Study On The Influence Of Climate Change On Runoff Of Tumen River Mainstream

Climate change affects the process of water circulation to a certain extent,thus the regional water balance is also destroyed correspondingly.Therefore,it is of great significance to rationally develop and utilize water resources to study the influence of climate change on runoff of Tumen River Mainstream.Based on the meteorological data and runoff data,such as temperature,precipitation,evaporation,sunshine hours,relative humidity and wind speed for 1960-2016 years in Tumen River,using linear method,Mann Kendall test,wavelet analysis,Pearson correlation analysis,redundancy analysis,grey correlation analysis,multiple regression analysis and BP neural network analysis,the variation characteristics,mutation time,change period and the relationship between main climatic factors and runoff of Tumen River Mainstream were studied.The main conclusions are as follows:(1)The results of analysis of change characteristics show that:From 1960 to 2016,the overall temperature change of Tumen River Mainstream showed a significant upward trend,with little interannual variation;Precipitation change is relatively stable in general,and he largest interannual variation of precipitation in winter;Evaporation,sunshine hours,relative humidity and wind speed fluctuated reduction generally,and the interannual variation was not significant.Runoff showed a decreasing trend and the interannual variability of autumn runoff was the largest.The trend of runoff is roughly the opposite of that of air temperature and evaporation.The trend of runoff is opposite to temperature and evaporation,and roughly consistent with precipitation and relative humidity.There is no obvious synchronism with sunshine hours and wind speed.(2)The results of mutation analysis and periodic analysis show that:The temperature abrupt change point of Tumen River Mainstream was 1987,the abrupt change in sunshinet hours was 1978,relative humidity abrupt point is 1975,the abrupt point of runoff is 1973,and there is no mutation in precipitation,evaporation and wind speed.There are two periodic scales of precipitation(about 2 years and 7 years),evaporation has a periodic scale(about 5 years),sunshine hours has a periodic scale(about 3 years),runoff has a periodic scale(about 2 years),which is consistent with precipitation(about 2 years),and the temperature,relative humidity and wind speed have no obvious periodic changes,(3)The result of correlation analysis shows that:The runoff of Tumen River Mainstream in 1960-2016 is significantly related to the annual average temperature,the spring temperature,the precipitation in the flood season,the summer precipitation,the annual precipitation,the annual relative humidity,the relative humidity in summer and autumn,the wind speed in the autumn and spring,and the annual wind speed,while is no obvious correlation between evaporation and sunshine hours;The correlation order between runoff and major climate factors is:R04(summer precipitation)>R03(flood season precipitation)>R05(annual precipitation>R10(spring wind speed)R08(autumn relative humidity)>R09(annual wind speed)>R11(autumn wind speed)R07(summer relative mumidity)>R06(annual relative humidity)>Rol(annual average temperature)>R02(spring temperature),that is,precipitation factors have the greatest impact on runoff,followed by humidity factors and wind speed factors,while temperature factors have a relatively small impact on runoff.(4)Runoff simulation results show that:The regression coefficient of the multiple regression model is 0.797,which is well representative,The variation curve of runoff fitting value is generally consistent with the measured value,which can be used as runoff prediction model;The regression coefficient of BP neural network is 0.896,and the runoff simulation curve is almost identical with the measured value curve.The simulation accuracy of BP neural network analysis is higher than that of multiple regression analysis,and is more suitable for runoff simulation and prediction of Tumen River Mainstream...