Spatiotemporal Precision Evaluation And Hydrological Effect Study Of Multi-source Satellite Precipitation Products

Due to the uneven spatial and temporal distribution of rainfall in China,wide spatial distribution and relatively concentrated rainstorm,millions of people in China are affected by heavy rain and floods to different degrees every year.Therefore,how to obtain timely and reliable precipitation data and further obtain hydrological data such as runoff at the outlet section of the river basin plays a key role,which is a more effective way to deal with and deal with such disasters.This article selects three satellite precipitation products(CHIRPS,TRMM 3B42 V7,PERSIANN-CDR),and three typical watersheds in China area(Beijiang,Huaihe,Liaohe)as the research area,respectively,from the day,month,season and so on comprehensive assessment of the three kinds of time scale satellite precipitation product precision and adaptability in China;On this basis,four representative extreme precipitation indexes were selected to systematically evaluate the spatial and temporal capture capability of three kinds of satellite precipitation products to extreme precipitation events.CREST based on distributed watershed hydrological model set up two kinds of hydrological simulation scenario scenario I: with precipitation station data calibrate CREST model,then using three kinds of satellite precipitation data driven calibrated models for hydrological simulation;scenario II: using three kinds of satellite precipitation data to calibrate CREST model parameters respectively,then the hydrological simulation comparison),the quantitative evaluation of the three kinds of satellite precipitation products in Beijiang,Huai,Liaohe basin hydrological effect,to explore the substitutability of satellite precipitation data for ground rainfall stations,.The main conclusions of this paper are following:(1)The spatial and temporal accuracy evaluation results of satellite precipitation show that,on a daily scale,CHIRPS is underestimated,PERSIANN's spatial distribution of precipitation is relatively smooth,without obvious extreme value phenomenon,TRMM 3B42 V7 is the closest to the observed value,with the highest accuracy in most basins,and PERSIANN's accuracy is the highest in Songliao and Haihe basins.On a seasonal scale,the applicability of the precipitation of the three satellites is higher in autumn and winter than in spring and summer.TRMM 3B42 V7 performed best in spring,autumn and winter.In summer,the prediction accuracy of the three precipitation products is poor,lower than that of other seasons.For most basins,TRMM 3B42 V7 has the highest accuracy on a seasonal scale,and the accuracy of TRMM 3B42 V7 and PERSIANN-CDR is higher than CHIRPS in Songliao and Haihe basins.However,CHIRPS has the highest accuracy in Southwest river basin and Neilu river basin.(2)In terms of the inversion effect of extreme precipitation,the spatial distribution of the extreme precipitation index predicted by PERSIANN-CDR is smoother,and the differences in spatial distribution can be well captured.In general,it is the closest to the observed value.The overestimation of TRMM 3B42 V7 and CHIRPS exists in most regions of China,especially in southeast China,which is more significant.By comprehensive comparison,PERSIANN-CDR is superior to other two kinds of satellite precipitation products in inverting extreme precipitation.(3)The scenarios hydrological effect assessment results show that different climate types of basin,satellite precipitation products shown the hydrological effects of obvious differences,but the accuracy of the three kinds of satellite products in scenario II were better than scenario I.PERSIANN-CDR and TRMM 3B42 V7 in the Beijiang river basin are close to the simulation effect based on guage stations,which can replace the observation data of ground stations for hydrological simulation and prediction research.In the Huaihe river basin,the three sets of satellite precipitation all have good hydrological simulation results(but there is a slight phenomenon of high flow underestimation and low flow overestimation),which can meet the hydrological simulation and forecast requirements to a certain extent.In Liaohe river basin,no matter what kind of scenario,the three precipitation products can not well simulate the extreme value of runoff,especially the peak simulation effect is not good,and there is a significant underestimation phenomenon,which is not suitable for flood peak flow simulation and prediction in the basin.