Assessment of uncertainty for coal quality-tonnage curves through minimum spatial cross-correlation simulation

Coal quality-tonnage curves are helpful tools in optimum mine planning and can be estimated using geostatistical simulation methods. In the presence of spatially cross-correlated variables, traditional co-simulation methods are impractical and time consuming. This paper investigates a factor simulation approach based on minimization of spatial cross-correlations with the objective of modeling spatial relations of coal quality data and estimating quality-tonnage curves in a part of the Ömerler sector of Tunçbilek coalfield (Turkey). Data come from core samples analyzed for lower calorific value, ash content and moisture content. Prior to simulation, composite data and coal seam are unfolded and the composites are also de-trended. The simulations of the original data are obtained by adding the trend values to the simulated residuals and transforming the unfolded coordinates into the original ones. 100 realizations of the coal attributes are jointly generated by Minimum Spatial Cross-correlation (MSC) simulation method. The MSC-simulations are compared to the results of a widely used joint simulation method based on the minimum/maximum autocorrelation factors (MAF) technique. The comparison shows advantage of the new proposed method over the MAF technique. MSC-simulations reproduce the original data well on the basis of correlation coefficient, cumulative histograms and auto / cross-variograms. This suggests that the MSC-simulation method can be used in simulation of spatially cross correlated coal data. Quality-tonnage curve for each realization is calculated and uncertainty associated with tonnage is assessed by using a 95% confidence interval. The assessments show that the tonnage uncertainty depends on the cutoff