Estimation of methane content in coal mines using supplementary physical measurements and multivariable geostatistics

•Cokriging and cosimulation were used to estimate methane content in coal mine.•Methane content, desorption factor, coal strength aptly correlated also spatially.•If no faults, unbiased methane content estimates with cokriging & cosimulation.•Near faults co-methods & desorption factor overestimate methane content up to 2%.•Near faults co-methods & coal strength underestimate methane up to 0.5 m3CH4/t

The volume of methane that can be released from coal during extraction is usually estimated only on the basis of measurements of methane content in coal seam. Usually, apart from methane measurements, numerous physical variables are also measured before coal extraction. Although physical variables do not produce straightforward information about methane concentration, they can be significantly correlated with it and therefore can also be used in methane content estimation. The goal of this paper was to analyze the possibility of using multivariable geostatistics to estimate methane content in coal seam using physical secondary measurements. To do so, two types of secondary measurements were used: the desorption factor and coal strength index, both of which were used in cokriging and sequential Gaussian co-simulation (SGcoS). Both secondary variables were sufficiently correlated with methane content, validating their use in the analyses. The use of the desorption factor as supplementary data produced an approximately 2% overestimation of values of methane content but only in the vicinity of the geological faults. The overestimation was lower in the case of SGcoS than in the case of cokriging. Application of the coal strength index as a secondary variable did not lead to overestimation. Methods that use supplementary data produced lower kriging variance and simulated standard deviation values of methane content compared to those obtained using kriging and sequential Gaussian simulation (SGS), respectively, where only methane measurements were used. Simulated standard deviations of methane content obtained using SGcoS with a desorption factor were significantly higher than those obtained using SGcoS with the coal strength index used as supplementary data.