Permeability evolution in variably saturated shales controlled by bimodal organic/inorganic pore structure and surface wetting characteristics

Moisture is distributed heterogeneously within shale with distinctive features in both organic and inorganic pores. Importantly, this influences gas flow, impacts reservoir dynamics, and affects the accuracy of production forecasts. We establish a variable pressure gradient-bimodal pore structure (VPG-BPS) model to capture these characteristics of methane transport in partially water-saturated shale. We determine permeability in both organic and inorganic pores by matching estimated to actual gas production. Dynamic production experiments are conducted on variably saturated shales from the Sichuan Basin to determine permeability. Methane adsorption capacities are categorized into three stages, delineated by two intermediate threshold moisture contents—lower (M_Lc) and upper (M_Uc) threshold moisture contents, respectively. We demonstrate that the moisture content significantly influences shale transport dynamics by controlling adsorption-desorption properties and thereby gas-water distribution and permeability. Ultimate gas production decreases with increasing moisture content (M_c). Gas permeability (K_in and K_o) and gas desorption rate coefficients (k_ads,in and k_ads,o) in inorganic and organic pores range from 1 × 10⁻⁸ μm² to 1 × 10⁻⁷ μm², and 1 × 10⁻⁴ to 1 × 10⁻³, respectively. With M_c increasing from 0% to 3.5% (M_Uc), K_in, K_o, and k_ads,in exhibit a consistent trend of initially increasing before decreasing although k_ads,in shows no definitive trend above the threshold peak (M_Lc). When M_c increases from 0% to 2.2% (M_Lc), the desorptive hysteretic effect gradually weakens and adsorption k_ads,in increases with increasing M_c, subsequently prompting an increase in the other three flow parameters. As M_c further increases to M_Uc, k_ads,in remains constant but the other three flow parameters decrease as the gas flow channel is occupied by the invading water. Even when moisture migration ignored, the VPG-BPS model provides superior predictions of time histories of gas production over existing models and thus provides a theoretical foundation for dynamic evaluation of shale gas production.