Improvement Of CLDAS-PRCP Fusion Precipitation And Its Application In Rainfall Erosivit

Soil erosion is a severe problem facing the global environment.Precipitation is a key factor affecting soil erosion,and its potential impact on soil erosion can be reflected by rainfall erosivity.Accurately estimating rainfall erosivity is crucial for dynamically monitoring soil erosion.Currently,meteorological station observed precipitation is the main source of data used to estimate rainfall erosivity,but it can only reflect the precipitation situation near the station and is not suitable for large areas.Although satellite and reanalysis precipitation products can effectively reproduce the spatiotemporal distribution of precipitation,their accuracy has long been a bottleneck in various professional fields due to the influence of inversion algorithms and numerical models.Therefore,this paper evaluates the accuracy of mainstream satellite and reanalysis precipitation products,improves the CLDAS Multi-source precipitation fusion dataset(CLDAS-prcp)by combining the advantages of different precipitation products,and then evaluates the accuracy of rainfall erosivity estimated based on the fusion precipitation,satellite precipitation,and reanalysis precipitation products.Finally,the spatiotemporal analysis of rainfall erosivity in mainland China is conducted.The main conclusions of the article are as follows.(1)Accuracy evaluation of different precipitation products: This study comprehensively evaluated the performance of satellite-based precipitation products,including Integrated Multi-satellite Retrievals for Global Precipitation Measurement(IMERG),Climate Prediction Center morphing Technique(CMORPH),and fifth generation of ECMWF atmospheric reanalysis of the global climate(ERA5),in mainland China.The results indicated that all three precipitation products could reflect the spatial distribution characteristics of "more in the east and less in the west" accurately at both hourly and daily scales,but all overestimated the actual ground precipitation.In the cold season(January to March and October to December)and high-latitude areas(Northeast,North China,and Northwest),ERA5 had better precipitation description ability than satellite-based precipitation products;in the warm season(April to September)and low-latitude areas(Southwest,Central South,and East China),the satellite-based precipitation products had advantages in reflecting precipitation events,and IMERG had a higher overall accuracy than CMORPH.ERA5’s snowfall index at the hourly scale was superior to IMERG and CMORPH in various regions,with a correlation coefficient close to 0.7 in Northeast and East China.IMERG’s rainfall accuracy at the hourly scale was better than that of ERA5 and CMORPH in different regions,with a stable correlation coefficient of around 0.45 and a root-mean-square error lower than 1.2 mm.(2)Improved scheme for multi-source merged precipitation: Based on the multi-grid variational analysis method combined with the rainfall information of Integrated Multisatellite Retrievals for Global Precipitation Measurement(IMERG)and the snowfall information of fifth generation of ECMWF atmospheric reanalysis of the global climate(ERA5),this study developed the CLDAS-prcp-new merged precipitation for 2018.The results showed that CLDAS-prcp-new effectively alleviated the problem of overestimation of lake precipitation in the original CLDAS-prcp in western China.Validation indicated that without incorporating station data,IMERG performed best for daily lake precipitation,with a correlation coefficient and root-mean-square error of 0.71 and 4.47 mm,respectively.CMORPH significantly overestimated lake precipitation from March to October,and the comprehensive ranking of the five precipitation products’ lake precipitation performance was CLDAS-prcp-new > IMERG > ERA5 > CLDAS-prcp > CMORPH.In the Tibet Autonomous Region,CLDAS-prcp-new showed significant improvement compared to the original CLDAS-prcp,with a daily scale correlation coefficient increased from 0.80 to 0.91,and the root-mean-square error reduced from 2.62 mm to 1.59 mm.Therefore,CLDAS-prcp-new has good potential for application in reflecting lake precipitation and plateau precipitation,among others.(3)Application of different precipitation products in rainfall erosion: This study evaluated the applicability of different precipitation products in estimating rainfall erosion and analyzed the spatio-temporal distribution characteristics of rainfall erosion in mainland China over the past 20 years.The results showed that CLDAS-prcp-new had the best overall performance in estimating rainfall erosion,followed by IMERG and CLDAS-prcp,while ERA5 underestimated rainfall erosion severely due to its poor accuracy at small scales and weak capturing ability for heavy rainfall in the warm season,making it difficult to accurately reflect the spatial distribution of rainfall erosion at large scales.Analysis of rainfall erosion in mainland China over the past 20 years revealed that erosive rainfall frequency and the synergy of rainstorm processes affected the magnitude of rainfall erosion.Rainfall erosion in China exhibited a characteristic of high in the southeast coastal area and low in the northwest inland area,with erosive rainfall concentrated from May to August both temporally and spatially.In addition,there was a positive correlation between annual rainfall,annual erosion density,and annual rainstorm frequency.This study provides important references for predicting rainfall erosion and soil erosion prevention and control.