Study On The Simplified Compuation And Forecasting Models Of The Reference Crop Evaporanspiration

As both fractal and chaos theory have greatly auto-adapted capacity when being used to solve the non-linear problem with multiple parameters, both theories are adopted to study the simplified computation and forecast models of the reference crop evapotranspiration( ET0 )under the default meteorological data condition. So the paper depended on National Key Technology R&D Program(2006BAD11B04) and National Natural Science Foundation of China(50879072). This thesis further verified the reasonable relationship between the net shortwave radiation and net radiation of plant surface temperature according to the self-similarity of fractal theory, consequently simplifies the computation of ET0 . Meanwhile, bilinear theory is used to compute the ET0 , and chaos theoryused to analyze and predict the characteristics of ET0 . The result can provide a simple compute and forecast method for solving the complicated ET0 problem. The results and conclusions are as follows:1)Establishing the temperature and radiation computation model with the fractal theory based on analyzing the fractal characteristic between the temperature and radiation. Through analyzing the tendency between the temperature and the theoretical solar radiation, the long-wave radiation of plant surface, the shortwave length radiation and net radiation, the models of temperature and net shortwave radiation and net radiation are established, and further tested the good relationship between the elements of both models with fractal theory.2) The established model is verified through analyzing the errow between the Penman-Monteith Equation and the computation result by the simplified model. The results predict that there is small absolute error and relative error between model and equation. What's more, the simplified model has high precision, therefore it can be used to calculate the reference crop evapotranspiration.3) The bilinear theory is used to calculate the reference crop evapotranspiration, and set up the bilinear model of reference crop evapotranspiration with temperature and net radiation. What's more, it also establishes the simplified bilinear model to predict the reference crop evapotranspiration through temperature. With the comparison of average absolute error and relative error between P-M equation and bilinear simplified model, it presents that there is no big difference from both results in Suining and Luzhou. Consequently, the bilinear model can be used to calculate the reference crop evapotranspiration.4) Taking Chengdu and other weather stations as example, the bilinear model is established by choosing two factors from temperature, the relative humidity and radiates, and it is adopted to compute the reference crop evapotranspiration. The new bilinear model is built by the use of the computing reference crop evapotranspiration, and so on, five times in all. The precisions are high except the model with temperature and relative humidity. Meanwhile, the model of temperature and net radiation has high precision to calculate the reference crop evapotranspiration in four sampling stations. So this model can be used to obtain reference crop evapotranspiration value shortly and correctly.5) Using the Welch method, the small data quantity method, G-P the connection peace-keeping C-C method, the time series of reference crop evapotranspiration is recognized and analysed by the chaos theory, and predicted by the RBF neural network forecast technique and the Volterra progression adaptive prediction model. The result indicated that the reference crop evapotranspiration has certain chaos characteristic and the complex polytropy, when the minimum embedding dimension is 6 and the corresponding dimension is 2.025, and the maximum Lyapunov index is bigger than zero. Through comparing the predicting errors by the RBF neural network and the Volterra progression auto-adapted method, a RBF neural network step forecasting result error is smaller and the forecast precision is high. Therefore, the RBF neural network model can be used to predict the reference crop evapotranspiration in practice.