Excess pore pressure resulting from methane hydrate dissociation in marine sediments: A theoretical approach

Excess pore pressure resulting from methane hydrate dissociation

in marine sediments: A theoretical approach

Wenyue XuSchool of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta, Georgia, USA

Leonid N. GermanovichSchool of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, Georgia, USA

Received 26 December 2004; revised 14 September 2005; accepted 12 October 2005; published 27 January 2006.

[1] This study quantifies the excess pore pressure resulting from gas hydrate dissociationin marine sediments. The excess pore pressure in confined pore spaces can be up toseveral tens of megapascals due to the tendency for volume expansion associated with gashydrate dissociation. On the other hand, the magnitude of excess pore pressure in well-connected sediment pores is generally smaller, depending primarily on thehydrate dissociation rate and the sediment permeability. Volume expansion due to gashydrate dissociation in well-connected pore spaces is related via Darcy’s law to an increasein pore pressure and its gradient in sediment, which drives an additional upward fluidflow through the sediment layer overlying the gas hydrate dissociation area. Themagnitude of this excess pore pressure is found to be proportional to the rate of gashydrate dissociation and the depth below seafloor and inversely proportional to sedimentpermeability and the depth below sea level. The excess pore pressure is the greatest at lowinitial pressures and decreases rapidly with increasing initial pressure. Excess porepressure may be the result of gas hydrate dissociation due to continuous sedimentation,tectonic uplift, sea level fall, heating or inhibitor injection. The excess pore pressure isfound to be potentially able (1) to facilitate or trigger submarine landslides in shallowwater environments, (2) to result in the formation of vertical columns of focusedfluid flow and gas migration, and (3) to cause the failure of a sediment layer confined bylow-permeability barriers in relatively deep water environments.

Citation: Xu, W., and L. N. Germanovich (2006), Excess pore pressure resulting from methane hydrate dissociation in marine

sediments: A theoretical approach, J. Geophys. Res., 111, B01104, doi:10.1029/2004JB003600.

1. Introduction

[2] It has long been suggested that gas hydrate dissocia-tion in marine sediment may lead to excess pore pressureresulting in sediment deformation or failure, such as sub-marine landslides [Field, 1990; Kayen, 1988; Kayen andLee, 1991; McIver, 1982; Mienert et al., 2005; Paull et al.,1996; Rothwell et al., 1998], sediment slumping [Dillon etal., 2001; Vogt et al., 1994], pockmarks and mud volcanoes[Van Rensbergen et al., 2002; Vogt et al., 1994, 1999;Zuhlsdorff and Spiess, 2004], soft-sediment deformation[Kennett and Fackler-Adams, 2000] and giant hummocks[Davies et al., 1999]. Rapid release of a large amount ofmethane from dissociating submarine gas hydrates to theatmosphere may drastically impact the global climate[Dickens et al., 1995; Kennett et al., 2003; MacDonald,1990; Nisbet, 1990]. However, it is shown [Xu et al., 2001]that without invoking additional mechanisms, an enhancedtransport of methane through marine sediment due to gashydrate dissociation resulting from a falling sea level or an

increase in seafloor temperature is probably not sufficientto cause the drastic increase in greenhouse gas released fromdissociating gas hydrates suggested by Dickens and cow-orkers [Dickens et al., 1997a, 1995]. On the other hand, thecreation or reactivation of fault zones and fluid flow channelsor other types of sediment deformation and failure caused byexcess pore pressure as the result of gas hydrate dissociationmay provide pathways that are highly efficient in transportinggas [Paull et al., 2003; Zuhlsdorff and Spiess, 2004]. Acareful quantification of expected excess pore pressure levelsrelated to various environmental changes is needed.[3] Efforts to quantify excess pore pressures related to gas

hydrates in marine sediments have been made using severaldifferent approaches. Flemings et al. [2003] tried to estimatethe magnitude of excess pore pressure beneath the gashydrate layer at Ocean Drilling Program Site 997, BlakeRidge offshore North Carolina. They used an empiricalinterpretation of measured porosity data. Therefore it isnot guaranteed that the dynamic feedback between sedimentcompaction and pore pressure is adequately accounted for.[4] Hornbach et al. [2004] suggested that a critically

pressured layer of free gas is present underneath most gashydrate provinces. Aside from the general lack of evidence

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 111, B01104, doi:10.1029/2004JB003600, 2006

Copyright 2006 by the American Geophysical Union.0148-0227/06/2004JB003600$09.00

B01104 1 of 12

supporting their suggestion, how such a critical pressuremay be maintained is questionable. First, the volumefraction of gas bubbles in sediment pore space is estimatedto be less than 5% in most cases with a higher value of up to12% at Blake Ridge [e.g., Dickens et al., 1997b; Holbrooket al., 1996; Singh and Minshull, 1994; Yuan et al., 1996].At these volume fractions, the formation of aninterconnected gas column can still be difficult because ofthe capillary effect [Henry et al., 1999]. Moreover, the factthat bottom-simulating reflectors (BSR) locations of manygas hydrate provinces are consistent with the gas hydratestability boundary predicted by assuming a normal temper-ature gradient and hydrostatic pressure is in contradictionwith the suggested critical pressure at the top of such a gascolumn.[5] Sultan et al. [2004] are among the first efforts to more

accurately quantifying the excess pore pressure related togas hydrate dissociation. However, their work considersonly the maximum excess pore pressure when gas hydratedissociation occurs in confined pore spaces. It does notdistinguish gas hydrate dissolution from dissociation andinappropriately uses experimental data of CO2 hydratedissolution as a proxy for methane hydrate dissolution.These simplifications lead to the conclusion that excess porepressure can result from methane hydrate dissolution nearthe top of the gas hydrate layer and contribute to the large-scale Storegga submarine landslide offshore mid-Norway.[6] Here we address this problem by first quantifying the

volume expansion associated with gas hydrate dissociation.The resultant excess pore pressure is then calculated eitherbased on the amount of dissociated gas hydrate in aconfined pore space or according to the rate of gas hydratedissociation in the interconnected pore space of an incom-pressible sediment by relating dissociation-released fluids toan enhanced upward fluid flow away from the horizonwhere gas hydrate dissociation takes place. The cases ofsea level fall, tectonic uplift, heating and inhibitor injectionin a horizontal layer of gas hydrate-bearing sediment arealso analyzed to obtain the associated rate of gas hydratedissociation. Finally, implications of the excess pore pres-sure caused by gas hydrate dissociation to marine sedimentdeformation or failure are discussed.[7] Previous studies [Clennell et al., 1999; Xu and

Ruppel, 1999; Zatsepina and Buffett, 1998] suggest that athree-phase zone in which gas hydrate, water and free gascoexist may occur in marine gas hydrate systems. Furtherinvestigations [Xu, 2002, 2004] indicate that at conditionsclose to thermodynamic equilibrium, gas hydrate dissocia-tion takes place within the three-phase zone until all gashydrates are dissociated. This equilibrium can be dynamicin the sense that the equilibrium conditions may changewith time. This idea is fundamentally different from earlierconcepts of gas hydrate dissociation along a three-phaseinterface separating an overlying zone of coexisting waterand gas hydrate from an underlying zone of coexistingwater and free gas. Unless explicitly stated otherwise, gashydrate dissociation and related processes consideredthrough out this study are assumed to occur at localconditions close to thermodynamic equilibrium. All calcu-lations of phase equilibrium and thermodynamic propertiesof free gas, liquid solution and gas hydrate are doneaccording to the methods described by [Xu, 2002, 2004;

Xu and Ruppel, 1999]. All calculations are done forstructure I methane hydrate, the dominant component innatural gas hydrate systems.[8] For simplicity, calculations are done assuming a

constant salinity of 3.5% and effects of a varying salinityon phase equilibrium and transition are neglected. In gen-eral, salinity of the liquid phase decreases as gas hydratesdissociate. This change raises the equilibrium temperatureand, consequently, stabilizes gas hydrates. It can also lowerthe solubility of methane in liquid water. Dissociation of asmall amount, say less then 10% of pore volume, of gashydrates in a mixture of gas hydrate and liquid solutionresults in a rather small, in the order of 10% or less, changein liquid salinity and thus has little effect on the processesstudied in this work. Parameters and properties used in thisstudy are listed in Table 1.

2. Volume Expansion Due to Gas HydrateDissociation

[9] To avoid misconceptions, we preface our gas hydratedissociation discussion whit a clarification of the distinctionbetween gas hydrate dissolution and dissociation. Theformer can take place within the whole gas hydrate stabilityzone as gas solubility of the coexistent liquid phaseincreases. In general, no free gas is produced during thecourse of gas hydrate dissolution and, since the density ofmethane hydrates is lower than that of the coexistent liquidwater-gas solution, gas hydrate dissolution does not result inexcess pore pressure in natural environments. Sultan et al.[2004] suggest that gas hydrate dissolution near the top ofthe gas hydrate occurrence zone might have caused consid-erable excess pore pressure and thus contributed to theoccurrence of the Storegga Slide offshore mid-Norway.Their speculation is based on laboratory experiments doneon CO2 hydrate [e.g., Zhang, 2003], which has a higherdensity than the coexistent liquid water-gas solution, andmay not be applicable to natural environments wheremethane is the most abundant gas component of gashydrates. In contrast, gas hydrate dissociation takes placenear its stability boundary, produces water plus free gas andtherefore results in excess pore pressure due to the volumeexpansion associated with the dissociation. Though thereare certain exceptions [Xu, 2004], in natural environments,gas hydrate dissociation occurs much more frequently nearthe base of a gas hydrate stability zone within a dynamicallyevolving natural gas hydrate system in a changing environ-ment. Here we analyze this volume change resulting fromgas hydrate dissociation.[10] Suppose that a mixture of liquid water, free gas and

gas hydrate resides in pore volume Vp, which is

Vp ¼ fADD ð1Þ

for a porous sediment column of cross sectional area A,length DD and porosity f. The volume of gas hydratesinside the pore space is

Vh ¼ VpSh; ð2Þ

where Sh is the volume fraction of gas hydrate averagedover Vp. The total volume change dV resulting from the

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

2 of 12

B01104

dissociation of a small dVh gas hydrates, which releases dVw

water and dVg free gas, includes the volume change due tothe fact that the densities of released water and free gas aredifferent from that of the dissociated gas hydrate and thevolume change due to the compressibility of extant free gas,liquid and gas hydrate.[11] The first volume change, (dV)1, related to density

differences is

dVð Þ1¼ dVw þ dVg þ dVh: ð3Þ

Compared to that of gas released by dissociating gashydrate, the amount of gas dissolved in the released water issmall. As long as the gas solubility of water does not changerapidly, the dissolved gas to the volume change is usuallynegligible. Therefore, for a fixed gas mass fraction rg of thegas hydrate,

dVg ¼ dMg=rg ¼ �rgdMh=rg ¼ �rg rh=rg� �

dVh ð4Þ

dVw ¼ dMw=rw ¼ � 1� rg� �

dMh=rw ¼ � 1� rg� �

rh=rwð ÞdVh;

ð5Þ

where rw, rg, rh andMw,Mg,Mh denote the densities and themasses of the liquid water, the free gas and the gas hydrate,respectively. Thus the volume change relative to porevolume Vp is

dVð Þ1Vp

¼ �Rv

dVh

Vp

¼ �RvdSh; ð6Þ

where

Rv ¼ 1� rg� �

rh=rw þ rgrh=rg � 1 ð7Þ

is the factor of volume expansion resulting from the gashydrate dissociation.[12] Variations in rw, rh and rg are usually negligible. Rv is

only weakly dependent on temperature, controlled insteadby density changes of the gas, which is usually highlycompressible. Gas hydrate dissociation leads to volumeexpansion when Rv > 0, whereas a volume contractionresults if Rv < 0. This fact is depicted in Figure 1 forstructure I methane hydrate, which is calculated usingSetzmann and Wagner’s [1991] equation of state and showsa dramatic increase in the tendency of expansion as pressure

Table 1. Parameters and Properties Used in This Study

Parameter Definition

A section area of sediment column (m2)As total area of dissociation interfaces within a sediment column with a unit

section area (m2)c heat capacity of hydrate-bearing sediment (J kg�1 K�1)cw, ch, cg, cs specific heat of liquid water, gas hydrate, free gas or sediment (J kg�1 K�1)D depth to the layer of gas hydrate dissociation below seafloor (m)DD thickness of the layer of gas hydrate dissociation (m)g gravitational acceleration (m s�2)DHdis enthalpy of gas hydrate dissociation (4.3 � 105 J kg�1)k permeability (m2)Mw, Mh, Mg mass of liquid water, gas hydrate or free gas (kg)P pressure (Pa)Pex excess pore pressure (Pa)Pex

0 excess pore pressure t = t0 (Pa)Pexmax maximum excess pore pressure (Pa)P0 initial pore pressure (Pa)Qe rate of heating (W m�2)Dqf change in mass flux of fluid flow (kg m�2 s�1)R rate coefficient of gas hydrate dissociation (2.227 � 10�4 kg (m2 K2 s)�1)rg mass fraction of gas in hydrate form, 0.1292 for methane hydrate with

100% filling of the hydrate cagesRv volume expansion factorSw, Sh, Sg volume fraction of liquid water, gas hydrate or free gas in pores

Sw + Sh + Sg = 1t time (s)T temperature (K)T0 temperature at t = 0 (K)T0 temperature at t = t0 (K)Te temperature along the stability boundary (K)Te0 temperature along the stability boundary at t = 0 (K)Vw, Vh, Vg, V volume of liquid water, gas hydrate, free gas or their mixture (m3)

V = Vw + Vh + Vg

vsd rate of sea level drop or tectonic uplift (m s�1)Vp = fADD volume of pore space (m3)z space (z)a ratio of the vertical extent of resultant hydrate dissociation zone to the

magnitude of sea level fall or tectonic upliftf sediment porosityk effective volume compressibility of the mixture of liquid water, gas

hydrate and free gas (Pa�1)l effective thermal conductivity of sediment column (W m�1 K�1)mf viscosity of fluids (Pa s)rf, rw, rh, rg, rs density of fluids, liquid water, gas hydrate, free gas or sediment (kg m�3)

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

3 of 12

B01104

decreases. For example, the dissociation of methane hydrateequilibrating with pure water or seawater at a pressure near34 MPa results in a volume expansion of �30% comparedto a �210% volume expansion if the dissociation takesplace at 6.5 MPa, and �500% for dissociation at 2.5 MPa.Because of higher pressures in marine sediment, the volumeexpansion due to gas hydrate dissociation in marine sedi-ment in general is much smaller than that at standardpressure and temperature conditions. Therefore referringto a much larger volume expansion at standard conditionswhen discussing the dissociation of gas hydrates in marinesediments can be misleading.[13] The second part of the total volume change is related

to the compressibility of existing free gas, gas hydrate andliquid solution. When volume compression resulting fromthe increasing pore pressure is considered, the magnitude ofvolume expansion becomes even smaller than what wouldbe when pore pressure is kept constant. Assuming thermo-dynamic equilibrium, the effective compressibility, k, of themixture of free gas, gas hydrate and liquid solution alongthe gas hydrate stability P-T boundary may be calculated as

k ¼ � 1

V

@V

@Pþ @V

@T

dTe

dP

� �¼ Sg

rg

@rg@P

þ@rg@T

dTe

dP

� �

þ Sw

rw

@rw@P

þ @rw@T

dTe

dP

� �þ Sh

rh

@rh@P

þ @rh@T

dTe

dP

� �; ð8Þ

where Te is the stability temperature of structure I methanehydrate corresponding to the given pressure and Sg, Sh andSw are the pore space volume fractions of free gas, gashydrate and liquid solution, respectively. Compared to thatof free gas, the compressibility of liquid solution or gashydrate can usually be neglected. However, the latterbecomes significant when the volume fraction of free gas issmall. The effective compressibility is larger when the

pressure is lower and the volume fraction of free gas ishigher (Figure 2). When there is no pressure change otherthan the excess pore pressure Pex resulting from gas hydratedissociation, the relevant volume change is

dVð Þ2Vp

¼ �kdPex: ð9Þ

Therefore the total volume change is

dV

Vp

¼ dVð Þ1Vp

þ dVð Þ2Vp

¼ �RvdSh � kdPex: ð10Þ

Note that dV 6¼ 0 when the pore space is not completelyclosed.

3. Excess Pore Pressure Resulting From GasHydrate Dissociation

3.1. In Confined Pore Space

[14] When host sediment permeability is sufficiently lowor gas hydrates dissociate rapidly, the dissociation processmay be treated as taking place in a constant pore volume.This may lead to a large excess pore pressure because fluidsreleased by dissociating gas hydrates do not have time toescape the pore space. Since the pore space is confined, thetotal volume change is zero and equation (10) becomes

RvdSh þ kdPex ¼ 0: ð11Þ

The magnitude of excess pore pressure can be calculatedby integrating equation (11) over the whole dissociationprocess, which gives

Pex ¼Z �DSh

0

Rv

kdSh; ð12Þ

Figure 1. Volume expansion factor Rv calculated as afunction of pressure and temperature. The stability bound-aries of methane hydrate in pure water and seawaterenvironments are also shown. Assuming thermodynamicequilibrium, the dissociation of gas hydrate takes placealong the stability boundary.

Figure 2. Effective compressibility of the mixture ofliquid, free gas and gas hydrate along the stability boundary.Calculations are for fixed 20% volume fraction of gashydrate. The sum of volume fractions of liquid and free gasphases is 80%.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

4 of 12

B01104

where DSh is the change in volume fraction of gas hydratein pore space, which is a negative number for gas hydratedissociation. The excess pore pressure, as calculated usingequation (12), is plotted in Figure 3 as a function of DShand the initial pore pressure before gas hydrate dissocia-tion starts. The excess pore pressure can be tens ofmegapascals even when only a small amount of gashydrate dissociates.

3.2. In Interconnected Pore Space

[15] In the following studies, fluid flow is assumed totake place in rigid or incompressible porous sediments andobey Darcy’s law with a relatively constant effective per-meability. To avoid the possibility of nonlinear fluid flowbehavior as pointed out by Clennell et al. [2000], thefollowing analyses do not apply to conditions close tocapillary leakage pressures or pressures sufficiently highto cause sediment fracturing. In addition, downward flow isassumed negligible compared to upward fluid flow associ-ated with excess pore pressure resulting from gas hydratedissociation. This is usually true when fluids outside of thedissociation layer are nearly incompressible and because theseafloor is the only place for fluid discharge. The downwardflow may be significant under certain circumstances, par-ticularly when an underlain sediment layer containing largeamount of free gas is present, as can occur at some BSRs.Then the compressibility of the free gas contained insediment under the dissociation layer will need to beaccounted for.[16] Considering a vertical one-dimensional system with

a closed lower end, the volume increase resulting from thedissociating gas hydrates is mostly accounted for by anincrease in the rate of upward fluid flow Dqf away from thegas hydrate dissociation area

ADqf ¼ rfdV

dt¼ �rf fADD Rv

dSh

dtþ k

dPex

dt

� �; ð13Þ

where rf is the density of the upward flowing fluid. Theflow rate increase can be directly related to a correspondingpressure gradient increase D(�@P/@z) across the overlainsediment layer according to Darcy’s law

ADqf ¼Akrfmf

D � @P

@z

� �

Akrf Pex

mf D; ð14Þ

where k denotes the permeability of the overlain sediment,mf is the fluid viscosity and D is the depth of the top of thedissociation area below the seafloor. Note that thesimplification for the right-hand side of equation (14) ismade based on the fact that in most circumstances, pressureequilibrium reestablishes itself in accordance to anyintroduced change in a relatively short time period. SeeFigures 5 and 6 of Xu [2004] for examples. The pressurereestablishment process may be slower for very lowpermeability sediments. In that case, gas hydrate dissocia-tion and pore pressure increase can be viewed approxi-mately as occurring in confined space.[17] Combining equations (13) and (14) yields

kfDDdPex

dtþ k

DmfPex ¼ �RvfDD

dSh

dt; ð15Þ

which can be rewritten as

dPex

dtþ aPex � b ¼ 0;

a ¼ k

kmf DfDD;

b ¼ �Rv

kdSh

dt:

ð16Þ

By assuming a constant dissociation rate, the solution ofequation (16) is

Pex ¼b

a1� e�atð Þ: ð17Þ

Solution (17) indicates that Pex increases with a character-istic time 1/a and its maximum is

Pexmax ¼b

a¼ �

mf RvDfDDk

dSh

dt: ð18Þ

[18] The magnitude of excess pore pressure Pex resultingfrom gas hydrate dissociation is proportional to the factor ofvolume increase Rv, the rate of dissociation fDDdSh/dt andthe depth below seafloor D, and inversely proportional tothe permeability k. Figure 4 plots the excess pore pressurePexmax as a function of initial pore pressure and a groupedparameter (fDDDdSh/dt) calculated for k = 10�16 m2 andmf = 10�3 kg m�1 s�1. Effect of permeability on Pex max isthe most significant (Figures 5 and 6). Unless explicitlyspecified otherwise, these values of k and mf for Figure 4 areused for the calculations thereafter. Note that rf and mf are ingeneral the density and viscosity, respectively, of a fluidmixture of free gas and liquid solution. However, if thevolume fraction of free gas is sufficiently small, say a few

Figure 3. Excess pore pressure caused by dissociation ofgas hydrates (initially 20% of pore space) is in confined,initially gas-free pore space plotted as a function of theinitial pore pressure and the amount of dissociated gashydrate expressed as the volume fraction of pore space.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

5 of 12

B01104

percent of the pore space, which is likely true within deeperparts of most natural gas hydrate systems, then it is probablysafe to assume that free gas released from the dissociatinggas hydrates tends to stay in the pores as individual gasbubbles [Henry et al., 1999] and the only mobile phase isliquid solution. Consequently, rf and mf can be approximatedas those of liquid water, which do not vary much withinnatural gas hydrate systems. In this case, only the methanedissolved in liquid water is transported into the overlainsediments. Since methane hydrate formation via this type ofprocesses is usually slow [Rempel and Buffett, 1997; Xu andRuppel, 1999] compared to the timescales of the processes

considered here, the effect of methane hydrate formation inoverlain sediments may be neglected.[19] If either a or b varies with time, then equation (16)

usually does not have a simple analytical solution and theexcess pore pressure at t = t0 + Dt may be approximatelycalculated for a sufficiently small time step Dt

Pex ¼2� aDtð ÞP0

ex þ 2bDt

2þ aDt; ð19Þ

where Pex0 is the excess pore pressure at time t = t0. Solution

(19) reveals that Pex decays with time when the dissociationrate decreases so that b becomes small or even reduces tozero after the dissociation ends.[20] Above analyses are done for excess pore pressure

resulting from a given rate of gas hydrate dissociation. Therate of dissociation is further investigated for gas hydratedissociation in marine sediments due to various geologicalor operational processes.

4. Examples of the Excess Pore Pressure

4.1. Due to Continuous Sedimentation At Seafloor

[21] Continuous sedimentation at the seafloor causes gashydrate near the base of hydrate stability zone to movedownward with the host sediment and start dissociatingwhen the stability boundary is reached. We want to knowthe magnitude of excess pore pressure caused by hydratedissociation after the whole process has reached a steadystate. Assuming that DSh amount of gas hydrates dissociatescompletely in time period Dt while getting buried anadditional depth DD, which is the thickness of the dissoci-ation layer with an average gas hydrate volume fraction ofapproximately DSh/2, and according to equation (18), theexcess pore pressure is

Pex mf RvDfvsDSh

k; ð20Þ

where vs = DD/Dt is the rate of sedimentation. For example,at a sedimentation rate of vs = 1 m kyr�1, f = 0.5, k =

Figure 4. Excess pore pressure caused by dissociation ofgas hydrates (initially 20% of pore space) sediments withinterconnected pores. (left) Maximum excess pore pressureplotted as a function of the initial pore pressure and agrouped parameter including the rate of dissociation. (right)Time-dependent excess pore pressure calculated for initialpore pressure of 20 MPa, (fDDDdSh/dt) = 2 � 10�8 m s�1,and dSh/dt = 4 � 10�10 s�1 is also plotted. Calculations aredone for permeability k = 10�16 m2.

Figure 5. Same as Figure 4, but for k = 10�17 m2.

Figure 6. Same as Figure 4, but for k = 10�18 m2.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

6 of 12

B01104

10�16 m2, D = 100 m and DSh = 1, (fDDDdSh/dt) <10�9 m2 s�1, and Pex can be found approximately onFigure 4 to be <0.01 MPa, except for shallow waterdepths.

4.2. Due to Sea Level Dropping or Tectonic Uplifting

[22] In cases of sea level fall or tectonic uplift, the porepressure in sediment in general decreases and the excesspore pressure is thus calculated relative to the decreasingpore pressure as if no gas hydrate dissociation is involved

Pex ¼ P � P0 � rf gDD� �

; ð21Þ

where P0 is the pore pressure before the sea level droppingor tectonic uplifting at t = 0, DD is the magnitude of the sealevel dropping or tectonic uplifting at time t and g is thegravitational acceleration.[23] The dissociation of gas hydrates near the base of the

hydrate stability zone takes certain amount of heat. Exceptfor situations where fluid flow is relatively focused and fast,heat transport within natural gas hydrate systems is usuallydominated by thermal conduction. Consequently, theamount of heat consumed by gas hydrate dissociation issupplied via the cooling of the dissociation area and thermalconduction to the dissociation front. As the dissociationregion cools, heat flow from surrounding sediments pro-vides the heat consumed by further hydrate dissociation.This heat balance can be described as

�rhDHdisfaDDdSh

dt¼ �l

DT

aDD� caDD

dT

dt; ð22Þ

where DHdis is the heat of gas hydrate dissociation, l and care the effective thermal conductivity and the effective heatcapacity, respectively, both consisting of contributions fromthe individual phases and the host sediment, averaged overthe dissociation layer. For example, the c can be expressedas

c ¼ f Shrhch þ Sgrgcg þ Slrlcl� �

þ 1� fð Þrscs; ð23Þ

where rs and cs are the density and the specific heat,respectively, of the sediment host, and rh, ch, Sh, rg, cg, Sg,and rl, cl and Sl are the density, the specific heat and thevolume fraction of gas hydrate, free gas and liquid solution,respectively. The thickness of the dissociation layer is aDD,where a can be calculated in accordance with the pressureand temperature regimes of the dissociation area. Insertingequation (15) into equation (22) results in

rhDHdis

Rv

kfaDDdPex

dtþ k

DmfPex

" #¼ �l

DT

aDD� caDD

dT

dt:

ð24Þ

DD is proportional to the time since the geologic eventbegan

DD ¼ vsdt; ð25Þ

where vsd is the rate of sea level drop or tectonic uplift.Since the phase transition occurs along the phase boundary,

the changes in temperature can be expressed approximatelyas

DT ¼ dTe

dPP � P0ð Þ ¼ dTe

dPPex � rf gDD� �

dT

dt¼ dTe

dP

dP

dt¼ dTe

dP

dPex

dt� rf gvsd

� �:

ð26Þ

Substituting equations (26) into equation (24) gives

dPex

dtþ C1

tþ C2

t2

� �Pex � C3 þ

C4

t

� �¼ 0;

C1 ¼ rhDHdisk

,aDmf vsd rhDHdiskfþ Rvc

dTe

dP

� �� �;

C2 ¼ RvldTe

dP

,a2v2sd rhDHdiskfþ Rvc

dTe

dP

� �� �;

C3 ¼ Rvrf gvsdcdTe

dP

,rhDHdiskfþ Rvc

dTe

dP

� �;

C4 ¼ Rvlrf gdTe

dP

,avsd rhDHdiskfþ Rvc

dTe

dP

� �� �:

ð27Þ

The solution to equation (27) is

Pex ¼Z t

0

C3 þC4

x

� �x

t

� �C1

expC2

t� C2

x

� �dx: ð28Þ

This excess pore pressure is plotted as a function of timeand initial pore pressure in Figure 7. Its magnitude isnegligible when the initial pressure and hence the waterdepth is large compared to that in shallower waterenvironments. In the context of paleoclimate, a sea levelfall is usually accompanied with a cooling at seafloor.However, such a cooling would tend to stabilize themethane hydrate at a different timescale. The characteristictimescale of heat conduction is D2/l/(rc), with l/(rc) �10�6 m2 s�1 for typical marine sediments. For instance,if the undisturbed base of methane hydrate stability zone isD � 100 m below the seafloor, a thermal signal would need�300 years to have a considerable effect on hydratestability at that depth. Therefore, when the base of methanehydrate stability is located at depths on the order of hundredsof meters or greater, the excess pore pressure due to the sealevel fall would stabilize well before the cooling signalwould reach the same depth. On the other hand, for a muchshallower depth of the base of methane hydrate stability, thethermal effect can be significant. A quantitative analysis ofthe effect of a seafloor cooling or warming is mathematicallymore complicated and is not dealt with in this study.

4.3. Caused by Heating

[24] Heating is one of the primary methods proposed forextracting natural gas from natural gas hydrate reservoirs.The basic idea is to induce gas hydrate dissociation byheating and collect the natural gas released by the dissoci-ating gas hydrates. To a certain degree, gas hydrate disso-ciation caused by a rapid increase in seafloor temperaturemay also be similarly considered. However, it requires amuch more complicated analysis and is beyond the scope ofthis study.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

7 of 12

B01104

[25] For simplicity, here we consider a horizontal heatingplane within or immediately below a gas hydrate layer suchthat the whole system can be viewed as one-dimensionalvertically. A heating rate of Qe is provided to dissociate gashydrates in a layer of DD thick while heating up thehydrate-bearing sediment within and near the dissociationarea, namely,

Qe ¼ cDDdT

dtþ lDT

DD� rhDHdisfDD

dSh

dt: ð29Þ

Assuming thermodynamic equilibrium and neglectingsalinity changes, temperature T is a function of porepressure within the hydrate dissociation area. Therefore

Qe ¼ cDDdTe

dP

dPex

dtþ lDD

dTe

dPPex � rhDHdisfDD

dSh

dt: ð30Þ

Combining equation (30) with equation (15) leads to a first-order ordinary differential equation of Pex in the same formof equation (16) but with differing coefficients

dPex

dtþ aHPex � bH ¼ 0;

aH ¼ 1

DD

RvlDD

dTe

dPþ krhDHdis

Dmf

!,Rvc

dTe

dPþ fkrhDHdis

� �;

bH ¼ QeRv

DD

,Rvc

dTe

dPþ fkrhDHdis

� �:

ð31Þ

The solution of equation (31) is of the same form of solution(17) and the maximum excess pore pressure is

Pexmax ¼bH

aH¼ Qe

,lDD

dTe

dPþ krhDHdis

DRvmf

!: ð32Þ

[26] Figure 8 plots the maximum excess pore pressureestimated using equation (32) as a function of the heatingrate Qe. For an order of magnitude estimation, using k �10�16 m2, Rv � 1, D � 102 m, l � 1 W m�1 K�1, DD � 10m, c � 106 J (m3 K)�1 and dTe/dP � 1 K MPa�1, thecharacteristic time 1/aH is �106 s. Therefore a couple ofweeks after the start of heating, the increase in pore pressureslows down as the excess pore pressure approaching its finalmaximum value. For relatively high permeabilities (k >10�16 m2), the second term of the denominator in equation(32) is usually larger than the first term, and therefore

Pexmax Rvmf DQe

krhDHdis

: ð33Þ

Compared to the maximum excess pore pressure (18), it isapparent that the rate of gas hydrate dissociation is now

�fDDdSh

dt¼ Qe

rhDHdis

: ð34Þ

Thus the magnitude of excess pore pressure may also beestimated from solution (18) using the rate of gas hydratedissociation calculated according to equation (34). Whenthe permeability k is low (k < 10�18 m2) or DD is smallduring the initial stage of gas hydrate dissociation, the firstterm of the denominator may be larger than the second oneand hence

Pexmax Qe

,lDD

dTe

dP

� �: ð35Þ

Again, this maximum excess pore pressure can be foundusing solution (18), but with

�fDDdSh

dt¼ QekDD

DRvmf l dTe=dPð Þ : ð36Þ

4.4. Caused by Inhibitor Injection

[27] Gas hydrate inhibition is another method proposedfor introducing gas hydrate dissociation via injection ofalcohols or glycols to destabilize of gas hydrate. Forexample, addition of 10 wt% methanol can lead to a�5�C decrease in stability temperature of pure methanehydrate at any given pressure, or an increase of severalmegapascals in stability pressure at a given temperature[Sloan, 1998]. Considering an inhibitor injection within alayer of thickness DD near the base of gas hydrate stability,as gas hydrates in contact with the inhibitor start todissociate spontaneously, local temperature decreases whilepressure increases as the result of dissociation. Assume thatthe rate of gas hydrate dissociation may be described by[Kamath, 1984]

�rhfDDdSh

dtffi rAs T � Teð Þ2

r ¼ 2:227� 10�4 kg= m2K2s� �

;

ð37Þ

where As is the total surface area of dissociation interfacesbetween hydrates and water within the vertical column witha unit section area and Te is the lowered stability

Figure 7. Excess pore pressure, as defined by equation(21), caused by dissociation of gas hydrates (initially 20%of pore space) resulting from a sea level dropping or atectonic uplifting of 30 m kyr�1 plotted (left) as afunction of the time and the initial pore pressure and(right) as a function of time for an initial pore pressure of20 MPa.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

8 of 12

B01104

temperature due to the addition of inhibitor. Althoughequation (37) could be overly simplified and may or maynot be applied to the field, it is still worth to using it as anexample to demonstrate how the method developed in thisstudy may be applied to problems in industrial operations.[28] Since the dissociation of gas hydrates takes place

away from the three-phase equilibrium, it is difficult torelate the change in temperature to the excess pore pressure.However, we can directly solve equation (15) for Pex,assuming k may be calculated approximately using (8),and equation (22) for T separately and numerically (similarto equation (19)) using equation (37) for the rate of hydratedissociation. At time t = t0 + Dt the solutions are

Pex ¼2� aIPDtð ÞP0

ex þ 2bIPDt

2þ aIPDt;

aIP ¼ k

kmf DfDD;

bIP ¼ RvrAs T � Teð Þ2

rhkfDD;

ð38Þ

T ¼ 2� aITDtð ÞT0 þ 2bITDt

2þ aITDt;

aIT ¼ l

c DDð Þ2;

bIT ¼ aITT0 �DHdisrAs T � Teð Þ2

cDD;

ð39Þ

where T0 and T0 are temperatures at t = 0 and t = t0,respectively. Figure 9 shows the excess pore pressure and thetemperature calculated for an initial pressure of P0 = 20 MPaand a Te0 � T0 = �1�C shifting of gas hydrate stabilitytemperature due to the injection. Initially the dissociation ofgas hydrates causes a rapid increase in pore pressure and a

rapid decrease in temperature. Consequently, the tempera-ture decrease and an increase in gas hydrate stabilitytemperature resulting from the rising pore pressure lead to adecrease in the rate of dissociation as defined by equation(37). At some point in time, the dissociation rate becomes solow that the pore pressure starts to decrease. Eventually, theexcess pore pressure approaches its steady state value,which is small since the heat needed for gas hydratedissociation is solely supplied by a slow conductive heatingby the surrounding sediment.[29] Heating may be applied together with inhibitor

injection to enhance the recovery of natural gas from gashydrates. In this case, instead of equation (22) as usedbefore, equation (29) is used to solve for the temperatureand the solution becomes

T ¼ 2� aITDtð ÞT0 þ 2bITDt

2þ aITDt;

aIT ¼ l

c DDð Þ2;

bIT ¼ aITT0 þQe � DHdisrAs T � Teð Þ2

cDD:

ð40Þ

Solution (40) reduces to equation (39) when Qe = 0. Theeffect of an additional heating of Qe = 5 W/m2 isdemonstrated in Figure 10. Because of the heating, theinitial decrease in temperature due to a rapid dissociation iseventually reversed and the pore pressure continues toincrease until reaches its final steady state.

5. Implications to Marine SedimentDeformation or Failure

[30] The calculations so far have shown that gas hydratedissociation induced by various geological or operational

Figure 8. Excess pore pressure caused by dissociation ofgas hydrates (initially 20% of pore space) resulting fromheating from a horizontal plane plotted (left) as a function ofthe rate of heating and the initial pore pressure and (right) asa function of time for a heating of 0.5 W m�2 and initialpore pressure of 20 MPa.

Figure 9. (left) Excess pore pressure and (right) tempera-ture variation caused by dissociation of gas hydrates(initially 20% of pore space) resulting from inhibitorinjection plotted as a function of time. The dashed linesare the temperatures of gas hydrate stability, and the dottedlines are the temperatures and the excess pore pressures atthe corresponding steady states.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

9 of 12

B01104

processes may lead to excess pore pressure in marineenvironments. The magnitude of the excess pore pressurevaries for different processes that cause the dissociation ofgas hydrate and depends on the properties of the marinesediment, particularly the permeability (Figures 4 to 6). Ingeneral, the excess pore pressure is high in sediments withlow permeability (lower than, say, k = 10�18 m2) andreaches it maximum in the extreme case of a confined porespace.[31] Sufficiently elevated excess pore pressure can lead

to growth of cracks or decementation and even liquefac-tion, which may form a weak or unstable zone within thehost sediment. Still higher excess pore pressure, such asthat in nearly confined pore spaces, may cause rapid andextensive mechanical failure of the sediment. These effectsmay be related to events of marine sediment deformationor failure observed over geologic history. Below weprovide a brief discussion of potential geologic consequen-ces of this type of excess pore pressure. Detailed analysisand quantification of failure mechanisms of marine sedi-ment related to gas hydrate dissociation are beyond thescope of this study and will be dealt with in anotherpublication.

5.1. Submarine Landslides

[32] Gas hydrate dissociation tends to take place near thebase of hydrate stability zone within natural gas hydratesystems. The excess pore pressure due to gas hydratedissociation can lower the effective stress and hence thestrength of sediment along the base of hydrate stabilityzone. If hydrate-bearing pores are well connected and thepermeability of the host sediment is not too low (k � 10�16

m2 or higher), the magnitude of excess pore pressure isusually moderate and by itself is unlikely to trigger, butcould facilitate, a submarine landslide. When sedimentpermeability is sufficiently low (k < 10�18 m2) or if poresin sediment with k > 10�18 m2 are already overpressured towithin several megapascals from lithostatic before gashydrate dissociation takes place due to, say, underconsoli-dation of the sediment, excess pore pressure resulting

from gas hydrate dissociation may be able to trigger alandslide.[33] Submarine landslides can reach lengths of �100 km,

with a length-to-thickness ratio as large as �1000. Puzrinand Germanovich [2005] reasoned that gas hydrate disso-ciation may form an initial weakened zone approximatelyparallel to the seafloor extending subhorizontally from tensof meters up to �1 km. They explained the evolution of alandslide on submarine slope by a catastrophic shear bandpropagation of this flaw. Our calculations suggest that thegrowth of fractures and sediment decementation or lique-faction resulting from the excess pore pressure due to gashydrate dissociation may indeed have lead to submarinelandslide events.[34] Figures 3 to 10 indicate that the magnitude of excess

pore pressure decreases rapidly with increasing initial porepressure and hence the seafloor depth. Therefore this type ofsediment weakening is more relevant to marine gas hydratesystems with a shallow seafloor depth of hundreds ofmeters. The lower boundary of hydrate-bearing sedimentin these systems is usually no deeper than several tens ofmeters. Consequently, this type of gas-hydrate-dissociation-related submarine landslides should most frequently takeplace in shallow water environments (less then 1000 metersof water depth) and involve a relatively thin sediment layer(usually �100 m or less). The thickness of the landslidesediment layer tends to be proportional to the depth ofseafloor.

5.2. Soft Sediment Deformation and VerticalColumns of Focused Flow and Transport

[35] As excess pore pressure builds up due to continuousgas hydrate dissociation, the host sediment within thedissociation area can be liquefied. In some circumstancesthe required pore pressure may be significantly lower thanthe lithostatic pressure. Since the horizontal stress sus-tained by sediment is usually smaller than the verticalstress due to overburden pressure, a partial liquefaction ofsediment occurs first when the increasing pore pressurereduces the effective horizontal stress to zero. On the otherhand, the dissociation of gas hydrates residing in porespaces can also lead to an interpore shear stress becausethe excess pore pressure is essentially nonuniformly distrib-uted. This may result in localized deformation or failure ofthe host sediment. Sediment that has undergone such aninternal disturbance has consequently a much lower shearstrength compared to the undisturbed sediment, and the resultcan be very similar to sediment liquefaction. The soft-sediment deformation observed by Kennett and Fackler-Adams [2000] might be related to this process.[36] The same excess pore pressure may also initiate

hydraulic fractures above the dissociation area [Zuhlsdorffand Spiess, 2004] and, since the area of gas hydratedissociation often has a lower density than the surroundingsediment, the liquefied sediment material tends to fill andfurther pressurize the fracture. In this scenario, the pressuredecrease due to fracturing further enhances gas hydratedissociation and fluid supply. Furthermore, such fracturesare likely to migrate along closely spaced trajectories.Eventually, a quasi-vertical elongated region of disturbedsediment forms and appears on seismic profiles as a verticalcolumn and as a pockmark at the seafloor. The gas-charged

Figure 10. Same as Figure 9, but with an additionalheating of Qe = 5 W m�2.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

10 of 12

B01104

soft sediment might provide a mechanism of gravitationalinstability that leads to the formation of giant hummocks[Davies et al., 1999]. Note that a temporary liquefaction ofsediment caused by traveling seismic waves is probablyirrelevant in this case since the duration of liquefaction isprobably too short.

5.3. Seafloor Collapse

[37] Compared to that in the case of well-connectedsediment pores, excess pore pressure resulting from disso-ciation of gas hydrates in confined space can be as high asseveral tens of megapascals (Figure 3). For instance, such alarge excess pore pressure may build up underneath ahorizontally extended massive gas hydrate layer, whichserves as a seal or caprock of the dissociation zone. Thedissociation of gas hydrate taking place underneath such alayer may be viewed as occurring in a more or less confinedenvironment and, in certain circumstances, may result in avery high excess pore pressure that is sufficient to overcomethe load of the overlying sediment. For instance, the rangeof pressures near the BSR at the Blake Ridge collapse is�33 MPa hydrostatic to �41 MPa lithostatic. A dissocia-tion of gas hydrates by just �1% of the pore space wouldlead to a �10 MPa excess pore pressure and a pore pressurethat exceeds the lithostatic pressure.

[38] Acknowledgments. This study is funded by NSF (USA) grantOCE-0242163. W.X. acknowledges support of the Chinese Academy ofSciences (project KZCX3-SW-224) and NSF (China) grants 40074033and 40274026. L.N.G.’s work is also supported by NSF (USA) grantCMC-0421090. Constructive comments by Ben Clennell, an anonymousreviewer, and the Associate Editor William Waite helped improve themanuscript significantly.

ReferencesClennell, M. B., M. Hovland, J. S. Booth, P. Henry, and W. J. Winters(1999), Formation of natural gas hydrates in marine sediments: 1. Con-ceptual model of gas hydrate growth conditioned by host sediment prop-erties, J. Geophys. Res., 104, 22,985–23,003.

Clennell, M. B., A. G. Judd, and M. Hovland (2000), Movement andaccumulation of methane in marine sediments: Relation to gas hydratesystems, in Natural Gas Hydrate in Oceanic and Permafrost Environ-ments, edited by M. D. Max, pp. 105–122, Springer, New York.

Davies, R., J. Cartwright, and J. Rana (1999), Giant hummocks in deep-water marine sediments: Evidence for large-scale differential compactionand density inversion during early burial, Geology, 27, 907–910.

Dickens, G. R., J. R. O’Neil, D. K. Rea, and R. M. Owen (1995), Dis-sociation of oceanic methane hydrates as a cause of the carbon isotopeexcursion at the end of the Paleocene, Paleoceanography, 10(6), 965–971.

Dickens, G. R., M. M. Castillo, and J. C. Wlakern (1997a), A blast of gas inthe latest Paleocene: Simulating first order effect of massive dissociationof oceanic methane hydrate, Geology, 25, 259–262.

Dickens, G. R., C. Paull, and P. Wallace (1997b), Direct measurement of insitu methane quantities in a large gas-hydrate reservoir, Nature, 385,426–428.

Dillon, W., J. Nealon, M. Taylor, M. Lee, R. Drury, and C. Anton (2001),Seafloor collapse and methane venting associated with gas hydrate on theBlake Ridge: Causes and implications to seafloor stability and methanerelease, in Natural Gas Hydrates, Occurrence, Distribution and Detec-tion, Geophys. Monogr. Ser., vol. 124, edited by C. K. Paull and W. P.Dillon, pp. 211–233, AGU, Washington, D. C.

Field, M. E. (1990), Submarine landslides associated with shallow seafloorgas and gas hydrates off northern California, AAPG Bull., 74(6), 971–972.

Flemings, P. B., X. L. Liu, and W. J. Winters (2003), Critical pressureand multiphase flow in Blake Ridge gas hydrates, Geology, 31, 1057–1060.

Henry, P., M. Thomas, and M. B. Clennell (1999), Formation of naturalgas hydrates in marine sediments 2: Thermodynamic calculations ofstability conditions in porous sediments, J. Geophys. Res., 104,23,005–23,022.

Holbrook, W. S., H. Hoskins, W. T. Wood, R. A. Stephen, and D. Lizarralde(1996), Methane hydrate and free gas on the Blake Ridge from verticalseismic profiling, Science, 273, 1840–1843.

Hornbach, M. J., D. A. Saffer, and W. S. Holbrook (2004), Criticallypressured free-gas reservoirs below gas-hydrate provinces, Nature, 427,142–144.

Kamath, V. A. (1984), Study of heat transfer characteristics during disso-ciation of gas hydrates in porous media, Ph.D. thesis, Univ. of Pittsburgh,Pittsburgh, Pa.

Kayen, R. (1988), Arctic ocean landslides by sea level fall-induced gashydrate decomposition, M.S. thesis, Calif. State Univ., Hayward.

Kayen, R. E., and H. J. Lee (1991), Pleistocene slope instability of gashydrate-laden sediment on the Beaufort Sea margin, Mar. Geotechnol.,10, 125–141.

Kennett, J. P., and B. N. Fackler-Adams (2000), Relationship of clathrateinstability to sediment deformation in the upper Neogene of California,Geology, 28, 215–218.

Kennett, J. P., K. G. Cannariato, I. L. Hendy, and R. J. Behl (2003),Methane Hydrates in Quaternary Climate Change: The Clathrate GunHypothesis, 224 pp., AGU, Washington, D. C.

MacDonald, G. J. (1990), Role of methane clathrates in past and futureclimates, Clim. Change, 16, 247–281.

McIver, R. D. (1982), Role of naturally occurring gas hydrates in sedimenttransport, AAPG Bull., 66(6), 789–792.

Mienert, J., M. Vanneste, S. Bunz, K. Andreassen, H. Haflidason, andH. P. Sejrup (2005), Ocean warming and gas hydrate stability on themid-Norwegian margin at the Storegga Slide, Mar. Pet. Geol., 22, 233–244.

Nisbet, E. (1990), The end of the ice-age, Can. J. Earth Sci., 27, 148–157.Paull, C., W. J. Buelow, W. Ussler, and W. S. Borowski (1996), Increasedcontinental-margin slumping frequency during sea-level lowstands abovegas hydrate- bearing sediments, Geology, 24, 143–146.

Paull, C. K., P. G. Brewer, W. Ussler, E. T. Peltzer, G. Rehder, andD. Clague (2003), An experiment demonstrating that marine slumpingis a mechanism to transfer methane from seafloor gas-hydrate depositsinto the upper ocean and atmosphere, Geo Mar. Lett., 22(4), 198–203.

Puzrin, A. M., and L. N. Germanovich (2005), The growth of shear bandsin the catastrophic failure of soils, Proc. R. Soc. Math. Phys. Eng. Sci.,461(2056), 1199–1228.

Rempel, A., and B. A. Buffett (1997), Formation and accumulation of gashydrate in porous media, J. Geophys. Res., 102, 10,151–10,164.

Rothwell, R. G., J. Thomson, and G. Kahler (1998), Low-sea-level empla-cement of a very large Late Pleistocene ‘megaturbidite’ in the westernMediterranean Sea, Nature, 392, 377–380.

Setzmann, U., and W. Wagner (1991), A new equation of state and tables ofthermodynamic properties for methane covering the range from the melt-ing line to 625 K at pressures up to 1000 MPa, J. Phys. Chem. Ref. Data,20, 1061–1151.

Singh, S. C., and T. A. Minshull (1994), Velocity structure of a gas hydratereflector at Ocean Drilling Program site 889 from a global seismic wave-form inversion, J. Geophys. Res., 99, 24,221–24,234.

Sloan, E. D., Jr. (1998), Clathrate Hydrates of Natural Gases, 2nd ed., rev.and expanded, 754 pp., CRC Press, Boca Raton, Fla.

Sultan, N., P. Cochonat, J. P. Foucher, and J. Mienert (2004), Effect of gashydrates melting on seafloor slope instability, Mar. Geol., 213(1 –4),379–401.

Van Rensbergen, P., M. De Batist, J. Klerkx, R. Hus, J. Poort, M. Vanneste,N. Granin, O. Khlystov, and P. Krinitsky (2002), Sublacustrine mudvolcanoes and methane seeps caused by dissociation of gas hydrates inLake Baikal, Geology, 30, 631–634.

Vogt, P. R., K. Crane, E. Sundvor, M. D. Max, and S. L. Pfirman (1994),Methane-generated(?) pockmarks on young, thickly sedimented oceaniccrust in the Arctic: Vestnesa ridge, Fram strait, Geology, 22, 255–258.

Vogt, P. R., J. Gardner, and K. Crane (1999), The Norwegian-Barents-Svalbard (NBS) continental margin: Introducing a natural laboratory ofmass wasting, hydrates, and ascent of sediment, pore water, and methane,Geo Mar. Lett., 19(1/2), 2–21.

Xu, W. (2002), Phase balance and dynamic equilibrium during formationand dissociation of methane gas hydrate, paper presented at 4th Interna-tional Conference on Gas Hydrates, Cent. for Gas Hydrate Res., Yoko-hama, Japan.

Xu, W. (2004), Modeling dynamic marine gas hydrate systems, Am. Miner-al., 89, 1271–1279.

Xu, W., and C. Ruppel (1999), Predicting the occurrence, distribution, andevolution of methane gas hydrate in porous marine sediments, J. Geo-phys. Res., 104, 5081–5096.

Xu, W., R. P. Lowell, and E. T. Peltzer (2001), Effect of seafloor tempera-ture and pressure variations on methane flux from a gas hydrate layer:Comparison between current and late Paleocene climate conditions,J. Geophys. Res., 106, 26,413–26,423.

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

11 of 12

B01104

Yuan, T., R. D. Hyndman, G. D. Spence, and B. Desmons (1996), Seismicvelocity increase and deep-sea hydrate concentration above a bottomsimulating reflector on the northern Cascadia continental slope, J. Geo-phys. Res., 101, 13,655–13,671.

Zatsepina, O. Y., and B. A. Buffett (1998), Thermodynamic conditions forthe stability of gas hydrate in the seafloor, J. Geophys. Res., 103,24,127–24,139.

Zhang, Y. (2003), Formation of hydrate from single-phase aqueoussolutions, internal report, 77 pp., Univ. of Pittsburgh, Pittsburgh,Pa.

Zuhlsdorff, L., and V. Spiess (2004), Three-dimensional seismic character-ization of a venting site reveals compelling indications of natural hydrau-lic fracturing, Geology, 32, 101–104.

�����������������������L. N. Germanovich, School of Civil and Environmental Engineering,

Georgia Institute of Technology, Atlanta, GA 30332, USA.W. Xu, School of Earth and Atmospheric Sciences, Georgia Institute of

Technology, 311 Ferst Drive, Atlanta, GA 30332, USA. ([email protected])

B01104 XU AND GERMANOVICH: HYDRATE DISSOCIATION AND PORE PRESSURE

12 of 12

B01104