Spatial analysis of reconstructed mine soils: Soil survey, statistical modeling and terrain analysis for land resource inventory

Mining companies and regulatory agencies need clearly defined methods by which sample data of reconstructed mine soils can be interpolated to determine soil spatial variability and suitability for reclamation. Objectives of this study were to examine the distribution of reconstructed mine soils from several perspectives: soil survey, spatial statistics, and terrain modeling. Mine soils, in this context, provided a special case for a larger discussion of soil resource inventory in general.;Initially, mine soils at the Rosebud Mine in Colstrip, Montana were mapped using standard soil survey procedures. Reconstructed mine soils are uniquely different from their native counterparts. They provide a uniformly deep soil substrate for plant roots and have been largely homogenized by soil salvaging. Erratic spatial variations in soil textures are the result of mixed sedimentary parent materials and the reclamation process.;Spatial statistics were used to assess the spatial distribution of mine soil attributes from data collected by Western Energy Company. Soil attributes, in all cases, were spatially independent at the 300 foot sample spacing used at the mine. Kriging was deemed unwarranted due to spatial independence of the data and more traditional statistical methods that rely on independent data assumptions were used to interpolate the data. For many soil properties, a constant surface through the overall sample mean provided the best prediction at unsampled locations.;Initial results were tested further using data collected exclusively for application of spatial statistics. Closer grid spacing resulted in semivariograms exhibiting weak to moderate spatial dependence for subsoil attributes. Despite the empirical evidence of spatial correlations, kriging estimates did not outperform use of the field mean in predicting measured values of an independent data set. Knowledge about the physical processes controlling spatial distributions of soil properties appears to be an important, yet often overlooked, consideration in decisions about the appropriateness of applying kriging techniques.;Terrain models provide a unique vantage point to study how mine soils and reconstructed landscapes will evolve in the future. A terrain model generated for the Area-E portion of the Rosebud Mine provides the basis for discussion of changes that are certain to occur in the reclamation resource.