| Water is the core element that restricts the protection and development of the Yellow River Basin.The precipitation in the Yellow River Basin is characterized by uneven distribution within the year,large inter-annual variation,and strong spatial variability.Floods and droughts occur frequently.As one of the regions with shortage of water resources in China,timely access to accurate and reliable precipitation data is of great significance to the water resources management of the Yellow River Basin and the prediction,prevention and monitoring of natural disasters such as floods and droughts.However,the density of rainfall stations in the Yellow River Basin is low and unevenly distributed,and there are problems such as scarcity of measured precipitation data or discontinuous data series.Satellite precipitation products have high temporal and spatial resolution and wide spatial coverage.They are potential precipitation substitutes,but their accuracy needs to be evaluated.In this study,the Yellow River Basin was taken as the research area,and the accuracy of seven satellite precipitation products(CHIRPS,CPC,GPCP 1DD,GPM,PERSIANN,MERRA and ITPCAS)were comprehensively evaluated from multi-scale based on the measured precipitation data of meteorological stations;SPEI(Standardized Precipitation Evapotranspiration Index)was selected to explore the applicability of different satellite precipitation products in the field of drought monitoring in the Yellow River Basin;In addition,the SWAT(Soil and Water Assessment Tool)hydrological model of typical sub-basins(Yanhe river basin)in the study area was built to quantitatively evaluate the applicability of different precipitation data in hydrological simulation in the basin scale.On this basis,credible simulation results were selected to analyze the spatial distribution characteristics of hydrological elements in the basin.The major conclusions are as follows:(1)Based on the measured precipitation data of meteorological stations,it is found that ITPCAS has the highest accuracy and the best applicability in the Yellow River Basin(correlation coefficient is 0.97,normalized central root mean square error is 0.23,normalized standard deviation is 1.02),followed by CPC and GPM,and PERSIANN has the worst accuracy(correlation coefficient is 0.32,normalized central root mean square error is 1.61,normalized standard deviation is 1.47).The accuracy of satellite precipitation products is significantly different on space.The accuracy in the southeast of the basin is better than that in the northwest,and the applicability in the middle and lower reaches of the basin is higher than that in the upper reaches.The accuracy of satellite precipitation products is different on different time scales,and the accuracy is the highest on monthly scale.The accuracy of satellite precipitation data varies greatly in different months.ITPCAS has the highest overall accuracy and the most stable performance.PERSIANN has the largest accuracy difference in different months,and its accuracy is poor from January to March and December(normalized standard deviation is between 5.47～9.13,normalized central root mean square error is between4.90～8.97).The comprehensive performance of satellite precipitation data is that the accuracy is relatively good in July and August.(2)The evaluation results of satellite precipitation products in the application of drought monitoring in the Yellow River Basin show that,ITPCAS shows the best applicability(correlation coefficient is 0.98,normalized central root mean square error is 0.21,normalized standard deviation is 1.01),followed by CPC and GPM,while PERSIANN shows the worst applicability(the correlation coefficient is 0.25,the normalized center root mean square error is 1.60,and the normalized standard deviation is 1.51).Satellite precipitation data overestimate the degree of drought,which is more suitable for drought monitoring in the southeast than in the northwest.The satellite precipitation data can better characterize the drought frequency characteristics and the drought trend in spring,summer and annual scales on the temporal scale,and the drought trend in autumn and winter is slightly worse;The distribution characteristics of drought trend and drought frequency captured by ITPCAS are similar to the measured precipitation data of the station,and the other satellite precipitation data are different from the measured precipitation data.(3)The application results of satellite precipitation products in SWAT hydrological model simulation show that the monthly scale simulation results of CHIRPS,CPC and ITPCAS are satiafied,the NSE is higher than 0.5 and the R~2 is higher than 0.6.Among them,the simulation results of ITPCAS precipitation data are the best,while the simulation results of GPCP 1DD,GPM,PERSIANN and MERRA precipitation data are poor.The accuracy of SWAT model simulation results of the measured precipitation data at the station is slightly lower than that of ITPCAS precipitation data,and better than other six kinds of satellite precipitation data.Comparing with the measured runoff,the simulated monthly average runoff based on SWAT model is higher in wet season and lower in dry season. |
