Stochastic Agent-Based Modelling of Hydrocarbon Migration in Stratigraphically Complex Basins

A stochastic agent-based modelling (ABM) framework has been developed to simulate hydrocarbon migration and accumulation in stratigraphically complex geological environments, advancing traditional deterministic approaches by integrating probabilistic elements to capture natural geological variability. By representing hydrocarbons as agents navigating a two-dimensional grid of facies-defined pathways, the model uses Move Probability Matrices (MPMs) and Markov Chain-driven facies transitions to simulate migration behaviour influenced by permeability contrasts across sandstone, silt, and shale facies. Environmental factors, such as sea level changes, drive facies transitions over time, creating a dynamic system that reflects the impact of depositional variability on fluid migration pathways.

The model employs iterative simulations to capture the probabilistic ‘Most Probable Path’ of hydrocarbon agents, revealing zones with a higher likelihood of accumulation. Multiple stochastic runs are aggregated into heatmaps, visualising regions where geological structures and stochastic variability intersect to provide insights into hydrocarbon distribution patterns. Results indicate that while deterministic models outline structured migration routes, the stochastic approach allows for the identification of additional potential accumulation zones by accounting for subtle geological variations and environmental fluctuations that influence fluid movement.

This stochastic ABM framework improves predictive accuracy in hydrocarbon exploration, supporting more informed decisions in resource management and well placement. Beyond hydrocarbon migration, the model offers a flexible framework applicable to other geoscientific challenges, such as groundwater modelling and carbon sequestration, where geological heterogeneity significantly impacts fluid dynamics.