Cloud-Precipitation Resources Forecasting And Warning Model Using Satellite Remote Sensing Information

Cloud and precipitation,which are two different physical forms of water resources,are an important part of regional water resources and play a crucial role in regional water resources management and early warning.In recent years,with the change of global environment,water resources have increasingly become a scarce resource,at the same time,it also brings frequent disasters.It has become an urgent problem to study the relationship between cloud and precipitation and establish a prediction and warning model of cloud-precipitation water resources.In this paper,we select Wuxijiang basin for the study,a cloud-precipitation water resources prediction and warning model is established based on the cloud parameters obtained by MODIS inversion,the precipitation data of the measured ground stations from 2000 to 2004,and BP neural network and PSO-LSSVM to provide methods and model for regional water resources management and early warning.Finally,the cloud precipitation water resources prediction and warning model was optimized based on the comparative analysis of the results of the two models.The results show that,(1)Taking the formation mechanism of cloud and precipitation,remote sensing inversion basis and instrument performance into comprehensive consideration,the five cloud parameters,namely the normalized cloud detection index,atmospheric water vapor content,and the cloud-top infrared bright temperature difference in MODIS band 29,31 and 32,can accurately establish the relationship between cloud and precipitation.(2)The relationship between cloud parameters and precipitation is analyzed,and it is difficult to construct a complex relationship between cloud and precipitation.(3)The prediction and warning model of cloud and precipitation in Wuxijiang basin based on BP neural network was established.The results show that the correlation coefficient between the rainy season and the dry season is 0.74 and 0.75 respectively.(4)The prediction and early warning model of cloud precipitation water resources in Wuxijiang basin established by PSO-LSSVM method has good performance.The correlation coefficients of training phase and verification phase are 0.90 and 0.92 respectively.The model can accurately simulate the relationship between cloud and precipitation.By comparing the performance of the two models,the PSO-LSSVM model is more suitable for the study of the prediction model of cloud precipitation water resources in the Wuxijiang basin.This study makes full use of the advantages of satellite remote sensing technology to establish the prediction and warning model of water resources,which provides a feasible method for the management and warning of regional water resources and which has important social and economic value and practical guidance for improving regional water resources management capacity and preventing floods and droughts caused by precipitation.