Modelling the effects of meteorological and geographical drivers on damage from late radiation frost on apple trees in Northeast Iran

Late frosts occurring in spring create significant bud damage and decrease yield of fruit trees, especially apples, in Northeast Iran. Assessment and risk modelling of late radiation frost damage would be useful to manage and decrease the damage. In order to model frost damage risk, 12 driving quantities were selected, including meteorological (minimum temperature, temperature decrease rate, temperature increase rate, Julian date of frost, cumulative degree days, area below zero line, and frost duration) and geographical variables (elevation, Longitude, Latitude, aspect, and slope). 3 damaging radiative frosts were selected in the period of apple flowering time, including 20 April 2003, 8 April 2005, and 28 March 2005 cases. Required meteorological data were collected from 9 meteorological standard stations in the apple cultivation area of Northeast Iran. Linear multiple regression was used to model the relationships. The map of each parameter was plotted by using interpolation methods including IDW (Inv. Distance Weighting), Spline, and Kriging. A 5 by 5 km grid map was defined in order to extract input data for the model. The regression equation is significant at the level of 5%. By using this equation, the predicted frost risk damage was calculated for each grid point. The regression equation of observed and predicted frost damage risk had a correlation of 0.93