Downscaling Research And Accuracy Assessment Of Satellite Precipitation Data In Inner Mongolia

TRMM and GPM satellite precipitation data are widely used satellite precipitation data at present,and the spatial precipitation information is of great significance to the study of the temporal and spatial variation characteristics of precipitation in areas with few meteorological stations.At present,it is still difficult to directly obtain high-precision spatial precipitation information.Based on this,this paper takes Inner Mongolia as the study area,reduces the scale of TRMM and GPM precipitation data to obtain high-resolution annual,seasonal and monthly precipitation data,and analyzes the temporal and spatial variation characteristics of precipitation in Inner Mongolia.The main conclusions are as follows:(1)The annual,seasonal and monthly precipitation data of TRMM and GPM are downscaled by simple downscaling model and GWR downscaling model,respectively.The results show that the downscaling result of TRMM is better than that of GPM,and the annual precipitation value of GPM is higher than that of TRMM.The precipitation change characteristics of the two kinds of data downscaling results in most areas of central and eastern Inner Mongolia are basically the same,that is,the annual precipitation decreases gradually from northeast to central.In western Inner Mongolia,the results of TRMM downscaling are consistent with the actual precipitation characteristics,while the results of GPM downscaling are obviously overestimated with high errors.(2)On the annual scale,the correlation coefficient of TRMM precipitation data is 0 ? 0.95,higher than 0.8 in Hulunbeier,Wulanchabu and Hohhot,lower than 0.3 in Alxa League,and the correlation in other areas is 0.3?0.8.RMSE of 45.2% stations is less than 60 mm,and that of 35.5%stations is between 60?100mm,only 19%The correlation coefficient of GPM precipitation data is between 0 and 0.54,RMSE of 40%stations is less than 200mm,RMSE of 49%stations is between 200 and500mm,11%stations are greater than 500 mm,annual precipitation of 84%stations is overestimated and 16%stations are underestimated.On the seasonal scale,the correlation coefficient of precipitation in TRMM is higher than that in GPM,and the proportion of overestimated precipitation in TRMM(GPM)is 80%(52%)in spring,83%(6%)in summer,60%(53%)in autumn and 88%(72%)in winter respectively.The precipitation data in TRMM are overestimated in four seasons,while GPM precipitation data were mainly underestimated in summer and overestimated in other seasons.The accuracy evaluation of simple downscaling results shows that the correlation of TRMM downscaling results is between 0 and 0.96 on the annual scale.RMSE of 16.1% stations is lower than 50 mm,73.1% stations are between 50 and 100 mm,and 9.3% stations are higher than 100 mm.92.4% of the sites are overvalued.GPM correlation coefficient is between0 and 0.55.The RMSE of 49% stations is lower than 200 mm,40% stations are between 200 mm and 500 mm,and 11% stations are higher than 500 mm.84% of the sites are overvalued.The accuracy evaluation of GWR downscaling results shows that the correlation of TRMM precipitation data is between 0?0.85 on the annual scale.The RMSE of 7.5% stations is lower than 50 mm,60.2% stations are between 50?100mm,and 32.2% stations are between 100-175mm;89% of the sites are overvalued.The correlation of GPM precipitation data is between 0 and 0.89.RMSE of 76% sites is between 70?200mm,RMSE of 14% sites is between 200?500mm and RMSE of 10% sites is higher than 500 mm.82% of the sites are overvalued.On the whole,the accuracy of simple downscaling results of TRMM precipitation data has been improved,and the accuracy is better than that of GWR downscaling results.GWR downscaling results of GPM precipitation data are better than simple downscaling results,but the accuracy is still lower than that of TRMM precipitation data.(3)Based on TRMM downscale precipitation data,the analysis of precipitation time variation in Inner Mongolia from 2001 to 2019 shows that precipitation in Inner Mongolia shows an overall increasing trend,increasing by 40 mm every 10 years.In terms of seasonal variation,the precipitation in spring,summer and autumn showed an increasing trend,with the trend rates of 1.2mm/10 a,27mm/10 a and 10mm/10 a,respectively.However,the precipitation in winter showed a decreasing trend with a trend rate of-0.02mm/10 a.The precipitation in March and April in spring and December and January in winter showed no significant decreasing trend,but increased in May in spring,June,July and August in summer,September,October and November in autumn and February in winter.From the spatial characteristics of precipitation change trend,the annual precipitation in most areas of Inner Mongolia increased from 2001 to 2019,with the largest increase in Hulunbeier,Xing 'an League and Tongliao City,with an increase trend of 60?118 mm/10 a,followed by the most areas of Xilin Gol League,southern Wulanchabu City,most areas of southeastern Ordos City and a small part of southeastern Alxa League,with an increase trend of 20 ? 60 mm/10 a.In the west of Hulunbeier City and the north of Bayannaoer City,the annual precipitation shows a decreasing trend of-15?0 mm/10 years.Spring precipitation increased significantly in the central and eastern areas of Tongliao in eastern Inner Mongolia.Summer precipitation increased significantly in a small part of Hulunbeier in eastern Inner Mongolia,eastern Xing 'an League and Tongliao City,and western Alashan League in western Inner Mongolia.Autumn precipitation increased significantly in the southeast of Hulunbeier City and the northeast of Xilingol League.Winter precipitation increased significantly in most areas of central and eastern Xilingol League.