Setup and test of a simple model for prediction of late frost events over complex terrain

simple algorithm for prediction of spring frost is presented along with results form its application to reproduce observed temperature minima at 24 meteorological stations in Trentino (Italy). The method is based on the model proposed by Reuter (1947, 1951) where the cooling process is assumed proportional to the square root of time after sunset by means of a decay coefficient. Data recorded during the periods March-May 2003 and 2004 have been used to determine an optimal value of the decay coefficient embedded in the algorithm. The value for this coefficient is obtained through a multiple linear regression which involves meteorological variables as wind speed, air temperature and relative humidity. The performance of the method is satisfactory in reproducing the time evolution of the process. For most of the stations the method provide a satisfactory tool for issuing a frost alert within 1.5°C of tolerance.