Dissociation of Sub-Seafloor Gas Hydrates and Seafloor Stability: What Thermobaric Models Show

Introduction: Gas hydrates, which are stable at the high pressures of seafloor more than a few hundred meters deep, are crystalline solids with a cage-like (clathrate) structure of water molecules enclosing larger host molecules, most commonly methane. The material is present between sediment grains or in larger veins and nodules under most continental margins, with enormous aggregate volumes. Ongoing research, some at NRL, is motivated by the potential use of hydrates as a fuel source, the possible release of greenhouse gas by hydrate dissociation, and the influence of such dissociation on seafloor sediment stability, which forms the basis for our study.

The association of undersea landslides with geophysical evidence for hydrates below the U.S. East Coast continental slope led G. Carpenter to hypothesize in 1981 that low sea levels of the last Ice Age (ca. 18,000 years ago, abbreviated 18ka (millenia ago)) may have caused hydrate to dissociate at depth, reducing sediment shear strength and facilitating failure. Other researchers elaborated on this idea-uggesting that the methane liberated by slides could have warmed the atmosphere, forming a "shut-off valve" for ice ages. While these ideas have gained wide acceptance, Norwegian scientists found the three major slides they studied along their margin (Fig. 10) to have happened when they should not—during the rapidly rising sealevels of post-glacial times. The largest (Storegga, 5,500 cubic kilometers!) and best known slide-nd one that NRL and Norwegian collaborators have studied, including 1999 dives by USN nuclear submarine NR-1-ccurred 8,150 years ago (8.15ka), generating a major tsunami. Intrigued by this contradiction but still suspicious that hydrate dissociation helped cause the failures, we began modeling the evolution of the gas hydrate stability zone at Storegga and other margins. If major slides can happen in post-glacial times, could they occur today or tomorrow?

FIGURE 10
Norwegian continental margin, showing relatively recent giant underwater landslides in green: SS, Storegga slide; TS, Traenadjupet slide; AS, Andoeya slide; and BIFS, Bear Island Fan slide. Gas hydrate stability models shown in Figs. 11 and 12 were computed along line #1. Black contours show water depths in 100s of meters; thickness of modern methane hydrate stability zone, in meters, is shown by dashed blue contours (600 and 0 m). Blue ovals and patches show areas where seismic surveying has detected hydrates to date.

Research Approach: Starting with a "slice" through the sea and seafloor in the middle of the Storegga slide scar (red line in Fig. 10), we calculated the thickness of the Methane Hydrate Stability Zone as a function of rising sea level and bottom water temperatures, which have opposing effects on hydrate stability: A rising sea level, due mostly to melting glaciers, increases pressure at all depths and makes hydrate stable to greater depths. However, seawater warming at the ocean floor has the opposite effect—once this warmer boundary condition diffuses to the base of the hydrate stability zone, any hydrate in that basal region would become unstable, liberating free gas and water, and thus reducing the shear strength of the mud.

As in other physical modeling studies, we need reasonably accurate estimates for various physical quantities—either from laboratory or field measurements. For example, sediment density determines the pressure exerted on the hydrate by the overlying mud. Previous NRL cores gave us near-bottom density, but at greater depths (10s to 100s of meters) we rely on published seismic and borehole results. The subbottom temperature gradient controls the pressure (depth) at which hydrate becomes unstable—the steeper the gradient, the thinner the stability zone. To get this gradient, we combined shallow (a few meters penetration) thermal probe data from past NRL cruises with other published values. The thermal diffusivity is needed to calculate the rate at which seafloor warming propagated ("diffused") down into the subbottom. Because we are interested in what happened in the Storegga slide area over time, we have to figure out what the topography looked like before the slide altered it. Such topographic reconstruction can be done by extrapolating depth contours from one side of the horseshoe-shaped slide scar to the other (Fig. 10 shows concept). One can think of the pre-slide margin as a cookie, from which a bite has been taken. Even if the bitten-off piece is missing, the cookie's original shape can be estimated. For Storegga, the volume of the "bite" has been checked by seismic reflection techniques against the amount of slide debris in the Norway Basin.

The phase boundary separating hydrate from methane+water depends on pore water chemistry. Laboratory studies show that if the water between the sediment grains is salty, the boundary is shifted to decreasing stability ("A" side of gray band in Fig. 11). On the other hand, if some higher hydrocarbons, for example ethane, are mixed with the methane, the stability field is expanded ("B" in Fig. 11). Uncertain pore water chemistry led us to show the boundary as a band, vs a sharp line. Fresh pore water with pure methane lies in the middle of the band. The "truth," likely varying with depth and locale, probably lies somewhere within. Future deep drilling and pore water chemistry will some day provide a more precise answer.

Since seawater sulfate ions oxidize methane in the upper sediments (light blue regions in Fig. 12), methane hydrate, although stable, is largely absent there.

A most critical dataset is the history (more precisely the PRE-history) of how sea levels and water temperatures rose starting 18,000 years ago, when the last Ice Age "maxed out." One of the two sea level rise curves we used was derived from a joint NRL-US Geological Survey sediment coring expedition to our own Chesapeake Bay, and similar curves have been derived by others. When a branch of the Gulf Stream entered the Norwegian Sea ca. 11,000 years ago, it left a different assemblage of planktonic fossils in the sediments below from which past ocean temperatures can be reconstructed.

Although we cannot go back in a time machine to directly measure past gas hydrate distribution, present (0ka) model validation can and has been performed by seismic determination of hydrate distribution.

Results and Implications: Our models predict how the Methane Hydrate Stability Zone must have changed during the last 18,000 years in the area of the great Storegga Slide. We show this change in two ways: Figure 11 shows how selected parcels of sediment (numbered in Fig. 12) moved in temperature-pressure (`thermobaric') space. Figure 12 shows conditions at various times along line #1 (Fig. 10) across the upper slide scar.

During the greatest extent of glaciation (18ka), sea level, and hence subbottom pressure, was too low for hydrate to be stable on the Storegga shelf. By 11ka, rising sea level had expanded the stability zone onto the shelf. Sometime after upper waters warmed, the stability zone was forced to retreat from the shelf once more. The time between 18ka until somewhat after 11ka years ago would not have been favorable for dissociation, and therefore slides. However, comparison of profile 11ka- and 8.15ka+ shows how hydrate dissociation could have promoted the Storegga slide when it happened - hydrate present in the dark blue zone at 11ka- would have dissociated, creating gassy, low-shear strength sediment (stippled at 8.15ka+), which then failed when the right earthquake came along. Failure most likely began where the dissociation zone (stippled in Fig. 12, 8.15ka+) intersects the known slide base.

FIGURE 11
Trajectories (from last Ice Age at 18ka, to the present) of seven sub-seafloor points (see Fig. 12 for locations) in temperature and pressure space, in relation to boundary of methane hydrate stability. Circled points correspond to 8.15ka, time of Storegga slide. Trajectory branches for points within the slide area and covered by slide debris are shown dashed. Because the phase boundary depends on pore water salinity and admixed higher hydrocarbons, it is shown here as a band. The left-hand edge (A) reflects pore water with the salinity of seawater, while the right-hand edge (B) reflects fresh water with 2% ethane.

Although we did not try to model the transient slide event, we know it stripped off most of the hydrate stability zone (8.15ka-), which then rethickened as the warm subbottom exposed by the slide was cooled by the overlying water. Some authors have suggested that reduced post-slide pressure would have caused additional hydrate dissociation and secondary slides but at Storegga, most hydrate-infested sediment, except in deep water, was removed. Even if some had been left, the large (negative) latent heat of dissociation would have cooled the remaining hydrate, delaying further dissociation.

The solid red tracks in Fig. 11 illustrate how shallower sediment parcels (#1,2,4,5) first moved upward toward hydrate stability as sea levels rose, and then veered back out towards instability due to post-11ka warming. Deep parcels (e.g.,#7) were unaffected by warming, while other parcels (e.g., #3,6) never approached the stability field, except in the slide scar, where they were abruptly jerked toward lower pressure and temperature by the slide event (dashed lines), only to warm again as thermal equilibrium was restored after 8.15ka.

Summary and Conclusion: Modeling the thermobaric evolution of the Storegga slide area may explain why this and other slides occurred during post-glacial times, when pressures were rising and hydrate stability should have been on the increase. The reason has to do with bottom water warming, which postdates most of the sea level rise, and which then took additional time to "diffuse" down far enough below the seafloor to cause hydrate to become unstable. Post-glacial slides could not have caused deglaciation, but were consequences of climate warming. Our results apply to other continental margins, except in polar and in deep waters, which have remained cold. Our study also shows that the risk of major slides (and resulting tsunamis) did not disappear after the last glaciation, but remains high today on the upper continental slope. Additionally, future ocean warming would exacerbate this risk.

[Sponsored by ONR]