Petrophysical properties of the Green Canyon Block 955 hydrate reservoir inferred from reconstituted sediments: Implications for hydrate formation and production

ABSTRACT

We explore the petrophysical behavior of the two interbedded lithofacies (sandy silt and clayey silt) that constitute the Green Canyon Block 955 hydrate reservoir in the deep-water Gulf of Mexico by performing experiments on reconstituted samples of the reservoir material. Sandy silts reconstituted to the in situ porosity have a permeability of 11.8 md (1.18 × 10−14 m2), which is similar to the intrinsic permeabilities measured in intact cores from hydrate reservoirs of similar grain size offshore Japan (Nankai Trough) and offshore India. Reconstituted clayey silts have a much lower intrinsic permeability of 3.84 × 10−4 md (3.84 × 10−19 m2) at the in situ stress. The reconstituted sandy silt is less compressible than the clayey silt. Mercury injection capillary pressure measurements demonstrate that the largest pores with the clayey silt are still smaller than the pores remaining after 90% hydrate saturation in sandy silt. We interpret that the methane solubility in pores of clayey silt is always less than that necessary to form hydrate, which explains why no hydrate is present in the clayey silt. We upscale the reservoir properties to estimate the behavior of interbedded sandy silt and clayey silt. We find the upscaled intrinsic horizontal and vertical permeabilities for the entire reservoir interval are 8.6 md (8.6 × 10−15 m2) and 1.4 × 10−3 md (1.4 × 10−18 m2). We estimate that during reservoir production, a maximum vertical strain of approximately 12% will result. Ultimately, this study will inform reservoir simulation models with petrophysical properties at scales of both individual lithofacies and reservoir formation.

INTRODUCTION

Methane hydrate is a crystalline solid composed of methane molecules trapped in cages of water molecules (Sloan and Koh, 2007). It is stable at low temperatures and high pressures, which in natural environments is commonly found in permafrost regions, outer continental margins, and shallow-basin sediments (T. S. Collett, 2002; T. S. Collett et al., 2011).

Concentrated hydrate deposits are commonly found in coarse-grained sediments (e.g., silt to sand size) (Boswell and Collett, 2011). In the deep-water Gulf of Mexico (GOM), multiple reservoirs of this type have been found (Boswell et al., 2012a). A full log suite across a part of the gas hydrate stability zone in the northern GOM at Alaminos Canyon (AC) Block 818 recorded pore-filling hydrate at a saturation of 80% within a 13-m-thick formation (Boswell et al., 2009). The GOM Gas Hydrate Joint Industry Project Leg II (JIP II) drilled two wells (well G and well H) at Walker Ridge 313 in Terrebonne Basin, and the logging while drilling (LWD) data suggested saturations above 80% in the higher-quality sands encountered (Boswell et al., 2012b). The JIP II program also drilled the Green Canyon Block 955 (GC 955) H001 well (H001) where high hydrate saturations (Sh = 60%–65%) were found in sediments with relatively low gamma-ray values that were interpreted to represent silt or sand size (T. S. Collett et al., 2012; Lee and Collett, 2012). Shipboard hydrate dissociation tests on pressure cores from the eastern Nankai Trough, Japan, revealed that the average hydrate saturation of the sandy cores ranges from 55% to 68% (Fujii et al., 2009). Recent studies from the India National Gas Hydrate Program Expedition 02 (NGHP-02) suggested that the hydrate saturations measured from controlled on-site depressurizations record various hydrate saturations over a wide range of grain sizes (e.g., super high saturation [∼100%] present in gas hydrate-bearing gravels, high saturation [∼65%–85%] within sand layers, moderate saturation [∼45%] in silty clays, and low or no saturation [0%–10%] in thin clay layers) (Boswell et al., 2018b).

Multiple offshore drilling and coring programs conducted in the past decade have evaluated reservoirs in the GOM for their potential as an energy resource (T. Collett et al., 2015). One such expedition, targeting a coarse-grained reservoir, was The University of Texas-Gulf of Mexico 2-1 (UT-GOM2-1) Hydrate Pressure Coring Expedition at GC 955 in the deep-water GOM (Flemings et al., 2018a, b, 2020, this issue). One of the primary goals of this expedition was to characterize the petrophysical properties of hydrate reservoirs and their bounding seals. Petrophysical characteristics, such as grain size, porosity, compressibility, permeability, and capillary behavior, all control both how these reservoirs form (Clennell et al., 1999; Liu and Flemings, 2006, 2007; Torres et al., 2008; Fujii et al., 2009; You and Flemings, 2018) and how they will behave during production (Moridis et al., 2007).

The most common way to study the petrophysical properties of hydrate-bearing sediments is through pressure core analysis. For example, previous studies measured permeability and compressibility of hydrate-bearing core and then dissociated the hydrate and repeated the measurement (Konno et al., 2015; Santamarina et al., 2015; Yoneda et al., 2019a). Extensive pressure coring programs have been performed in offshore Japan (Yoneda et al., 2017, 2018b), offshore Korea (Yun et al., 2011; Ryu et al., 2013), offshore India (Yoneda et al., 2019b), and more recently in offshore China (S. Yang et al., 2017).

Pressure core analysis is beginning to illuminate the mechanical and hydraulic properties of methane hydrate reservoirs, but it has its own challenges. First, the coring, recovery, and measurement process may disturb the sample and thus alter its properties. For example, during pressure core recovery, the friction between the pressure coring tool and the sediment may fracture the core sample. When the confining pressure of the earth is removed, the stress unloading may expand the sample altering its property (Santagata and Germaine, 2005; Casey et al., 2016; Thomas et al., 2020, this issue). In addition, during depressurization, the fine-grained sediment may move as the hydrate is dissociated, thus altering the material that is being measured (Jung et al., 2012; Han et al., 2018). Second, the intact core may be heterogeneous at the scale of the core. In this case, any measured property will reflect a mixture of the thin-bedded materials. Third, although this method illuminates the present petrophysical nature of the reservoir, it is not possible to provide the evolution of petrophysical properties from old in situ conditions to its current in situ conditions.

In this study, we took a different approach to analyzing petrophysical properties of hydrate-bearing reservoirs. We took sediment from the two characteristic lithofacies present in the GC 955 hydrate reservoir. We then reconstituted these sediments to the approximate state they were at in the subsurface. Reconstitution experiments disaggregate original sediments to their constituent materials and then recompress them in a manner that is similar to the burial process (Casey et al., 2015). We then studied the petrophysical properties of these reconstituted materials to gain insight into the properties of the in situ reservoir. Ideally, this approach can illuminate the intrinsic permeability of the reservoir (the permeability of the porous media with only water present in the pores), compressibility of the reservoir without hydrate present, and pore structure of the reservoir and its associated capillary properties. These are all necessary parameters for reservoir engineering simulators because they require both reported experimental data (i.e., intrinsic permeability, compressibility, and capillary behavior) and an understanding of these data so that they can ultimately build a reliable model and predict accurate reservoir behavior in hydrate production.

The approach of reconstituting sediments has strengths and limitations. It eliminates sample variability, produces uniform specimens, and isolates the effects of composition, effective stress, and stress history on petrophysical properties through systematic experiments. This is not possible with the use of intact samples because no two intact samples will possess an identical composition and stress history. In addition, analysis of reconstituted samples allows one to study behavior over a range of in situ effective stresses (Casey et al., 2016). However, the reconstitution approach homogenizes the sediments. Thus, any layering or local heterogeneity originally present will be lost. Furthermore, there is an open discussion on whether the compression approach used successfully generates the pore-scale structure that is formed by geological deposition and burial. That said, for materials that have not been significantly affected by diagenesis, cementation, or aging effect, a large amount of experimental work has demonstrated that the mechanical behavior of intact samples can be closely reproduced by the laboratory reconstituted material (Burland, 1990; Berman et al., 1993; Betts, 2014). No cementation or diagenesis has been observed in the GC 955 sediments (Flemings et al., 2018d).

This paper analyzes the two lithofacies (sandy silt and clayey silt) encountered at two particular depth intervals within the GC 955 hydrate reservoir. First, we separately selected the sandy silt and clayey silt materials and describe their index properties (e.g., Atterberg limits, particle size distribution, grain density, and porosity). We then describe how we reconstituted samples of sandy silt and clayey silt. We then describe the intrinsic permeability, compressibility, and capillary behavior of the reconstituted materials from each lithofacies. Finally, we integrate these data to infer possible mechanisms for hydrate formation in these reservoirs and describe the evolution of permeability and reservoir compaction during depressurization.

GEOLOGICAL AND PETROPHYSICAL SETTING

The hydrate-bearing, coarse-grained reservoir interval at GC 955 on the Texas–Louisiana continental rise was originally located based on industry drilling and geophysical characterization (D. R. McConnell, 2000; Hutchinson et al., 2008; D. R. McConnell et al., 2012). In 2009, JIP II used LWD to confirm the hydrate reservoir between 414 and 550 m below the sea floor (mbsf) (1358–1804 ft) from the high resistivity, high P-wave velocity, and low gamma-ray density values (D. McConnell et al., 2010; Boswell et al., 2012a, b; T. S. Collett et al., 2012). The reservoir was interpreted to be composed of fine to very fine sands interbedded with clays on the basis of LWD log character (Boswell et al., 2012b), which is similar to the thinly bedded and laminated deposits from the Viosca Knoll area, eastern GOM (Shew et al., 1994).

The GC 955 hydrate reservoir was cored in 2017 during the UT-GOM2-1 expedition (Flemings et al., 2018d). Two wells (hole GC 955 H002 and hole GC 955 H005) were drilled to retrieve pressurized gas hydrate–saturated cores. The pressure core analysis and transfer system (PCATS) (Schultheiss et al., 2011) was used to measure the P-wave velocity (Vp) and bulk gamma-ray density (ρb) of pressure cores through the core liner at a resolution of 0.5 cm. We obtained x-ray computed tomography (CT) images (voxel resolution of 112 μm). Twenty-one 1.2-m-long pressurized storage vessels were transported to the shore-based laboratory (UT Pressure Core Center) for further analysis. Details regarding the stratigraphy and sedimentology of the reservoir are available in Santra et al. (2020, this issue) and Meazell et al. (2020, this issue).

The GC 955 hydrate reservoir was pressure cored to acquire samples (Thomas et al., 2020, this issue), and from these samples, we interpreted the composition, hydrate concentration, and mechanical properties of this reservoir. Meazell et al. (2020, this issue) defined two lithofacies, sandy silt and clayey silt, within the primary hydrate-bearing interval (415–449 mbsf) (Figure 1A–D). Sandy silt lithofacies, with abundant trough cross-bedding structures (Meazell et al., 2020, this issue), has a low bulk density (ρb = 1.7 to 1.9 g/cm3) and high P-wave velocity (Vp = ∼2700–3200 m/s) (Figure 1E–G). Quantitative degassing indicated that sandy silt lithofacies contains high hydrate saturation (Sh = 66%–87%) and clayey silt lithofacies contains little to no hydrate (Flemings et al., 2020, this issue; Phillips et al., 2020, this issue). The hydrate saturation (Sh) of the samples was identified by quantitative degassing experiments on intact samples of the formation. Details of the estimation method are described in Phillips et al. (2020, this issue). In contrast, clayey silt lithofacies has relatively higher bulk density (ρb = 1.9 – 2.1 g/cm3) and lower P-wave velocity (Vp = ∼1700 m/s) (Figure 1H–J) (Flemings et al., 2018d, 2020, this issue; Meazell et al., 2020, this issue).

Figure 1. (A) Depth in meters below seafloor (mbsf) at the hole Green Canyon (GC) 955 H005. (B) The ring resistivity log from the Gulf of Mexico Joint Industry Project (JIP II) hole GC 955 H001 acquired during JIP II depth shifted to the estimated position in the H005 well. (C) Cored intervals of the H005 well. (D) Recovered and not recovered zones within the cored interval. All recovered core is assumed to reside at the top of the cored interval. (E) Pressure core analysis transfer system (PCATS) x-ray image of sandy silt core interval (H005-4FB-8 [423.56–423.99 mbsf]) studied in this paper; the yellow box outlines the interval of the sample analyzed. (F) Gamma-ray attenuation bulk density (ρ∗b) from the PCATS taken at 0.5-cm intervals. (G) The PCATS P-wave velocity (Vp). (H) An x-ray image of clayey silt lithofacies interval analyzed in this paper (H005-11FB-1 [441.27–441.55 mbsf]); the yellow box outlines the interval of the sample analyzed. (I) The ρ∗b for clayey silt. (J) The PCATS Vp for clayey silt. See Flemings et al. (2018c) for a discussion of PCATS analysis methods.

The overburden stress was calculated by integrating the LWD bulk density log from the GOM JIP hole GC 955 H001. Seawater salinity was assumed to be 3.5 wt. %. The pore pressure (u) was assumed to be hydrostatic (pore pressure gradient = 8.6 ppg [0.447 psi/ft]) because there was no evidence of any elevated pore pressures during previous drilling of H001 (Flemings et al., 2018d). The vertical effective stress (σ′v), calculated by subtracting the hydrostatic pore pressure (u) from the overburden (σv) is 3.76 MPa at the top of the hydrate-bearing interval (415 mbsf) and 4.01 MPa at the bottom (449 mbsf) (Flemings et al., 2020, this issue). The in situ temperature of GC 955 hydrate-bearing sediments was estimated to be between 18.6°C (at a depth of 414 mbsf) and 19.8°C (at a depth of 449 mbsf) based on the estimated temperature at the base of the hydrate stability zone (20.4°C) and the geothermal gradient of 34.7°C/km (Flemings et al., 2020, this issue).

METHODS

Sample Selection and Preparation

We analyzed material from an interval within core H005-4FB-8 that represented typical examples of the sandy silt lithofacies (Figure 1E, yellow box) and from core H005-11FB-1 (Figure 1H, yellow box) that represented a typical example of the clayey silt lithofacies. We chose these specific intervals (i.e., Figure 1E, H, location range highlighted in the yellow box) because they are characteristic of sandy silt lithofacies and clayey silt lithofacies as described by Flemings et al. (2020, this issue) and Meazell et al. (2020, this issue) (Figure 1F, G, I, J). For each sample interval, we mixed all of the sediments within the interval into a single batch. The clayey silt lithofacies is composed both of a few thin-bedded coarse layers and a majority of finer-grained material (see figure 10 in Phillips et al. (2020, this issue); there was very little sandy silt in 11FB-1); when we mixed this material together, any of the local layering was removed (see the Reconstituted Versus In Situ Behavior section).

We used two different techniques to reconstitute the material. For sandy silt lithofacies, we used the undercompaction technique (Germaine and Germaine, 2009). This technique is commonly used for nonplastic material or coarse sandy soils (Ladd, 1978). The material is oven dried at 40°C for 48 hr and disaggregated using a pestle and mortar. It is then uniformly mixed and moistened with 10 wt. % deaired brine (3.5% salinity for reducing the expandability of smectitic clays). Progressive layers of the sediments are then placed in a cylindrical sample, and each layer is packed to the desired thickness to achieve a specific porosity. To estimate the intrinsic permeability of sandy silt lithofacies, we packed the material in a 35.6-mm-diameter and 37.1-mm-long cylinder. We also used the undercompaction method to measure the compression behavior and the permeability at lower porosities (comparing to in situ porosity) of sandy silt. In this case, sediments were uniformly packed into a fixed-diameter metal ring with an initial height of 16.58 mm.

We could not use the undercompaction technique for the clayey silt lithofacies because the presence of clumped clay aggregates at low moisture content (10 wt. % deaired brine) would result in a nonuniformly reconstituted sample. Instead, we prepared the specimen by simulating natural sedimentation using resedimentation (Sheahan, 1991; Santagata and Kang, 2007; Schneider et al., 2011; Casey et al., 2015). In this approach, the clayey silt sediment is removed from the core liner and is mixed using a particular water content (twice the liquid limit) and salinity (3.5%). The sediment slurry is fully disaggregated and uniformly mixed using an electrical stand mixer. The slurry is then vacuumed and uniaxially incrementally loaded in a core liner to an initial vertical effective stress (0.1 MPa in this study). The resedimented specimen is then extruded from the core liner and trimmed into a steel ring for constant-rate-of-strain (CRS) consolidation experiment (see the Constant-Rate-of-Strain Consolidation Experiment section).

Index Properties Measurements

We measured index properties of each lithofacies before we reconstituted them. We measured the grain density (ρg) in a pycnometer (ASTM International, 2014). We also measured the Atterberg limits (the liquid limit [wL] and plastic limit [wP]) following ASTM guidelines (ASTM International, 2017b). We analyzed the particle size distribution of each lithofacies with the hydrometer technique (Germaine and Germaine, 2009). This method is particularly reliable in measuring clay-size particles and can capture particle sizes ranging from 0.0002 to 0.1 mm (Flemings et al., 2020, this issue; Meazell et al., 2020, this issue). Laser diffraction particle size analysis was also used for coarse-grained silts when the mass of sediments was insufficient for hydrometer analysis. Meazell et al. (2020, this issue) shows that the laser and the hydrometer approaches estimate similar particle size distributions within the sandy silt lithofacies but the hydrometer approach predicts a larger clay-size fraction than the laser-based approach in the clayey silt lithofacies, which is a well-understood phenomenon (Di Stefano et al., 2010).

We use four different porosity measurements. The PCATS porosity (nPCATS) is calculated from the best-estimated bulk density (ρb) of the core sediments assuming only water and hydrate are in the pores:

bltn19165eqd1

The grain density (ρg) is measured directly; the pore fluid density is assumed to equal that of seawater (ρw = 1.035 g/cm3) and the hydrate density (ρh) is assumed to equal 0.9 g/cm3. The measured PCATS bulk density (ρ∗b) (Figure 1F, I) is an apparent value because the core sediments do not occupy all of the volume within the core liner (Figure 1E, H). The best-estimated bulk density (ρb) of sediments is therefore adjusted from ρ∗b based on the bulk core volume (Vb), using mass balance:

bltn19165eqd2

The core volume (Vb) is calculated by measuring the cross-sectional area of the core for individual PCATS CT slices (each 112 μm thick) using ImageJ software (Abràmoff et al., 2004) and integrating the areas along the core length (Phillips et al., 2020, this issue). The volume of seawater (Vw) surrounding the core within the core liner is obtained by subtracting the bulk core volume (Vb) from the total volume of the core liner (Vt).

The PCATS porosity is measured at zero effective stress and has thus undergone expansion caused by unloading from the in situ stress conditions (Flemings and Saffer, 2018). In fact, the clayey silt lithofacies expands in the core liner; thus, the observed volume is larger than the in situ volume. This is known because the diameter of core (Dc) estimated by PCATS CT images is larger than the inner diameter (ID) of the coring tool (Dtool = 50.8 mm) (Thomas et al., 2020, this issue). However, we cannot necessarily assume that the cored sample (both sandy silt and clayey silt lithofacies) diameter at in situ (Do) is equal to the tool diameter (Dtool). This is because rotary coring disturbance can reduce the actual core diameter. We have observed that the actual core diameter ranges from approximately 45 to approximately 48 mm based on the PCATS CT images, which is significantly less than the ID of the coring tool. To best estimate the in situ diameter of the core, we have assumed that the sandy silt underwent no expansion and used the observed diameter of the sandy silt as the original diameter of the clayey silt core. This could overestimate the core volume for both the sandy silt and the clayey silt. With the above assumptions, the in situ porosity of clayey silt lithofacies (n) is further corrected, using mass balance as

bltn19165eqd3

where nPCATS is the estimated porosity of a clayey silt after volume expansion, Dc is calculated by dividing the Vb of a clayey silt core section over its core length (Lc), D0 is the core diameter of the sandy silt core section measured with x-ray, and n is thus a result of PCATS porosity calibrated for expansion. We obtained an average Dc (48.98 mm) and an average nPCATS (0.40) from multiple clayey silt samples (4FB-3, 4FB-5, 8FB-2, and 11FB-1) and also obtained an average D0 (46.15 mm) from multiple sandy silt samples (3FB-3, 4FB-2, 4FB-4, 4FB-7, and 7FB-1). These data sets are listed in Phillips et al. (2020, this issue).

The LWD porosity (nLWD) is calculated from equation 1 using the LWD bulk density. However, this estimate is subject to limitations because the LWD porosity is based on LWD values that average over a considerable vertical data sampling interval resolution (T. S. Collett et al., 2012).

The moisture and density (MAD) porosity (nMAD) is calculated in our laboratory experiments where no hydrate is present. This approach is described by (Blum, 1997). It assumes only water is present and the sample is 100% saturated; it is calculated as

bltn19165eqd4

where mw is the mass difference (i.e., water mass) between the measured dry mass and wet mass of the subsampled sediments. The grain density (ρg) is the same as used in equation 1, and the water density (ρw) is assumed to be 1.0 g/cm3.

The mercury injection capillary pressure (MICP) porosity (nHg) is the porosity from our mercury porosimetry measurements. Such porosity is the volume of mercury intruded into the reconstituted sample divided by the bulk volume of the sample (see the Mercury Injection Capillary Pressure Measurement section).

Steady-State Permeability Measurement

We consider hydrate to be an immobile, infinitely viscous fluid phase with a saturation Sh. Hence, porosity does not change as a function of hydrate saturation. The intrinsic permeability is the permeability of the sediments when fully saturated with a single fluid phase. In hydrate reservoirs, it is measured by disassociating any hydrate present in cores or by measuring permeability in samples in which only water is present, and no hydrate was ever formed. The effective permeability is the permeability estimated by Darcy’s law when multiple fluid phases are present. A permeability measured with hydrate present is an effective permeability and not an intrinsic permeability. Both intrinsic and effective permeability have dimensions of length squared. Relative permeability is distinct and is the ratio of effective to intrinsic permeability. This paper focuses entirely on measurements of intrinsic permeability.

We measured the intrinsic permeability of the sandy silt lithofacies material with a steady-state method both in a triaxial cell (ASTM International, 2016) and within a rigid-walled consolidation cell (ASTM International, 2015). In the triaxial cell, we used the undercompaction technique (see the Sample Selection and Preparation section) to reconstitute a sample to its estimated in situ porosity of 0.38. In fact, the resultant sample had an approximately uniform porosity of 0.39 as calculated from dry mass (md = 60.6 g), bulk volume (Vb = 37.04 cm3), and grain density (ρg = 2.675 g/cm3) (Figure 2C). We also measured the permeability of sandy silt lithofacies material at lower porosities of approximately 0.32 and approximately 0.34 to explore the relationship between porosity and permeability. In this case, we prepared the material to an initial porosity of 0.43 and then uniaxially compressed the material to lower porosities (see the Constant-Rate-of-Strain Consolidation Experiment section). We stopped compression and measured the steady-state permeability and then continued compression.

Figure 2. Shown are x-ray images of pressure cores and reconstituted core samples. (A) The 8-cm sandy silt lithofacies part of hydrate-bearing core 4FB-8 labeled with the yellow box in Figure 1E; (B) the 15-cm clayey silt lithofacies part of pressure core 11FB-1 labeled with the yellow box in Figure 1H. See Flemings et al. (2018c) for a description of these core data. (C) Sandy silt lithofacies sample reconstituted from parent sediments as shown in (A) by undercompaction technique (see the Sample Selection and Preparation section). (C) The sample prepared for steady-state permeability measurement (see the Steady-State Permeability Measurement section). (D) Clayey silt lithofacies sample reconstituted from parent sediments as shown in (B) by resedimentation technique (see the Sample Selection and Preparation section). The sample was prepared for constant-rate-of-strain (CRS) consolidation test and was scanned by computed tomography (CT) post-CRS test. The bright edge of the bottom as shown in (C) is caused by beam hardening during CT scanning (Brooks and Di Chiro, 1976). Comparison between reconstituted sample (C, D) and in situ material (A, B) is discussed in the Reconstituted versus In Situ Behavior section.

After the samples were prepared, they were saturated with deaired brine (3.5% salinity), and a hydrostatic confining pressure of 0.1 MPa was applied. We measured permeability by pumping water at a constant flow rate (q) while measuring the pressure difference (Δu) across the specimen at room temperature (25°C). Multiple flow rates were applied from 0.01 to 0.07 ml/min. The fluid pressure of each pump was monitored by a pressure transducer with a resolution of 0.1 kPa. The hydraulic gradient was selected to be between 1 and 10 to generate a sufficient pressure gradient to be observed accurately with the transducer yet also avoid specimen compaction. The intrinsic permeability (k) of the sample was calculated with Darcy’s law:

bltn19165eqd5

where q is the flow rate (m3/s), μ is dynamic viscosity of pore fluid (8.9 × 10−4 Pa·s at 25°C), H is specimen height (m), A is the cross section area of the specimen (m2), and Δu (Pa) is the measured pressure difference between the upstream and downstream ends of the specimen.

Constant-Rate-Of-Strain Consolidation Experiment

We measured the compression behavior of the sandy silt lithofacies and both the compression and permeability of the clayey silt lithofacies with uniaxial CRS consolidation experiments. These tests were performed at room temperature (25°C) following ASTM standards (ASTM International, 2012).

The specimen and its surrounding metal ring were loaded into the consolidation chamber with porous stones, and filter paper was placed on the top and base of the sample. The metal ring maintained uniaxial strain during consolidation. The consolidation chamber was sealed, filled with deaired brine (3.5% salinity), and pressurized with a constant pressure of 386 kPa for at least 16 hr to ensure full saturation. After this saturation stage, the drain valve at the base of the sample was locked, and the sample was consolidated vertically at a constant strain rate (bltn19165inf1). A strain rate of 2.5%/hr for the sandy silt lithofacies specimen and 0.4%/hr for the clayey silt lithofacies specimen was applied. In fact, significant excess pressure (Δu) could not be generated for the high-permeability (sandy silt lithofacies) specimen during consolidation. For this reason, we used a steady-state flow permeability measurement within the CRS cell (see the Steady-State Permeability Measurement section).

The sandy silt lithofacies specimen was consolidated to a vertical effective stress of 12 MPa, then unloaded to 3.8 MPa, then reloaded to 15 MPa and finally unloaded to 0.05 MPa. The stress levels 12 and 15 MPa were selected as two reference effective stresses. The clayey silt lithofacies specimen was first loaded to 3.8 MPa and held for 12 hr to allow excess pore pressure to dissipate to zero. The stress 3.8 MPa was the approximate representative of the in situ effective stress in the hydrate-bearing interval (see the Geological and Petrophysical Setting section). The specimen was then unloaded to 0.1 MPa under the same strain rate used for loading.

Axial load (σv), specimen height (H), and pore fluid pressure (u) at the top and base of the specimen were monitored throughout the test. The base excess pressure (Δu), when coupled with strain rate (bltn19165inf2), was used to calculate the vertical permeability (k) and coefficient of consolidation (Cv) (ASTM International, 2012):

bltn19165eqd6



bltn19165eqd7

where k is permeability (m2), bltn19165inf3 is strain rate (1/s), H is specimen height (m), H0 is initial specimen height (m), μ is dynamic viscosity of pore fluid (9.321 × 10−4 Pa·s at 22°C), and Δu is base excess pressure (Pa). Variable mv is the coefficient of volume compressibility determined by −Δε/Δσ′v and is also expressed as −Δnσ′v, which indicates the change in porosity n as a result of the application of vertical effective stress Δσv.

Void ratio (e) during the consolidation test was determined from strain data and the initial void ratio (e0). During virgin consolidation (the consolidation part that occurs when the new effective stress is larger than its previously experienced maximum effective stress), the stress–strain relationship is defined by the compression index Cc:

bltn19165eqd8

where Δe is the change in void ratio when vertical effective stress increases from σ’v1 to σ’v2.

Mercury Injection Capillary Pressure Measurement

We performed MICP measurements to characterize the pore-throat size distribution and capillary behavior of each lithofacies. Samples were first oven dried at 105°C for at least 24 hr, then measured at room temperature (20°C) using a Micrometrics AutoPore IV device. During the mercury intrusion experiment, mercury is injected at a progressively increasing pressure into the dried and evacuated sample. Detailed procedures are described in Daigle and Dugan (2014). The data were corrected for conformance following the method of Comisky et al. (2011). The mercury pressure is assumed equal to the capillary pressure in the smallest pore throats that the mercury can enter, that is, the smallest pore throats that are connected to the exterior of the sample through a pathway of mercury-filled pores. The pore-throat radius (rp) was calculated from Washburn’s equation:

bltn19165eqd9

where PHg is absolute mercury injection pressure (Pa), σHg is air–mercury interfacial tension (0.485 N/m), and θHg is the contact angle between mercury and the particle surface (130°). Mercury injection pressure (PHg) was converted to gas–water capillary pressure (Pcgw) by

bltn19165eqd10

We assumed that gas–water interfacial tension (σgw) is 0.072 N/m, and the gas–water contact angle (θgw) is 180°. The height above free water level (hfwl) was then calculated as

bltn19165eqd11

where hfwl is vertical distance from free water level to the top of reservoir (also interpreted to be the minimum hydrocarbon column height required to be accumulated for entering reservoir pores), ρw is water density (1035 kg/m3) with 3.5% salinity, and ρgas is the in situ methane density (194 kg/m3) (Starling and Savidge, 1992) at an in situ fluid pressure of 24.8 MPa and an in situ temperature of 19°C (Flemings et al., 2018d).

We determined three characteristic points on the capillary curves that we term the displacement pressure (Pd), the extrapolated displacement pressure (Pde), and the modal displacement pressure (Pmodal). The displacement pressure Pd was selected as the minimum pressure when mercury started entering the pore network during primary drainage, which corresponded, approximately, to the first data point when Sw was below 1. The extrapolated displacement pressure was determined using the model in Thomeer (1960): a hyperbola is fit to the displacement curve and used to project its horizontal asymptote for constraining the Pde and vertical asymptote for constraining the residual wetting saturation Srw (see the example of figure 1 in Thomeer, 1960). This Pde value is approximately equal to the point at which the plateau of the capillary curve is extended to 100% wetting saturation. Finally, the modal displacement pressure Pmodal (i.e., capillary pressure at modal pore-throat radius rmodal) for each sample was calculated by determining the maximum intruded volume for a given increase in capillary pressure. This is marked by a rapid increase in intruded mercury volume with a small increase in mercury pressure or the point of maximum slope on a plot of intruded mercury volume versus mercury pressure.

RESULTS

Pressure Core Analysis and Transfer System Analysis of Lithofacies

The core section from 4FB-8 (Figure 1E, yellow box) is characteristic of the sandy silt lithofacies. It has centimeter-scale layers of light (low density) and dark (higher density) that Meazell et al. (2020, this issue) infer to record ripple cross-laminations (Figure 2A). The darker (more dense) layers are inferred to be finer-grained material (Meazell et al., 2020, this issue). The average PCATS P-wave velocity in the section was 2947.7 m/s (Figure 1G), and the hydrate saturation was approximately 83% as measured from quantitative degassing experiments (Phillips et al., 2020, this issue) (Table 1).

The core section (5–20 cm) from 11FB-1 (Figure 1H, yellow box) is characteristic of the clayey silt lithofacies. It has an average P-wave velocity of 1665.8 m/s in PCATS and has large intervals with x-ray–imaged dark (high density) and structureless sediments (Figure 2B). This core section has a few sparsely distributed submillimeter planar light-color laminations, which may likely represent thin bands of coarse-grained material, which is a typical component in clayey silt lithofacies (see an example of Figure 5 in Meazell et al., 2020, this issue). Core 11FB-1 upon recovery in PCATS had low hydrate saturations (∼2%) (Table 1).

Figure 3. The Casagrande plasticity chart of core sediments. The clay fraction of each sample is color coded. The background colors record the soil classifications of the Unified Soil Classification System (USCS), which are defined and interpreted in ASTM International (2017a). Atterberg limits of our samples are compared to other gas hydrate reservoirs: India National Gas Hydrate Program Expedition 02 (NGHP-02) samples are from NGHP B and C areas, Krishna–Godavari Basin of Eastern India (Dai et al., 2018); Ulleung Basin Gas Hydrate (UBGH) samples are from Ulleung Basin, Sea of Japan (Yun et al., 2011). Characteristic samples from nonhydrate reservoir locations in the Gulf of Mexico (GOM) are also included for comparison: GOM Eugene Island (GOM-EI) is a clay from Eugene Island, GOM, and GOM-Ursa is a siltstone from Ursa Basin, GOM (Casey et al., 2019). A-Line = the line that separates silt from clay; CH = fat clay; CL = lean clay; MH = elastic silt; ML = silt; OH = organic silt or clay with high liquid limit; OL = organic silt or clay with low liquid limit; U-Line = the uppermost extent of limits for naturally occurring soils; UT-GOM2-1 = The University of Texas-Gulf of Mexico 2-1 Hydrate Pressure Coring Expedition.

Both reconstituted sandy silt lithofacies and clayey silt lithofacies materials have a homogenous structure in comparison to the intact samples (Figure 2C vs. Figure 2A and compare Figure 2D to Figure 2B). This is not surprising because we homogenized the material over each highlighted yellow-box interval in Figure 1E, H.

Index Properties

Liquid limit (wL[%]) reflects the quantity and type (e.g., fat vs. lean and organic vs. inorganic) of clay minerals present in natural sediments, combining the mechanical effects of these two attributes into a single number, providing a first and fast approximate prediction of engineering properties such as permeability (k) and compression index (Cc) at a given effective stress (Casey et al., 2013, 2019). Sandy silt lithofacies material has very low plasticity, plots above the A-Line, and is an inorganic silt when described by the Casagrande plasticity chart (Figure 3; Table 2). Clayey silt lithofacies material lies at the low- to high-plasticity boundary, lies above the A-Line, and is characterized as a lean clay (Figure 3). Sandy silt lithofacies has a small clay fraction (∼4% particle-size clay by mass; Figures 3, 4) and is dominated by silt, whereas clayey silt lithofacies has 47% clay by mass (see Figures 3, 4 and Meazell et al., 2020, in press). The clay composition is approximately the same in both cases (Meazell et al., 2020, in press). Therefore, the primary driver for this difference in liquid limit is the fraction of silt and sand present. This behavior is also reflected in Figure 3 in which samples with a higher silt or sand fraction plot to the lower left.

The PCATS porosity of sandy silt lithofacies in 4FB-8 core (Figure 2A) is approximately 0.38 using equations 1 and 2 and the parameters listed in Table 1. For sandy silt lithofacies, we assumed Sh = 80% based on Phillips et al. (2020, in press). The LWD porosity of the sandy silt is also approximately 0.38 based on equation 1 using the LWD density between 423 and 424 mbsf and parameters in Table 1. We interpret that the in situ porosity of sandy silt lithofacies in this location is approximately 0.38. The fact that the PCATS porosity and the LWD porosity are so similar implies that the volume expansion of sandy silt sample in PCATS is not significantly enough to be distinguished by PCATS CT scan and LWD log.

Figure 4. Grain size distributions on a semilog scale. The particle size fraction is expressed as a percentage of the total dry weight. Silts 4FB-4, 11FB-1 (Figure 1H, labeled with the yellow box), and 4FB-3 were analyzed using the hydrometer method (Germaine and Germaine, 2009), whereas 4FB-8 (Figure 1E, labeled with the yellow box) was analyzed by laser diffraction (Flemings et al., 2020, in press). The sand versus silt boundary is defined at 62.5 μm, and the silt versus clay boundary is defined at 2 μm. Silt 4FB-8 has a clay-size fraction of 4%, and 11FB-1 has a clay-size fraction of 47%. Sediments of 4FB-8 and 11FB-1 were used for reconstituting artificial samples of sandy silt lithofacies and clayey silt lithofacies, respectively. Inset shows sedimentological classification of the lithofacies based on the Shepard scale (Shepard, 1954). D50 = median grain diameter.

The average in situ porosity of clayey silt is estimated to be approximately 0.33 using the approach described in the Index Properties Measurements section that accounts for core expansion (equations 1–3). The LWD porosity of material in 11FB core cannot be directly measured because in this interval the LWD log was significantly affected by borehole washout (figure 15 in Boswell et al., 2012b).

Figure 5. The permeabilities (k) of reconstituted sandy silt (4FB-8) and clayey silt (11FB-1). The sandy silt k were obtained at n = 0.39, 0.34, and 0.32, respectively (orange dots). The clayey silt k were measured by the constant-rate-of-strain experiment (bluish-green dots). The porosity–k behavior of the Gulf of Mexico (GOM) Ursa siltstone is marked by the blue line, and GOM Ursa mudstone with clay from 50% to 70% are marked in the tannish-yellow zone (Reece et al., 2012). The black lines are the predicted intrinsic k using k–liquid limit (wL[%]) correlations (wL[%] = 23 for sandy silt lithofacies and wL[%] = 49.8 for clayey silt lithofacies) summarized from all mud rocks in Casey et al. (2013). σ′v = vertical effective stress.

Bulletin Article

Outcrop-based reservoir characterization of ...

This study documents that Danian-aged sand remobilization o...

Bulletin Article

Subsalt pressure prediction in the Miocene M...

Sandstone pressures follow the hydrostatic gradient in Mioc...

Article

Friendly Geology, Tough Crowd

Fighting the 'good fight' ... this geologist helps the environmentally conscious understa...

Bulletin Article

Integrating borehole image logs with core: A...

Image-logs provide critical data for characterizing fractur...

Emphasis Article

Gulf of Mexico: Where Big Fields Are Getting...

Cindy Yeilding of BP America discusses the production histo...

Article

NAPE Crowd Shows 'Savvy' Side

Once again, the North American Prospect Expo can be declare...

Article

Katrina Recovery to Take Awhile

After the storm: Geologists, like everyone else in the post...

Emphasis Article

Deep Gulf Targets Get New Tactics

Seismic technology continues to drive the deep gas play in ...

Emphasis Article

Industry Must Face Tough Issues

If you could listen to a heartbeat of the geophysical indu...

Emphasis Article

Underexplored Rub' al Khali Basin Could Hold...

It is one of the most remote places on the planet and home ...

Bulletin Article

Petrography, fluid-inclusion, isotope, and t...

Feldspar dissolution, quartz cementation, and clay cementat...

Bulletin Article

Handling natural complexity in three-dimensi...

Volumetric restoration can provide crucial insights into th...

Article

Flexibility Rules in Shale Fields

Independent companies, both large and small, are widely ack...

Emphasis Article

Meteorites Make Good Impression

Heaven-sent blessings? Geologists discuss the exploration ...

Bulletin Article

Diagenesis and quartz cement distribution of...

Reservoir properties of Upper Triassic–Middle Jurassic sand...

Article

Memoir Explores Southern GOM Geology

A new view of a familiar area: A recently published AAPG me...

Emphasis Article

Visualization: Not a Parlor Trick

It used to be a 'gee whiz' extravagance, but 3-D computer v...

Emphasis Article

A Couple of Deep GOM Projects Get Go-Ahead

It’s been little more than a year since the infamous ...

Bulletin Article

Measuring and modeling fault density for CO2...

Emission of carbon dioxide (CO2) from fossil-fueled power g...

Emphasis Article

Science Curricula Under Threat?

So-called 'academic freedom' bills at various stages in sta...

Bulletin Article

Mechanisms of shale gas storage: Implication...

This article reviews the mechanisms of shale gas storage an...

Bulletin Article

The equivalent pore aspect ratio as a tool f...

The equivalent pore aspect ratio (EPAR) provides a tool to ...

Article

New ACE Committee Structure Welcomes Diversi...

With technological advancements driving exploration toward ...

Article

Davy Jones: Shallow Water, Deep Treasure

Shallow water, deep treasure: The recent Davy Jones discove...

Please log in to read the full article

ABSTRACT

We explore the petrophysical behavior of the two interbedded lithofacies (sandy silt and clayey silt) that constitute the Green Canyon Block 955 hydrate reservoir in the deep-water Gulf of Mexico by performing experiments on reconstituted samples of the reservoir material. Sandy silts reconstituted to the in situ porosity have a permeability of 11.8 md (1.18 × 10−14 m2), which is similar to the intrinsic permeabilities measured in intact cores from hydrate reservoirs of similar grain size offshore Japan (Nankai Trough) and offshore India. Reconstituted clayey silts have a much lower intrinsic permeability of 3.84 × 10−4 md (3.84 × 10−19 m2) at the in situ stress. The reconstituted sandy silt is less compressible than the clayey silt. Mercury injection capillary pressure measurements demonstrate that the largest pores with the clayey silt are still smaller than the pores remaining after 90% hydrate saturation in sandy silt. We interpret that the methane solubility in pores of clayey silt is always less than that necessary to form hydrate, which explains why no hydrate is present in the clayey silt. We upscale the reservoir properties to estimate the behavior of interbedded sandy silt and clayey silt. We find the upscaled intrinsic horizontal and vertical permeabilities for the entire reservoir interval are 8.6 md (8.6 × 10−15 m2) and 1.4 × 10−3 md (1.4 × 10−18 m2). We estimate that during reservoir production, a maximum vertical strain of approximately 12% will result. Ultimately, this study will inform reservoir simulation models with petrophysical properties at scales of both individual lithofacies and reservoir formation.

INTRODUCTION

Methane hydrate is a crystalline solid composed of methane molecules trapped in cages of water molecules (Sloan and Koh, 2007). It is stable at low temperatures and high pressures, which in natural environments is commonly found in permafrost regions, outer continental margins, and shallow-basin sediments (T. S. Collett, 2002; T. S. Collett et al., 2011).

Concentrated hydrate deposits are commonly found in coarse-grained sediments (e.g., silt to sand size) (Boswell and Collett, 2011). In the deep-water Gulf of Mexico (GOM), multiple reservoirs of this type have been found (Boswell et al., 2012a). A full log suite across a part of the gas hydrate stability zone in the northern GOM at Alaminos Canyon (AC) Block 818 recorded pore-filling hydrate at a saturation of 80% within a 13-m-thick formation (Boswell et al., 2009). The GOM Gas Hydrate Joint Industry Project Leg II (JIP II) drilled two wells (well G and well H) at Walker Ridge 313 in Terrebonne Basin, and the logging while drilling (LWD) data suggested saturations above 80% in the higher-quality sands encountered (Boswell et al., 2012b). The JIP II program also drilled the Green Canyon Block 955 (GC 955) H001 well (H001) where high hydrate saturations (Sh = 60%–65%) were found in sediments with relatively low gamma-ray values that were interpreted to represent silt or sand size (T. S. Collett et al., 2012; Lee and Collett, 2012). Shipboard hydrate dissociation tests on pressure cores from the eastern Nankai Trough, Japan, revealed that the average hydrate saturation of the sandy cores ranges from 55% to 68% (Fujii et al., 2009). Recent studies from the India National Gas Hydrate Program Expedition 02 (NGHP-02) suggested that the hydrate saturations measured from controlled on-site depressurizations record various hydrate saturations over a wide range of grain sizes (e.g., super high saturation [∼100%] present in gas hydrate-bearing gravels, high saturation [∼65%–85%] within sand layers, moderate saturation [∼45%] in silty clays, and low or no saturation [0%–10%] in thin clay layers) (Boswell et al., 2018b).

Multiple offshore drilling and coring programs conducted in the past decade have evaluated reservoirs in the GOM for their potential as an energy resource (T. Collett et al., 2015). One such expedition, targeting a coarse-grained reservoir, was The University of Texas-Gulf of Mexico 2-1 (UT-GOM2-1) Hydrate Pressure Coring Expedition at GC 955 in the deep-water GOM (Flemings et al., 2018a, b, 2020, this issue). One of the primary goals of this expedition was to characterize the petrophysical properties of hydrate reservoirs and their bounding seals. Petrophysical characteristics, such as grain size, porosity, compressibility, permeability, and capillary behavior, all control both how these reservoirs form (Clennell et al., 1999; Liu and Flemings, 2006, 2007; Torres et al., 2008; Fujii et al., 2009; You and Flemings, 2018) and how they will behave during production (Moridis et al., 2007).

The most common way to study the petrophysical properties of hydrate-bearing sediments is through pressure core analysis. For example, previous studies measured permeability and compressibility of hydrate-bearing core and then dissociated the hydrate and repeated the measurement (Konno et al., 2015; Santamarina et al., 2015; Yoneda et al., 2019a). Extensive pressure coring programs have been performed in offshore Japan (Yoneda et al., 2017, 2018b), offshore Korea (Yun et al., 2011; Ryu et al., 2013), offshore India (Yoneda et al., 2019b), and more recently in offshore China (S. Yang et al., 2017).

Pressure core analysis is beginning to illuminate the mechanical and hydraulic properties of methane hydrate reservoirs, but it has its own challenges. First, the coring, recovery, and measurement process may disturb the sample and thus alter its properties. For example, during pressure core recovery, the friction between the pressure coring tool and the sediment may fracture the core sample. When the confining pressure of the earth is removed, the stress unloading may expand the sample altering its property (Santagata and Germaine, 2005; Casey et al., 2016; Thomas et al., 2020, this issue). In addition, during depressurization, the fine-grained sediment may move as the hydrate is dissociated, thus altering the material that is being measured (Jung et al., 2012; Han et al., 2018). Second, the intact core may be heterogeneous at the scale of the core. In this case, any measured property will reflect a mixture of the thin-bedded materials. Third, although this method illuminates the present petrophysical nature of the reservoir, it is not possible to provide the evolution of petrophysical properties from old in situ conditions to its current in situ conditions.

In this study, we took a different approach to analyzing petrophysical properties of hydrate-bearing reservoirs. We took sediment from the two characteristic lithofacies present in the GC 955 hydrate reservoir. We then reconstituted these sediments to the approximate state they were at in the subsurface. Reconstitution experiments disaggregate original sediments to their constituent materials and then recompress them in a manner that is similar to the burial process (Casey et al., 2015). We then studied the petrophysical properties of these reconstituted materials to gain insight into the properties of the in situ reservoir. Ideally, this approach can illuminate the intrinsic permeability of the reservoir (the permeability of the porous media with only water present in the pores), compressibility of the reservoir without hydrate present, and pore structure of the reservoir and its associated capillary properties. These are all necessary parameters for reservoir engineering simulators because they require both reported experimental data (i.e., intrinsic permeability, compressibility, and capillary behavior) and an understanding of these data so that they can ultimately build a reliable model and predict accurate reservoir behavior in hydrate production.

The approach of reconstituting sediments has strengths and limitations. It eliminates sample variability, produces uniform specimens, and isolates the effects of composition, effective stress, and stress history on petrophysical properties through systematic experiments. This is not possible with the use of intact samples because no two intact samples will possess an identical composition and stress history. In addition, analysis of reconstituted samples allows one to study behavior over a range of in situ effective stresses (Casey et al., 2016). However, the reconstitution approach homogenizes the sediments. Thus, any layering or local heterogeneity originally present will be lost. Furthermore, there is an open discussion on whether the compression approach used successfully generates the pore-scale structure that is formed by geological deposition and burial. That said, for materials that have not been significantly affected by diagenesis, cementation, or aging effect, a large amount of experimental work has demonstrated that the mechanical behavior of intact samples can be closely reproduced by the laboratory reconstituted material (Burland, 1990; Berman et al., 1993; Betts, 2014). No cementation or diagenesis has been observed in the GC 955 sediments (Flemings et al., 2018d).

This paper analyzes the two lithofacies (sandy silt and clayey silt) encountered at two particular depth intervals within the GC 955 hydrate reservoir. First, we separately selected the sandy silt and clayey silt materials and describe their index properties (e.g., Atterberg limits, particle size distribution, grain density, and porosity). We then describe how we reconstituted samples of sandy silt and clayey silt. We then describe the intrinsic permeability, compressibility, and capillary behavior of the reconstituted materials from each lithofacies. Finally, we integrate these data to infer possible mechanisms for hydrate formation in these reservoirs and describe the evolution of permeability and reservoir compaction during depressurization.

GEOLOGICAL AND PETROPHYSICAL SETTING

The hydrate-bearing, coarse-grained reservoir interval at GC 955 on the Texas–Louisiana continental rise was originally located based on industry drilling and geophysical characterization (D. R. McConnell, 2000; Hutchinson et al., 2008; D. R. McConnell et al., 2012). In 2009, JIP II used LWD to confirm the hydrate reservoir between 414 and 550 m below the sea floor (mbsf) (1358–1804 ft) from the high resistivity, high P-wave velocity, and low gamma-ray density values (D. McConnell et al., 2010; Boswell et al., 2012a, b; T. S. Collett et al., 2012). The reservoir was interpreted to be composed of fine to very fine sands interbedded with clays on the basis of LWD log character (Boswell et al., 2012b), which is similar to the thinly bedded and laminated deposits from the Viosca Knoll area, eastern GOM (Shew et al., 1994).

The GC 955 hydrate reservoir was cored in 2017 during the UT-GOM2-1 expedition (Flemings et al., 2018d). Two wells (hole GC 955 H002 and hole GC 955 H005) were drilled to retrieve pressurized gas hydrate–saturated cores. The pressure core analysis and transfer system (PCATS) (Schultheiss et al., 2011) was used to measure the P-wave velocity (Vp) and bulk gamma-ray density (ρb) of pressure cores through the core liner at a resolution of 0.5 cm. We obtained x-ray computed tomography (CT) images (voxel resolution of 112 μm). Twenty-one 1.2-m-long pressurized storage vessels were transported to the shore-based laboratory (UT Pressure Core Center) for further analysis. Details regarding the stratigraphy and sedimentology of the reservoir are available in Santra et al. (2020, this issue) and Meazell et al. (2020, this issue).

The GC 955 hydrate reservoir was pressure cored to acquire samples (Thomas et al., 2020, this issue), and from these samples, we interpreted the composition, hydrate concentration, and mechanical properties of this reservoir. Meazell et al. (2020, this issue) defined two lithofacies, sandy silt and clayey silt, within the primary hydrate-bearing interval (415–449 mbsf) (Figure 1A–D). Sandy silt lithofacies, with abundant trough cross-bedding structures (Meazell et al., 2020, this issue), has a low bulk density (ρb = 1.7 to 1.9 g/cm3) and high P-wave velocity (Vp = ∼2700–3200 m/s) (Figure 1E–G). Quantitative degassing indicated that sandy silt lithofacies contains high hydrate saturation (Sh = 66%–87%) and clayey silt lithofacies contains little to no hydrate (Flemings et al., 2020, this issue; Phillips et al., 2020, this issue). The hydrate saturation (Sh) of the samples was identified by quantitative degassing experiments on intact samples of the formation. Details of the estimation method are described in Phillips et al. (2020, this issue). In contrast, clayey silt lithofacies has relatively higher bulk density (ρb = 1.9 – 2.1 g/cm3) and lower P-wave velocity (Vp = ∼1700 m/s) (Figure 1H–J) (Flemings et al., 2018d, 2020, this issue; Meazell et al., 2020, this issue).

Figure 1. (A) Depth in meters below seafloor (mbsf) at the hole Green Canyon (GC) 955 H005. (B) The ring resistivity log from the Gulf of Mexico Joint Industry Project (JIP II) hole GC 955 H001 acquired during JIP II depth shifted to the estimated position in the H005 well. (C) Cored intervals of the H005 well. (D) Recovered and not recovered zones within the cored interval. All recovered core is assumed to reside at the top of the cored interval. (E) Pressure core analysis transfer system (PCATS) x-ray image of sandy silt core interval (H005-4FB-8 [423.56–423.99 mbsf]) studied in this paper; the yellow box outlines the interval of the sample analyzed. (F) Gamma-ray attenuation bulk density (ρ∗b) from the PCATS taken at 0.5-cm intervals. (G) The PCATS P-wave velocity (Vp). (H) An x-ray image of clayey silt lithofacies interval analyzed in this paper (H005-11FB-1 [441.27–441.55 mbsf]); the yellow box outlines the interval of the sample analyzed. (I) The ρ∗b for clayey silt. (J) The PCATS Vp for clayey silt. See Flemings et al. (2018c) for a discussion of PCATS analysis methods.

The overburden stress was calculated by integrating the LWD bulk density log from the GOM JIP hole GC 955 H001. Seawater salinity was assumed to be 3.5 wt. %. The pore pressure (u) was assumed to be hydrostatic (pore pressure gradient = 8.6 ppg [0.447 psi/ft]) because there was no evidence of any elevated pore pressures during previous drilling of H001 (Flemings et al., 2018d). The vertical effective stress (σ′v), calculated by subtracting the hydrostatic pore pressure (u) from the overburden (σv) is 3.76 MPa at the top of the hydrate-bearing interval (415 mbsf) and 4.01 MPa at the bottom (449 mbsf) (Flemings et al., 2020, this issue). The in situ temperature of GC 955 hydrate-bearing sediments was estimated to be between 18.6°C (at a depth of 414 mbsf) and 19.8°C (at a depth of 449 mbsf) based on the estimated temperature at the base of the hydrate stability zone (20.4°C) and the geothermal gradient of 34.7°C/km (Flemings et al., 2020, this issue).

METHODS

Sample Selection and Preparation

We analyzed material from an interval within core H005-4FB-8 that represented typical examples of the sandy silt lithofacies (Figure 1E, yellow box) and from core H005-11FB-1 (Figure 1H, yellow box) that represented a typical example of the clayey silt lithofacies. We chose these specific intervals (i.e., Figure 1E, H, location range highlighted in the yellow box) because they are characteristic of sandy silt lithofacies and clayey silt lithofacies as described by Flemings et al. (2020, this issue) and Meazell et al. (2020, this issue) (Figure 1F, G, I, J). For each sample interval, we mixed all of the sediments within the interval into a single batch. The clayey silt lithofacies is composed both of a few thin-bedded coarse layers and a majority of finer-grained material (see figure 10 in Phillips et al. (2020, this issue); there was very little sandy silt in 11FB-1); when we mixed this material together, any of the local layering was removed (see the Reconstituted Versus In Situ Behavior section).

We used two different techniques to reconstitute the material. For sandy silt lithofacies, we used the undercompaction technique (Germaine and Germaine, 2009). This technique is commonly used for nonplastic material or coarse sandy soils (Ladd, 1978). The material is oven dried at 40°C for 48 hr and disaggregated using a pestle and mortar. It is then uniformly mixed and moistened with 10 wt. % deaired brine (3.5% salinity for reducing the expandability of smectitic clays). Progressive layers of the sediments are then placed in a cylindrical sample, and each layer is packed to the desired thickness to achieve a specific porosity. To estimate the intrinsic permeability of sandy silt lithofacies, we packed the material in a 35.6-mm-diameter and 37.1-mm-long cylinder. We also used the undercompaction method to measure the compression behavior and the permeability at lower porosities (comparing to in situ porosity) of sandy silt. In this case, sediments were uniformly packed into a fixed-diameter metal ring with an initial height of 16.58 mm.

We could not use the undercompaction technique for the clayey silt lithofacies because the presence of clumped clay aggregates at low moisture content (10 wt. % deaired brine) would result in a nonuniformly reconstituted sample. Instead, we prepared the specimen by simulating natural sedimentation using resedimentation (Sheahan, 1991; Santagata and Kang, 2007; Schneider et al., 2011; Casey et al., 2015). In this approach, the clayey silt sediment is removed from the core liner and is mixed using a particular water content (twice the liquid limit) and salinity (3.5%). The sediment slurry is fully disaggregated and uniformly mixed using an electrical stand mixer. The slurry is then vacuumed and uniaxially incrementally loaded in a core liner to an initial vertical effective stress (0.1 MPa in this study). The resedimented specimen is then extruded from the core liner and trimmed into a steel ring for constant-rate-of-strain (CRS) consolidation experiment (see the Constant-Rate-of-Strain Consolidation Experiment section).

Index Properties Measurements

We measured index properties of each lithofacies before we reconstituted them. We measured the grain density (ρg) in a pycnometer (ASTM International, 2014). We also measured the Atterberg limits (the liquid limit [wL] and plastic limit [wP]) following ASTM guidelines (ASTM International, 2017b). We analyzed the particle size distribution of each lithofacies with the hydrometer technique (Germaine and Germaine, 2009). This method is particularly reliable in measuring clay-size particles and can capture particle sizes ranging from 0.0002 to 0.1 mm (Flemings et al., 2020, this issue; Meazell et al., 2020, this issue). Laser diffraction particle size analysis was also used for coarse-grained silts when the mass of sediments was insufficient for hydrometer analysis. Meazell et al. (2020, this issue) shows that the laser and the hydrometer approaches estimate similar particle size distributions within the sandy silt lithofacies but the hydrometer approach predicts a larger clay-size fraction than the laser-based approach in the clayey silt lithofacies, which is a well-understood phenomenon (Di Stefano et al., 2010).

We use four different porosity measurements. The PCATS porosity (nPCATS) is calculated from the best-estimated bulk density (ρb) of the core sediments assuming only water and hydrate are in the pores:

bltn19165eqd1

The grain density (ρg) is measured directly; the pore fluid density is assumed to equal that of seawater (ρw = 1.035 g/cm3) and the hydrate density (ρh) is assumed to equal 0.9 g/cm3. The measured PCATS bulk density (ρ∗b) (Figure 1F, I) is an apparent value because the core sediments do not occupy all of the volume within the core liner (Figure 1E, H). The best-estimated bulk density (ρb) of sediments is therefore adjusted from ρ∗b based on the bulk core volume (Vb), using mass balance:

bltn19165eqd2

The core volume (Vb) is calculated by measuring the cross-sectional area of the core for individual PCATS CT slices (each 112 μm thick) using ImageJ software (Abràmoff et al., 2004) and integrating the areas along the core length (Phillips et al., 2020, this issue). The volume of seawater (Vw) surrounding the core within the core liner is obtained by subtracting the bulk core volume (Vb) from the total volume of the core liner (Vt).

The PCATS porosity is measured at zero effective stress and has thus undergone expansion caused by unloading from the in situ stress conditions (Flemings and Saffer, 2018). In fact, the clayey silt lithofacies expands in the core liner; thus, the observed volume is larger than the in situ volume. This is known because the diameter of core (Dc) estimated by PCATS CT images is larger than the inner diameter (ID) of the coring tool (Dtool = 50.8 mm) (Thomas et al., 2020, this issue). However, we cannot necessarily assume that the cored sample (both sandy silt and clayey silt lithofacies) diameter at in situ (Do) is equal to the tool diameter (Dtool). This is because rotary coring disturbance can reduce the actual core diameter. We have observed that the actual core diameter ranges from approximately 45 to approximately 48 mm based on the PCATS CT images, which is significantly less than the ID of the coring tool. To best estimate the in situ diameter of the core, we have assumed that the sandy silt underwent no expansion and used the observed diameter of the sandy silt as the original diameter of the clayey silt core. This could overestimate the core volume for both the sandy silt and the clayey silt. With the above assumptions, the in situ porosity of clayey silt lithofacies (n) is further corrected, using mass balance as

bltn19165eqd3

where nPCATS is the estimated porosity of a clayey silt after volume expansion, Dc is calculated by dividing the Vb of a clayey silt core section over its core length (Lc), D0 is the core diameter of the sandy silt core section measured with x-ray, and n is thus a result of PCATS porosity calibrated for expansion. We obtained an average Dc (48.98 mm) and an average nPCATS (0.40) from multiple clayey silt samples (4FB-3, 4FB-5, 8FB-2, and 11FB-1) and also obtained an average D0 (46.15 mm) from multiple sandy silt samples (3FB-3, 4FB-2, 4FB-4, 4FB-7, and 7FB-1). These data sets are listed in Phillips et al. (2020, this issue).

The LWD porosity (nLWD) is calculated from equation 1 using the LWD bulk density. However, this estimate is subject to limitations because the LWD porosity is based on LWD values that average over a considerable vertical data sampling interval resolution (T. S. Collett et al., 2012).

The moisture and density (MAD) porosity (nMAD) is calculated in our laboratory experiments where no hydrate is present. This approach is described by (Blum, 1997). It assumes only water is present and the sample is 100% saturated; it is calculated as

bltn19165eqd4

where mw is the mass difference (i.e., water mass) between the measured dry mass and wet mass of the subsampled sediments. The grain density (ρg) is the same as used in equation 1, and the water density (ρw) is assumed to be 1.0 g/cm3.

The mercury injection capillary pressure (MICP) porosity (nHg) is the porosity from our mercury porosimetry measurements. Such porosity is the volume of mercury intruded into the reconstituted sample divided by the bulk volume of the sample (see the Mercury Injection Capillary Pressure Measurement section).

Steady-State Permeability Measurement

We consider hydrate to be an immobile, infinitely viscous fluid phase with a saturation Sh. Hence, porosity does not change as a function of hydrate saturation. The intrinsic permeability is the permeability of the sediments when fully saturated with a single fluid phase. In hydrate reservoirs, it is measured by disassociating any hydrate present in cores or by measuring permeability in samples in which only water is present, and no hydrate was ever formed. The effective permeability is the permeability estimated by Darcy’s law when multiple fluid phases are present. A permeability measured with hydrate present is an effective permeability and not an intrinsic permeability. Both intrinsic and effective permeability have dimensions of length squared. Relative permeability is distinct and is the ratio of effective to intrinsic permeability. This paper focuses entirely on measurements of intrinsic permeability.

We measured the intrinsic permeability of the sandy silt lithofacies material with a steady-state method both in a triaxial cell (ASTM International, 2016) and within a rigid-walled consolidation cell (ASTM International, 2015). In the triaxial cell, we used the undercompaction technique (see the Sample Selection and Preparation section) to reconstitute a sample to its estimated in situ porosity of 0.38. In fact, the resultant sample had an approximately uniform porosity of 0.39 as calculated from dry mass (md = 60.6 g), bulk volume (Vb = 37.04 cm3), and grain density (ρg = 2.675 g/cm3) (Figure 2C). We also measured the permeability of sandy silt lithofacies material at lower porosities of approximately 0.32 and approximately 0.34 to explore the relationship between porosity and permeability. In this case, we prepared the material to an initial porosity of 0.43 and then uniaxially compressed the material to lower porosities (see the Constant-Rate-of-Strain Consolidation Experiment section). We stopped compression and measured the steady-state permeability and then continued compression.

Figure 2. Shown are x-ray images of pressure cores and reconstituted core samples. (A) The 8-cm sandy silt lithofacies part of hydrate-bearing core 4FB-8 labeled with the yellow box in Figure 1E; (B) the 15-cm clayey silt lithofacies part of pressure core 11FB-1 labeled with the yellow box in Figure 1H. See Flemings et al. (2018c) for a description of these core data. (C) Sandy silt lithofacies sample reconstituted from parent sediments as shown in (A) by undercompaction technique (see the Sample Selection and Preparation section). (C) The sample prepared for steady-state permeability measurement (see the Steady-State Permeability Measurement section). (D) Clayey silt lithofacies sample reconstituted from parent sediments as shown in (B) by resedimentation technique (see the Sample Selection and Preparation section). The sample was prepared for constant-rate-of-strain (CRS) consolidation test and was scanned by computed tomography (CT) post-CRS test. The bright edge of the bottom as shown in (C) is caused by beam hardening during CT scanning (Brooks and Di Chiro, 1976). Comparison between reconstituted sample (C, D) and in situ material (A, B) is discussed in the Reconstituted versus In Situ Behavior section.

After the samples were prepared, they were saturated with deaired brine (3.5% salinity), and a hydrostatic confining pressure of 0.1 MPa was applied. We measured permeability by pumping water at a constant flow rate (q) while measuring the pressure difference (Δu) across the specimen at room temperature (25°C). Multiple flow rates were applied from 0.01 to 0.07 ml/min. The fluid pressure of each pump was monitored by a pressure transducer with a resolution of 0.1 kPa. The hydraulic gradient was selected to be between 1 and 10 to generate a sufficient pressure gradient to be observed accurately with the transducer yet also avoid specimen compaction. The intrinsic permeability (k) of the sample was calculated with Darcy’s law:

bltn19165eqd5

where q is the flow rate (m3/s), μ is dynamic viscosity of pore fluid (8.9 × 10−4 Pa·s at 25°C), H is specimen height (m), A is the cross section area of the specimen (m2), and Δu (Pa) is the measured pressure difference between the upstream and downstream ends of the specimen.

Constant-Rate-Of-Strain Consolidation Experiment

We measured the compression behavior of the sandy silt lithofacies and both the compression and permeability of the clayey silt lithofacies with uniaxial CRS consolidation experiments. These tests were performed at room temperature (25°C) following ASTM standards (ASTM International, 2012).

The specimen and its surrounding metal ring were loaded into the consolidation chamber with porous stones, and filter paper was placed on the top and base of the sample. The metal ring maintained uniaxial strain during consolidation. The consolidation chamber was sealed, filled with deaired brine (3.5% salinity), and pressurized with a constant pressure of 386 kPa for at least 16 hr to ensure full saturation. After this saturation stage, the drain valve at the base of the sample was locked, and the sample was consolidated vertically at a constant strain rate (bltn19165inf1). A strain rate of 2.5%/hr for the sandy silt lithofacies specimen and 0.4%/hr for the clayey silt lithofacies specimen was applied. In fact, significant excess pressure (Δu) could not be generated for the high-permeability (sandy silt lithofacies) specimen during consolidation. For this reason, we used a steady-state flow permeability measurement within the CRS cell (see the Steady-State Permeability Measurement section).

The sandy silt lithofacies specimen was consolidated to a vertical effective stress of 12 MPa, then unloaded to 3.8 MPa, then reloaded to 15 MPa and finally unloaded to 0.05 MPa. The stress levels 12 and 15 MPa were selected as two reference effective stresses. The clayey silt lithofacies specimen was first loaded to 3.8 MPa and held for 12 hr to allow excess pore pressure to dissipate to zero. The stress 3.8 MPa was the approximate representative of the in situ effective stress in the hydrate-bearing interval (see the Geological and Petrophysical Setting section). The specimen was then unloaded to 0.1 MPa under the same strain rate used for loading.

Axial load (σv), specimen height (H), and pore fluid pressure (u) at the top and base of the specimen were monitored throughout the test. The base excess pressure (Δu), when coupled with strain rate (bltn19165inf2), was used to calculate the vertical permeability (k) and coefficient of consolidation (Cv) (ASTM International, 2012):

bltn19165eqd6



bltn19165eqd7

where k is permeability (m2), bltn19165inf3 is strain rate (1/s), H is specimen height (m), H0 is initial specimen height (m), μ is dynamic viscosity of pore fluid (9.321 × 10−4 Pa·s at 22°C), and Δu is base excess pressure (Pa). Variable mv is the coefficient of volume compressibility determined by −Δε/Δσ′v and is also expressed as −Δnσ′v, which indicates the change in porosity n as a result of the application of vertical effective stress Δσv.

Void ratio (e) during the consolidation test was determined from strain data and the initial void ratio (e0). During virgin consolidation (the consolidation part that occurs when the new effective stress is larger than its previously experienced maximum effective stress), the stress–strain relationship is defined by the compression index Cc:

bltn19165eqd8

where Δe is the change in void ratio when vertical effective stress increases from σ’v1 to σ’v2.

Mercury Injection Capillary Pressure Measurement

We performed MICP measurements to characterize the pore-throat size distribution and capillary behavior of each lithofacies. Samples were first oven dried at 105°C for at least 24 hr, then measured at room temperature (20°C) using a Micrometrics AutoPore IV device. During the mercury intrusion experiment, mercury is injected at a progressively increasing pressure into the dried and evacuated sample. Detailed procedures are described in Daigle and Dugan (2014). The data were corrected for conformance following the method of Comisky et al. (2011). The mercury pressure is assumed equal to the capillary pressure in the smallest pore throats that the mercury can enter, that is, the smallest pore throats that are connected to the exterior of the sample through a pathway of mercury-filled pores. The pore-throat radius (rp) was calculated from Washburn’s equation:

bltn19165eqd9

where PHg is absolute mercury injection pressure (Pa), σHg is air–mercury interfacial tension (0.485 N/m), and θHg is the contact angle between mercury and the particle surface (130°). Mercury injection pressure (PHg) was converted to gas–water capillary pressure (Pcgw) by

bltn19165eqd10

We assumed that gas–water interfacial tension (σgw) is 0.072 N/m, and the gas–water contact angle (θgw) is 180°. The height above free water level (hfwl) was then calculated as

bltn19165eqd11

where hfwl is vertical distance from free water level to the top of reservoir (also interpreted to be the minimum hydrocarbon column height required to be accumulated for entering reservoir pores), ρw is water density (1035 kg/m3) with 3.5% salinity, and ρgas is the in situ methane density (194 kg/m3) (Starling and Savidge, 1992) at an in situ fluid pressure of 24.8 MPa and an in situ temperature of 19°C (Flemings et al., 2018d).

We determined three characteristic points on the capillary curves that we term the displacement pressure (Pd), the extrapolated displacement pressure (Pde), and the modal displacement pressure (Pmodal). The displacement pressure Pd was selected as the minimum pressure when mercury started entering the pore network during primary drainage, which corresponded, approximately, to the first data point when Sw was below 1. The extrapolated displacement pressure was determined using the model in Thomeer (1960): a hyperbola is fit to the displacement curve and used to project its horizontal asymptote for constraining the Pde and vertical asymptote for constraining the residual wetting saturation Srw (see the example of figure 1 in Thomeer, 1960). This Pde value is approximately equal to the point at which the plateau of the capillary curve is extended to 100% wetting saturation. Finally, the modal displacement pressure Pmodal (i.e., capillary pressure at modal pore-throat radius rmodal) for each sample was calculated by determining the maximum intruded volume for a given increase in capillary pressure. This is marked by a rapid increase in intruded mercury volume with a small increase in mercury pressure or the point of maximum slope on a plot of intruded mercury volume versus mercury pressure.

RESULTS

Pressure Core Analysis and Transfer System Analysis of Lithofacies

The core section from 4FB-8 (Figure 1E, yellow box) is characteristic of the sandy silt lithofacies. It has centimeter-scale layers of light (low density) and dark (higher density) that Meazell et al. (2020, this issue) infer to record ripple cross-laminations (Figure 2A). The darker (more dense) layers are inferred to be finer-grained material (Meazell et al., 2020, this issue). The average PCATS P-wave velocity in the section was 2947.7 m/s (Figure 1G), and the hydrate saturation was approximately 83% as measured from quantitative degassing experiments (Phillips et al., 2020, this issue) (Table 1).

The core section (5–20 cm) from 11FB-1 (Figure 1H, yellow box) is characteristic of the clayey silt lithofacies. It has an average P-wave velocity of 1665.8 m/s in PCATS and has large intervals with x-ray–imaged dark (high density) and structureless sediments (Figure 2B). This core section has a few sparsely distributed submillimeter planar light-color laminations, which may likely represent thin bands of coarse-grained material, which is a typical component in clayey silt lithofacies (see an example of Figure 5 in Meazell et al., 2020, this issue). Core 11FB-1 upon recovery in PCATS had low hydrate saturations (∼2%) (Table 1).

Figure 3. The Casagrande plasticity chart of core sediments. The clay fraction of each sample is color coded. The background colors record the soil classifications of the Unified Soil Classification System (USCS), which are defined and interpreted in ASTM International (2017a). Atterberg limits of our samples are compared to other gas hydrate reservoirs: India National Gas Hydrate Program Expedition 02 (NGHP-02) samples are from NGHP B and C areas, Krishna–Godavari Basin of Eastern India (Dai et al., 2018); Ulleung Basin Gas Hydrate (UBGH) samples are from Ulleung Basin, Sea of Japan (Yun et al., 2011). Characteristic samples from nonhydrate reservoir locations in the Gulf of Mexico (GOM) are also included for comparison: GOM Eugene Island (GOM-EI) is a clay from Eugene Island, GOM, and GOM-Ursa is a siltstone from Ursa Basin, GOM (Casey et al., 2019). A-Line = the line that separates silt from clay; CH = fat clay; CL = lean clay; MH = elastic silt; ML = silt; OH = organic silt or clay with high liquid limit; OL = organic silt or clay with low liquid limit; U-Line = the uppermost extent of limits for naturally occurring soils; UT-GOM2-1 = The University of Texas-Gulf of Mexico 2-1 Hydrate Pressure Coring Expedition.

Both reconstituted sandy silt lithofacies and clayey silt lithofacies materials have a homogenous structure in comparison to the intact samples (Figure 2C vs. Figure 2A and compare Figure 2D to Figure 2B). This is not surprising because we homogenized the material over each highlighted yellow-box interval in Figure 1E, H.

Index Properties

Liquid limit (wL[%]) reflects the quantity and type (e.g., fat vs. lean and organic vs. inorganic) of clay minerals present in natural sediments, combining the mechanical effects of these two attributes into a single number, providing a first and fast approximate prediction of engineering properties such as permeability (k) and compression index (Cc) at a given effective stress (Casey et al., 2013, 2019). Sandy silt lithofacies material has very low plasticity, plots above the A-Line, and is an inorganic silt when described by the Casagrande plasticity chart (Figure 3; Table 2). Clayey silt lithofacies material lies at the low- to high-plasticity boundary, lies above the A-Line, and is characterized as a lean clay (Figure 3). Sandy silt lithofacies has a small clay fraction (∼4% particle-size clay by mass; Figures 3, 4) and is dominated by silt, whereas clayey silt lithofacies has 47% clay by mass (see Figures 3, 4 and Meazell et al., 2020, in press). The clay composition is approximately the same in both cases (Meazell et al., 2020, in press). Therefore, the primary driver for this difference in liquid limit is the fraction of silt and sand present. This behavior is also reflected in Figure 3 in which samples with a higher silt or sand fraction plot to the lower left.

The PCATS porosity of sandy silt lithofacies in 4FB-8 core (Figure 2A) is approximately 0.38 using equations 1 and 2 and the parameters listed in Table 1. For sandy silt lithofacies, we assumed Sh = 80% based on Phillips et al. (2020, in press). The LWD porosity of the sandy silt is also approximately 0.38 based on equation 1 using the LWD density between 423 and 424 mbsf and parameters in Table 1. We interpret that the in situ porosity of sandy silt lithofacies in this location is approximately 0.38. The fact that the PCATS porosity and the LWD porosity are so similar implies that the volume expansion of sandy silt sample in PCATS is not significantly enough to be distinguished by PCATS CT scan and LWD log.

Figure 4. Grain size distributions on a semilog scale. The particle size fraction is expressed as a percentage of the total dry weight. Silts 4FB-4, 11FB-1 (Figure 1H, labeled with the yellow box), and 4FB-3 were analyzed using the hydrometer method (Germaine and Germaine, 2009), whereas 4FB-8 (Figure 1E, labeled with the yellow box) was analyzed by laser diffraction (Flemings et al., 2020, in press). The sand versus silt boundary is defined at 62.5 μm, and the silt versus clay boundary is defined at 2 μm. Silt 4FB-8 has a clay-size fraction of 4%, and 11FB-1 has a clay-size fraction of 47%. Sediments of 4FB-8 and 11FB-1 were used for reconstituting artificial samples of sandy silt lithofacies and clayey silt lithofacies, respectively. Inset shows sedimentological classification of the lithofacies based on the Shepard scale (Shepard, 1954). D50 = median grain diameter.

The average in situ porosity of clayey silt is estimated to be approximately 0.33 using the approach described in the Index Properties Measurements section that accounts for core expansion (equations 1–3). The LWD porosity of material in 11FB core cannot be directly measured because in this interval the LWD log was significantly affected by borehole washout (figure 15 in Boswell et al., 2012b).

Figure 5. The permeabilities (k) of reconstituted sandy silt (4FB-8) and clayey silt (11FB-1). The sandy silt k were obtained at n = 0.39, 0.34, and 0.32, respectively (orange dots). The clayey silt k were measured by the constant-rate-of-strain experiment (bluish-green dots). The porosity–k behavior of the Gulf of Mexico (GOM) Ursa siltstone is marked by the blue line, and GOM Ursa mudstone with clay from 50% to 70% are marked in the tannish-yellow zone (Reece et al., 2012). The black lines are the predicted intrinsic k using k–liquid limit (wL[%]) correlations (wL[%] = 23 for sandy silt lithofacies and wL[%] = 49.8 for clayey silt lithofacies) summarized from all mud rocks in Casey et al. (2013). σ′v = vertical effective stress.

The sandy silt lithofacies material (clay fraction = 4%; Table 2) is well sorted, with a median grain size D50 = 48 μm (Figure 4, orange symbols). In contrast, clayey silt lithofacies material (clay fraction = 47%; Table 2) is more poorly sorted, with a median grain size D50 = 2.8 μm (Figure 4, bluish-green symbols). These grain size distributions are similar to other descriptions of sandy silt lithofacies and clayey silt lithofacies discussed by Meazell et al. (2020, in press) and (Flemings et al., 2020, in press).

Permeability

The intrinsic permeability of the reconstituted sandy silt lithofacies material in the triaxial cell is 11.8 md (1.18 × 10−14 m2) at a porosity of 0.39 (Figure 5, filled circle). We also measured the intrinsic permeability of reconstituted sandy silt lithofacies material in the CRS cell at a porosity of 0.34 (σ′v = 12 MPa) and found k = 2.90 md (2.90 × 10−15 m2) and at a porosity of 0.32 (σ’v = 15 MPa) and found k = 2.02 md (2.02 × 10−15 m2). These data follow a log–linear trend:

bltn19165eqd12

Here, γ = 11.20, and β = −18.30 for reconstituted sandy silt lithofacies. We suggest that the permeability of the sandy silt lithofacies material measured at a porosity of 0.39 is an estimate of the in situ permeability of sandy silt lithofacies.

) in Casey et al. (2019). For this case, nref = 0.38, σ′vref = 1 MPa, empirical compression constant (Ccn) = 0.12, and coefficient of determination = 0.999. Ref = reference."> Figure 6. Evolution of void ratio with vertical effective stress during compression. Orange dots show the compression curve of reconstituted sandy silt (4FB-8). The black dashed line is extrapolated assuming a constant compression index (Cc = 0.21). Bluish-green dots show the compression curve of resedimented clayey silt (11FB-1). The solid black line is a fit of the clayey silt lithofacies compression curve described by the log–linear relationship between porosity (n) and vertical effective stress (σ′v) (bltn19165inf4) in Casey et al. (2019). For this case, nref = 0.38, σ′vref = 1 MPa, empirical compression constant (Ccn) = 0.12, and coefficient of determination = 0.999. Ref = reference.

The intrinsic permeability of reconstituted clayey silt lithofacies material ranges from 2.7 × 10−2 md (2.7 × 10−17 m2) to 3.84 × 10−4 md (3.84 × 10−19 m2) over a porosity range from 0.52 (0.02 MPa) to 0.31 (3.8 MPa) (Figure 5, bluish-green line). These data also follow a log–linear trend with γ = 8.38 and β = −21 (Figure 5). The intrinsic permeability of the reconstituted material at the in situ effective stress (σ′v = 3.8 MPa) is 3.84 × 10−4 md (3.84 × 10−19 m2).

Compression Behavior

The porosity (n) of reconstituted sandy silt lithofacies decreased from 0.43 to 0.38 as the loading stress increased from zero to the interpreted in situ vertical effective stress of 3.8 MPa (Figure 6). We then increased the load to 12 MPa, and the porosity declined to 0.34. In this stress range (3.8–12 MPa), the compression index Cc is 0.21 (Table 3). The stress was then unloaded to 3.8 MPa and reloaded to 15 MPa, resulting in a porosity of 0.32. After unloading to a stress less than 0.1 MPa, the final porosity was approximately 0.35 (Figure 6; Table 4).

The reconstituted clayey silt lithofacies was compressed from a porosity of 0.52 at a stress of 0.01 MPa to a porosity of 0.31 at 3.8 MPa (Figure 6), at which 3.8 MPa is interpreted to be the in situ effective stress. The recompression index Cr was approximately 0.09 (0.01–0.1 MPa) (Table 3). The compression index Cc decreased slightly from 0.36 to 0.28 during the compression loading (0.1–3.8 MPa) (Table 3). After unloading the stress from 3.8 MPa to less than 0.1 MPa, the porosity rebounded to 0.36 (Figure 6; Table 4).

Mercury Porosimetry

The two capillary curves for sandy silt are very similar, although one has a porosity, measured by the MAD technique, of 0.39 (equivalent to the in situ porosity) (Figure 7A, filled orange circles) and the other has a porosity of 0.36 because it was compressed to 15 MPa (Figure 7A, empty orange circles). In fact, the sandy silt lithofacies sample that was loaded to a higher stress (empty orange circles) has a slightly higher displacement PdeHg–air (0.10 vs. 0.087 MPa) and an identical modal pore-throat radius (5.09 vs. 5.09 μm) (Figure 7B; Table 4). We ran our mercury injection tests to a pressure of approximately 420 MPa. At this pressure, the volume of mercury we intruded into sandy silt lithofacies was approximately equal to the pore volume calculated from the measured masses and grain density (i.e., MAD porosity technique). The residual wetting saturation (Srw) of sandy silt lithofacies is 3.9%, implying that the mercury had accessed most of the pore volume (Table 4).

Figure 7. Mercury injection capillary pressure measurement of reconstituted sandy silt and clayey silt. (A) Capillary pressure curves. The wetting-phase saturation is calculated as Sw = 1 − VHg/Vpore, where VHg is the injected Hg volume and Vpore is the bulk pore volume. For sandy silt (orange curves), we used the VHg and Hg volume injected at infinite pressure (VinfHg), both of which were adopted from the best-fitted values in Thomeer (1960). The VinfHg value is nearly equal to Vpore for sandy silt lithofacies (i.e., porosity measured by mercury porosimetry [nHg] = porosity measured by moisture and density technique [nMAD]; Table 4). For clayey silt (bluish-green curves), we only used the measured VHg and the Vpore calculated from nMAD and grain density because Vpore is much larger than VHg (i.e., nHg < nMAD; Table 4). The small windows show the displacement pressure (Pd) and the extrapolated displacement pressures (Pde) for each lithofacies. (B) Incremental mercury injection volume with pore-throat radius. Values of modal pore-throat radius (rmodal), Pd, Pde, and modal displacement pressure (Pmodal) are listed in Table 4. CRS = constant rate of strain; Sh = hydrate saturation.

We explored the capillary behavior of one clayey silt sample that was compressed to only 0.1 MPa and a second one that was compressed to the inferred in situ vertical effective stress of 3.8 MPa (Figure 7A; Table 4). At high wetting-phase saturations (Sw > 80%), the capillary behavior of these two samples is very similar (Figure 7A). The extrapolated displacement pressure is more than 60 times greater than that inferred for sandy silt lithofacies. As with the sandy silt, the sample that was loaded to a higher stress (empty bluish-green circles) has a slightly higher displacement PdeHg–air (6.84 vs. 5.84 MPa) and a smaller modal pore-throat radius (0.054 vs. 0.068 μm) (Figure 7B; Table 4). Both samples have a relatively low non–wetting-phase saturation and a lower mercury (apparent) porosity (nHg = 0.30 vs. nMAD = 0.52 in a low-stressed sample and nHg = 0.28 vs. nMAD = 0.37 in a high-stressed sample; Table 4), which is because a large fraction of the pore volume was not accessed by the mercury. This discrepancy is likely caused by three physical processes: (1) before mercury intrusion, pore space shrinkage has occurred during oven drying (Diamond, 1970); (2) in mercury intrusion, the pore throats are compressed by the effective stress developed by mercury pressure and capillary resistance (Klaver et al., 2015); and (3) the capillary pressure might be insufficient to enter the smaller pores. These processes have more influence on weaker mud rocks than stronger mud rocks (Daigle et al., 2019), which may explain why the sample with high stress (σ′v = 3.8 MPa) has lower residual wetting saturation (Srw = 35%) than that of sample with high stress (σ′v = 0.1 MPa, Srw = 60%). As with the capillary curves, the pore size distributions were also very similar for samples pre- and post-CRS tests (Figure 7B). For the sandy silt, the modal pore-throat radius distributions pre- and post-CRS tests were identical, and the high stress only compressed the largest pores. For the clayey silt, high stress resulted in a compression of the modal pore size, whereas the micro–pore-throat sizes remained unchanged.

DISCUSSION

Reconstituted versus In Situ Behavior

We reconstituted sediments to gain insight into the petrophysical properties of the lithofacies that compose the hydrate reservoir at GC 955. Reconstitution eliminates sample variability, produces uniform specimens, and allows us to study the effects of composition, effective stress, and stress history on petrophysical properties through systematic experiments (Casey et al., 2016). However, the approach homogenizes any heterogeneity present in the original sample. For example, the in situ core sandy silt has thin laminae that record waxing and waning flow (Meazell et al., in press) (Figure 2A). This heterogeneity is obliterated when we homogeneously mix the bulk sediments and reconstitute the uniformly mixed material across a 3.7-cm section (Figure 2C). In contrast, the in situ core fine-grained clayey silt is fairly homogeneous (Figure 2B). Therefore, resedimentation of the clay-rich material (Figure 2D) may more successfully capture the in situ behavior of the clayey silt lithofacies. Casey et al. (2013) has shown good agreement between in situ permeabilities of intact and resedimented mudstones for both Ursa mudstones and Boston blue clay. Betts (2014) also concluded a good similarity between in situ compression and permeability behavior of intact materials and resedimented sediments.

In summary, the physical processes of forming reconstituted samples are not identical with their natural counterparts but are sufficiently analogous so that these methods can provide insight into the in situ petrophysical behavior when the sample has minimal diagenesis process in its structure (Germaine and Germaine, 2009). In this light, it should be understood just how little insight there is into these properties in hydrate reservoirs, particularly in the GOM. For example, the only estimate of the permeability of hydrate-bearing reservoirs in the GOM that we are aware of are based on percussion sidewall cores sampled from AC 818 where sediments hosting the hydrate were composed of volcanoclastic sand with grain sizes ranging from 40 to 60 μm. Boswell et al. (2009) estimated the intrinsic permeability of these sediments to range from 550 to 1500 md determined from algorithms based on the grain size of recovered sidewall cores, measured porosity, and other data.

Permeability

The intrinsic permeability of reconstituted sandy silt is 11.8 md (1.18 × 10−14 m2) at the in situ porosity of 0.38. This is our best estimate of the in situ intrinsic permeability of the sandy silt. This permeability may be a lower bound for the horizontal permeability; this is because we mixed the coarser laminae that potentially have higher permeability (Figure 2A, lighter layers) with the finer and lower-permeability laminae (Figure 2A, darker layers) to ultimately make a sample of homogenous composition (Figure 2C). In contrast, the permeability measured on the reconstituted sandy silt may be an upper bound for vertical permeability; this is because the removal of the low permeability finer laminae might increase vertical permeability.

The reconstituted clayey silt has a much lower permeability 3.84 × 10−4 md (3.84 × 10−19 m2) at a porosity of 0.31, which is the experimental porosity at the in situ effective vertical stress (3.8 MPa) (Figure 5). This permeability is an estimate of the vertical permeability of the mud rock that dominates clayey silt material. We do not consider the effect of the very thin coarse-grained layers locally present in what is mapped as clayey silt lithofacies (e.g., figure 10D in Phillips et al., in press).

The measured permeabilities are consistent with those observed for these types of lithologies. Reece et al. (2012) described a range of permeabilities for mud rocks from the shallow GOM Ursa Basin that have a clay-size fraction from 50% to 70% (Figure 5, tannish-yellow region). The reconstituted clayey silt lithofacies has a slightly lower clay content (47%) and plots on the upper end of the relationship demonstrated by Reece et al. (2012). The reconstituted clayey silt lithofacies has a slightly lower permeability than we would predict by correlation of the liquid limit to permeability described in Casey et al. (2013) (Figure 5, clayey silt-liquid limit prediction). The sandy silt lithofacies (∼4% clay) has a grain size at the boundary of silt and sandstone (Figure 4). Not surprisingly, it has a greater permeability than the siltstone at Ursa, which has a much higher clay fraction (32%).

The log of the intrinsic permeability is linearly correlated to the log of the median grain size (D50) for our two samples and for previous measurements in hydrate reservoirs (Figure 8). The intrinsic permeability of our reconstituted sandy silts is similar to those measured in hydrate reservoirs with similar grain size (D50 = 48 μm), whereas the clayey silt has the finest grain size (D50 = 2.8 μm) and lowest intrinsic permeability among all hydrate reservoirs tested. The clayey silt is most likely not a hydrate reservoir at all because it contains little to no hydrate (see hydrate saturation of clayey silt core samples in Table 5).

; coefficient of determination = 0.89). The dashed lines define the silt versus clay boundary and silt versus sand boundary (also see Figure 4). The bluish-green color refers to clayey silt, and orange refers to silt, sandy silt, silty sand, or sand. The details (i.e., lithofacies, sample handling, effective stress at measurement, consolidation condition, porosity at measurement, intrinsic permeability, permeability measurement method, D50, and grain size method) of each data set are listed in Table 5. The figure is modified from Yoneda et al. (2019a). AT = core label; NGHP-02 = India National Gas Hydrate Program Expedition 02; UT-GOM2-1 = The University of Texas-Gulf of Mexico 2-1 Hydrate Pressure Coring Expedition."> Figure 8. The vertical intrinsic permeabilities (k) of reconstituted sediments correlate with the median grain sizes (D50) of sediments and are compared to k measurements on hydrate-dissociated cores from eastern Nankai Trough, Japan (Konno et al., 2015; Yoneda et al., 2015, 2017, 2018a, b) and from Krishna–Godavari Basin, Indian Ocean (Yoneda et al., 2019a). The solid black line is a fit of the log–log linear relationship between k and D50 (bltn19165inf5; coefficient of determination = 0.89). The dashed lines define the silt versus clay boundary and silt versus sand boundary (also see Figure 4). The bluish-green color refers to clayey silt, and orange refers to silt, sandy silt, silty sand, or sand. The details (i.e., lithofacies, sample handling, effective stress at measurement, consolidation condition, porosity at measurement, intrinsic permeability, permeability measurement method, D50, and grain size method) of each data set are listed in Table 5. The figure is modified from Yoneda et al. (2019a). AT = core label; NGHP-02 = India National Gas Hydrate Program Expedition 02; UT-GOM2-1 = The University of Texas-Gulf of Mexico 2-1 Hydrate Pressure Coring Expedition.

The intrinsic permeabilities of other hydrate reservoirs discussed below were measured by steady-state flow tests (Konno et al., 2015; Dai et al., 2018; Yoneda et al., 2019a) or consolidation tests (i.e., consolidation theory) (Yoneda et al., 2015, 2017, 2018a, b) either after hydrate dissociation of hydrate-bearing core in the core that did not have hydrate originally (Table 5). Dai et al. (2018) also measured the intrinsic permeability (k = 2.7 md) of a hydrate-free core (NGHP-02-23C-10P) that has a very similar mean grain size (D50 = 44.1 μm) to our sandy silt lithofacies material (D50 = 48 μm).

We arranged the reservoir quality into five ranks based on the kD50 relationship in Figure 8. First, the sample core (i.e., NGHP-02-9B-35P in Table 5) that contains coarse-grained sands and gravel (see figure 13 in T. S. Collett et al., 2019) has the highest reservoir quality (Figure 8, red box A). Cores (e.g., AT1-C 8P, 13P, and 20P) that are sandy siltstones with no bedding of different characteristics have the second-highest intrinsic permeability (Figure 8, red box B). According to Konno et al. (2015), these represent the highest-quality reservoir at the Nankai location. Third, in red box C of Figure 8, most NGHP-02 samples are sandy silts from site NGHP-02-16B (Table 5), which is considered a moderate quality with a high degree of fine grains (Boswell et al., 2018a). Fourth, in the location of area B from NGHP-02 off the eastern coast of India, more than 60% of the hydrate-bearing cores recovered have a similar mean grain size and intrinsic permeability to our sandy silt lithofacies material (Figure 8, red box C) (Yoneda et al., 2019a). The hydrate-bearing sediment samples with finer median grain size (17–30 μm) showing much lower intrinsic permeabilities (6.4 × 10−4 to 0.63 md) are grouped as low reservoir quality (Figure 8, red box D). Fifth, the lowest-quality reservoir samples grouped in red box E in Figure 8 are clayey silts, which are deemed a non–hydrate-reservoir lithofacies.

Compression Behavior

The reconstituted sandy silt has a low initial porosity of 0.43 at low effective stress (0.01 MPa) and does not compress significantly with increasing stress (Figure 6, red line). In contrast, the reconstituted clayey silt has a higher initial porosity (n = 0.52) at low effective stress and then compresses abruptly with stress (Figure 6, bluish-green line). This behavior has been repeatedly described in uniaxial compression tests of silt and clay mixtures (Burland, 1990; Y. Yang and Aplin, 2004; Reece et al., 2013; Casey et al., 2019) and is also observed in NGHP hydrate-bearing and seal cores (Dai et al., 2018). The compression behavior of clayey silt lithofacies is of a concave upward form on a void ratio versus log stress plot (Figure 6); this behavior is common in GOM mud rocks (Long et al., 2011) and is commonly described with a power-law regression (Butterfield, 1979; Long et al., 2011; Casey et al., 2019) (Figure 6, solid black line). In contrast, the compression curve of sandy silt lithofacies has a concave downward form with increasing effective stress, which is interpreted to be caused by particle sliding, rolling at low stresses, and grain-to-grain breakdown at high stresses (Pestana and Whittle, 1999). We assume that the curve is linear at high stress to extrapolate the compression behavior (Figure 6, dashed line). The compression parameters as a function of stress are listed in Table 3. These values are compared to a correlation of liquid limit (wL[%]) to Cc proposed by Casey et al. (2019) (Figure 9), and the agreement is remarkable. However, this wLCc-σ′v correlation does not apply to the NGHP-02 hydrate-bearing sediments in Dai et al. (2018) (Figure 9), implying that hydrate in pores has possibly changed the stiffness of the sediments.

Figure 9. Compression indices (Cc) as a function of liquid limit (wL(%)) and vertical effective stress (σv) (MPa). The lines are the summarized CcwL(%) correlations of all mud rocks in Casey et al. (2019). Compression parameters of The University of Texas-Gulf of Mexico 2-1 (UT-GOM2-1) Hydrate Pressure Coring Expedition samples are present in Table 3. The Cc values of India National Gas Hydrate Program Expedition 02 (NGHP-02) are approximately estimated from Figure 4 in Dai et al. (2018). Symbols represent different stress levels (triangle: 0.1–1 MPa; circle: 1–10 MPa; square: 10–100 MPa).

At the in situ effective stress of 3.8 MPa, the reconstituted porosities for sandy silt lithofacies is 0.38 (e = 0.61) (Figure 6; Table 6). The reconstituted sandy silt lithofacies porosity (Figure 6) is equal to the LWD porosity and the PCATS porosity (n = 0.38) (Table 1). The reconstituted clayey silt lithofacies at the in situ effective stress has a porosity of 0.31, which agrees with the average in situ porosity (n = 0.33) of clayey silt calibrated from PCATS porosity using the approach in the Index Properties Measurements section (Table 1).

Capillary Behavior

The measured capillary pressures of reconstituted lithofacies materials are reasonably constrained by those measured for the same lithologies. For instance, Shew et al. (1994) reported that siltstone plug retrieved from the OCS-G-6884 well in the Viosca Knoll 780 area of northeastern GOM, which was recovered from a deeper depth (∼1000 mbsf) than our sandy silt sediment, has a low porosity (nHg = 0.29 vs. nHg = 0.40) and a relatively high extrapolated displacement pressure (Pde = 0.175 MPa vs. Pde = 0.087 MPa). The clay-rich shale plug, from interbedded core interval (Shew et al., 1994), has a much higher extrapolated displacement pressure (Pde = 19 MPa vs. Pde = 6.84 MPa) than our clayey silt material because of its lower porosity (nHg = 0.24 vs. nHg = 0.31).

The capillary pressures for clayey silt are approximately 100 times greater than those of sandy silt over wetting-phase saturations (Sw) that range from 20% to 100% (Figure 7A). These capillary pressures are also expressed as an equivalent pore-throat size (Figure 7A, right axis; equation 5). We wish to determine the capillary pressure at which a connected filament of nonwetting fluid traverses the sample. This is the point at which hydrocarbons (gas phase) begin to migrate through the rock and is commonly termed the percolation pressure. Katz and Thompson (1986, 1987) interpret the percolation pressure to occur when the curvature of the capillary curve changes from concave downward to concave upward (Pmodal; Figure 7A), which is also the pressure at which the maximum injected pore volume occurs for a given increase in capillary pressure (Figure 7B). The Pmodal can be expressed as a pore-throat radius (equation 5), which we term rmodal. An alternative approach is to assume the percolation occurs at or near the displacement pressure (Pd), which is commonly inferred to occur at a non–wetting-phase saturation of approximately 10% (Schowalter, 1979). Thomeer (1960) suggested the extrapolated displacement pressure Pde as the most accurate value, which was modeled from a hyperbola fitting to the displacement curve (see the Mercury Injection Capillary Pressure Measurement section). We view Pmodal as an upper bound and Pde as a lower bound for percolation. With this approach, we find that sandy silt lithofacies, with no hydrate present, can trap a column of methane gas between a height of 2.44 and 3.86 m before percolation (Figure 7A, right axis; Table 4), whereas the clayey silt will seal a column of gas between 191.80 and 366.95 m (Table 4). At the observed hydrate concentration of 90% in the sandy silt, a column of gas equal to 63.5 m can be trapped. This supports the interpretation that hydrate in permeable reservoirs can self-create a significant seal for trapping free gas beneath it (Liu and Flemings, 2007; You and Flemings, 2018).

The capillary curves (Figure 7) provide insight into why the saturation in sandy silt is high (Sh = ∼80%–90%) although there is little or no hydrate in the clayey silt (Table 1) (Flemings et al., 2020, this issue; Phillips et al., 2020, this issue). Methane solubility is the dissolved methane concentration in liquid water when dissolved methane and methane hydrate are at thermodynamic equilibrium. According to Henry et al. (1999), methane solubility increases with temperature and decreases with pressure and salinity. The hydrate-water capillary pressure increases the methane hydrate solubility in pores and is described by the Gibbs–Thomson equation (Clennell et al., 1999; Henry et al., 1999; Liu and Flemings, 2011; You et al., 2019). At the observed 90% hydrate saturation in sandy silt lithofacies (Figure 7A, vertical dashed line), the methane solubility is defined by the smallest pores filled with hydrate (e.g., Figure 10, red filled circle, where Sw = 10% and dp = ∼0.35μm). This solubility is less than that necessary to form hydrate in the very largest pores of clayey silt lithofacies (e.g., Figure 10, bluish-green filled circle, where Sw = 100% and dp = 0.18 μm). Thus, at the observed methane concentration, it is impossible for hydrates to form in clayey silt lithofacies.

Figure 10. The liquid + gas (L + G) and L + hydrate (H) solubilities (black lines) in fine pores (pore-throat diameter [dp] = 17.28, 0.35, 0.18, 0.07, 0.06 μm). At any depth in hydrate-bearing sediment zone (light green zone), the hydrate solubility in fine pores of sandy silt lithofacies (orange, filled circle; wetting-phase saturation [Sw] = 10%, dp = 0.35 μm) is always less than that in the largest pore of clayey silt lithofacies (bluish-green empty circle; Sw = 100%, dp = 0.18 μm). Calculations of L + H and L + G solubilities are described in Liu and Flemings (2011). GC 955 = Green Canyon Block 955; mbsf = meters below seafloor; mbsl = meters below sea level.

We envision two possible models by which hydrate formed to the current state. If the methane source is biogenic degradation of organic matter within clayey silt (Malinverno et al., 2018), then the dissolved methane concentration never builds enough to deposit hydrate within clayey silt; instead, there is diffusion of methane from clayey silt lithofacies to sandy silt lithofacies (Figure 10, path 1 in inset). Alternatively, free gas from below invades sandy silt because it has a lower percolation pressure than clayey silt and subsequently forms hydrate. In this case, even given very high gas pressures within sandy silt lithofacies, there will not be percolation into clayey silt. The dissolved gas in the sandy silt will diffuse to clayey silt lithofacies, gradually increasing the concentration in clayey silt lithofacies. However, it will never exceed the hydrate solubility in the pores to form hydrate in clayey silt lithofacies (Figure 10, path 2 in inset).

Implications for Production and Reservoir Simulation

We hope that studies like this will inform reservoir simulation models with grid cells much larger than individual beds of sandy silt and clayey silt. Simulation models require both an understanding of intrinsic permeability and relative permeability. Thus, although the simulations will be made with multiple phases present, an understanding of the intrinsic permeability is a critical input. We predict the bulk horizontal and vertical permeability for a formation composed of interbedded sandy silt and clayey silt. These upscaled horizontal and vertical intrinsic permeabilities depend on the fraction of each lithofacies present, or net-to-gross ratio, which is the ratio of the thickness of sandy silt lithofacies (Lss) to the total thickness (Lss + Lcs), with Lcs referring to the total thickness of the clayey silt lithofacies. When fluid flows in series through a combination of beds (e.g., sandy silt lithofacies vs. clayey silt lithofacies), the vertical permeability is weighted by the low permeability (harmonic-averaged permeability). In contrast, when flow occurs parallel to the bedding, the horizontal permeability is weighted by high permeability (arithmetic-averaged permeability). The net-to-gross ratio ranges from 59% to 97%, and the average net-to-gross ratio over the entire interval is 73% in the hydrate reservoir interval at GC 955 (Meazell et al., 2020, this issue). Thus, the upscaled intrinsic horizontal permeability for the entire reservoir interval is 8.6 md (8.6 × 10−15 m2), and the upscaled intrinsic vertical permeability is 1.4 × 10−3 md (1.4 × 10−18 m2) (Figure 11).

Figure 11. Predicted upscaled intrinsic permeabilities under in situ effective stress at different net-to-gross ratios (NG). Black solid curve: upscaled horizontal intrinsic permeability (kH); black dashed curve: upscaled vertical intrinsic permeability (kV). Blue vertical dashed lines mark the NG range (59%–97%) of the Green Canyon Block 955 hydrate reservoir, which has an average NG of 73%. Lcs = thickness of clayey silt lithofacies; Lss = thickness of sandy silt lithofacies.

We also consider the permeability reduction caused by the reservoir compaction, which results from drawdown of the reservoir pressure and hence increases in the effective stress during production. The in situ hydrostatic pore pressure is 24.8 MPa (Flemings et al., 2018d). We assume a minimum drawdown pressure to be that necessary to dissociate the hydrate. The hydrate phase boundary pressure is 20.2 MPa for brine and 16.4 MPa for fresh water, assuming the in situ temperature (19°C) is maintained. We chose 18.6 MPa as an approximate pore pressure at which the hydrates could be dissociated. In this case, pressure is drawn down by 6.2 MPa (i.e., from 24.8 to 18.6 MPa), which increases the effective stress from 3.8 to 10 MPa (Table 6). As an extreme example, we also consider the case in which the pressure is drawn down to atmospheric pressure (i.e., from 24.8 to 0 MPa). In this case, the effective stress increases to 28.6 MPa (Table 6).

When the reservoir pressure is drawn down to 18.6 MPa, the porosity of reconstituted sandy silt lithofacies declines to 0.35, whereas that of clayey silt lithofacies declines to 0.26 (Table 6). At this condition, the intrinsic permeability of the reconstituted material is 4.15 md (4.15 × 10−15 m2) and 1.51 × 10−4 md (1.51 × 10−19 m2) for sandy silt lithofacies and clayey silt lithofacies, respectively (Table 6). When the reservoir is drawn down to atmospheric pressure, the porosity of sandy silt lithofacies declines to 0.3, whereas that of clayey silt lithofacies declines to 0.21 (Figure 6, black dashed curve vs. black solid curve). The resultant permeability for sandy silt lithofacies is 1.15 md (1.15 × 10−15 m2) while that of clayey silt lithofacies is 5.24 × 10−5 md (5.24 × 10−20 m2) (Table 6).

Finally, we predict the upscaled vertical volumetric strains of the reservoir with interbedded lithofacies during the gas production considering the above assumed end-member production scenarios. For the first scenario, the vertical strain of sandy silt lithofacies and clayey silt lithofacies is 4.6% and 6.8%, respectively (Table 6). Given a net-to-gross ratio of 0.73, the upscaled static-state vertical volume strain of the reservoir is 5.2% (Figure 12). For the second scenario, if the reservoir is depressurized to atmospheric pressure, the vertical strain of sandy silt lithofacies and clayey silt lithofacies is 11.4% and 13.2%, respectively (Table 6). The upscaled vertical strain at a net-to-gross ratio of 0.73 is 11.9% (Figure 12). Thus, the static-state vertical volume compaction of the reservoir during gas production at GC 955 is expected to be between 5% and 12%.

Figure 12. Predicted upscaled net vertical volumetric strain of the reservoir during gas production. The strain is estimated at five vertical effective stress (σ′v) levels (black solid lines). At σ′v = 5 MPa, the hydrate is still within the hydrate stability zone. At σ′v = 10 MPa, the hydrate is barely out of hydrate stability zone for dissociation. At σ′v = 28.6 MPa, the reservoir pore pressure is assumed to be pumped down to atmospheric pressure, which is an extreme scenario for gas production. The blue vertical dashed lines mark the net-to-gross ratio range (59%–97%). Lcs = thickness of clayey silt lithofacies; Lss = thickness of sandy silt lithofacies.

CONCLUSIONS

The GC 955 hydrate reservoir is composed of interbedded sandy silt lithofacies with high hydrate saturation and clayey silt lithofacies with no hydrate saturation. Reconstituted sandy silts have an intrinsic permeability of 11.8 md (1.18 × 10−14 m2), whereas clayey silts reconstituted to the in situ effective stress are found to have a much lower intrinsic permeability of 3.84 × 10−4 md (3.84 × 10−19 m2). This intrinsic permeability is similar to the intrinsic permeabilities measured in intact cores from hydrate reservoirs of similar grain size offshore Japan (Nankai Trough) and offshore India. The reconstituted sandy silt is less compressible than the clayey silt. Through capillary analysis, the largest pores of the clayey silt are still smaller than the open pores present in the sandy silt after 90% hydrate saturation, which provides a physical explanation for why hydrates are only present in the sandy silt. We upscaled the reservoir properties to estimate the behavior of interbedded sandy silt and clayey silt. We find the upscaled intrinsic horizontal permeability for the entire reservoir interval is 8.6 md (8.6 × 10−15 m2) and the upscaled intrinsic vertical permeability is 1.4 × 10−3 md (1.4 × 10−18 m2). We estimate that during reservoir production, a maximum vertical strain of approximately 12% will result.

Reconstitution is one important method to study petrophysical properties of sediment materials. Analysis of reconstituted sediments from hydrate-bearing pressure cores provides an approach to systematically explore some petrophysical components of the hydrate reservoir. Ultimately, this study will inform reservoir simulation models with intrinsic petrophysical properties at scales of both individual lithofacies and reservoir formation.

APPENDIX: NOMENCLATURE

REFERENCES CITED

Abràmoff, M. D., P. J. Magalhães, and S. J. Ram, 2004, Image processing with ImageJ: Biophotonics International, v. 11, p. 36–42.

ASTM International, 2012, ASTM D4186/D4186M-12e1: Standard test method for one-dimensional consolidation properties of saturated cohesive soils using controlled-strain loading: West Conshohocken, Pennsylvania, ASTM International, 18 p., doi:10.1520/D4186_D4186M-12E01.

ASTM International, 2014, ASTM D854-14: Standard test methods for specific gravity of soil solids by water pycnometer: West Conshohocken, Pennsylvania, ASTM International, 8 p., doi: 10.1520/D0854-14.

ASTM International, 2015, ASTM D5856-15: Standard test method for measurement of hydraulic conductivity of porous material using a rigid-wall, compaction-mold permeameter: West Conshohocken, Pennsylvania, ASTM International, 9 p., doi:10.1520/D5856-15.

ASTM International, 2016, ASTM D5084-16a: Standard test methods for measurement of hydraulic conductivity of saturated porous materials using a flexible wall permeameter: West Conshohocken, Pennsylvania, ASTM International, 24 p., doi:10.1520/D5084-16A.

ASTM International, 2017a, ASTM D2487-17: Standard practice for classification of soils for engineering purposes (Unified Soil Classification System): West Conshohocken, Pennsylvania, ASTM International, 10 p., doi:10.1520/D2487-17.

ASTM International, 2017b, ASTM D4318-17e1: Standard test methods for liquid limit, plastic limit, and plasticity index of soils: West Conshohocken, Pennsylvania, ASTM International, 20 p., doi:10.1520/D4318-17E01.

Berman, D. R., J. T. Germaine, and C. C. Ladd, 1993, Characterization of engineering properties of Boston Blue Clay for the MIT campus: Cambridge, Massachusetts, Massachusetts Institute of Technology Department of Civil and Environmental Engineering, 146 p.

Betts, W. S., 2014, Compressibility and permeability of Gulf of Mexico mudrocks, resedimented and in-situ, Master’s thesis, The University of Texas at Austin, Austin, Texas, 122 p.

Blum, P., 1997, Physical properties handbook: A guide to the shipboard measurement of physical properties of deep-sea cores: College Station, Texas, Ocean Drilling Program Technical Note 26, 117 p., doi:10013/epic.40058.

Boswell, R., and T. S. Collett, 2011, Current perspectives on gas hydrate resources: Energy & Environmental Science, v. 4, no. 4, p. 1206–1215, doi:10.1039/C0EE00203H.

Boswell, R., T. S. Collett, M. Frye, W. Shedd, D. R. McConnell, and D. Shelander, 2012a, Subsurface gas hydrates in the northern Gulf of Mexico: Marine and Petroleum Geology, v. 34, no. 1, p. 4–30, doi:10.1016/j.marpetgeo.2011.10.003.

Boswell, R., M. Frye, D. Shelander, W. Shedd, D. R. McConnell, and A. Cook, 2012b, Architecture of gas-hydrate-bearing sands from Walker Ridge 313, Green Canyon 955, and Alaminos Canyon 21: Northern deepwater Gulf of Mexico: Marine and Petroleum Geology, v. 34, no. 1, p. 134–149, doi:10.1016/j.marpetgeo.2011.08.010.

Boswell, R., E. Myshakin, G. Moridis, Y. Konno, T. S. Collett, M. Reagan, T. Ajayi, and Y. Seol, 2018a, India National Gas Hydrate Program Expedition 02 summary of scientific results: Numerical simulation of reservoir response to depressurization: Marine and Petroleum Geology, v. 108, p. 154–166, doi:10.1016/j.marpetgeo.2018.09.026.

Boswell, R., D. Shelander, M. Lee, T. Latham, T. Collett, G. Guerin, G. Moridis, M. Reagan, and D. Goldberg, 2009, Occurrence of gas hydrate in Oligocene Frio sand: Alaminos Canyon Block 818: Northern Gulf of Mexico: Marine and Petroleum Geology, v. 26, p. 1499–1512, doi:10.1016/j.marpetgeo.2009.03.005.

Boswell, R., J. Yoneda, and W. F. Waite, 2018b, India National Gas Hydrate Program Expedition 02 summary of scientific results: Evaluation of natural gas-hydrate-bearing pressure cores: Marine and Petroleum Geology, v. 108, p. 143–153, doi:10.1016/j.marpetgeo.2018.10.020.

Brooks, R. A., and G. Di Chiro, 1976, Beam hardening in x-ray reconstructive tomography: Physics in Medicine and Biology, v. 21, no. 3, p. 390–398, doi:10.1088/0031-9155/21/3/004.

Burland, J. B., 1990, On the compressibility and shear strength of natural clays: Géotechnique, v. 40, no. 3, p. 329–378, doi:10.1680/geot.1990.40.3.329.

Butterfield, R., 1979, A natural compression law for soils (an advance on e–log p′): Géotechnique, v. 29, no. 4, p. 469–480, doi:10.1680/geot.1979.29.4.469.

Casey, B., J. T. Germaine, P. B. Flemings, and B. P. Fahy, 2015, Estimating horizontal stresses for mudrocks under one-dimensional compression: Marine and Petroleum Geology, v. 65, p. 178–186, doi:10.1016/j.marpetgeo.2015.02.001.

Casey, B., J. T. Germaine, P. B. Flemings, and B. P. Fahy, 2016, In situ stress state and strength in mudrocks: Journal of Geophysical Research. Solid Earth, v. 121, no. 8, p. 5611–5623, doi:10.1002/2016JB012855.

Casey, B., J. T. Germaine, P. B. Flemings, J. S. Reece, B. Gao, and W. Betts, 2013, Liquid limit as a predictor of mudrock permeability: Marine and Petroleum Geology, v. 44, p. 256–263, doi:10.1016/j.marpetgeo.2013.04.008.

Casey, B., J. S. Reece, and J. T. Germaine, 2019, One-dimensional normal compression laws for resedimented mudrocks: Marine and Petroleum Geology, v. 103, p. 397–403, doi:10.1016/j.marpetgeo.2019.02.023.

Clennell, M. B., M. Hovland, J. S. Booth, P. Henry, and W. J. Winters, 1999, Formation of natural gas hydrates in marine sediments: 1. Conceptual model of gas hydrate growth conditioned by host sediment properties: Journal of Geophysical Research. Solid Earth, v. 104, no. B10, p. 22985–23003, doi:10.1029/1999JB900175.

Collett, T., J.-J. Bahk, R. Baker, R. Boswell, D. Divins, M. Frye, and D. Goldberg, et al., 2015, Methane hydrates in nature—Current knowledge and challenges: Journal of Chemical & Engineering Data, v. 60, no. 2, p. 319–329, doi:10.1021/je500604h.

Collett, T. S., 2002, Energy resource potential of natural gas hydrates: AAPG Bulletin, v. 86, no. 11, p. 1971–1992, doi:10.1306/61EEDDD2-173E-11D7-8645000102C1865D.

Collett, T. S., R. Boswell, W. F. Waite, P. Kumar, S. K. Roy, K. Chopra, and S. K. Singh, et al., 2019, India National Gas Hydrate Program Expedition 02 summary of scientific results: Gas hydrate systems along the eastern continental margin of India: Marine and Petroleum Geology, v. 108, p. 39–142, doi:10.1016/j.marpetgeo.2019.05.023.

Collett, T. S., M. W. Lee, W. F. Agena, J. J. Miller, K. A. Lewis, M. V. Zyrianova, R. Boswell, and T. L. Inks, 2011, Permafrost-associated natural gas hydrate occurrences on the Alaska North Slope: Marine and Petroleum Geology, v. 28, no. 2, p. 279–294, doi:10.1016/j.marpetgeo.2009.12.001.

Collett, T. S., M. W. Lee, M. V. Zyrianova, S. A. Mrozewski, G. Guerin, A. E. Cook, and D. S. Goldberg, 2012, Gulf of Mexico Gas Hydrate Joint Industry Project Leg II logging-while-drilling data acquisition and analysis: Marine and Petroleum Geology, v. 34, no. 1, p. 41–61, doi:10.1016/j.marpetgeo.2011.08.003.

Comisky, J. T., M. Santiago, B. McCollom, A. Buddhala, and K. E. Newsham, 2011, Sample size effects on the application of mercury injection capillary pressure for determining the storage capacity of tight gas and oil shales: Canadian Unconventional Resources Conference, Calgary, Alberta, Canada, November 15–17, 2011, SPE-149432-MS, 23 p., doi:10.2118/149432-MS.

Dai, S., J. Kim, Y. Xu, W. F. Waite, J. Jang, J. Yoneda, T. S. Collett, and P. Kumar, 2018, Permeability anisotropy and relative permeability in sediments from the National Gas Hydrate Program Expedition 02, offshore India: Marine and Petroleum Geology, v. 108, p. 705–713, doi:10.1016/j.marpetgeo.2018.08.016.

Daigle, H., and B. Dugan, 2014, Data report: Permeability, consolidation, stress state, and pore system characteristics of sediments from Sites C0011, C0012, and C0018 of the Nankai Trough: Proceedings of the Integrated Ocean Drilling Program, v. 333, 23 p., doi:10.2204/iodp.proc.333.201.2014.

Daigle, H., J. S. Reece, and P. B. Flemings, 2019, Evolution of the percolation threshold in muds and mudrocks during burial: Geophysical Research Letters, v. 46, no. 14, p. 8064–8073, doi:10.1029/2019GL083723.

Diamond, S., 1970, Pore size distributions in clays: Clays and Clay Minerals, v. 18, no. 1, p. 7–23, doi:10.1346/CCMN.1970.0180103.

Di Stefano, C., V. Ferro, and S. Mirabile, 2010, Comparison between grain-size analyses using laser diffraction and sedimentation methods: Biosystems Engineering, v. 106, no. 2, p. 205–215, doi:10.1016/j.biosystemseng.2010.03.013.

Flemings, P. B., S. C. Phillips, R. Boswell, T. S. Collett, A. E. Cook, T. Dong, and M. Frye, et al., 2020, Pressure coring a Gulf of Mexico deep-water turbidite gas hydrate reservoir: Initial results from The University of Texas-Gulf of Mexico 2-1 (UT-GOM2-1) Hydrate Pressure Coring Expedition: AAPG Bulletin, v. 104, no. 9, p. 1847–1876, doi:10.1306/01062019165.

Flemings, P. B., S. C. Phillips, T. Collett, A. Cook, and R. Boswell, UT-GOM2-1 Expedition Scientists, eds., 2018a, Hole GC 955 H002, in UT-GOM2-1 Hydrate Pressure Coring Expedition: Austin, Texas, University of Texas Institute for Geophysics, 85 p., accessed February 1, 2019, http://www-udc.ig.utexas.edu/gom2/Chapter%203%20-%20H002.pdf.

Flemings, P. B., S. C. Phillips, T. Collett, A. Cook, and R. Boswell, UT-GOM2-1 Expedition Scientists, eds., 2018b, Hole GC 955 H005, in UT-GOM2-1 Hydrate Pressure Coring Expedition: Austin, Texas, University of Texas Institute for Geophysics, 164 p., accessed February 1, 2019, http://www-udc.ig.utexas.edu/gom2/Chapter%204%20-%20H005.pdf.

Flemings, P. B., S. C. Phillips, T. Collett, A. Cook, and R. Boswell, UT-GOM2-1 Expedition Scientists, eds., 2018c, Expedition methods, in UT-GOM2-1 Hydrate Pressure Coring Expedition: Austin, Texas, University of Texas Institute for Geophysics, 41 p., accessed February 1, 2019, http://www-udc.ig.utexas.edu/gom2/Chapter%202%20-%20Methods.pdf.

Flemings, P. B., S. C. Phillips, T. Collett, A. Cook, and R. Boswell, ; UT-GOM2-1 Expedition Scientists, 2018d, Expedition summary, in UT-GOM2-1 Hydrate Pressure Coring Expedition: Austin, Texas, University of Texas Institute for Geophysics, 138 p., accessed February 1, 2019, https://ig.utexas.edu/files/2018/02/1.0-UT-GOM2-1-Expedition-Summary.pdf.

Flemings, P. B., and D. M. Saffer, 2018, Pressure and stress prediction in the Nankai Accretionary Prism: A critical state soil mechanics porosity-based approach: Journal of Geophysical Research. Solid Earth, v. 123, no. 2, p. 1089–1115, doi:10.1002/2017JB015025.

Fujii, T., M. Nakamizu, Y. Tsuji, T. Namikawa, T. Okui, M. Kawasaki, K. Ochiai, M. Nishimura, and O. Takano, 2009, Methane-hydrate occurrence and saturation confirmed from core samples, eastern Nankai Trough, in T. Collett, A. Johnson, C. Knapp, and R. Boswell, eds., Natural gas hydrates—Energy resource potential and associated geologic hazards: AAPG Memoir 89, p. 385–400.

Germaine, J. T., and A. V. Germaine, 2009, Geotechnical laboratory measurements for engineers: Hoboken, New Jersey, John Wiley & Sons, 351 p., doi:10.1002/9780470548790.

Han, G., T.-H. Kwon, J. Y. Lee, and T. J. Kneafsey, 2018, Depressurization-induced fines migration in sediments containing methane hydrate: X-ray computed tomography imaging experiments: Journal of Geophysical Research. Solid Earth, v. 123, no. 4, p. 2539–2558, doi:10.1002/2017JB014988.

Henry, P., M. Thomas, and M. B. Clennell, 1999, Formation of natural gas hydrates in marine sediments: 2. Thermodynamic calculations of stability conditions in porous sediments: Journal of Geophysical Research. Solid Earth, v. 104, no. B10, p. 23005–23022, doi:10.1029/1999JB900167.

Hutchinson, D. R., D. Shelander, J. Dai, D. McConnell, W. Shedd, M. Frye, and C. D. Ruppel, et al., 2008, Site selection for DOE/JIP gas hydrate drilling in the northern Gulf of Mexico: 6th International Conference on Gas Hydrates, Vancouver, British Columbia, Canada, July 6–10, 2008, 12 p.

Jung, J. W., J. Jang, J. C. Santamarina, C. Tsouris, T. J. Phelps, and C. J. Rawn, 2012, Gas production from hydrate-bearing sediments: The role of fine particles: Energy & Fuels, v. 26, no. 1, p. 480–487, doi:10.1021/ef101651b.

Katz, A. J., and A. H. Thompson, 1986, Quantitative prediction of permeability in porous rock: Physical Review B, v. 34, p. 8179–8181, doi:10.1103/PhysRevB.34.8179.

Katz, A. J., and A. H. Thompson, 1987, Prediction of rock electrical conductivity from mercury injection measurements: Journal of Geophysical Research: Solid Earth, v. 92, p. 599–607, doi:10.1029/JB092iB01p00599.

Klaver, J., S. Hemes, M. Houben, G. Desbois, Z. Radi, and J. L. Urai, 2015, The connectivity of pore space in mudstones: Insights from high-pressure Wood’s metal injection, BIB-SEM imaging, and mercury intrusion porosimetry: Geofluids, v. 15, no. 4, p. 577–591, doi:10.1111/gfl.12128.

Konno, Y., J. Yoneda, K. Egawa, T. Ito, Y. Jin, M. Kida, K. Suzuki, T. Fujii, and J. Nagao, 2015, Permeability of sediment cores from methane hydrate deposit in the Eastern Nankai Trough: Marine and Petroleum Geology, v. 66, p. 487–495, doi:10.1016/j.marpetgeo.2015.02.020.

Ladd, R., 1978, Preparing test specimens using undercompaction: Geotechnical Testing Journal, v. 1, no. 1, p. 16–23, doi:10.1520/GTJ10364J.

Lee, M. W., and T. S. Collett, 2012, Pore- and fracture-filling gas hydrate reservoirs in the Gulf of Mexico Gas Hydrate Joint Industry Project Leg II Green Canyon 955 H well: Marine and Petroleum Geology, v. 34, no. 1, p. 62–71, doi:10.1016/j.marpetgeo.2011.08.002.

Liu, X., and P. B. Flemings, 2006, Passing gas through the hydrate stability zone at southern Hydrate Ridge, offshore Oregon: Earth and Planetary Science Letters, v. 241, no. 1–2, p. 211–226, doi:10.1016/j.epsl.2005.10.026.

Liu, X., and P. B. Flemings, 2007, Dynamic multiphase flow model of hydrate formation in marine sediments: Journal of Geophysical Research. Solid Earth, v. 112, B03101, 23 p., doi:10.1029/2005jb004227.

Liu, X., and P. B. Flemings, 2011, Capillary effects on hydrate stability in marine sediments: Journal of Geophysical Research. Solid Earth, v. 116, B07102, 24 p., doi:10.1029/2010jb008143.

Long, H., P. B. Flemings, J. T. Germaine, and D. M. Saffer, 2011, Consolidation and overpressure near the seafloor in the Ursa Basin, Deepwater Gulf of Mexico: Earth and Planetary Science Letters, v. 305, no. 1–2, p. 11–20, doi:10.1016/j.epsl.2011.02.007.

Malinverno, A., A. E. Cook, H. Daigle, and B. Oryan, 2018, Glacial cycles influence marine methane hydrate formation: Geophysical Research Letters, v. 45, no. 2, p. 724–732, doi:10.1002/2017GL075848.

McConnell, D., R. Boswell, T. Collett, M. Frye, W. Shedd, G. Guerin, and A. Cook, et al., 2010, Gulf of Mexico Gas Hydrate Joint Industry Project Leg II: Green Canyon 955 site summary: Morgantown, West Virginia, National Energy Technology Laboratory, 23 p.

McConnell, D. R., 2000, Optimizing deepwater well locations to reduce the risk of shallow-water-flow using high-resolution 2D and 3D seismic data: Offshore Technology Conference, Houston, Texas, May 1-4, 2000, OTC-11973-MS, 11 p., doi:10.4043/11973-MS.

McConnell, D. R., Z. Zhang, and R. Boswell, 2012, Review of progress in evaluating gas hydrate drilling hazards: Marine and Petroleum Geology, v. 34, no. 1, p. 209–223, doi:10.1016/j.marpetgeo.2012.02.010.

Meazell, P. K., P. B. Flemings, M. Santra, and J. E. Johnson, 2020, Sedimentology and stratigraphy of a deep-water gas hydrate reservoir in the northern Gulf of Mexico: AAPG Bulletin, v. 104, no. 9, p. 1945–1969, doi:10.1306/05212019027.

Moridis, G. J., M. B. Kowalsky, and K. Pruess, 2007, Depressurization-induced gas production from class-1 hydrate deposits: SPE Reservoir Evaluation & Engineering, v. 10, no. 5, p. 458–481, doi:10.2118/97266-PA.

Oshima, M., K. Suzuki, J. Yoneda, A. Kato, M. Kida, Y. Konno, M. Muraoka, Y. Jin, J. Nagao, and N. Tenma, 2019, Lithological properties of natural gas hydrate–bearing sediments in pressure-cores recovered from the Krishna–Godavari Basin: Marine and Petroleum Geology, v. 108, p. 439–470, doi:10.1016/j.marpetgeo.2019.01.015.

Pestana, J. M., and A. J. Whittle, 1999, Formulation of a unified constitutive model for clays and sands: International Journal for Numerical and Analytical Methods in Geomechanics, v. 23, no. 12, p. 1215–1243, doi:10.1002/(SICI)1096-9853(199910)23:12<1215::AID-NAG29>3.0.CO;2-F.

Phillips, S. C., P. B. Flemings, M. E. Holland, P. J. Schultheiss, W. F. Waite, J. Jang, E. G. Petrou, and H. Hammon, 2020, High concentration methane hydrate in a silt reservoir from the deep-water Gulf of Mexico: AAPG Bulletin, v. 104, no. 9, p. 1971–1995, doi:10.1306/01062018280.

Reece, J. S., P. B. Flemings, B. Dugan, H. Long, and J. T. Germaine, 2012, Permeability-porosity relationships of shallow mudstones in the Ursa Basin, northern deepwater Gulf of Mexico: Journal of Geophysical Research. Solid Earth, v. 117, B12102, 13 p., doi:10.1029/2012JB009438.

Reece, J. S., P. B. Flemings, and J. T. Germaine, 2013, Data report: Permeability, compressibility, and microstructure of resedimented mudstone from IODP Expedition 322, Site C0011: Proceedings of the Integrated Ocean Drilling Program, v. 322, 26 p., doi:10.2204/iodp.proc.322.205.2013.

Ryu, B.-J., T. S. Collett, M. Riedel, G. Y. Kim, J.-H. Chun, J.-J. Bahk, J. Y. Lee, J.-H. Kim, and D.-G. Yoo, 2013, Scientific results of the Second Gas Hydrate Drilling Expedition in the Ulleung Basin (UBGH2): Marine and Petroleum Geology, v. 47, p. 1–20, doi:10.1016/j.marpetgeo.2013.07.007.

Santagata, M., and J. T. Germaine, 2005, Effect of OCR on sampling disturbance of cohesive soils and evaluation of laboratory reconsolidation procedures: Canadian Geotechnical Journal, v. 42, no. 2, p. 459–474, doi:10.1139/t04-104.

Santagata, M., and Y. I. Kang, 2007, Effects of geologic time on the initial stiffness of clays: Engineering Geology, v. 89, no. 1–2, p. 98–111, doi:10.1016/j.enggeo.2006.09.018.

Santamarina, J. C., S. Dai, M. Terzariol, J. Jang, W. F. Waite, W. J. Winters, and J. Nagao, et al., 2015, Hydro-bio-geomechanical properties of hydrate-bearing sediments from Nankai Trough: Marine and Petroleum Geology, v. 66, p. 434–450, doi:10.1016/j.marpetgeo.2015.02.033.

Santra, M., P. B. Flemings, E. Scott, and P. K. Meazell, 2020, Evolution of gas hydrate–bearing deep-water channel-levee system in abyssal Gulf of Mexico: Levee growth and deformation: AAPG Bulletin, v. 104, no. 9, p. 1921–1944, doi:10.1306/04251918177.

Schneider, J., P. B. Flemings, R. J. Day-Stirrat, and J. T. Germaine, 2011, Insights into pore-scale controls on mudstone permeability through resedimentation experiments: Geology, v. 39, no. 11, p. 1011–1014, doi:10.1130/G32475.1.

Schowalter, T. T., 1979, Mechanics of secondary hydrocarbon migration and entrapment: AAPG Bulletin, v. 63, no. 5, p. 723–760, doi:10.1306%2F2F9182CA-16CE-11D7-8645000102C1865D.

Schultheiss, P. J., M. Holland, J. Roberts, Q. Huggett, M. Druce, and P. Fox, 2011, PCATS: Pressure core analysis and transfer system: Proceedings of the 7th International Conference on Gas Hydrates, Edinburgh, Scotland, United Kingdom, July 17–21, 2011, 10 p.

Sheahan, T. C., 1991, An experimental study of the time-dependent undrained shear behavior of resedimented clay using automated stress path triaxial equipment, Sc.D. thesis, Massachusetts Institute of Technology, Cambridge, Massachusetts, 952 p.

Shepard, F. P., 1954, Nomenclature based on sand-silt-clay ratios: Journal of Sedimentary Research, v. 24, p. 151–158, doi:10.1306/d4269774-2b26-11d7-8648000102c1865d.

Shew, R. D., D. R. Rollins, G. M. Tiller, C. J. Hackbarth, and C. D. White, 1994, Characterization and modeling of thin-bedded turbidite deposits from the Gulf of Mexico using detailed subsurface and analog data,in P. Weimer, A. H. Bouma, and B. F. Perkins, eds., Submarine fans and turbidite systems—Sequence stratigraphy, reservoir architecture, and production characteristics: Tulsa, Oklahoma, Gulf Coast Section SEPM Annual Research Conference 15, p. 327–334.

Sloan, E. D., Jr., and C. A. Koh, 2007, Clathrate hydrates of natural gases, 3rd ed.: Boca Raton, Florida, CRC Press, 752 p., doi:10.1201/9781420008494.

Starling, K. E., and J. L. Savidge, 1992, Compressibility factors of natural gas and other related hydrocarbon gases, 2nd ed.: Arlington, Virginia, American Gas Association Transmission Measurement Committee Report 8, 205 p.

Thomas, C., S. C. Phillips, P. B. Flemings, M. Santra, H. Hammon, T. S. Collett, and A. E. Cook, et al., 2020, Pressure-coring operations during Expedition UT-GOM2-1 in Green Canyon Block 955, northern Gulf of Mexico: AAPG Bulletin, v. 104, no. 9, p. 1877–1901, doi:10.1306/02262019036.

Thomeer, J. H. M., 1960, Introduction of a pore geometrical factor defined by the capillary pressure curve: Journal of Petroleum Technology, v. 12, no. 3, p. 73–77, doi:10.2118/1324-G.

Torres, M. E., A. M. Tréhu, N. Cespedes, M. Kastner, U. G. Wortmann, J.-H. Kim, and P. Long, et al., 2008, Methane hydrate formation in turbidite sediments of northern Cascadia, IODP Expedition 311: Earth and Planetary Science Letters, v. 271, no. 1–4, p. 170–180, doi:10.1016/j.epsl.2008.03.061.

Yang, S., Y. Lei, J. Liang, M. Holland, P. Schultheiss, J. Lu, and J. Wei, 2017, Concentrated gas hydrate in the Shenhu area, South China Sea: Results from drilling expeditions GMGS3 & GMGS4: 9th International Conference on Gas Hydrates, Denver, Colorado, June 25–30 2017, paper no. 105, 16 p.

Yang, Y., and A. C. Aplin, 2004, Definition and practical application of mudstone porosity-effective stress relationships: Petroleum Geoscience, v. 10, no. 2, p. 153–162, doi:10.1144/1354-079302-567.

Yoneda, J., A. Masui, Y. Konno, Y. Jin, K. Egawa, M. Kida, T. Ito, J. Nagao, and N. Tenma, 2015, Mechanical properties of hydrate-bearing turbidite reservoir in the first gas production test site of the Eastern Nankai Trough: Marine and Petroleum Geology, v. 66, p. 471–486, doi:10.1016/j.marpetgeo.2015.02.029.

Yoneda, J., A. Masui, Y. Konno, Y. Jin, K. Egawa, M. Kida, T. Ito, J. Nagao, and N. Tenma, 2018a, Corrigendum to mechanical properties of hydrate-bearing turbidite reservoir in the first gas production test site of the Eastern Nankai Trough JMPG 66 (Sep 2015); Pages 471-486: Marine and Petroleum Geology, v. 91, p. 659, doi:10.1016/j.marpetgeo.2018.02.025.

Yoneda, J., A. Masui, Y. Konno, Y. Jin, M. Kida, J. Katagiri, J. Nagao, and N. Tenma, 2017, Pressure-core-based reservoir characterization for geomechanics: Insights from gas hydrate drilling during 2012–2013 at the eastern Nankai Trough: Marine and Petroleum Geology, v. 86, p. 1–16, doi:10.1016/j.marpetgeo.2017.05.024.

Yoneda, J., A. Masui, Y. Konno, Y. Jin, M. Kida, J. Katagiri, J. Nagao, and N. Tenma, 2018b, Corrigendum to pressure-core-based reservoir characterization for geomechanics: Insights from gas hydrate drilling during 2012e2013 at the eastern Nankai Trough JMPG 86 (Sep 2017) pages 1-16: Marine and Petroleum Geology, v. 91, p. 658, doi:10.1016/j.marpetgeo.2018.02.026.

Yoneda, J., M. Oshima, M. Kida, A. Kato, Y. Konno, Y. Jin, J. Jang, W. F. Waite, P. Kumar, and N. Tenma, 2019a, Permeability variation and anisotropy of gas hydrate-bearing pressure-core sediments recovered from the Krishna–Godavari Basin, offshore India: Marine and Petroleum Geology, v. 108, p. 524–536, doi:10.1016/j.marpetgeo.2018.07.006.

Yoneda, J., M. Oshima, M. Kida, A. Kato, Y. Konno, Y. Jin, J. Jang, W. F. Waite, P. Kumar, and N. Tenma, 2019b, Pressure core based onshore laboratory analysis on mechanical properties of hydrate-bearing sediments recovered during India’s National Gas Hydrate Program Expedition (NGHP) 02: Marine and Petroleum Geology, v. 108, p. 482–501, doi:10.1016/j.marpetgeo.2018.09.005.

You, K., and P. B. Flemings, 2018, Methane hydrate formation in thick sandstones by free gas flow: Journal of Geophysical Research. Solid Earth, v. 123, no. 6, p. 4582–4600, doi:10.1029/2018JB015683.

You, K., P. B. Flemings, A. Malinverno, T. S. Collett, and K. Darnell, 2019, Mechanisms of methane hydrate formation in geological systems: Reviews of Geophysics, doi:10.1029/2018rg000638.

Yun, T. S., C. Lee, J.-S. Lee, J. J. Bahk, and J. C. Santamarina, 2011, A pressure core based characterization of hydrate-bearing sediments in the Ulleung Basin, Sea of Japan (East Sea): Journal of Geophysical Research. Solid Earth, v. 116, B02204, 12 p., doi:10.1029/2010jb007468.

ACKNOWLEDGMENTS

This work is the result of support provided by the US Department of Energy (DOE) under Contract No. DE-FE0023919. This support is gratefully acknowledged. We thank John Germaine from Tufts University for his help and discussion of the experiments and Glen Baum from The University of Texas at Austin for running the mercury injection capillary pressure tests. We thank Aaron Price (The University of Texas at Austin) for the assistance with computed tomography images and data processing. We thank Athma Bhandari (The University of Texas at Austin), William Waite (US Geological Survey [USGS]), Junbong Jang (USGS), Liang Lei (DOE National Energy Technology Laboratory [DOE-NETL]), Sheng Dai (Georgia Institute of Technology), and Jun Yoneda (National Institute of Advanced Industrial Science and Technology, Japan) for the discussion of capillary behaviors in this work. We also thank Jun Yoneda for his help in organizing the data set in Table 5. We also thank the very helpful edits and comments from Ray Boswell (DOE-NETL) and Timothy Collett (USGS) on this paper. We thank the support from The University of Texas at Austin Jackson School of Geosciences, and Institute for Geophysics. Reviews by two anonymous reviewers helped strengthen this paper.

You may also be interested in ...