Study On Soil Moisture Change Rules And Prediction Models Base On GIS In Jiefangzha Irrigation District

Although water shortage is a serious problem in our country, the problems of lower water use efficiency still widely exists, especially in agricultural water use. The water usage volume of agriculture accounts for 75% of the total amount. In arid or semi arid zone, growth of farm crop mainly depends on irrigation. Hetao Irrigation District is one of China's three major irrigation district, the ultimate purpose of establishing promotes the efficiency of water use, to increase the yield and improve the quality of crop, which is the inexorable trend of the future develonment of the irrigation agticulture. Technicians, both home and abroad, have already done some exploratory and practical work in this field, and have also developed some systems. However, these systems still have some defects in operation, user-machine interface, and the reusability of the program.Soil moisture forecast is the forecast of increment and depletion of field soil moisture based on soil water simulation models. It is the basis for irrigation forecasting and is important for the rational regulation of field water.Based on the observed data of soil moisture and rainfall from 2004 to 2005 in Jiefangzha irrigation district(2006,2007 as reference),this paper studies the spatial-temporal variety regularity of soil moisture in cropland during the growth period of main crops.The author pointed out that the seasonal variety of soil moisture can be divided into three phases as fluctuations phase,moisture loss phase,and moistur-storage phase,and periodic changes in interannual,the vertical variety of soil moisture clearly emerges as three layers,it is lower moisture contents in surface layer, active in middle layer, and relative steady in bottom layer.Creating the overall trend surface,to analyze trend of measured value from space latitude and longitude,Irrigation domain can be seen in the southwest of the low soil moisture, soil moisture value of the central region generally on the high side, the northeast soil moisture values slightly lower than the central region of soil moisture. In the overall trend analysis based on the application of geostatistical software, rationally determine the future soil moisture monitoring cloth point distance.In the paper, we establish the experience model of soil moisture forecast, stepwise regression model, the soil water balance model and BP neural network model, and test these models by substantive data. The results showed that experience model have hight forecast accuracy in the whole irrigation pevioel;Stepwise regression model as a multiple regression model, the sampling factor and the volume of data, this paper only partial irrigation domain summer of moisture forecast, forecast accuracy higher will be able to meet forecast demand; Forecasting accuracy of the soil water balance model and BP neural network mode are slightly lower than experience model, these maybe relate to more coefficients and less information, experimental means and precision of equipments.