Research On The Relationship Between Water Supply And Demand Of Irrigation Area With Multi-time Scales

In recent years, human activity and climate change make water supply and demand imbalance, which leads to the agricultural water safety issues highlighted. Agriculture water safety is the result of water resources supply and demand uncertainty relation of cognitive science. Irrigation district as the guarantee of food security, the relationship of supply and demand of agricultural water resources is important for regional economic development and ecological security. At the same time, irrigation water is one of the main economic indicators of the irrigation district management, the forecast of irrigation water use can make the limited water resources for the maximum benefit. Based on the relationship of irrigation water supply and demand, scientific forecasting irrigation water can provide a scientific basis for the internal rational distribution of irrigation water and provide a scientific basis for guiding irrigation agriculture.This paper took Luhun Irrigation district as the research region, using the method of multi-time scales to analysis the relationship between water supply and demand of irrigation area, at the same time, calculate the uncertainty relation between them, the main conclusions are as follows:(1) Using 10 a moving average and MK test to examine the analyze the spatial and temporal variations of water supply and demand from 1951 to 2013 in the irrigation area. The results show that: In recent years, irrigation area rainfall is a trend of first increased after decreased, the mutation point is more, and more obvious mutation occurred in 1958, 1981 and 1986; crop water requirementshowed a trend of increase after the first reduce and mutation is less, occurred in 1974 and 2012; irrigation water have been reduced after the trend of increas, and the trend changing until the 1980 s, there are more point mutations, have been reduced after the trend of increase,obvious mutation occurred in the year 1979 and 1983.(2) The use of wavelet analysis to multiple time scale decomposition of water supply and demand in the irrigation area,the results show that: the fluctuation cycle of precipitation is 2-4 years, 4-9 years, 9-12 years, 25 years and 30 years; Crop water requirement of the fluctuation cycle for 2-4 years, 4-7 years, 8-14 years, 15-24 years and 38 years; the volatility of the irrigation water cycle for for 2-4 years, 5-7 years, 8-17 years, and 26 years.(3) Using set pair analysis to make multi-scale set pair analysis, the analysis results showed that: rainfall and crop water demand on multiple time scales and overall performance for the different state; crop water requirements and irrigation water can be over a shorter time scale mainly for the different state, over a longer time scale mainly for the same state.(4) By using a three-variable linear cointegration method to buildthe linear modle of precipitation and the error correction model, the results showed: the using of raw data of three the variables linear cointegration model to forecast the irrigation water,the prediction error is bigger, and the absolute error is as high as 50% above; Using wavelet decomposition to establish multiple time scales linear cointegration model, the prediction results are relatively good, the absolute error is around 30%, but linear cointegration model for the forecast of irrigation water is poorer.(5) Through the wavelet neural network to building the nonlinear cointegration modeling of rainfall,crop water requirement and irrigation water,the results show that :there is a nonlinear cointegration relationship among the rainfall, crop water requirement and irrigation water, and the use of wavelet neural network model of nonlinear cointegration forecasting effect is good, the absolute error for the forecast of irrigation water basic control within 20%.