Interpolation of daily temperatures for a complex terrain of Slovenia

Weather station measurements, while valuable for localized information, often fall short in applications requiring detailed spatial temperature assessments, particularly in cases with complex terrain. There is a growing demand for gridded time series of daily temperature data across various scientific fields. Daily temperature grids are essential as reference datasets for analyses of climate models, climate change monitoring, spatial assessment of natural hazards (such as spring frost) and studies on climate impacts. 

Classical interpolation methods (e.g. universal kriging) that were used in the past often produce insufficient results for a complex terrain of Slovenia. Especially interpolation of daily minimum temperature has been proven very difficult with these methods. The most challenging cases involve small-scale and regionally varying inversions. To address this issue, deterministic method for the spatial interpolation of daily temperature by Frei (2014), originally applied in Switzerland, has been adapted for Slovenia. The methodology, which is based on the superimposition of non-linear vertical profile fields with non-Euclidean distance weighted residual fields, primarily focuses on complex terrains with mountains and intervening valleys, addressing the specific challenges of interpolation in Slovenia. Adaptations of Frei’s method included the interpolation of daily extreme temperatures alongside daily mean values, different spatial resolution and several interpolation settings.  

The above-mentioned interpolation method was tested for daily mean, minimum, and maximum temperatures. The primary data source for interpolation consisted of homogenized measurements from 183 stations (159 within Slovenia and 24 from the neighbouring countries). Spatial grids with resolution of 1 x 1 km were constructed for the territory of Slovenia.  

Preliminary qualitative comparison with classical interpolation techniques (e. g. universal kriging) shows that the adapted interpolation method results in improved representation of temperature fields especially in challenging scenarios involving small-scale inversions.  

We are currently working on extending the temperature analysis by integrating land use effects to represent even more local temperature patterns.