Spatialization of climatic data at the Italian national level by local regressive models

Abstract

The availability of spatialised climatic data is an essential pre-requisite for the implementation of GIS-based analysis in many application fields. Among the different methodologies for the spatialization of climatic data collected in weather-stations the most used are those based on geostatistical approaches, on parametric correlative models or on neural networks. Within the “Completamento delle Conoscenze Naturalistiche di Base” project, funded by the Italian Ministry for the Environment (Department of Nature Protection) a database of 403 weather-stations distributed across Italy with a time series of thirty years was collected. Data of mean monthly temperature (minimum and maximum) and rainfalls were spatialized by a local linear univariate regressive method based on elevation as independent variable. A total of 36 monthly maps with a geometric resolution of 250 m was generated. The present paper introduces the adopted methodology and the accuracy results estimated by leave-one-out cross validation.

Keywords

• Climate
• Meteorological data
• Spatialization
• Regressive models
• Italy.