Geostatistics. Modeling spatial uncertainty / Геостатистика. Моделирование пространственной неопределенности

Geostatistics aims at providing quantitative descriptions of natural variables distributed in space or in time and space. Examples of such variables are 

Ore grades in a mineral deposit 

Depth and thickness of a geological layer 

Porosity and permeability in a porous medium 

Density of trees of a certain species in a forest 

Soil properties in a region 

Rainfall over a catchment area 

Pressure, temperature, and wind velocity in the atmosphere
Concentrations of pollutants in a contaminated site
These variables exhibit an immense complexity of detail that precludes a description by simplistic models such as constant values within polygons, or even by standard well-behaved mathematical functions. Furthermore, for economic reasons, these variables are often sampled very sparsely. In the petroleum industry, for example, the volume of rock sampled typically represents a minute fraction of the total volume of a hydrocarbon reservoir. The following figures, from the Brent field in the North Sea, illustrate the orders of magnitude of the
volume fractions investigated by each type of data (“cuttings” are drilling debris, and “logging” data are geophysical measurements in a wellbore):

Cores 0.000 000 001
Cuttings 0.000 000 007
Logging 0.000 001 <...>