As noted in the introduction, the remaining portion of this paper deals with the assessment of the geologic, engineering, economic, and political factors that control the ultimate resource potential of gas hydrates.

Where Do Gas Hydrates Occur?

The geologic occurrence of gas hydrates has been known since the mid-1960s, when gas-hydrate accumulations were discovered in Russia (reviewed by Makogon, 1981).  As discussed in the previous section of this paper, gas hydrates are widespread in permafrost regions and beneath the sea in sediment of outer continental margins (reviewed by Kvenvolden, 1993).  Cold surface temperatures at high latitudes on earth are conducive to the development of onshore permafrost and gas hydrate in the subsurface.  Onshore gas hydrates (Figure 4) are known to be present in the West Siberian Basin (Makogon et al., 1972) and are believed to occur in other permafrost areas of northern Russia, including the Timan-Pechora province, the eastern Siberian Craton, and the northeastern Siberia and Kamchatka areas (Cherskiy et al., 1985).  Permafrost-associated gas hydrates are also present in the North American Arctic.  Direct evidence for gas hydrates on the North Slope of Alaska comes from a core-test in the Northwest Eileen State-2 well, and indirect evidence comes from drilling and open-hole industry well logs that suggest the presence of numerous gas hydrate layers in the area of the Prudhoe Bay and Kuparuk River oil fields (as discussed previously in this paper).  Well-log responses attributed to the presence of gas hydrates have been obtained in about one-fifth of the wells drilled in the Mackenzie Delta, and more than half of the wells in the Arctic Islands are inferred to contain gas hydrates (Judge and Majorowicz, 1992).  The recently completed Mallik 2L-38 gas hydrate research well, confirmed the presence of a relatively thick, highly concentrated, gas hydrate accumulation on Richards Island in the outer portion of the Mackenzie River Delta (Dallimore et al., 1999).

The presence of gas hydrates in offshore continental margins (Figure 4) has been inferred mainly from anomalous seismic reflectors (i.e., BSRs) that coincide with the predicted phase boundary at the base of the gas-hydrate stability zone.  Gas hydrates have been recovered in gravity cores within 10 m of the sea floor in sediment of the Gulf of Mexico (Brooks et al., 1986), the offshore portion of the Eel River Basin of California (Brooks et al., 1991), the Black Sea (Yefremova and Zhizhchenko, 1974), the Caspian Sea (Ginsburg et al., 1992), and the Sea of Okhotsk (Ginsburg et al., 1993).  Also, gas hydrates have been recovered at greater sub-bottom depths during research coring along the southeastern coast of the United States on the Blake Ridge (Kvenvolden and Barnard, 1983; Shipboard Scientific Party, 1996), in the Gulf of Mexico (Shipboard Scientific Party, 1986), in the Cascadia Basin near Oregon (Shipboard Scientific Party, 1994), the Middle America Trench (Kvenvolden and McDonald, 1985), offshore Peru (Kvenvolden and Kastner, 1990), and on both the eastern and western margins of Japan (Shipboard Scientific Party, 1990, 1991).

How Do Gas Hydrates Occur in Nature?

Little is known about the nature of gas hydrate reservoirs.  For example, do hydrates occur as pore-filling constituents or are they only found in massive form (Figure 15).  Information about the nature and texture of reservoired gas hydrates is needed to accurately determine the amount of gas hydrate and associated gas in a given gas hydrate accumulation.  The textural nature of gas hydrate in the reservoir also controls the production potential and characteristics of a gas hydrate accumulation.  The physical and chemical conditions that result in different forms (disseminated, nodular, layered, massive) and distributions (uniform or heterogeneous) of gas hydrates are not understood (reviewed by Sloan, 1998).  It is necessary, therefore, to systematically review descriptions of known gas hydrate occurrences and evaluate existing gas hydrate reservoir models at both microscopic and macroscopic scales in order to assess the nature of gas-hydrate-bearing reservoirs.  This section of the paper begins with a review of published gas hydrate sample descriptions from both marine and permafrost environments, which is followed by an interpretive discussion of existing and proposed microscopic and macroscopic gas hydrate reservoir models.

Recovered Gas Hydrate Samples

For this review of the nature of gas hydrate occurrences, I have relied extensively on the offshore gas hydrate sample database recently published by Booth et al. (1996).  In this database, Booth et al. (1996) systematically review and describe more than 90 marine gas hydrate samples recovered from 15 different geologic regions.  The individual descriptions of the gas hydrate occurrences include information on the number of recovered samples, physiographic province, tectonic setting, geographic position, water depth, sub-sea-floor depth of the recovered sample, geothermal gradient and temperature conditions, depth to the base of the gas hydrate stability zone, presence of a bottom simulating seismic reflector, thickness of the gas-hydrate-bearing sedimentary interval, thickness and size of pure gas hydrate layers and grains, habit or mode of occurrence, host sediment lithologic description, and the origin of the included gas.

In general, most of the recovered gas hydrate samples consist of individual grains or particles, which are often described as inclusions or disseminated in the sedimentary section (Figure 15).  Gas hydrates also occur as, what has been described as, a cement, nodules, or as laminae and veins, which tend to be characterized by dimensions of a few centimeters or less.  In several cases, thick, pure gas hydrate layers measuring as much as 3- to 4-m-thick have been sampled (DSDP Site 570; Shipboard Scientific Party, 1985).  In both marine and terrestrial permafrost environments, the thickness of identified gas-hydrate-bearing sedimentary sections varies from a few centimeters to as much as 30 m (Collett, 1993; Booth et al., 1996; Dallimore et al., 1999).  Most pure gas hydrate laminae and layers, however, are often characterized by thicknesses of millimeters to centimeters (Booth et al., 1996; Dallimore and Collett, 1995; Dallimore et al., 1999).  Booth et al. (1996) conclude that gas-hydrate-bearing sedimentary sections tend to be tens of centimeters to tens of meters thick, but thick zones of pure hydrate are relatively rare and only represent a minor constituent of potential gas hydrate accumulations.

The Booth et al. (1996) review along with recently published gas hydrate sample descriptions from the Mackenzie Delta (Dallimore and Collett, 1995; Dallimore et al., 1999) and the Blake Ridge (ODP Leg 164, Shipboard Scientific Party 1996), confirm that gas hydrates are usually uniformly distributed within sediments as mostly pore-filling constituents.

Gas Hydrate Reservoir Models

Most discussions on the nature or texture of gas hydrate occurrences deal with macroscopic issues (reviewed by Booth et al., 1996).  However, information on the occurrence of gas hydrates at the pore-scale are needed, since many gas hydrate reservoir physical properties are controlled by microscopic parameters (Dvorkin and Nur, 1993).  Of particular concern is the acoustic nature and the fluid-flow permeability characteristics of gas-hydrate-bearing sediments (Lee et al., 1993; Dvorkin and Nur, 1993).

Dvorkin and Nur (1993) along with Ecker et al. (1996) have proposed and examined two "micromechanical" models that represent the two extreme cases of gas hydrate occurrence at the pore-scale: (Model-1) gas hydrate cement grain contacts and increases the stiffness of the sediment; and (Model-2) gas hydrate is located away from grain contacts in the "bulk" pore volume, and it does not affect the stiffness of the sediment frame.  Dvorkin and Nur (1993) experimentally demonstrated that even small amounts of intergranular cementation, such as proposed by gas hydrate Model-1, can dramatically increase the stiffness of granular material.  Dvorkin and Nur (1993) used the intergranular gas hydrate cementation model (Model-1) to explain the occurrence of seismic bottom-simulating-reflectors (BSR's), which they attributed to a strong increase of the elastic moduli of the rock due to the occurrence of gas hydrates at the base of the gas hydrate stability zone.  Ecker et al. (1996) amplitude-versus-offset (AVO) analyses of the BSR on the Blake Ridge, however, concluded that only reservoir Model-2 could qualitatively reproduce the observed BSR, and that gas hydrates at the pore-scale are located away from the intergranular contacts, in large pores.  Ecker et al. (1996) further concluded that the sediment above the BSR is uncemented and mechanically weak.  However, they do not explain the acoustic parameters that control the occurrence of the BSR on the Blake Ridge.  At this time, we must consider the conclusions of Dvorkin and Nur (1993) and Ecker et al. (1996) preliminary until additional laboratory field observations become available.

Before attempting to assess the volume of gas hydrate in a particular reservoir, we need to develop and define a series of reservoir models for the occurrence of gas hydrates in nature.  Most reservoir models are based on simple mixing rules, where complex multi-component systems consist of simple mixtures of rock matrix (consisting of quartz, calcite, and/or clay), water (including clay-bound- and free-water), and hydrocarbons (gas and/or oil).  In permafrost and relatively deep marine environments, however, other reservoir constituents can include gas hydrates and permafrost ice.  The first two reservoir models to be considered represent complex gas-hydrate-bearing reservoirs both below (Model A; Figure 16a) and above (Model B; Figure 16b) the base of ice-bearing permafrost in a terrestrial setting.  In both of these models the sediment matrix consists of a simple mixture of quartz, calcite, and a relatively small amount of clay.  Gas hydrate reservoir Models A and B assume no free-gas phase, since all of the available gas is in the gas hydrate.  The only difference between Models A and B is that Model B assumes that all of the free-water and some of the clay-bound-water are frozen.  Reservoir Model C (Figure 16c) has been designed to represent a clay-rich marine gas hydrate reservoir.  Reservoir Models C and A are similar, but Model C assumes the clay content of the sediment and associated volume of bound-water are higher in most marine gas-hydrate reservoirs.  The last gas hydrate reservoir model to be considered may not occur in nature.  Reservoir Model D (Figure 4d) assumes that a free-gas phase exists, and that all of the available water is included in the gas hydrate.  Water, being relatively abundant in nature, should not be a gas hydrate limiting factor in most reservoirs.

Why Do Gas Hydrates Occur in a Particular Setting?

Review of previous gas hydrate studies indicates that the formation and occurrence of gas hydrates is controlled by formation temperature, formation pore-pressure, gas chemistry, pore-water salinity, availability of gas and water, gas and water migration pathways, and the presence of reservoir rocks and seals (reviewed by Collett, 1995).  In the following section, these geologic controls on the stability and formation of gas hydrates are reviewed and assessed.

Formation-Temperature, Formation Pore-Pressure, Gas Chemistry

Gas hydrates exist under a limited range of temperature and pressure conditions such that the depth and thickness of the zone of potential gas-hydrate stability can be calculated.  Depicted in the temperature/depth plot of Figure 17 are a series of subsurface temperature profiles from an onshore permafrost area and two laboratory-derived gas-hydrate stability curves for different natural gases (modified from Holder et al., 1987).  This gas-hydrate phase-diagram (Figure 17) illustrates how variations in formation-temperature and gas composition can affect the thickness of the gas-hydrate stability zone.  In Figure 17, the mean-annual surface temperature is assumed to be -10°C; however, the depth to the base of permafrost (0°C isotherm) is varied for each temperature profile (assumed permafrost depths of 305 m, 610 m, and 914 m).  Below permafrost, three different geothermal gradients (4.0°C/100 m, 3.2°C/100 m, and 2.0°C/100 m) are used to project the sub-permafrost temperature profiles.  The two gas-hydrate stability curves represent gas hydrates with different gas chemistries.  One of the stability curves is for a 100% methane hydrate, and the other is for a hydrate that contains 98% methane, 1.5% ethane, and 0.5% propane.

The zone of potential gas-hydrate stability in Figure 17 lies in the area between the intersections of the geothermal gradient and the gas-hydrate stability curve.  For example, in Figure 17, which assumes a hydrostatic pore-pressure gradient, the temperature profile projected to an assumed permafrost base of 610 m intersects the 100% methane-hydrate stability curve at about 200 m, thus marking the upper boundary of the methane-hydrate stability zone.  A geothermal gradient of 4.0°C/100 m projected from the base of permafrost at 610 m intersects the 100% methane-hydrate stability curve at about 1,100 m; thus, the zone of potential methane-hydrate stability is approximately 900 m thick.  However, if permafrost extended to a depth of 914 m and if the geothermal gradient below permafrost is 2.0°C/100 m, the zone of potential methane-hydrate stability would be approximately 2,100 m thick.

Most gas-hydrate stability studies assume that the pore-pressure gradient is hydrostatic (9.795 kPa/m; 0.433 psi/ft).  Pore-pressure gradients greater than hydrostatic will correspond to higher pore-pressures with depth and a thicker gas-hydrate stability zone.  A pore-pressure gradient less than hydrostatic will correspond to a thinner gas-hydrate stability zone.  For example, in Figure 17, which assumes a hydrostatic (9.795 kPa/m; 0.433 psi/ft) pore-pressure gradient, the thickness of the 100% methane-hydrate stability zone with a 610 m permafrost depth and a sub-permafrost geothermal gradient of 2.0°C/100 m would be about 1,700 m.  However, if a pore-pressure gradient of 11.311 kPa/m (0.500 psi/ft) is assumed, the thickness of the methane-hydrate stability zone would be increased to about 1,850 m.

The gas-hydrate stability curves in Figure 17 were obtained from laboratory data published in Holder et al. (1987).  The addition of 1.5% ethane and 0.5% propane to the pure methane gas system shifts the stability curve to the right, thus deepening the zone of potential gas-hydrate stability.  For example, assuming a hydrostatic pore-pressure gradient (Figure 17), a permafrost depth of 610 m, and a sub-permafrost geothermal gradient of 4.0°C/100 m, the zone of potential methane (100% methane) hydrate stability would be about 900 m thick; however, the addition of ethane (1.5%) and propane (0.5%) would thicken the potential gas-hydrate stability zone to 1,100 m.

Pore-Water Salinity

Salt, such as NaCl, when added to a gas-hydrate system, lowers the temperature at which gas hydrates form.  Pore-water salts in contact with the gas during gas hydrate formation can reduce the crystallization temperature by about 0.06°C for each part per thousand (ppt) of salt (Holder et al., 1987).  Therefore, a pore-water salinity similar to that of seawater (32 ppt) would shift the gas-hydrate stability curves in Figure 17 to the left about 2°C and reduce the thickness of the gas-hydrate stability zone.

Availability of Gas and Water

Most naturally occurring gas hydrates are characterized by two crystal structures known as Structure I and Structure II (reviewed by Sloan, 1998).  The ideal gas/water ratio of Structure I gas hydrate is 8/46, whereas the ideal gas/water ratio of Structure II gas hydrate is 24/136.  These ideal ratios confirm the observation that gas hydrates contain a substantial volume of gas.  The ideal hydrate gas/water ratios also indicate that there is a substantial amount of water stored in the gas-hydrate structure.  These high gas and water concentrations demonstrate that the formation of gas hydrate requires a large source of both gas and water.  Thus, it becomes necessary to quantify the potential sources of gas and water when assessing a potential gas-hydrate accumulation.  In previous studies, this evaluation is based on assessing a set of minimum source-rock criteria that includes organic richness (total organic carbon), sediment thickness, and thermal maturity.  It has been shown that the availability of large quantities of hydrocarbon gas from both microbial and thermogenic sources is an important factor controlling the formation and distribution of natural gas hydrates (Kvenvolden, 1988; Collett, 1993).  Carbon isotope analyses indicate that the methane in many oceanic hydrates is derived from microbial sources.  However, molecular and isotopic analyses indicate a thermal origin for the methane in several offshore Gulf of Mexico and onshore Alaskan gas-hydrate occurrences.

Gas and Water Migration Pathways

Other factors controlling the availability of gas and water are the geologic controls on fluid migration.  As previously shown, gas hydrates contain a substantial volume of gas and water that must be supplied to a developing gas-hydrate accumulation.  If effective migration pathways are not available, it is unlikely that a significant volume of gas hydrates would accumulate.  Therefore, geologic parameters such as rock permeability and the nature of faulting must be evaluated to determine if the required gas and water can be delivered to the potential hydrate reservoir.

Presence of Reservoir Rocks and Seals

The study of gas-hydrate samples recovered during research coring operations in oceanic sediments suggests that the physical nature of in-situ gas hydrates may be highly variable (Figure 15, as previously discussed in this paper).  Gas hydrates were observed to be (1) occupying pores of coarse-grained rocks; (2) nodules disseminated within fine-grained rocks; (3) a solid, filling fractures; or (4) a massive unit composed mainly of solid gas hydrate with minor amounts of sediment.  This review suggests that porous rock intervals serve as reservoir rocks in which gas and water can be concentrated in the amounts necessary for gas-hydrate formation.  Therefore, the presence of reservoir rocks may play a role in gas-hydrate formation, particularly in well-consolidated rock intervals.

It is also speculated that the presence of effective reservoir seals or traps may play a role in gas-hydrate formation.  Gas generated at depth moves upward, generally along tilted permeable carrier beds, until it either seeps at the surface or meets an impermeable barrier (trap) that stops or impedes its flow.  As migrating gas accumulates below an effective seal, the total gas concentrations may reach the critical amounts necessary for the formation of gas hydrates.  Thus, impermeable seals can provide a mechanism by which the required gas can be concentrated within reservoir rocks.

Besides conventional reservoirs and trapping mechanisms, it is possible for gas hydrate to form its own reservoir and trap.  As gas migrates into the zone of gas-hydrate stability, it may interact with the available pore water to generate gas hydrate.  With the appropriate volumes of gas and water, the pore space within the reservoir rock could be completely filled, thus making the rock impermeable to further hydrocarbon migration.  The plugging of gas pipelines and production tubing by gas hydrates is testimony to the sealing potential of gas hydrates (Sloan, 1998).  It has also been shown that, in marine environments, gas hydrates can mechanically displace sediments to form their own reservoir.  Thus, the availability of reservoir quality rocks may not always be a limiting factor.

Natural Gas Hydrates: Resource of the 21st Century?