| Soil moisture content which is briefly called soil moisture is the status of soil water in given soil. It is an important component element of soil and composes of the fertility of soil. It is the basic condition of field crop growing, for the move of the soil components include plant assimilate nutrient from soil doesn't leave the function of soil moisture. When the soil moisture can't meet the need of crop natural growth, irrigation becomes the best efficient way to improve soil moisture status and be content the crop water requirement. So with the more and more shortages of water resources precise forecasting the change of soil moisture and fact crop water requirement becomes more important, at the same time it is very important to protect the general agriculture yield and to make certain the ration of farmland irrigation. The article takes the Shanxi Fenhe irrigation district for example and bases on the field experiment and analysis to study the forecast of soil moisture and gets the follow conclusion: (1) Based on the formula of Penman and Penman-Monteith, set up visual module to calculate crop water requirement. (2) Applying of the data such as soil moisture content, irrigation, etc. of Fenhe district from 1993 to 2002 analyzed the annual change, seasonal change, vertical change of the field soil moisture content. And set up two types three layers BP artificial neural network to forecast soil moisture. The model BP(6,8,1) and BP(7,8,1) of forecasting 0~80cm had a maximal relative error which was 12.6%, 13.2%, and average relative error 4.9%, 3.7%, respectively; and 2.7%, 3.0% was the maximal absolute error and 0.9%, 0.7% was the average absolute error respectively. The model BP(5,8,1) of forecasting 20~40cm had the maximal relative error, average relative error, maximal absolute error, average absolute error shoes were 14.5%, 4.7%, 3.0%, 0.9%,respectively. They can be used to practice. (3) Use the field experiment data of wheat root density and e exponential function to get the winter wheat root density function. (4) Based on the winter wheat root density function set up mended Feddes wheat root water uptake model; Set up the wheat root water uptake model with the field experiment data of soil moisture; citing Kang Shaozhong model of wheat root water uptake with parameter 's testing and adjusting. With the three models it get the compare of forecasting value and measure. For mended Feddes model, Kang Shaozhong model and the other there were maximal relative error which were 14.7%, 14.25%, 14.36% and average relative error were 3.33%, 4.21%, 3.75%, respectively. (7) Put up the visual procedure which uses the soil moisture data to calculate the ration of root water uptake and numerical method to simulate the soil　moisture.　(8) The simulated system is very simple in design. The forms of model and the parameters needed are given in each widow and at the time there is a help window to tell the user the model's shortcoming and strongpoint. The user only needs to choose the parameters then clicks mouse to get the result.
|
PDF Full Text Request