Khorasan Province is known as the major Saffron production area of Iran. In this study, Crop Weather Models were applied to predict Saffron yield in main production areas of the province (Birjand, Torbat Heidariye, Ferdows and Gonabad cities) by using 20 years data. Four predictors including temperature, rainfall, relative humidity and sunshine hours were considered for Model , and two weather parameters including temperature and rainfall were used for Model . These two models were applied using stepwise regression analysis and JMP statistical Software. The results indicated that the coefficients of determination of model were higher than of model . The results also showed that the models can predict Saffron yield in Gonabad city more precisely. The coefficients of determination for Gonabad were 0.77 and 0.74 for model and model respectively. The least coefficient of determination was 0.38 for model in Torbat Heidariye. It was also concluded that during recent years the increasing trend in mean air temperature during the spring season was the main cause of Saffron yield decline across the province.

Keywords