The modeling of germination and seedling emergence is required for the construction of a simulation model of three species of millet （panicum miliaceum, pennisetum galucum and setaria italica）. This study provides t...The modeling of germination and seedling emergence is required for the construction of a simulation model of three species of millet （panicum miliaceum, pennisetum galucum and setaria italica）. This study provides the necessary temperature parameters to model these processes. For this purpose, different non-linear regression models including fiat, logistic, quadratic, sigmoidal, dent-like, segmented, beta and curvilinear were used. Root Mean Square of Errors, coefficient of determination and regression coefficients of predicted values versus observed were used to find the appropriate model. Investigating regression coefficients indicated that dent-like model has the least RMSE and a coefficient （RMSE=0.000009, a=0.0006） and the biggest R2 and b coefficient （R2=0.96, b=0.98） in common millet. These coefficients were （RMSE=0.01, a=0.005） and （R2=0.94, b=0.97）, and （RMSE=0.004, a=0.05） and （R2=0.99, b=0.99）, for beta in foxtail and pearl millet, respectively. According to these coefficients, dent-like, was chosen as the best model to describe the response of common millet germination to temperature （Tb=7~C and Tc=49.50℃）. Also beta, was chosen for foxtail millet （Tb=7℃, Tc=49.50℃）. Beta, was chosen as the best model for pearl millet （Tb=6.5 ℃ and To=4 ℃ ）. These parameters can be used in millet simulation models to predict sowing to emergence duration based on a thermal time concept. Also, required biological days from sowing to emergence using these models varied from 3.57, 4.29 and 5.54, for common millet, foxtail millet and pearl millet, respectively.展开更多
文摘The modeling of germination and seedling emergence is required for the construction of a simulation model of three species of millet （panicum miliaceum, pennisetum galucum and setaria italica）. This study provides the necessary temperature parameters to model these processes. For this purpose, different non-linear regression models including fiat, logistic, quadratic, sigmoidal, dent-like, segmented, beta and curvilinear were used. Root Mean Square of Errors, coefficient of determination and regression coefficients of predicted values versus observed were used to find the appropriate model. Investigating regression coefficients indicated that dent-like model has the least RMSE and a coefficient （RMSE=0.000009, a=0.0006） and the biggest R2 and b coefficient （R2=0.96, b=0.98） in common millet. These coefficients were （RMSE=0.01, a=0.005） and （R2=0.94, b=0.97）, and （RMSE=0.004, a=0.05） and （R2=0.99, b=0.99）, for beta in foxtail and pearl millet, respectively. According to these coefficients, dent-like, was chosen as the best model to describe the response of common millet germination to temperature （Tb=7~C and Tc=49.50℃）. Also beta, was chosen for foxtail millet （Tb=7℃, Tc=49.50℃）. Beta, was chosen as the best model for pearl millet （Tb=6.5 ℃ and To=4 ℃ ）. These parameters can be used in millet simulation models to predict sowing to emergence duration based on a thermal time concept. Also, required biological days from sowing to emergence using these models varied from 3.57, 4.29 and 5.54, for common millet, foxtail millet and pearl millet, respectively.
