Overburden Volume Estimation Assisted by Geostatistics in Open Cast Coal Mine

This paper compares different results for overburden volume estimation assisted by geoestatistics, particularly using kriging methods. The main goal is to obtain the most reliable digital terrain model for the calculation of the overburden volume. The data was collected on an irregularly spaced grid and interpolated values were calculated to provide a denser, more regular data set before volume estimation. The field work was developed at the Butia-Leste Coal Mine, southern most Brazil, where an accurate survey was done using a total station theodolite with electronic data collector. The study was developed in five main steps: data acquisition, exploratory data analysis, spatial continuity measurements, point and block kriging estimation and verification of the results. Data acquisition was carried out on a 5 x 5 in grid. This dense grid provides an exhaustive data set of the real terrain against which data subsets can be compared for the various interpolation procedures and volume calculation. This provides an interesting insight into the influence of the number of data on the estimation error. Univariate statistics were used to compare each data subset and the exhaustive dataset. Ordinary kriging on the data subsets followed by contouring was carried out using Geoeas and Surfer software respectively. A subroutine written in C provided the link between the files generated in Geoeas and subsequent contouring routines embedded into the Surfer package. Omnidirectional and directional variograms were modelled defining the parameters for kriging estimation using Geoeas. Different variogram models were checked through cross validation prior to kriging. A comparison between the magnitude of the differences in volumes calculated with the different DTMs was attained. These final result indicates the sensitivity of the volume obtained to the number of data points. As the number of data points increases the estimation error decreases, but at some point there is no significant decrease in the final volume with the collection of additional data.