The project goal is to evaluate whether the transport of methane, and the specific mechanism by which it is transported, are the primary controls on the development of persistent, massive hydrate accumulations in sediments below the seabed. The dissolved methane flux and time required to develop the accumulations observed at Walker Ridge Block 313 (WR313) in the northern Gulf of Mexico by long-distance updip migration or by short-distance local migration will be studied and defined within the scope of this research. Researchers will also determine whether there is enough methane in the dissolved phase in the fine-grained sediments to form the observed hydrate deposits or whether a gas phase is present and, if so, what the conditions are for three-phase equilibrium.
University of Texas at Austin, Austin, TX 78713-7726
Ohio State University, Columbus, OH 43210
Lamont-Doherty Earth Observatory (Columbia University), Palisades, NY 10964
Massive hydrate deposits, defined as thick (>5 feet) accumulations of high hydrate saturation (>50 percent), have been encountered in many regions worldwide. This project will focus specifically on accumulations found at WR313 during the Gulf of Mexico Gas Hydrate Joint Industry Project Leg 2. Hydrates may be thought of broadly within a petroleum systems framework, requiring a methane source, migration mechanisms, a reservoir, and an appropriate seal. Hydrate reservoirs and seals are defined by thermodynamics rather than buoyancy as in the case of conventional oil and gas. Hydrates form most easily within coarse-grained sediments within the methane hydrate stability zone, the depth interval at which pressure and temperature favor hydrate as the stable phase. Methane sources may include microbial activity as well as thermogenic sources. The focus will be on migration mechanisms in marine hydrate reservoirs as they represent some of the least understood processes in hydrate systems, but at the same time represent a crucial link between methane generation sites and hydrate reservoirs.
Successful completion of this project will provide valuable insight into conditions necessary for the development of massive gas hydrate accumulations and the role of free gas in their persistence. This, in turn, will advance understanding of the transport and fate of methane in the subsurface, carbon cycling associated with hydrates, and the role of a free gas phase in the formation and persistence of hydrate deposits.
An existing reservoir simulator has been altered to include methanogenesis, sedimentation, salinity, and pore size effects on hydrate stability. The modified simulator has been successfully benchmarked against published 1-D and 2-D simulations and found to be consistent within 1% in time and hydrate saturation given the same input parameters. A series of 2-D simulations were conducted in a dipping sand layer at WR313 to investigate the conditions that led to the development of the shallow, hydrate saturated sand layers at WR313. Preliminary 2-D model results suggested that microbial methanogenesis may be necessary to form the observed hydrate accumulations in the shallow sands at WR313; however, long range methane migration might be necessary to form hydrate saturations in the deeper sand layers. Metabolizable organic content was found to be an important parameter in determining hydrate saturations associated with microbial methanogenesis. Additional simulations will be conducted using realistic sediment properties (such as metabolizable organic content) calculated from the 1-D reactive transport model discussed below.
A one-dimensional (1-D) reactive transport model was created to couple sedimentation and microbial methanogenesis to model hydrate accumulation in a dipping, subsiding sand layer at WR313. Because sediment porosity changes during compaction and the amount of methane dissolved in the pore fluid increases as sediment is buried, sedimentation is a key component for modeling hydrate accumulation. The reactive transport model has been fully developed and benchmarked against existing data. The model uses numerical integration to compute steady-state dissolved methane concentrations and gas hydrate or gas bubble volume fractions as a function of depth in a 1-D geometry. The numerical method is computationally fast and is suitable for Monte Carlo simulations. In these simulations, model outputs are computed in a large number of iterations where uncertain input parameters are allowed to vary within realistic limits. By retaining the results where the predicted gas hydrate amounts agree with observations, these simulations measure the uncertainty in the input model parameters. The reactive transport simulations were applied to quantify the uncertainty of the intensity and depth distribution of microbial methanogenesis. The results have been used to constrain the available organic carbon at the seafloor, reaction rate of microbial methanogenesis, and pore water advection rate in the 3-D, time-dependent reservoir simulator. The time-dependent reactive transport model has also been used to predict gas hydrate contents in fine-grained sediments. Microbial methanogenesis in these sediments can contribute to the accumulation of hydrates in adjacent coarse-grained layers. The goal of this modeling work was to test whether a transient period of high organic carbon deposition at the seafloor can result in a sediment interval with more intense methane generation and enhanced hydrate formation. The model was successful in the prediction of gas hydrate amounts close to those estimated at Walker Ridge 313.
In order to further constrain the sediment properties for reservoir simulations at WR313, detailed log analysis was conducted on wells in the Gulf of Mexico. In particular, data collected in the Keathley Canyon 151 research well were used to derive a method of generating synthetic pore size distributions from gamma ray and porosity logs. The method is still being refined, but preliminary results indicate that the pore size contrast between the sands and clays at WR313 may only be one order of magnitude, which might not be enough to drive sufficient diffusive flux into the sand layers. Furthermore, a journal article was published in the Marine and Petroleum Geology journal describing a new method to calculate permeability from log data using porosity, clay-sized particle mass fractions, and appropriate endmember permeabilities for clay and sand (Daigle et al., 2015). The permeability calculations have been verified against measured permeability data from three wells in the Gulf of Mexico and extrapolated to WR313 where permeability log data are unavailable. Sediment properties calculated from log data at WR313 have been applied as input parameters to the reservoir simulator and initial simulations have been completed.
Using the physical properties derived from well logs and realistic sand layer geometries (Figure 1), multi-dimensional reservoir simulations were conducted to show how the transport of methane and the mechanism by which it is transported control the development of hydrate accumulations at WR313. Simulations of diffusive methane migration and advective transport of methane were conducted to assess the resulting hydrate accumulations as well as the methane flux and time scales required to develop the massive hydrate accumulations observed at WR313.
Reservoir simulation results indicate that diffusion of microbial methane into sand layers can result in high hydrate saturations, but the organic inputs necessary to achieve these hydrate saturations may not be sustainable over geologic time. Simulations of long-distance advective transport indicate that updip advection may be able to produce the observed hydrate saturations but the resulting hydrate accumulations are heterogeneous with the highest hydrate saturations at the downdip limits of the sand layers (Figure 2). However, model results suggest that a hybrid migration mechanism “short-advection” may be responsible for the hydrate accumulations at WR313. In this scenario, methane is produced microbially in clays but pervasive compaction-drive flow transports the methane into higher-permeability clays. The combination of advection and microbial methanogenesis can result in a more uniform hydrate distribution in the sands, with high hydrate saturation (Figure 3). In the case of horizontal sand layers, hydrate will only accumulate at the base of the sand, but in the case of dipping or vertical sands, as hydrate accumulates in the downdip portion of the sand, fluid flow will be deflected around the lower-permeability hydrate-bearing sands and supply methane farther updip in the sand. This work was published in a Geophysical Research Letters journal article proposing short-range advective migration as a methane transport mechanism for hydrate formation in coarse-grained marine sand layers surrounded by fine-grained sediments within the MHSZ (Nole et al., 2016).
A rock physics model was developed to investigate the relationship between the character of bottom simulating refection and the amount of methane present as gas, hydrate, or a dissolved phase at the base of hydrate stability. The change in elastic properties with phase saturation has been modeled in a three-phase zone at Walker Ridge 313. Results indicate that gas content of the three-phase zone is the primary driver of changes in the elastic properties and that hydrate saturation only plays a critical role in the elastic properties of the sediments when the gas saturation is very low.
Work is also ongoing to model capillary pressure within the three-phase zone at the base of hydrate stability. In this zone, capillary effects cause hydrate, water, and gas to occupy distinct regions of the pore size distribution of the host sediment. Because host sediments have a distribution of pore sizes, capillary effects cause the base of hydrate stability to be located at different depths in pores of different sizes. The result is a zone over which gas replaces hydrate as the phase occupying the largest pores. This model has been applied to Blake Ridge and Hydrate Ridge as test cases, and will ultimately be applied to Walker Ridge to investigate the effects of methane recycling at the base of hydrate stability.
A 3-D basin-scale reservoir simulator was modified to include methanogenesis, sedimentation, salinity, and pore size effects on hydrate stability. The model domain was created for a realistic representation of a hydrate sand layer at WR313 and sediment properties and has been modified to reflect log data calculations. The 1-D reactive transport model has been designed and benchmarked against existing simulations. Physical properties of the sediments at WR313 have been refined based on reactive transport model results, pore size distributions, and nuclear magnetic resonance log data. Simulations of dissolved methane migration mechanisms (short-distance diffusive migration, short-distance advective transport, and long-distance advective transport) have been conducted to assess the methane flux time scales required to develop the massive hydrate accumulations observed at WR 313. Simulation results suggest that a combination of diffusive and advective methane flux may be responsible for the hydrate saturations at WR313. Work is ongoing to (1) assess the methane budget required for the presence of a gas phase by defining methane availability, local solubility, and the amount of methane that may be present in the gas phase; (2) assess whether the gas phase accumulated beneath the methane hydrate stability zone can contribute significantly to hydrate saturations and evaluate the conditions and geologic settings in which significant updip migration can be expected; and (3) evaluate the fate of hydrate that moves below the base of the MHSZ as a result of sedimentation.