Downscaling Correction Of Multi-source Satellite Remote Sensing Precipitation Products In Pingtang Basin And Its Applicability In Runoff Simulation

The release of various satellite remote sensing precipitation products has provided a new source of precipitation data for the field of precipitation observation.However,different types of satellite remote sensing precipitation products are affected by factors such as the geographical location,climate characteristics and topography of the region,as well as differences in inversion algorithms.The accuracy of their performance in different regions is different,and when they applied to small watersheds,there is also the problem that the spatial resolution is too coarse.This paper takes the Pingtang Basin as the research area,and selects the latest satellite remote sensing precipitation products of GPM,GSMAP and MSWEP with high spatial and temporal resolution as the research objects.Firstly,the spatial and temporal accuracy characteristics of the three satellite remote sensing precipitation products and their ability to invert extreme precipitation in the basin were analyzed by using multiple accuracy evaluation indicators;then the three satellite remote sensing precipitation products were downscaled respectively.Finally,the application effects of the three satellite remote sensing precipitation products and their downscaled corrected precipitation data in the runoff simulation application of the Pingtang Basin are evaluated.The main results of this study are as follows:(1)The accuracy of GPM,GSMAP and MSWEP in the Pingtang watershed has its own advantages and disadvantages.On the time scale,GPM and GSMAP perform better than MSWEP in error control,but all three have the phenomenon of error accumulation over time scale accumulation.On the spatial scale,the three satellite remote sensing precipitation products can more accurately express the spatial distribution characteristics of precipitation in Pingtang Basin,but they all underestimate the actual precipitation in the basin.In terms of precipitation event recognition ability,GPM and GSMAP have better ability to identify precipitation events in the rainy season,while MSWEP’s ability to identify precipitation events is weak,which showing a high hit rate and a high false alarm rate.Therefore,MSWEP is not suitable for identifying precipitation events in Pingtang Basin.(2)The three satellite remote sensing precipitation products all underestimate the number of extreme precipitation days in the basin,and most of the extreme precipitation indices represented by MSWEP are quite different from the station precipitation data,so a better extreme precipitation inversion effect cannot be obtained.The inversion ability of GPM and GSMAP for extreme precipitation indices in the basin is better than that of MSWEP,but GPM and GSMAP also have their own advantages and disadvantages for the characterization of different extreme precipitation indices.Therefore,we should be paid more attention to the actual selection of satellite remote sensing precipitation products for inversion research of extreme precipitation in the basin.(3)The downscaled correction of GPM,GSMAP and MSWEP was carried out respectively,and the correlation between the obtained DC-GPM,DC-GSMAP and DC-MSWEP and the station precipitation data was significantly improved,and the error was also significantly improved.The advantage is that the spatial distribution of precipitation in the basin is more detailed,and the accuracy of daily precipitation estimation in the basin is improved,indicating that the downscaling correction is effective.The ability of DC-GPM and DC-GSMAP to identify precipitation events is greatly improved;while DC-MSWEP shows higher hit rate and false alarm rate,and the ability to identify precipitation events is poor.Overall,DC-GSMAP has the lowest root mean square error and small deviation,and has relatively good error performance.Which has relatively good error performance among the three downscaled precipitation data.(4)According to the runoff simulation results of the Grid-XAJ model,the runoff simulations of GPM,GSMAP and MSWEP can all capture the time variation characteristics of the measured runoff at the outlet of the basin,but they all underestimate the measured runoff,and the peak simulation effect is not good.The runoff simulation curves of DC-GPM,DCGSMAP and DC-MSWEP are in good agreement with the measured runoff and peak discharge curves,and the underestimation has been improved.It shows good runoff simulation performance.The monthly-scale runoff simulation effects of all precipitation products are better than that of the daily-scale,but there are certain deviations in the runoff simulations of the two timescales.In conclusion,the downscaled corrected precipitation data has higher application value in the runoff simulation of the Pingtang Basin,especially the simulation results of DCGSMAP have a high Nash-Sutcliffe efficiency coefficient(NSE)and the lowest deviation(RB)in Pingtang Basn,which has great application potential in runoff simulation in Pingtang Bain.