Probabilistic mapping and uncertainty quantification of the tectonically deformed coal ratio: A case study of the No. 15 coal seam in the Yangquan Mining Area

Abstract

The tectonically deformed coal (TDC) ratio is important for coal-mine safety assessment and coalbed methane development. Reliable prediction remains difficult because sparse well control, thin-bed tuning, seismic noise, and geological heterogeneity introduce substantial interpretive uncertainty. This study presents an uncertainty-aware workflow for probabilistic mapping of the TDC ratio in the No. 15 coal seam at Xinjing Coal Mine, Yangquan Mining Area. Well-log-derived total coal-seam thicknesses and TDC ratios were integrated with horizon-based seismic attributes. Principal component analysis reduced attribute dimensionality, and Gaussian process regression predicted total coal-seam thickness across the study area. Stochastic elastic forward modeling then established a thickness-dependent tuning function that reduced the contribution of thin-bed tuning and enabled derivation of a calibrated TDC-related amplitude attribute. Calibration increased the correlation between seismic amplitude and TDC ratio from 0.71 to 0.82. Sequential Gaussian co-simulation using Markov Model 2 generated 100 conditional TDC-ratio realizations, with well-derived TDC ratios as hard data and the calibrated amplitude attribute as secondary information. The ensemble mean showed higher ratios in the northeastern area and isolated central-to-southern zones, broadly consistent with the spatial pattern of the earlier deterministic TDC-thickness estimate. Lower ratios were more likely in the southwestern and central-western areas. Percentile, threshold-probability, and interquartile-range maps quantified spatial uncertainty. Within this case study, the workflow converted legacy post-stack seismic data and sparse well control into decision-oriented probabilistic products. Application to other geological settings requires local recalibration and external validation.