Abstract
The optimization of agricultural productivity by means of numerical weather models requires a detailed analysis of historical daily weather data in order to determine which areas best meet the climatic requirements of a particular crop. The determination of the profiles of climatic factors, in order to identify climatologically similar and closely correlated zones, provides an adequate method of transferring locally tested numerical crop-weather models from one area to others with the same agroclimatic conditions. The study was made on a small but meteorologically, orographically and agriculturally varied area. It is based on daily measurements of precipitation and of minimum and maximum temperatures, taken by 42 weather recording stations over the period 1961-1977. The climatic profiles are made by analyzing the spatial and temporal variability of meteorological paramaters in order to select the similarity function to be used in models of hierarchical clustering.