Simulation Analysis Of Soil Moisture During Summer Maize Winter Wheat Rotation In Huaibei Region

Huaibei,located at the boundary between the north and the south of China,is an important crop planting and grain production base in China,due to the large interannual variation of rainfall,uneven distribution within the year,and the alternation of drought and flood,the water resources in this area are relatively scarce.Soil water,as the main source of water absorption by crops,is an important basis for promoting the normal development of crops and ensuring the normal operation of agricultural production activities.In today's water shortage,in order to improve the effective utilization rate of farmland irrigation water,avoid serious waste of water resources,timely and accurately grasp the dynamic law of farmland soil moisture change is of great significance to formulate timely and appropriate irrigation.Based on Wudaogou hydrological experimental station,the soil moisture during summer maize-winter wheat rotation was studied.The field soil moisture,groundwater table depth and meteorological data of the experimental station from June 1991 to may 2020 were used.The relationship between soil water and groundwater during the growth period of summer maize-winter wheat was analyzed,and 10 kinds of curve equations were fitted;The correlation between soil moisture and hydrometeorological factors during summer-maize winter wheat rotation was analyzed,and the multiple linear regression equation of soil moisture was established;BP neural network and genetic algorithm were used to optimize the BP neural network method to establish the prediction model of soil moisture during crop rotation.The results show that:(1)The relationship between soil moisture and groundwater depth during the growth period of summer maize and winter wheat was analyzed,ten kinds of functions are selected to simulate the transformation of soil water and groundwater,The results show that:The Fourier function has the highest fitting accuracy,the fitting accuracy of summer maize growth period was 0.8185,and that of winter wheat was0.9505.(2)The correlation between soil moisture and hydrometeorological factors during summer maize-winter wheat rotation was analyzed,taking the most relevant influence factor as the initial variable,the four factors were added gradually,the prediction models of soil moisture in different growth stages of summer maize were established,with the increase of variables,the model fitting accuracy is higher,when there are four variables,R~2 is 0.742?0.979.Except for the surface soil layer,the average relative error can be controlled within 0.1.(3)Ten meteorological factors which have direct influence on soil moisture content were selected,BP neural network model was used to establish soil moisture prediction models of different soil layers(0-20cm,20-40cm,40-60cm,60-80cm,80-100cm),the models were BP(10-16-1),BP(10-11-1),BP(10-8-1),BP(10-14-1),BP(10-14-1).The prediction results show that the relative error of soil surface layer is large,but with the increase of soil layer,the error is controlled within 0.09.(4)In order to improve the prediction accuracy,genetic algorithm with global search ability is introduced to optimize the initial weights and thresholds of BP neural network,so it can better predict the output.The results show that the relative error of soil moisture decreases with the increase of soil layer,and the relative error is controlled within 0.065.