Using nonlinear regression approach for prediction of cardinal temperature of canola and four common weeds

This study was done to evaluate different nonlinear regression models to describe response of germination rate to temperature in Brassica napus L. (canola), Sinapis arvensis L. (wild mustard), Raphanus raphanistrum L. (wild radish), Descurania sophia L., Phalaris minor L. (phalaris). This experiment based on completely block randomized design with 3 replications at Plant Research Institute (Dep. of Weed Research) laboratory in 2011. The seeds treated with different temperatures (0, 5, 10, 15, 20, 25, 30, 35, 40 and 45 oC) and the germination rate were calculated. The result showed that temperature had significant effect on germination rate of all plants. Tree regression models (Dent-like, Segmented and Beta) used to predict germination rate and cardinal temperature. Root mean square of error (RMSE), coefficient of determination and corrected AKAIKE index (AICc) were used to find the appropriate model(s). Therefore segmented model were superior compared to other models in canola and all weeds. It was concluded that this model can be used to quantify response of common weeds of canola field germination to temperature and to obtain cardinal temperature of germination. Base (Tb), optimum (To) and ceiling (Tc) temperature were predicted with appropriate models. Tb, To, Tc were for canola 1.8, 25, 40.9; wild mustard 2.01, 15, 30.6, wild radish 1.9, 15, 30.1, D. sophia 1.2, 26.9, 35 and phalaris 1.9, 15.5, 30.3 respectively. These parameters are required to predict common weeds of canola field germination and emergence.