Computer modelling the map of average annual temperature in the ground layer of air is performed for the southern part of Western Siberia (Russia). Four methods for data interpolation were used in ArcMap 9. Procedure of creation of digital model is described in detail. An original mathematic way for ranking of methods is proposed. According to results, Ordinary Kriging method gives the best approximation to the initial data.
 Tsaregorodcev, V. G., & Pogrebnaya, N. A. (1998). Neural Network Method of Information Processing in the Prediction Problem Climatic Characteristics and Silvicultural Properties of Landscape Zones. Methods of Neuroinformatics. Collection of Scientific Works, Krasnoyarsk, 65-110.