The flow of gas through shallow marine sediments is an important component of the global carbon cycle and affects methane release to the ocean and atmosphere as well as submarine slope stability. Seafloor methane venting is often linked to dissociating hydrates or gas migration from a deep source, and subsurface evidence of gas-driven tensile fracturing is abundant. However, the physical links among hydrate dissociation, gas flow, and fracturing has not been rigorously investigated. We used mercury intrusion data to model the capillary drainage curves of shallow marine muds as a function of clay content and porosity. We combined these with estimates of in situ tensile strength to determine the critical gas saturation at which the pressure of the gas phase would exceed the pressure required to generate tensile fractures. Our work demonstrates that tensile fracturing is more likely as clay content increases due to decreased pore sizes and increased capillary pressure, but tends to be restricted to the shallowest portion of the sediment column (<130 m below seafloor) except when the clay-sized fraction exceeds 50%. Dissociating hydrate may supply sufficient quantities of gas to cause fracturing, but this is only likely near the updip limit of the hydrate stability zone, where release of methane bubbles from discrete vents is to be expected due to the combination of weak sediments and significant gas expansion. Gas-driven tensile fracturing is probably a common occurrence near the seafloor, does not require much gas, and is not necessarily an indication of hydrate dissociation.

At thermodynamic equilibrium, gas hydrates are arranged in the pore space of host sediments to minimize free energy, including the energy of interfaces. Through an analogy with frozen soil, we show that free energy minimization in hydrate-bearing sediments requires the presence of a water film of finite thickness separating hydrate from the sediment grains. The thickness of this premelted layer may be predicted from a balance of intermolecular forces acting across the film. Temperature and porewater salinity are the strongest determiners of premelted layer thickness. We show that, at temperatures and salinities typical of the subsurface or commonly used in laboratory investigations of hydrate-bearing porous media, the premelted layer varies in thickness from microns to sub-nanometer, with thicker layers corresponding to lower salinities and/or higher temperatures. Balance of intermolecular forces predicts that hydrate will be completely nonwetting on hydrophilic surfaces, including silica. We also show that flow through premelted layers may be a significant component of the permeability of hydrate-bearing sediments, particularly at moderate to high hydrate saturation (>60%); and that the electrical conductivity of the premelted layer at needs to be accounted for in assessments of hydrate abundance from subsurface resistivity logs. This work highlights the importance of considering premelted layers when predicting the effects of hydrate on sediment properties.

The surface of ice in contact with water contains sites that undergo deprotonation 6 and protonation, and can act as adsorption sites for aqueous ions. Therefore, an electrical double layer should form at this interface, and existing models for describing the electrical double layer at metal oxide-water interfaces should be able to be modified to describe the surface charge, surface potential, and ionic occupancy at the ice-water interface. I used a surface complexation model along with literature measurements of zeta potential of ice in brines of various strength and pH to constrain equilibrium constants. I then made predictions of ion site occupancy, surface charge density, and partitioning of counterions between the Stern and diffuse layers. The equilibrium constant for cation adsorption is more than 5 orders of magnitude larger than the others constants, indicating that this reaction dominates even at low salinity. Deprotonated OH sites are predicted to be slightly more abundant than dangling O sites, consistent with previous work. Surface charge densities are on the order of ±0.001 C/m^2 and are always negative at the moderate pH values of interest to atmospheric and geophysical applications (6-9). In this pH range, over 99% of the counterions are contained in the Stern layer. This suggests that diffuse layer polarization will not occur because the ionic concentrations in the diffuse layer are nearly identical to those in the bulk electrolyte, and that electrical conduction and polarization in the Stern layer will be negligible due to reduced ion mobility.

Mudrocks serve as geological seals for carbon sequestration or hydrocarbon formation where mudrock capillary seals having high capillary entry pressure prevent leakage of underlying fluids. However, seal failure can occur if the trapped nonwetting fluid escapes by porous flow or by induced tensile fractures caused by elevated nonwetting phase pressures. Since mudrocks are mainly composed of silt and clay size grains, a silt bridging effect has been observed when there are sufficiently abundant silt size grains. This effect creates force chains across the rock to help preserve large pores and throats and can reduce the sealing capacity of a mudrock. We used network models and discrete element (DEM) models to determine the conditions under which silt abundance will cause a mudrock seal to fail and allow a non-wetting fluid like CO2 or natural gas to flow. We show that when larger grains in a grain pack become 40-60 % of total grain volume, the drainage capillary pressure curves display two percolation thresholds, and the percolation threshold transitions to a lower value allowing seal failure even below tensile fracture pressure. The DEM compaction simulations found that strong force chains are mostly formed across grain contacts between large grains and their neighbors and not between small grains, which decreases coordination numbers and shields pore space from compaction before reaching a stress limit. Thus, through better understanding of grain concentrations and sizes on fluid flow behavior, we can improve risk management efforts in anthropogenic storage and estimates of reserve capacity of reservoirs.