the southwest monsoon,with a better representation of the spatial pattern and spatial variability produced by IMERG.To improve precipitation estimation,the effects of complex terrain should be explicitly incorporated into the retrieval algorithms,with more gauged observations in a denser network and at a finer temporal scale needed to robustly calibrate the satellite-based estimates.