Analysis and modeling of spatial environmental data / Анализ и моделирование пространственных данных окружающей среды (экология)

For the analysis of spatially distributed data, a fundamental aspect is first considered: can the data be interpolated, or are they of such a type that an interpolation between two values measured at different locations makes no sense? In the case of interpolated data, one will usually encounter variables like chemical grades or thickness which vary more or less slowly from one location to the other, and for which some assumption of stepwise variability appears relevant. The methods presented in this book apply mainly to this type of data, either for pollution measurements or for ore bodies grades.

Another case of interpolated data is constituted by categorical data, where the value inferred at a non-measured location is not a value contained in the surrounding values, but a value to be chosen among a given set of categories, for instance a soil type. \Vhen interpolating between two points on a map which present different categories of soils, one has usually not to define a "percentage" of category between the categories of each point, but to define the spatial frontier where the soil category would change from one to the other decision boundary between classes should be found using available data and knowledge <...>