The paper presents some contemporary approaches to spatial environmental data analysis. The main topics are concentrated on the decision-oriented problems of environmental spatial data mining and modeling: valorization and representativity of data with the help of exploratory data analysis, spatial predictions, probabilistic and risk mapping, development and application of conditional stochastic simulation models. The innovative part of the paper presents integrated/hybrid model—machine learning (ML) residuals sequential simulations—MLRSS.

From its inception as a separate discipline, geostatis tics sought recognition from practitioners, not from math ematicians or physicists. and rightfully so. Indeed, the theory was essentially established by the 1950's by Kol mogorov and Wiener and exposed by Matern (1960), Whit tle (1963), and Matheron (1965), among others. But there is a long, hard way between a concept expressed by matrix notations in a Hilbert space and its implementation and routine application. It is my opinion that the main con tribution of geostatistics has been and still is implementa. tion, an essential follow-up step much too often forsaken by theoreticians. <...>

Geological data, notably geochemical data, often take the form of a regionalized composition. The concept of regionalized composition combines the concepts of composition and coregionalization. A composition, also known in the literature as a closed array (Chayes 1962), is a random vector whose components add up to a constant. A coregionalization is a set of two or more regionalized variables defined over the same spatial domain, which is modeled as a realization of a vector random function. Here the term regionalized composition is used both for the vector random function used to model a composition and for the realization that we can observe <...>

This text intends to be a technical one. This means that techniques to solve identified problems will be presented. As the theory which serves as a basis for these techniques is very new, and relatively unfamiliar to the mineral industry, several chapters or sections will be devoted to it. These two ideas of a technique and a theory have been my guideline in preparing this course on the geostatistical estimation of mineral resources. The main target was to stay, as much as possible, close to the practical problems. This is the reason for the many examples which are intermeshed with the text; however, in many cases, staying t o o close t o a problem obscures the broader frame into which a question has to be asked before finding a correct answer. This is the reason for some theoretical digressions, which may seem to some as an attempt to try and make things look complicated. Certainly, in a particular mine, many problems can be solved without a total understanding of the complete theory. On the other hand, when one considers all the problems occurring in different mines, one cannot hope to solve them without having a good grasp, a synthetic view of the theory of regionalized variables as developed by G. Matheron in France, the most advanced developments of which have just been published in the Proceedings of a N.A.T.O. Advanced Study Institute (Guarascio, Huijbregts, David, 1976) <...>