Experimental Study On Dynamic Simulation Of Farmland Soil Moisture In Huaibei Plain

Soil moisture is one of the important parameters in regulating the soil-plant-atmosphere(SPAC)feedback system.At the same time,soil moisture is also the "four-water" conversion link,which is the main source of crop water absorption.It is very important to grasp the dynamic changes of soil moisture in the growing stage of crops for crop growth and to formulate a reasonable irrigation system.The dynamic movement of soil moisture is closely related to transpiration evaporation,phreatic evaporation,ground-water depth,meteoro-logical factors and crop physio-logical indicators.Due to the lack of multi-factor and multi-process monitoring,soil moisture dynamic simulation and estimation are uncertain.In order to further clarify the dynamic changes of soil moisture in the Huaibei Plain,and quantitatively identify the driving factors of soil moisture,the Wudaogou Hydrological Experimental Station 1984-2019 long series of soil moisture and hydrometeorological measured data was used as the key data support.The soil moisture prediction model of this model quantitatively evaluates the winter wheat-summer corn soil moisture prediction model.Based on a comprehensive analysis of soil moisture prediction models at home and abroad,this paper used the measured data of soil moisture and hydrometeorology from the Wudaogou Hydrographic Experimental Station series,using neural network method,grey prediction GM(1,N)method and time series method.In order to study the applicability of each model in the Huaibei Plain,the soil moisture prediction models for each growth period of winter wheat and summer corn were established.The specific research results are as follows:(1)Through a variety of experimental facilities,prototype observations of soil moisture,hydrometeorological elements and other data,combined with relevant data over the years,with reference to domestic and foreign research status,lay a theoretical foundation for the study of farmland soil moisture dynamic simulation in Huaibei Plain.(2)According to the principle of water balance,the dynamic changes of soil moisture under different scenarios are clarified.During the growth period of winter wheat-summer corn,the soil moisture regression coefficient is roughly inversely proportional to the buried depth.As the buried depth increases,the regression coefficient tends to be a fixed value,and the dispersion degree weakens;0.3m is the thickness of the soil layer where the soil moisture exchange is most active;Thedynamicchange trend of soil moisture in the 1m and 2m burial depths of the bare ground is the same.The soil moisture values in the 1m burial depth 0-20 cm,20-40 cm and40-60 cm soil layers are all greater than the 2m burial depth,60-80 cm vice versa;The dynamic changes of soil moisture at the depth of 2m and the depth of 2m are basically the same.In the early stage of winter wheat growth,the soil moisture at each depth of1 m and 2m is basically the same from the middle of the month to the score.Buried depth greater than 2m,from the beginning of March to the end of May,the soil moisture values of the 1m buried depth and the 2m buried depth of 0-20 cm,20-40 cm and40-60 cm soil layers are basically the same;Small,non-bare land soil moisture value fluctuates greatly;2m submerged deep bare land soil moisture value is greater than non-bare land,except for wheat overwintering period.(3)Based on path analysis and grey correlation analysis,this paper identifies soil water driving factors.Establish the correlation between soil moisture and hydrometeorological elements of different soil layers during the growth period of winter wheat and summer corn.Path analysis of wheat shows significant effects of T,GT,P,and A.Grey correlation analysis shows significant effects of GT,RH,△E,aH,and T.Analysis of path analysis of corn shows significant effects of RH,A,and △E Grey correlation analysis shows that GT,△E,RH,V,d et al are significant.(4)Discussion on Soil Moisture Prediction Model Based on Multi-element Driving and Quantitative Evaluation of Model.During the wheat growth period,the prediction accuracy of the BP neural network,GM(1,N)and ARIMA models are 0.823,0.676 and 0.992,respectively,and the ARIMA model is the best.During the growth period of corn,the prediction accuracy of BP neural network,GM(1,N)and ARIMA models are 0.926 and 0.724,respectively.The ARIMA model is not applicable,and the prediction accuracy of BP neural network is the highest.