Chernobyl Fallout: Review of Advanced Spatial Data Analysis

The paper is based on the long experience in applying various spatial data analysis methods and approaches to the radioactive pollution data after the Chernobyl accident. Application of three basic approaches are described in the work: geostatistics (both parametric and non-parametric models), artificial neural networks (ANN) and monitoring network analysis, including fractals and cluster analysis. There have been developed hybrid geostatistical-ANN models, which show better performance then one approach models in many complex non-stationary cases.