Introduction to Disjunctive Kriging and Non-Linear Geostatistics

Although linear geostatistics was first developed to tackle problems posed in evaluating bodies of ore in mines, it has since been extended to other fields of activity. Tools such as the variogram and kriging are now in everyday use for estimating the mean grade of material in an area from values achieved from measured samples. Not all the problems in the management of natural resources can be tackled by the methods of geostatistics, however. Non-linear geostatistics must be used if the probability of an actual but unknown concentraion at a given location exceeding a particular threshold needs to be assessed. Although this is in the forefront of statistical development, it has a potential application over a wide range of practical problems and is likely to occupy an important place in the statutory monitoring and control of air, water, and land pollution. Dr Rivoirard addresses the lack of any comprehensive book on this subject and the very mathematical nature of the publications that do exist by giving a description in detail of the basic methods and models of non-linear geostatistics, with the emphasis on concepts rather than on mathematics. This book is intended for academics, practitioners, and postgraduate students familiar with linear geostatistics. Likely departments are physical (hydrology, meteorology, oceanography, soil sciences) and economic (agriculture, mining, fishing) geography, natural resources management, environmental sciences, mining engineering.