Geostatistic modeling of temperature data for agricultural land evaluation

Abstract

For low spatial resolution climate data (in Romania case) and relatively small study areas (used in land evaluation) classical methods of interpolation of climate data does not give significant spatial resolution results. To illustrate the use of geostatistic in solving this issue, we have compiled a database of annual average temperatures for the period 1947-2010 in Plateau of Moldavia, further used for applying kriging combined with regression of residues. Using this method and additional data for modeling multiple regression, we have been able to obtain a statistically and spatial valid resolution for an area with a surface of 949.3 km² located in the proximity of Iasi.