Abstract
In this study, we evaluated the usefulness of six interpolation methods selected to model statistical surfaces using forest cover and population density. The three criteria of consistency of the mapped range of data values, preservation of the volume of the phenomenon, and visual perception of the maps by users were used. The first two criteria were verified experimentally, and the third by interviewing a group of advanced map users. The study results represent a compromise between the mathematical correctness of the maps and their graphical properties. Therefore, they indicate that the optimal interpolation is linear interpolation followed by natural neighbourhood and IDW. This was investigated by interviewing a group of advanced map users, who preferred linear interpolation followed by natural neighbour and IDW maps and their graphical properties.