Prediction of spring radiation frost by an empirical model (Case study of Qazvin Plain)

Late spring frost cause a lot of damage to the agricultural sector every year. Prediction of this phenomenon is needed to active protection of plants. In this research, using FAO experimental method, daily and hourly data of two meteorological stations were used to determine the coefficients of the experimental model for prediction radiation frost in Qazvin Plain. also, in order to investigate the climatic condition of spring frost, the daily minimum temperature data of Qazvin and Buinzahra stations were used. The analysis of sixty years data in Qazvin stations showed that the intensity of frost has decreased during these years, but frequency of frost in Ordibehesht month has increased. Air, dew-point at two hours after sunset and minimum temperature relate 25 events of radiation frost at Simorgh station were used for regional coefficient calculation based on two models. These models were evaluated using 14 events of radiation frost at Tat stations. The mean absolute error(MAE) for testing and evaluating of Model1 was 0.71℃ and 1.21℃ and for Model2 was 0.67℃ and 1.09℃. The findings also showed that both models have acceptable accuracy in estimating the minimum temperature of the next day. It is proposed that these two models can be used for prediction of radiation frost in other regions.