Analysis of the Soil Properties Spatial Heterogeneity by Means of Engineering Geological Modelling
Authors
Gulmira S. Amanova, Eugene A. Voznesensky
DOI
Abstract
The problem of soil properties heterogeneity in the massif and possible approaches of evaluating its regularities by means of three-dimensional numerical modelling of spatial fields of the studied parameters within the framework of geostatistical approach is considered.
It can be demonstrated that the structure of geological parameters fields (a) differs for different stratigraphic-genetic complexes of soils and rocks, (b) can be simpler than the combination of engineering-geological units, (c) the proposed approach is able to identify and explain significant deviations from the normal or lognormal distribution in its upper and lower quartiles, presenting practically in any geological subdivisions of the section. These outbursts in the maximum and minimum ranges are themselves an additional characteristic of soil heterogeneity; they testify to the non-randomness of such "outbursts", their typicality, which should be taken into account, though usually rejected in standard statistical processing due to the smallness of data samples. Such analysis of heterogeneity allows to see the overall picture of the massif structure or part of it, and can be especially useful in case of mottled structure and frequent alternation of soil varieties, complicating detailed division of the soil profile. It can also provide additional important information like cracking distribution and potential for the development of unfavorable geological processes in them, like karst and suffusion.
Unlike a formal approach of subdividing the soil profile into a set of engineering geological units the proposed differentiation is based initially on composing the geological model consisting of bodies (strata, lenses, dykes, etc.) with varying origin, age and lithology and separated by distinct boundary surfaces. Additional internal boundaries in these 3D bodies are, on the contrary, diffusive and based on the analysis of spatial fields of the studied soil parameters (composition, density, moduli, etc.) by means of appropriate geostatistical procedures within a specialized software.