Formation of Lithology-Dependent Hydrate Distribution by Capillary-Controlled Gas Flow Sourced from Faults

Analysis of seismic profiles at Green Canyon 955, Gulf of Mexico shows that hydrocarbons (hydrate and free gas) are concentrated in the crest of an anticlinal structure intersected by deeply rooted normal faults. Drilling, coring and analysis of pressurized samples at this location documented lithology-dependent hydrate concentration distribution, with high concentration in layers composed of sandy silt and low in layers composed of clayey silt. We show with a two-dimensional multiphase flow reactive transport model that such heterogeneity is caused by the capillary-driven preferential flow of gas, and capillary effect on methane hydrate solubility has negligible influence on this heterogeneity with the free gas flow model. In our model, free gas is deeply sourced and flows upward along faults under buoyancy. If the capillary entry pressure of the reservoir is lower than the fault, gas redirects to preferentially flow laterally into the sandy silt layer from the base of the reservoir where it intersects the fault. This drives preferential hydrate formation in the sandy silt layers above the base of hydrate stability zone. Hydrate formation gradually increases the local capillary pressure, decreases the sediment effective permeability and elevates the surrounding salinity. Eventually capillary pressure within the sandy silt layers is sufficient to push gas vertically into the adjacent clayey silt layers to form hydrate where the salinity is already above seawater. The limited amount of methane that can be transported into the clayey silt layers, together with the elevated salinity, leads to much less hydrate formation in clayey silt layers. If the capillary entry pressure of the entire reservoir is higher than the fault, gas preferentially flows upward along the fault under buoyancy and passes through the reservoir depth at early time. Later, enough hydrate formation within the fault redirects gas to preferentially flow laterally into the sandy silt layers of the reservoir where hydrate preferentially forms. Our model provides an alternate explanation of lithology-dependent hydrate distribution, and demonstrates the importance of capillary-controlled gas flow in the development of hydrate reservoirs.