Parameter Sensitivity Analysis and Optimization of Sesame Phenology Simulation Model
Phenological development is critical for crop field management.In order to accurately predict the development process of sesame by quantifying the response of sesame physiological processes to temperature and photoperiod,a sesame phenology simulation model(SPSM)based on clock model theory was constructed.The sensitivity of parameters was analyzed by extened fourier amplitude sensitivity test(EFAST)to determine the high sensitivity parameters affecting the development process.The parameters of four sesame varieties were determined by the Monte Carlo optimization method and verified in the main sesame producing areas of Henan,Hubei,Anhui and Jiangxi provinces in China.The results showed that the average global sensitivity indexes of optimum temperature(Topt),basic development coefficient(k),and critical day length(Dc)were the highest among the seven cultivar parameters of SPSM.The R2 of the simulated and observed values of Yuzhi 4,Zhongzhi 22,Wanzhi 10 and Ganzhi 7 were 0.999,RMSE was 0.50 to 1.32 d,NRMSE was 1.06%—3.02%,and d value was 0.999.By independent verification,the standard deviations of the development period simulations and observations of the above four sesame varieties were 0-7.65 d,R2 were 0.978,0.988,0.992,0.989,RMSE were 5.07,4.43,3.07,4.24 d,NRMSE and Willmott consistency(d)were 5.87%—9.44%and 0.996-0.997,respectively.The optimized SPSM can well simulate the development dynamics of sesame in the main producing areas of sesame in China.
万方数据
维普
