Research On The Response Of Hydrological And Meteorological Elements In Jinsha River Basin To Climate Change

Exploring the evolution law of hydrological and meteorological elements in the river basin and mastering the water cycle process of the river basin can provide scientific decision support for river basin water resources management,water conservancy project planning and construction,and safe and economic operation of reservoirs.However,under the background of global climate change,extreme hydrological and climatic events occur frequently,and water problems are increasingly constraining the sustainable development of my country's society and economy.Therefore,how to use scientific methods to analyze the future evolution trend of hydrometeorological processes in the river basin and clarify its response to climate change is one of the key issues that need to be resolved in the field of water resources management.This paper takes the Jinsha River basin as the research object,carries out the analysis of the temporal and spatial changes of historical hydrological and meteorological elements in the basin,establishes a variety of statistical downscaling models,and evaluates the applicability of each downscaling model in the Jinsha River basin.On this basis,through downscaling of global climate model data,the evolution process of future meteorological elements in the basin was predicted.Furthermore,a runoff forecast model driven by hydrometeorological elements was constructed,and the response of the runoff process to climate change was deeply analyzed.The main research content and achievements of the thesis are as follows:(1)The evolution law of the key hydrometeorological elements under the changing environment is taken as the starting point,combining the climate tendency rate,MannKendall test,cumulative anomaly and Morlet wavelet function to explore the hydrological cycle elements in the Jinsha River basin for more than 40 years Change trend,sudden change characteristics and periodic characteristics under different scales.The analysis results show that the annual precipitation,annual average temperature,average annual maximum temperature,average annual minimum temperature and average annual runoff in the basin all show an insignificant upward trend.Except for the insignificant decline in summer runoff,the rest of the seasons showed an upward trend,among which the spring runoff trend is significant;each hydrometeorological element has a single mutation year in the 1990s;all hydrometeorological elements in the whole basin have a multi-scale significant cycle,and most periodic time-frequency structures are similar.The annual runoff is mainly concentrated in June to October.Generally speaking,the rainfall and runoff in the basin have been relatively stable in recent years.(2)In order to explore the future trends of hydrometeorological changes in the basin,based on the measured meteorological data and NCEP reanalysis data,support vector machines(SVM),Gaussian process regression(GPR),and multiple linear regression(MLR)are introduced to construct a basin statistical downscaling model System and assess its applicability to various meteorological elements at various stations in the Jinsha River Basin.At the same time,the CanESM2 climate model data is used to drive the downscaling model to obtain the predicted values of meteorological elements under RCP2.6,RCP4.5 and RCP8.5 emission scenarios,and then to analyze the future climate change trend of the basin.The results show that the MLR downscaling model performs better in the downscaling method of precipitation elements,and the optimization rate is 69.2%.In the downscaling methods of the three temperature elements,the GPR model has the highest probability of being better,which are 84.6%,76.9% and 92.3% respectively;the annual precipitation and the three temperature elements under each scenario in the future will show an upward trend,and Compared with the base period,the year has increased by 4.5%,19.9%,10.4%,and34.9% respectively;and the order of the significant degree of the element trend from weak to strong is RCP2.6,RCP4.5,RCP8.5,of which RCP8.5 is significant The degree is obviously stronger than other scenarios.(3)In order to analyze the trend of future runoff changes in the basin under climate change,based on the measured meteorological and runoff data,long and short-term memory neural networks(LSTM)are introduced to construct a runoff forecast model for the basin;coupled with the simulated values of meteorological elements output by the CanESM2 climate model after downscaling to predict the basin From 2016 to 2100(divide 2016 to 2050 as the early forecast period,2051 to 2080 as the middle forecast period,and 2081 to 2100 as the late forecast period.)Monthly runoff and explore the change process of river basin runoff under various future scenarios.The results of the study show that the established LSTM runoff forecasting model has a rate of determination coefficient greater than 0.9 for regular and verification periods,and the average relative error is controlled within 13%.The model has a high degree of reliability;the annual runoff will show an upward trend in future scenarios,RCP8.The upward trend is relatively obvious under the 5 scenarios,and relatively flat under the RCP2.6 and RCP4.5 scenarios.Under the three emission scenarios,the inter-annual fluctuations of the runoff in some periods are relatively large;the annual runoff in each period is reduced relative to the base period,and the rate of change is between-8.5%and-16.6%;the monthly average runoff in most months is high in the base period,the rate of change in each period is between-37.8% to 19.8%,36.7% to 20.7%,and-37.4%to 31.7%.In the future,the peak of runoff will shift from September to July,and the flood season may appear earlier.