Andreas Persson

This study has evaluated the sampling density for creation of high-resolution digital elevation models (DEMs) for precision agriculture purposes. The relationships between yield and topographical factors were investigated in a study area located in the central Sweden province of Dalarna. The DEM data sampling was carried out with a RTK-GPS system. A dense sampling scheme was employed and data was divided into two for both interpolation and validation. Kriging interpolation was used for DEM generation. From the DEM, topographical parameters were extracted and topographical indices were estimated. The indices were calculated with slope length and its vertical and horizontal components. The drainage area for a point of interest and the relationship of this area to the total drainage area were also estimated. The relationship of yield and the topographical parameters and indices was investigated using
both circular and spatial statistics. A spatial regression was used to calculate a model for the relationship. Up to 20% of the yield could be explained in the final model for one of the fields.