Identification Of Critical State Of Drought Disaster And Drought Prediction In Future Climate Scenarios Of The Upper Basin Of The Hanjiang River

Drought disasters are natural disasters that have a huge impact on human society and economy.The occurrence and development of drought is a gradual creep process.The development of drought is difficult to be identified.The process from drought to disaster has a critical state.The identification of critical state is of great significance for drought prediction and early warning.The occurrence of future drought disasters is closely related to climate change,especially for climate-sensitive areas.The impact of climate change is more significant.Based on the identification of drought-induced disasters and the future climate change,the drought prediction of the upper reaches of the Shijiang River in Hanjiang River is carried out.Research provides a scientific basis for drought prevention and control in the region.Combining domestic and international definitions of drought disasters and understanding of drought indicators,comprehensive consideration of the effects of precipitation,runoff,evaporation and soil water content,and selecting multiple indicators that reflect the duration,severity and extreme characteristics of drought,identify potential droughts Samples,and using principal component analysis to reduce the dimensionality of these multiple indicators,remove redundant information generated by the correlation of indicators,and generate low-dimensional comprehensive indicators that are not related to each other to more objectively reflect the characteristic attributes of drought.Combined with the historical drought record,the support vector machine is used to find the optimal classification plane,and the potential drought samples containing the low-dimensional comprehensive index are classified.The classification plane determined by this is the phenomenological drought-induced critical surface.It can more intuitively reveal the development process and trends of drought,and facilitate the promotion and application in drought warning.In this paper,SDSM and SVR are used to predict regional meteorological elements under climate change.Based on the NECP reanalysis data from 1986 to 2015 and the actual observation data of nine rainfall stations and five evaporation stations in the study area,combined with the principal component analysis method,the daily precipitation PCA-SDSM model and the daily evaporation PCA-SVR model were established.According to the measured data from 1986 to 2015,the model was validated and verified.The established PCA-SDSM model and PCA-SVR model were applied to the CanESM2 climate model under the RCP4.5 emission scenario,respectively predicting future climate change under 2021-2040 daily precipitation and evaporation in the study area.Based on the Xin'anjiang model,the hydrological model of the study area was established.Combined with the measured data of precipitation,evaporation and runoff from 2009 to 2015,the parameters of Xin'anjiang model were optimized and verified by SCE-UA algorithm.The daily precipitation and evaporation in the future climate scenario were input into the Xin'anjiang model,and the future runoff changes in the study area were obtained.Through the potential drought sample selection method,the established drought-induced critical equation was used to predict and analyze the future drought situation in the study area.