Research On Water Use Efficiency Of Corn In Beijing Area Based On Crop Model

In order to select a simple and efficient method for replacing PM with the reference crop evapotranspiration method in Beijing and to explore the long-term changes and influencing factors of corn water use efficiency,the meteorological data of Beijing from January 1957 to December 2019 were used,10 different empirical models and different combinations of 2 machine learning models were used to estimate the evapotranspiration(ET₀)of reference crops in Beijing,and the long-term leaf area index,yield,elements such as soil moisture,water consumption were used to calculate the long-term water use efficiency,and the WOFOST model was used to simulate the change of water use efficiency.turn out:The sunshine hours,relative humidity and average wind speed in Beijing decreased significantly with the increase of the year,and the change rates were-0.18h/10a,-1.5%/10a,-0.04m/s/10a.The highest temperature,the lowest temperature,The average temperature increased significantly with the increase of the year,and the change rates were 0.41?/10a,0.23?/10a,0.50?/10a.The reference crop evapotranspiration showed an increasing trend with the change of the year.The correlation between each meteorological element and the reference crop evapotranspiration is significant,and the absolute value of the correlation coefficient between each meteorological element and the reference crop evapotranspiration is n(0.76)>T_max(0.67)>T_a(0.61)>T_min(0.47)>RH(0.43)>u₂(0.36).Among the reference crop evapotranspiration in Beijing calculated by various models,in Beijing area,when the input parameters are maximum temperature,minimum temperature,relative humidity,sunshine hours and wind speed,the application effect of SVM(support vector machine)model is better than that of XGB(Extreme Gradient boosting)model;when the input data is only the highest temperature,the lowest temperature and the number of sunshine hours,the simulation accuracy of the Priestley-Taylor model is better than the machine learning model and other radiation methods,and in the machine learning model,XGB is better than SVM;When there is only temperature data,the simulation accuracy of the machine learning model is better than that of the empirical model,and the simulation accuracy of the XGB model is better than that of the SVM model;in the absence of wind speed data,the simulation accuracy of the machine learning model is better than that of the empirical model.Among the models,the XGB model is better than the SVM model;in the case of incomplete input parameters,the simulation accuracy of the XGB model is always better than that of the SVM model.In the past 10 years,the water consumption of crops in Daxing District of Beijing has shown a significant upward trend.The peak water consumption of corn is mainly concentrated in the middle growth period,jointing stage and heading stage.The crop coefficient has shown a significant upward trend,the leaf area index has shown a downward trend,and the yield has been slow.The water use efficiency showed a significant downward trend,and the yield was significantly affected by the water use efficiency.Reasonable fertilization can effectively improve crop yield and water use efficiency.When the fertilization amount reaches 0.0225kg/m2,the improvement of yield and water use efficiency basically reaches the highest value.Irrigation of 60 mm before sowing does not necessarily improve yield and water use efficiency.Only in years with less effective rainfall during the growth period can irrigation be beneficial to improve yield and water use efficiency.The effect of fertilization and irrigation on soil water content is obvious,but the law is not very obvious.The WOFOST model can effectively simulate the corn yield in Daxing District of Beijing after the parameters are calibrated.The leaf area index changes dynamically,and the error of the simulated leaf area index is only 24%.The evapotranspiration calculated by the crop coefficient simulated by the dynamic leaf area index is closest to the evapotranspiration measured by the eddy correlation method,with an error of only31%;The factors are average temperature,followed by leaf area index,and finally average wind speed.