Spatial Interpolation of Mean Yearly Precipitation using Universal Kriging

Damijana Kastelec and Katarina Košmelj

Abstract

The objective of our work was to find an appropriate method for the spatial interpolation of mean yearly precipitation (MYP) into a regular grid with 1 km resolution, for Slovenia. A geostatistical approach and the universal kriging method was chosen. In the analysis, the dependence of MYP on geographical variables like altitude, latitude and longitude was considered.

In the first step, we used exploratory methods to examine spatial variability of MYP and its dependence on geographical variables. The positive trend of MYP with altitude and the negative trend of MYP with longitude were taken into account in spatial interpolation.

The spherical variogram model was chosen, considering variogram anisotropy. On the basis of cross-validation technique we chose for the influential surrounding the ellipse with the longer axis 80 km. The universal kriging results were encouraging and comparable with the subjectively obtained map (Zupančič, 1996; Zupančič, 1998). The analysis showed that the results for the eastern part of Slovenia are better than for the western mountainous part. This is due to the lack of observational stations in the mountainous part of Slovenia where the spatial variability of MYP is high.