| In recent years, new changes appeared on regional evaporation features with theinfluence of climate change and human activity. Objective to carry out the research ofspatio-temporal variability analysis, simulation and prediction of the reference cropevapotranspiration can provide scientific basis to the laws of the impact of climate changeupon water resource, the comprehensive planning and management of agricultural waterresource, the sustainable development model of formulating regional social and economic.Using data from all the major sites of Baojixia Irrigation, this paper makes a systematicanalysis of Winter wheat crop water demand and water shortage and spatial variation lawsand its influencing factors in Bjixia Irrigation, the prediction of the reference crop’sevapotranspiration were studied which based on gray system. The main conclusions in thisthesis have been summarized as follows:(1) Calculated the Baojixia Irrigation27sites reference crop’s water requirement andwinter wheat water requirement, winter wheat water deficit. Researched the interannualdistribution of the27sites reference crop water requirement, winter wheat water requirement,winter wheat water deficit, between31years. Selected the irrigation area5on behalf of thesites， Researched interannual variability and annual variability of the site’s eachcalculation,more clear understanding the distribution regularity of each calculation.(2) Due to the limitations of the observational sites, led directly to a low interpolationaccuracy for certain areas in Baojixia Irrigation. Take latitude, longitude, elevation, slope andaspect as the input items of the site to get the reference crop evapotranspiration data with BPneural network, and the estimation accuracy is better.(3) In this paper, the Overall Polynomial interpolation method, the Inverse DistanceWeighted interpolation method, Kriging method is used to do the spatial interpolation of33sites reference crop water requirement and winter wheat water requirement, winter wheat waterdeficit, got the space interpolation contour map and the mean error and the mean square errorof each interpolation method. In addition，using the interpolation contour map read theprediction of five sites, and calculate the error. through compare the result show that, the bestinterpolation method is not the same of each area, but In general Universal Kriginginterpolation method is the best,its error is minimum. With Universal Kriging interpolationmethod for interpolation and prediction of reference crop water requirement, winter wheatwater requirement and winter wheat water deficit is the most effective. （4）SPSS is used to analysis the correlation between each meteorological factor andwinter wheat water requirement and water deficit, obtain the correlation coefficient between5representative sites and winter wheat water requirement and water shortage, it is clear thatprecipitation, relative humidity and winter wheat water requirement is a negative correlation,sunshine hours mean temperature and winter wheat water requirement is a positive correlation,Wind speed has little effect on winter wheat water requirement. In the correlation analysisbetween water deficit and meteorological factor can be seen the greatest impact ofmeteorological factors to water deficit in different growth cycle is different, it played aguiding role in the irrigation area of reasonable irrigation.（5）This paper using site5as an example, used the genetic algorithm to optimize BPneural network to forecast the lack of sunshine duration data from September1,1983to10based on the site meteorological data（precipitation, humidity, sunshine hours, averagetemperature） during the period of1980~2010. Still using site5as an example, themeteorological data which has a maximum correlation with water deficit is used to forecastwinter wheat water deficit data of each quarter in irrigation area by grey neural network,because of the high prediction accuracy, it played a key role for improvement the real data. |
PDF Full Text Request