Modeling hydrocarbon pathways from accumulation to source: A stochastic agent-based approach with permeability-driven matrices

Abstract A stochastic agent-based framework is applied to model hydrocarbon migration by tracing pathways from accumulation zones back to probable source regions using an inverted move probability matrix (MPM) system. This study employs a two-stage process: initial upward migration modeling incorporating stochastic variability and subsequent deterministic downward tracing through geologic layers, guided by permeability-based inverted MPMs. Repeated simulation iterations reveal high-probability pathways and likely hydrocarbon origins. Incorporating stochastic variability enhances realism compared to deterministic models, better reflecting the complexity of real-world subsurface environments. Geologic heterogeneity is captured through microscale variations and facies transitions, creating multiple migration scenarios that provide a broader, probabilistic picture of hydrocarbon pathways. The resulting heatmaps, generated from aggregated agent paths across multiple runs, highlight consistent migration routes and probable source zones. The findings underscore the advantage of a stochastic approach for predicting high-likelihood migration paths and accumulation areas often overlooked by deterministic models. By identifying these pathways, the framework enhances exploration strategies, optimizes well placement, and reduces exploration risks, especially in stratigraphically complex basins. The methodology also holds potential for applications beyond hydrocarbon exploration, including groundwater modeling, contaminant tracking, and carbon sequestration, where geologic variability plays a critical role. Through this probabilistic approach, the model contributes valuable insights into resource management and subsurface modeling in complex geologic settings.