Influence of sedimentation, local and regional ...afisher/CVpubs/pubs/Hutnak... · rapid...

Influence of sedimentation, local and regional hydrothermal

circulation, and thermal rebound on measurements

of seafloor heat flux

M. Hutnak1 and A. T. Fisher1,2

Received 28 February 2007; revised 10 July 2007; accepted 29 August 2007; published 20 December 2007.

[1] We quantify the influence of environmental processes on measurements of seafloorheat flux with a new one-dimensional thermal model that includes time-varyingsedimentation and boundary conditions, represents several scales of hydrothermal andtransient conductive processes within basement, and allows fluid seepage throughaccumulating sediments. Variations in basement thermal conductivity, the extent ofhydrothermal mixing in upper basement, and fluid seepage through sediments eachinfluence seafloor heat flux by �2–8%. Conductive thermal rebound following thecessation of advective heat loss from the crust may lower seafloor heat flux values by�5–10%, even after several to several tens of million years. The new models indicate thatthermal rebound takes much longer than suggested by earlier (analytical) calculations,mainly because earlier models did not account for the heat capacitance of the conductivelithosphere. Application of the new model to 3.5–3.6 Ma seafloor on the eastern flank ofthe Juan de Fuca Ridge suggests that anomalously low heat flux in this area is bestexplained by incomplete conductive thermal rebound in the last 100–200 ka, followingthe burial of numerous basement outcrops. Application of the new model to the easternflank of the East Pacific Rise and extrapolation out to the age of some of the oldestremaining seafloor indicate that sedimentation corrections may be important even whereaccumulation rates are typical of global values. Model results also suggest that conductivethermal rebound through thick sediments may bias measurements made on moderateto old seafloor, even where there is little evidence at present for ridge-flank hydrothermalcirculation.

Citation: Hutnak, M., and A. T. Fisher (2007), Influence of sedimentation, local and regional hydrothermal circulation, and thermal

rebound on measurements of seafloor heat flux, J. Geophys. Res., 112, B12101, doi:10.1029/2007JB005022.

1. Introduction

[2] Measurements of seafloor heat flux elucidate thethermal evolution of oceanic lithosphere, rates and patternsof hydrothermal circulation, the occurrence and condition ofgas hydrates, and the magnitude and timing of volcanic andseismic activity [e.g., Davis et al., 1996; Davis and Lister,1977; Fisher et al., 2003a; Kastner et al., 1991; Lister,1977; Newman et al., 2002; Parsons and Sclater, 1977;Ruppel and Kinoshita, 2000; Sakai et al., 1990; Stein andStein, 1992]. Interpretation of heat flux data is facilitated bycollocation of measurements on bathymetric swath mapsand along seismic reflection profiles, which help to delin-eate relations between heat transport, subseafloor relief andstructure, and sediment thickness [Davis et al., 1989, 1992a,1997b; Johnson et al., 1993; Kimura et al., 1997; Langseth

et al., 1992]. This last approach has been particularlyeffective at sedimented spreading centers and on the flanksof mid-ocean ridges, where the thickness and distribution ofsediments allows spatially detailed heat flux surveys [e.g.,Anderson and Hobart, 1976; Davis and Villinger, 1992;Langseth et al., 1988; Lawver and Williams, 1979].[3] Much as satellite gravity data are compared to the

geoid and deep seismic velocity structures are compared tothe Preliminary Reference Earth Model, seafloor heat fluxdata are most commonly compared to one of several(similar) predictions for heat flux versus age based onone-dimensional lithospheric cooling, as constrained byregional and global compilations of bathymetric and heatflux data [e.g., Lister, 1977; Parsons and Sclater, 1977;Stein and Stein, 1992]. These compilations and analyseshave shown that seafloor heat flux data (1) are highlyscattered and (2) have mean values that generally fall belowconductive lithospheric cooling predictions for crustal agesless than �65 Ma. Both of these observations are generallyinterpreted to result from hydrothermal circulation: the for-mer represents the local redistribution of heat by vigorouscirculation, and the latter results from regional advectiveheat extraction from the plate. The relative consistency of

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112, B12101, doi:10.1029/2007JB005022, 2007ClickHere

for

FullArticle

1Department of Earth and Planetary Sciences, University of California,Santa Cruz, California, USA.

2Institute for Geophysics and Planetary Physics, University ofCalifornia, Santa Cruz, California, USA.

Copyright 2007 by the American Geophysical Union.0148-0227/07/2007JB005022$09.00

B12101 1 of 19

http://dx.doi.org/10.1029/2007JB005022

lithospheric cooling predictions and seafloor observationsfor crust older than 65 Ma is sometimes misinterpreted toindicate that the upper oceanic crust becomes ‘‘sealed’’ tohydrothermal circulation beyond this age; in fact, thevariability in observed surface heat flux at many older sitesdemonstrates continuing, thermally significant hydrother-mal circulation within basement [e.g., Von Herzen, 2004].[4] As more high-resolution data are collected, subtle

variations in seafloor heat flux can help to resolve a rangeof dynamic processes. Variations that once would have beendismissed as experimental noise now justify more rigorousinterpretation, but only after application of accurate correc-tions for the influence of sedimentation, fluid seepage, andother environmental processes. Earlier studies have quanti-fied sedimentation corrections to seafloor heat flux on thebasis of idealizations that are generally most appropriate formoderate to old seafloor or sedimented spreading centers,including lower boundary conditions (Dirichlet or Neumann)held constant with time, and purely conductive, steady stateconditions within basement below accumulating sediments[e.g., Benfield, 1949; Hutchison, 1985; Wang and Davis,1992]. We have developed a thermal model that can be usedto calculate sedimentation corrections while allowing tem-poral variability of boundary conditions. The new modelrepresents several scales of hydrothermal and transientconductive processes within basement below sediments,and permits assessment of the influence of fluid seepage

through accumulating sediments; all of these environmentalconditions are common to ridge flanks and can influencemeasurements of seafloor heat flux (Figure 1).[5] Quantitative evaluation of environmental corrections

to seafloor heat flux data from young ridge flanks isimportant for several reasons. First, heat flux data fromyoung seafloor are often influenced by the cooling effects ofrapid sedimentation, particularly where high sedimentationrates produce sufficiently thick and broadly distributedsediment so as to allow detailed surveys. Although conven-tional lithospheric cooling models can be used to predictseafloor heat flux in these settings, observational thermaldata from young sites are typically omitted from theanalyses used to constrain key cooling parameters becausethese data are often influenced by hydrothermal processes.Thus the extent of heat flux anomalies on young sites isoften ambiguous.[6] We present heat transport calculations using the new

model to explore conditions common to ridge flanks acrossa range of ages and sedimentation rates. After describingdesign and operation of the model, we present a series ofsensitivity analyses to evaluate the relative influence ofdistinct environmental processes and conditions: sedimen-tation, differences in basement thermal conductivity, localand regional hydrothermal circulation (including the cessa-tion of circulation), and fluid seepage through accumulatingsediments (Figure 1). We apply the model to two field

Figure 1. Conceptual model showing the effects of sedimentation, fluid seepage, regional advectiveheat extraction, and local heat redistribution by vigorous convection in permeable basement. The plotabove each cartoon shows relative heat flux (ratio of observed to basal) versus time that would bemeasured at locations shown. (a) Basement outcrops are abundant, and sedimentation, advective heatextraction, and local convection in permeable basement suppress surface heat flux on a regional basis.Heat flux would be higher adjacent to discharging outcrops and lower adjacent to recharging outcrops[e.g., Fisher et al., 2003a]. (b) Following accumulation of sufficient sediment, many outcrops becomeburied and the regional advective extraction of lithospheric heat ends. Upper basement and the overlyingsediment section rebound thermally as sediments accumulate.

B12101 HUTNAK AND FISHER: SEAFLOOR HEAT FLUX

2 of 19

B12101

examples representing a broad range of sediment accumu-lation rates (10–500 m/Ma): (1) young (up to 3.6 Ma)seafloor of the Juan de Fuca plate, east of the Juan de FucaRidge, northeastern Pacific Ocean, and (2) older (up to24 Ma) seafloor of the Cocos Plate east of the East PacificRise, central Pacific Ocean. In the latter analyses, weextrapolate the influence of continuing sedimentation andthe cessation of regional hydrothermal heat extraction out tothe age of some of the oldest remaining seafloor, to evaluatehow these processes may influence measurements of sea-floor heat flux more generally. The numerical model craftedfor this study is coded in MATLAB and is available, alongwith a user guide and example files, at http://es.ucsc.edu/�afisher/Research/Appen/SlugSed.

2. Model of Sedimentation and Heat Transport

2.1. Physical Basis

[7] In the absence of fluid seepage, steady state conduc-tive heat transport through a sedimented ridge flank gener-

ally results in a thermal gradient that decreases with depth,because deeper sediments usually have lower porosity andhigher thermal conductivity than shallower sediments(Figure 2). The thermal conductivity of upper oceanicbasement also tends to increase with depth, but the varia-tions are often associated with abrupt lithologic transitions,for example, from extrusive basalt pillows to more massiveflows or sills [e.g., Bartetzko et al., 2001; Busch et al., 1992;Karato, 1983]. Rapid sedimentation results in the heat fluxthrough sediments and upper basement being suppressed,with the magnitude and depth extent of suppression depend-ing on the sedimentation history (magnitude and duration)and on the thermal and physical properties of sediment andbasement. Upward fluid seepage through sediments causesthe thermal gradient to be elevated in the shallow subsurface(Figure 2), but seepage through sediments at thermallysignificant rates is generally limited to areas with relativelythin sediment cover (tens to a few hundreds of meters,depending on sediment type), on the basis of the consider-ation of typical driving forces and sediment properties[Mottl and Wheat, 1994; Spinelli et al., 2004b; Wheat andMottl, 2004].[8] Analytical studies of sedimentation and heat transport

are often based on a semi-infinite half-space with constantheat flux at depth and a moving upper boundary thatrepresents accumulating sediment [e.g., Benfield, 1949;Von Herzen and Uyeda, 1963]. Hutchison [1985] and Wangand Davis [1992] developed one-dimensional numericalmodels using a thick (but finite) plate, most of whichcomprised crustal basement, that ‘‘fell away’’ from theupper boundary, allowing for the deposition of sediments.The sediments consolidated after deposition and burialaccording to a series of property-depth relations. Thisone-dimensional approach is particularly well justified forseafloor that is distant from fluid recharge and dischargesites, where lateral temperature variations are generallymodest. Developing a detailed sedimentation history intwo or three dimensions would require an unusually highdensity of cored boreholes from which accumulation ratesand material physical properties have been determined. Inmost field studies, one is fortunate to have a single nearbyborehole from which detailed paleontological and otherphysical data necessary to develop a sedimentation historyis available.[9] The models developed in the present study comprise a

logical extension of earlier work, allowing several importantenvironmental processes to be simulated simultaneously.The accumulating sediment layer follows a porosity-depthrelation, f(z), with bulk sediment thermal conductivity (l)and permeability (k) depending on porosity (Figure 3). Thisapproach is based on the assumption that sediment porositydecreases with depth only (independent of sedimentationrate or elapsed time), and that the compressibility of porewater and sediment grains is smaller by several orders ofmagnitude than bulk sediment compressibility [Wang andDavis, 1992].[10] The sediment section may be underlain by a crustal

aquifer through which heat transfer is dominated by advec-tion, and one or two basement layers through which heat istransferred conductively (Figure 4a). The use of two con-ductive basement layers allows representation of distinctlithologic regions (having different thermal properties),

Figure 2. Illustration of the influence of fluid seepage andsedimentation on the crustal geothermal gradient, incorpor-ating one sediment layer and three basalt layers, eachhaving different physical properties. Basal heat flux isassumed constant. Solid line shows conductive, steady stateprofile; dotted curve shows heat flux elevation due toseepage through the sediment section; and dash-dottedcurve shows heat flux suppression due to sedimentation.Each of the latter two curves would evolve through time assediments accumulate.


3 of 19

B12101

whereas having a separate basement aquifer allows repre-sentation of several aspects of advective heat transport inupper basement. Hydrothermal circulation can extract sig-nificant quantities of lithospheric heat on a regional basisthrough (dominantly) rapid lateral fluid transport in theupper crust [e.g., Davis et al., 1992a; Fisher and Becker,2000; Langseth et al., 1984]. For the purposes of the presentstudy, we simulate this process by distributing heat sinkswithin the basement aquifer, where the collective magnitudeof the heat sinks removes a desired fraction of lithosphericheat before it can reach the accumulating (overlying)sediment layer.[11] In addition, vigorous convection can homogenize

basement temperatures on a local basis. This processes isrepresented using a high Nusselt number (Nu) approxima-tion [e.g., Davis et al., 1997a; Spinelli and Fisher, 2004],whereby the effective thermal conductivity of the basement

aquifer is increased. Within a two- or three-dimensionalcrustal system, high-Nu convection can elevate the seafloorheat flux above buried basement highs, and lower the heatflux above buried basement lows. Within a one-dimensionalmodel, we simulate only the increase in vertical heattransport associated with vigorous, local convection.Regional heat extraction and local heat redistribution canoccur simultaneously on ridge flanks within a ‘‘well-mixed’’ aquifer [e.g., Davis et al., 1999; Langseth andHerman, 1981; Rosenberg et al., 2000; Stein and Fisher,2001], and both processes can be represented independentlywith the new model.[12] Hydrothermal circulation can also influence seafloor

heat flux immediately adjacent to basement outcrops thatare fluid recharge and discharge sites [e.g., Fisher et al.,2003a; Hutnak et al., 2006; Lonsdale and Becker, 1985;Villinger et al., 2002], but this effect is not included in thepresent study. Instead, we focus on heat transport processescommon to ridge-flank areas relatively distant fromrecharge and discharge sites, where the one-dimensionalapproximation is most applicable (Figure 1).

2.2. Implementation

[13] The numerical model employs a deforming finitedifference grid to model heat transport within a lithosphericplate upon which sediment is added, using mixed Eulerianand Lagrangian reference frames for the sediment andbasement sections, respectively (Figure 4b). As with otherone-dimensional sedimentation models using a deformingmesh to simulate heat transport [e.g., Hutchison, 1985;Lucazeau and Le Douaran, 1985; Wang and Davis,1992], we first determine the kinematics of the massmovement and then solve the heat transfer problem.[14] The model domain evolves with time through the

creation of space at the sediment-basement interface, withdepths referenced to the seafloor (z = 0). Time-dependentsedimentation causes the sediment-basement interface (z =B) to move downward, and the grid near this interface isdeformed and extended through the addition of nodesaccording to a series of user-specified rules.[15] Conservation of mass allows sediment and fluid

velocities to be uniquely determined from the porosity-depth profile. For a pair of adjacent nodes at depths z1and z2 [Wang and Davis, 1992]:

vs z1; tð Þ 1� f z1ð Þ½ � ¼ vs z2; tð Þ 1� f z2ð Þ½ �vw z1; tð Þf z1ð Þ ¼ vw z2; tð Þf z2ð Þ ; ð1Þ

where f is porosity, v is velocity, and the subscripts w and sdenote the water and sediment phases, respectively.Sediment and fluid velocities associated with depositionand compaction vary spatially and temporally on the basisof user-specified rates of sediment accumulation and/ordeposition (Figure 5). In the case of sediment accumulationthe velocity of basement is known: [vb = vs(B, t) = vw(B, t)],leading to direct calculation of basement depth with time[db/dt = vb]. Because of compaction, the velocity ofsediment grains entering the model domain [vo = vs(0, t)]must increase with time in order to keep the model domainfilled. In the case of sediment deposition, the velocity ofsediment entering the model domain is known and B is

Figure 3. Example of a typical relation between sedimentporosity (f), thermal conductivity (l), and permeability (k),as applied with the new model. Porosity decreases withdepth, thermal conductivity is calculated using a geometricmean mixing model, and permeability varies with porosity(Table 1). Curves shown are appropriate for turbidites andhemipelagic mud found on the eastern flank of the Juan deFuca Ridge [e.g., Davis et al., 1997b; Giambalvo et al.,2000; Spinelli et al., 2004b].


4 of 19

B12101

determined by boot strapping along the porosity-depthprofile. In this situation, vb decreases as sediment thicknessincreases.[16] Sediment porosity decreases with depth using a

flexible analytical representation, including (if desired)constant, linear, polynomial, exponential, and logarithmicterms. Because we are particularly interested in conditionson young ridge flanks, some of which are modified soonafter the start of sedimentation, we begin the models with aminimum sediment thickness of 6 m and a minimum nodespacing of 2 m. Bulk thermal conductivity (lb) isdetermined using a geometric mean mixing model, lb =lwf ls

1�f, where the subscripts w and s denote water andsediment phases, respectively. Sediment permeability (k) iscalculated using an empirical analytical relation for theparticular lithology of interest, once again including (ifdesired) constant, linear, polynomial, exponential, andlogarithmic terms [e.g., Spinelli et al., 2004b].[17] Heat transfer is modeled using a standard conduc-

tion-advection equation:

rc@T

@t¼ @

@zlb

@T

@z� rcv

@T

@zþ Q; ð2Þ

where rc is the specific heat capacity (density times specificheat), v is velocity, T is temperature, t is time, z is depth, andQ is heat production or loss rate. The bar above selected

parameters denotes average properties for the fluid-grainmixture. We extend this standard equation by embedding aseparate fluid seepage term (qseep) independent of compac-tion in the conduction-advection equation:

rc ¼ rcwfþ rcs 1� fð Þ;rcv ¼ rcwvwfþ rcwqseep þ rcsvs 1� fð Þ: ð3Þ

It is implicit in this formulation that fluid seepage does notalter the porosity-depth profile, consistent with observationsfrom seepage sites on ridge flanks suggesting that thedriving pressures are modest (on the order of a few tens ofkPa) and there is little associated modification of sedimentcompaction trends [e.g., Giambalvo et al., 2000; Spinelli etal., 2004b; Spinelli and Saffer, 2004]. This representation offluid seepage includes no local fluid storage within thesediment column and permits time-varying fluid seepage atarbitrary (user specified) rates or based on a specifiedsequence of excess fluid pressures in basement.[18] When vertical fluid seepage is driven by an over-

pressure (DP, pressure in excess of hydrostatic), the modelfirst determines the hydraulic impedance, I, of the sedimentcolumn [Karato and Becker, 1983]:

I ¼Z z¼B

0

dz

k zð Þ : ð4Þ

Figure 4. Cartoons illustrating key aspects of the physical and numerical models used to calculate heattransport. (a) Layering of the model domain shows that sediment accumulates on top of a permeablebasement aquifer within which heat is transported by conduction and/or advection and may be extractedby regional hydrothermal circulation. Heat is transported by conduction within the underlying basementlayers. (b) The numerical representation of the physical model uses Eulerian and Lagrangian referenceframes for the sediment and basement sections, respectively. Node depths are represented by horizontaldashed lines, whereas time steps are represented by vertical solid lines. The solid circles on each verticalline denote finite difference node points used for that time step. (c) Node and block properties used innumerical calculations include the following: z is depth, T is temperature, l is thermal conductivity, Q isheat production or removal, vw and versus are the water and sediment velocities, respectively, f isporosity, and k is permeability. The thermal conductivity of adjacent blocks lh is calculated as thethickness-scaled, harmonic mean of that in surrounding nodes.


5 of 19

B12101

The seepage flux, qseep, is calculated according to Darcy’slaw:

qseep ¼ � keff

mDP

B; ð5Þ

where keff = B/I and m is dynamic fluid viscosity. Althoughm and r vary slightly with the range of temperaturestypically encountered on ridge flanks, changes in r arenegligible and we use the minimum temperature within thesediment column to calculate the maximum value m,

minimizing the rate of pressure-driven seepage (as willoccur in the natural system).[19] Basement porosity and thermal properties assigned

to various layers in the models that follow are based onobservations from drilling and associated experiments [e.g.,Bartetzko et al., 2001; Becker et al., 1983; Busch et al.,1992; Jarrard and Broglia, 1991; Karato et al., 1983;Pezard and Anderson, 1989; Shipboard Scientific Party,2005]. The upper 100–1000 m of basement is thought tocomprise the hydrothermal aquifer most responsible foradvective heat loss on ridge flanks, and is assigned thehighest porosity and lowest thermal conductivity (Table 1).Deeper sections of basement are assigned lower porosityand higher thermal conductivity. We explore the sensitivityof calculations to the choice of these and other parameters insection 3.

3. Sensitivity Analyses

3.1. Model Configuration

[20] After verifying that the numerical model replicatedpublished analytical and numerical solutions for steady stateand transient heat transport [Benfield, 1949; Bredehoeft andPapadopulos, 1965; Carslaw and Jaeger, 1959; Wang andDavis, 1992], we completed a series of sensitivity analysesto determine processes and parameters that may be mostimportant in influencing seafloor heat flux. We use aconstant heat flux lower boundary condition for mostsensitivity analyses, but also explore the influence of atime-varying lower boundary condition, which is mostimportant for very young (<1 Ma) sites.[21] We compare results initially for sediment accumula-

tion rates of 100 and 500 m/Ma. This range of accumulationrates is higher than values in much of the deep ocean, but isconsistent with rates found in sedimented spreading centers,rift valleys, and near many continental margins [e.g.,Berger, 1978; Bryant and Bennett, 1988; Emiliani andMilliman, 1966; Hay et al., 1988; Spinelli et al., 2004b].Sediment properties used for all sensitivity analyses arethose documented on the eastern flank of the Juan de FucaRidge, sandy to fine-grained turbidites and hemipelagicmud [Davis et al., 1992b; Underwood et al., 2005]. Section4 of this study includes analyses based on more globallyrepresentative accumulation rates of 10–50 m/Ma.

Figure 5. Relations between sediment accumulation rates(vb) and deposition rates (v0). vb is the velocity of the nodeat the sediment-basement interface, whereas v0 is thevelocity of material entering the domain at the seafloor.Curves illustrate the relative magnitude of v0/vb (in the caseof accumulation) and vb/v0 (in the case of deposition) forrates of 100 and 500 m/Ma, rates used for most of thesensitivity analyses shown in this study. For a constantaccumulation rate (vb), sediment consolidation requires thatthe flux across the seafloor increase with time. For aconstant deposition rate (v0), sediment consolidationrequires that the rate of basement subsidence decrease withtime.

Table 1. Physical Property Relations and Values Used in Thermal Models of Sedimentation and Conductive

Rebound

Model Propertya SedimentsBasementAquifer

ConductiveBasementLayer 1

ConductiveBasementLayer 2

Porosity (f) sensitivity and Juan de Fuca flank:b

f = AeBz, A = 0.7, B = �1/1200East Pacific Rise, Cocos Plate:b

f = CzD, C = 0.909, D = �0.073

0.10 0 0

Thermal conductivity(l), W/m-K

l = lwflg

(1�f)

lw = 0.6, lg = 2.741.7 2.0 2.9

Permeability (k), m2 k = EeF(f (1�f))

E = 3.7 � 10�18, F = 1.7NAc NAc NAc

aSediment and basement properties are based on site-specific and global compilations, with references cited in the text.bFractional porosity is given, with depth in meters.cFluid flow in basement is not explicitly modeled.


6 of 19

B12101

[22] In the discussion that follows, we consider variationsof a parameter to be significant when they result in differ-ences in calculated seafloor heat flux �2%. Differences inmeasured heat flux values smaller than this, with respect toeach other or to a lithospheric reference, are often difficultto resolve on the basis of field measurements. Differenceslarger than 2% could be important particularly if multipleeffects are superimposed.

3.2. Basement Thermal Conductivity

[23] In an initial set of simulations, there is no basementaquifer and the thermal conductivity of the remainingbasement layers is either 2.0 or 2.9 W/m-K. Thermalconductivity values �2.0 W/m-K in the uppermost crustare indicated by rock recovered from seafloor boreholes andinferences from geophysical logging [e.g., Becker et al.,1985; Busch et al., 1992; Shipboard Scientific Party, 2005],whereas higher values for the entire basement layer havebeen used in earlier thermal models of sedimentation [e.g.,Davis et al., 1999; Hutchison, 1985; Wang and Davis,1992]. Varying the thermal conductivity of basement from2.0 to 2.9 W/m-K results in differences in seafloor heat fluxduring sedimentation at the 2–4% level for sediment thick-nesses up to 2000 m (equivalent to crustal ages of 4–20 Mafor the accumulation rates used) (Figure 6). Higher values ofbasement thermal conductivity result in greater heat fluxsuppression during rapid sedimentation, because the thermalinfluence of sedimentation extends to greater depth. Theinfluence of the thermal conductivity of basement is moreimportant to seafloor heat flux values at higher accumula-tion rates.

3.3. Aquifer Thickness and Vigor of Local Convection

[24] Two of the least-constrained physical properties andprocesses within oceanic crust on ridge flanks are the depthextent of hydrothermal circulation, and the ease with whichconvecting fluids redistribute heat locally. The thickness ofthe hydrothermal aquifer is often associated with part or all

of the uppermost (extrusive) basaltic crust, and this layervaries greatly in thickness, generally 100–1000 m [e.g.,Becker et al., 1983; Davis et al., 1996; Fisher, 2004, 1998;Morin et al., 1992; Rohr, 1994; Rohr et al., 1994; Wilson etal., 2005].[25] Vigorous convection in shallow basement can locally

homogenize temperatures within the upper crust on ridgeflanks, despite large differences in basement elevation[Davis et al., 1989, 1999, 1997b; Fisher et al., 1990,2003b; Zuehlsdorff et al., 2005]. Nearly complete homog-enization of upper basement temperatures in these settingsindicates that convection is extremely efficient in redistrib-uting heat on a local basis, implying Nu values of 500 ormore [e.g., Davis et al., 1997c; Hutnak et al., 2006; Spinelliand Fisher, 2004]. We tested a range of Nu = 2 to 104, andshow results for the case when Nu is 1000; values of Nu >1000 yield identical results.[26] Variations in the thickness of the crustal aquifer and

in the efficiency of local convection can result in seafloorheat fluxes that differ by up to 6% when sedimentation rates

Figure 6. Sensitivity of models to the thermal conductiv-ity of basement. Heat flux fraction (seafloor heat fluxrelative to lithospheric input) is shown for basement thermalconductivities of 2.0 and 2.9 W/m-K, with sedimentaccumulation rates of 100 and 500 m/Ma. The selectionof thermal conductivity for basement influences themagnitude of the sedimentation correction to seafloor heatflux at the 2–4% level for the accumulation rates shown.

Figure 7. Sensitivity of sedimentation models to thethickness of the upper basement aquifer (100–1000 m),assuming extremely efficient local redistribution of heat byvigorous convection, as represented with a high-Nu proxy(Nu = 1000). Values of Nu > 1000 yield identical results.The thick solid curveon eachplot is a reference forwhich thereis no well-mixed basement aquifer (i.e., Nu = 1). (a) Heatflux suppression is shown for a sediment accumulation rateof100m/Ma. (b)Heat fluxsuppression is shownforasedimentaccumulation rateof500m/Ma.Local convection inbasementhas the greatest influence when the aquifer is thicker andthe sedimentation rate is greater. In general, the presence of awell-mixed aquifer increases the suppression of seafloor heatflux by sedimentation at the 1–6% level.


7 of 19

B12101

are 100–500 m/Ma (Figure 7). Higher values of Nu and athicker basaltic aquifer lead to larger differences in seafloorheat flux, and the differences are greatest for higher sedi-mentation rates, particularly while the first few hundredmeters of sediment are deposited. When accumulation ratesare �100 m/Ma, model results are virtually insensitive toaquifer thickness and vigor of local heat redistribution.

3.4. Hydrothermal Rebound Following AdvectiveExtraction of Lithospheric Heat

[27] Previous studies have quantified the heating of a one-dimensional slab, initially cooled, with a constant heat fluxlower boundary [e.g., Carslaw and Jaeger, 1959; Fisher,2003; Hobart et al., 1985]. This configuration includessome aspects of oceanic crust that is initially cooled hydro-thermally, after the end of advective heat loss (Figure 8a).However, the configuration of this conductive-reboundproblem lacks two key features common to the ridge-flankenvironment following cessation of hydrothermal cooling.First, oceanic lithospheric includes a thick conductiveboundary layer under the hydrothermally cooled layer.The conductive layer is important because it is also cooledby hydrothermal circulation in the crustal aquifer and acts asa thermal capacitor; the depth and extent of cooling dependson the efficiency of advective heat loss from the aquifer andthe thermal properties of the conductive layer below. Ingeneral, the rebound of seafloor heat flux at the top of ahydrothermally cooled layer takes longer if that layer isunderlain by a thick conductive layer. For example, in theabsence of an underlying conductive layer, rebound ofseafloor heat flux to 90% of the basal value takes 40 ka,whereas the presence of a thick, conductive layer below therebounding layer delays heat flux rebound by >1 Ma(Figure 8a).[28] Second, the presence of a conductive sediment layer

above the rebounding basement layer leads to an additionaldelay in the recovery of seafloor heat flux: a 500 msediment layer delays recovery to 90% of the basal valuefor nearly 4 Ma, and a 1000 m sediment layer delaysrebound to this extent for almost 10 Ma. These calculationshave important implications for interpretation of seafloorheat flux measurements on a global basis, even in caseswhere there is little or no evidence for hydrothermaladvection of lithospheric heat at present. The efficiency ofadvective cooling of the upper oceanic crust across much ofthe seafloor depends primarily on two factors: (1) thepermeability distribution of the crust, particularly withrespect to rapid lateral fluid flow, and (2) the availabilityof seamounts, fracture zones, and other basement outcropsthrough which hydrothermal fluids can enter and exit thecrust [e.g., Davis et al., 1992a; Fisher and Becker, 2000;Villinger et al., 2002]. Although there is evidence for asystematic reduction in upper crustal permeability with age,at least during the first few million years of lithosphericevolution, the extent of the global heat flow anomaly,limitations on fluid driving forces, and the presence ofhydrothermal circulation even within some of the oldestremaining seafloor indicate that permeability remains highthroughout much of the basaltic ocean crust [e.g., Beckerand Fisher, 2000; Davis and Becker, 2004; Fisher andBecker, 2000; Parsons and Sclater, 1977; Von Herzen,2004].[29] At the same time, the progressive burial of basement

outcrops is likely to be highly heterogeneous, both spatiallyand temporally. The elevation of basement outcrops relativeto the surrounding seafloor varies enormously, as do out-crop spacing and sedimentation rates [e.g., Harris et al.,2004; Spinelli et al., 2004b; Wessel, 2001]. Sediments areoften limited initially from accumulating on elevated base-ment outcrops by bottom currents, diagenesis, preferential

Figure 8. Conductive thermal rebound (fractional recov-ery of seafloor heat flux) following cessation of regionaladvective heat extraction. (a) Thermal rebound is shown ofa 1 km section of hydrothermally cooled basement with nosedimentary cover. Dash-dotted line shows rebound (calcu-lated analytically) when constant heat flux is applieddirectly to the base of the hydrothermally cooled layer[e.g., Hobart et al., 1985]. Solid line shows numericalcalculation of the same process when there is a thick sectionof conductive basement below the hydrothermally cooledlayer. Thermal rebound is greatly delayed because of theheat capacity of the lower conductive layer. Recovery ofseafloor heat flux to 90% of the basal value requires only�40 ka if there is no lower basement, but requires >1 Ma ifthere is lower basement. (b) Thermal rebound is shownfollowing accumulation of sediments above the (formally)hydrothermally active layer, with a thick, conductive layerbelow. There is no sedimentation in these calculations, justan initial condition in which basement and overlyingsediments are initially chilled, after which basement andsediments are heated from below and allowed to thermallyrebound. Recovery of seafloor heat flux to 90% of the basalvalue requires �1–20 Ma.


8 of 19

B12101

deposition of turbidites in low lying areas, and otherprocesses [e.g., Giambalvo et al., 2000; Harrington, 1985;Mayer, 1981; Moore et al., 2007]. However, if sedimenta-tion continues (and particularly if sedimentation is rapid),even large basement edifices may eventually become buried.

[30] It is not necessary for all basement outcrops in aregion to be buried in order for the efficiency of regionalheat extraction by hydrothermal fluids to be reduced. Oncethe number of outcrops is sufficiently small and the spacingbetween remaining exposed outcrops becomes too great,conduction becomes the primary mechanism by whichcrustal sections lose heat to the overlying ocean. Of partic-ular interest in the present study is the nature (extent,timing) of the thermal transition between crustal systemsfrom which a significant amount of heat is initially extractedadvectively, and one in which the dominant transfer mode isconductive.[31] Sedimentation has an important influence on the rate

of thermal rebound once significant advective heat lossends, as illustrated with two sets of models (Figure 9).For the purposes of illustration, the efficiency of advectiveheat extraction prior to the cessation of regional hydrother-mal circulation is assumed to be 80%, similar to that seen inmany ridge-flank areas [e.g., Davis et al., 1992a; Fisher etal., 2003b; Villinger et al., 2002]. Lithospheric heat loss byadvection is assumed to cease abruptly or gradually (linearlywith time) when sediment thickness is 200, 500, 1000, or2000 m (at different crustal ages, depending on the sedimentaccumulation rate), and sediment accumulation continuesduring thermal rebound. Other patterns of hydrothermalcessation would result in different rebound histories, butthese simple end-members illustrate important character-istics. Heat flux at the seafloor is compared to that whichwould be observed in the case of sedimentation alone(Figure 9).[32] When sediment accumulation occurs at 100–500m/Ma,

abrupt cessation of advective heat loss leads to conductivethermal rebound that requires several million years toachieve 90% of the (sedimentation-only) reference curve,even when the sediment thickness at the time advective heatextraction ceases is only 200 m. If advective heat extractionends when the sediment thickness is 1000 m (not anunreasonable limit in many areas given typical seamountgeometries), recovery to sedimentation-only thermal con-ditions may require >10 Ma. A gradual reduction in theextent of advective heat extraction results in a commensu-rately delayed rebound of seafloor heat flux (Figure 9).[33] In all cases, thermal rebound of a previously active

hydrothermal layer in basement slows the establishment ofconductive lithospheric heat flux at the seafloor (Figure 10).More rapid sedimentation rates lead to more rapid reboundfor a given sediment thickness, and the longest recovery inseafloor heat flux occurs in the case where sedimentationends when rebound begins. Rebound is faster when sedi-mentation is rapid because the reduction in seafloor heatflux resulting from sedimentation reduces the extent ofrebound needed to meet the sedimentation-only referencecurve. As will be shown later using examples from two fieldareas, the duration of conductive thermal rebound thatfollows the end of advective heat extraction from the upperoceanic crust creates challenges for the correction andinterpretation of seafloor heat flux values.

3.5. Vertical Fluid Seepage Through the SedimentColumn

[34] The thermal effects of vertical fluid seepagethrough sediments during rapid sediment accumulation

Figure 9. Calculated seafloor heat flux during andfollowing a period of regional advective heat extraction(with 80% efficiency) by hydrothermal circulation.(a) Seafloor heat flux fraction is shown following cessationof regional advective heat extraction after deposition of 200,500, and 1000 m of sediment at an accumulation rate of100 m/Ma. Sedimentation continues at this rate throughoutthe simulation. Thick lines are for instantaneous cessation,whereas thin lines are for gradual cessation over 2.5 and5.0 Ma, as shown. Thick solid curve is a reference for whichthere is no regional advective heat extraction (i.e., heat fluxsuppression is due solely to sedimentation). (b) Fractionalextraction is shown of advective heat by hydrothermalcirculation within the crustal aquifer, where line patternscorrespond to calculations plotted above. (c) Seafloor heatflux fraction is shown following cessation of regionaladvective heat extraction after deposition of 500 and 2000 mof sediment at an accumulation rate of 500 m/Ma.Sedimentation continues at this rate throughout the simula-tion. Thick lines are for instantaneous cessation, whereas thethin line is for gradual cessation over 3.0 Ma, as shown.(d) Fractional extraction is shown of advective heat byhydrothermal circulation within the crustal aquifer, whereline patterns correspond to calculations plotted above.


9 of 19

B12101

were explored across a range of seepage rates and drivingpressures (Figure 11). One set of simulations maintainsconstant seepage rates, whereas another set holds thebasement overpressure constant as sediments accumulate.Seepage fluxes tested are �2 mm/a (2 mm/a = 2000 m/Ma),consistent with rates observed in several ridge-flank settings[e.g., Giambalvo et al., 2000; Mottl, 1989; Wheat, 1990;Wheat et al., 2004; Wheat and Mottl, 1994]. Althoughhigher fluid seepage rates have been inferred in somesettings [Abbott et al., 1981; Anderson et al., 1979; Wheatand McDuff, 1995], it is very difficult to maintain higherrates once sediment thickness exceeds a few tens tohundreds of meters [Mottl and Wheat, 1994; Spinelli etal., 2004b].[35] Modeling the influence of simultaneous seepage and

sedimentation on seafloor heat flux requires manipulation ofthermal conditions at the base of the sediment layer. Even incases of relatively rapid seepage on ridge flanks [e.g.,Spinelli et al., 2004b; Wheat et al., 2004], the amount ofheat removed from the underlying plate by seeping fluids isgenerally small relative to the lithospheric reference value.Our modeling approach was to run an initial series ofsimulations that included only the thermal influence ofsedimentation in order to define a temperature-time seriesfor upper basement, and then to force temperatures at thebase of the sediment layer to follow this series in additionalsimulations that include fluid seepage. In this way, heat isadvected from basement to the seafloor during fluid seepage(using equation (3)), but seepage has no influence on thethermal state of underlying basement [e.g., Davis et al.,1992a; Langseth et al., 1984; Mottl, 1989; Spinelli et al.,2004b].

[36] Fluid seepage through sediments counteracts thethermal influence of rapid sedimentation, with seepage at2 mm/a reducing the influence of sedimentation on seafloorheat flux by up to 50%, turning a 16% sedimentationcorrection into an 8% correction (Figure 11a). Higherseepage rates could completely remove the thermal influ-ence of sedimentation, elevating seafloor heat fluxes abovethe lithospheric reference if they can be sustained [e.g.,Wheat et al., 2004].[37] In a second set of simulations, fixed excess fluid

pressures in basement drive seepage upward through sedi-ments at rates that decrease with time owing to the increas-ing hydraulic resistance of the accumulating material(Figure 11b). Excess fluid pressures available to driveseepage through sediments on ridge flanks are modest,

Figure 10. Time required for thermal rebound of seafloorheat flux to be completed to the 5% and 10% levels (squaresand bold lines, and crosses and thin lines, respectively).Numbers shown are sediment accumulation rates in m/Mafollowing cessation of hydrothermal heat extraction. Thedelay in thermal rebound is longest when the sedimentationrate is low and sediments are thick.

Figure 11. Influence of fluid seepage on the sedimentationcorrection to seafloor heat flux at accumulation rates of100 m/Ma (thick lines) and 500 m/Ma (thin lines). Solidlines are reference curves without fluid seepage, andsediment properties are those for turbidites and hemi-pelagic mud (Figure 5). (a) Constant vertical fluid seepageat 0–2 mm/a (0–2000 m/Ma) is shown. (b) Seepagethrough accumulating sediments is shown, driven byconstant excess fluid pressures in upper basement aslabeled. (c) Vertical fluid seepage rates are shown foroverpressures of 100 and 500 kPa, as labeled. Seepage ratesdrop dramatically with time owing to the increasinghydrological resistance of accumulating sediments.


10 of 19

B12101

generally no more than tens to a few hundred kPa [e.g.,Davis and Becker, 2002; Fisher et al., 2003a; Giambalvo etal., 2000; Giambalvo, 2001; Hutnak et al., 2006; Spinelli etal., 2004a]; extreme values of 100–500 kPa are used in thefollowing examples.[38] A fluid overpressure of 500 kPa initially drives

seepage at rates >300 mm/a, overcoming the chilling effectsof rapid sedimentation and raising seafloor heat flux by 7%above the lithospheric reference (Figure 11c). Seepage ve-locities drop rapidly as sediments accumulate, to�100 mm/aafter 15 m of sediment is deposited (in 150 ka and 30 ka ataccumulation rates of 100 and 500 m/Ma, respectively). Theseepage rate drops to <10 mm/a once 100 m of sediment hasaccumulated (requiring 1 and 0.2 Ma for accumulation ratesof 100 and 500 m/Ma, respectively). Surface heat flux iselevated relative to the sedimentation-only reference by<1% after 440 m of sediment has accumulated. A fluidoverpressure of 100 kPa, still somewhat greater than iscommon on ridge flanks, elevates the surface heat flux by>1% only until 100 m of sediment has accumulated. Thesesimulations were all completed using sediment propertiesappropriate for turbidites and hemipelagic mud, as found onthe eastern flank of the Juan de Fuca Ridge. Other sedimenttypes, particularly siliceous ooze, tend to have greaterpermeability for a given depth of accumulation [Spinelli etal., 2004b], and would allow for somewhat greater seepagerates, resulting in a longer-lived thermal influence.

3.6. Thermal Transient Associated With Cooling ofYoung Lithosphere

[39] Although some lithospheric cooling models arebased on a plate having constant thickness, it is widelyrecognized that heat flux decreases and the lithosphere coolsand thickens with age, particularly when young [e.g., Davisand Lister, 1974; Parker and Oldenberg, 1973; Parsonsand Sclater, 1977; Stein and Stein, 1992]. Most thermalsedimentation models used to correct seafloor heat flux dataneglect the influence of decreasing heat flux and lithospher-ic thickening with age, processes thought to be negligiblerelative to the influence of rapid sedimentation [e.g., Daviset al., 1999; Hutchison, 1985; Wang and Davis, 1992]. Wetested the influence of lithospheric thickening of youngseafloor with the new model by defining the thermallithospheric thickness to be the depth of the 1200�C isotherm[e.g., Conrad and Lithgow-Bertelloni, 2006; Parsons andSclater, 1977; Turcotte and Schubert, 1982; Watts andZhong, 2000].[40] Our model does not formally allow the lower,

conductive basement layer to thicken with time. Insteadwe test the influence of lithospheric cooling on sedimenta-tion corrections by using a thin lithosphere (10 km conduc-tive layer, appropriate for seafloor <1 Ma in age) with atime-varying lower boundary condition. The temperature-time history at the base of the domain was that required toproduce heat flux at the top of a conductive plate thatfollowed a standard 1/

page cooling history, in the absence

of sedimentation. This approach takes into account the finitethermal diffusivity of the plate. Results from simulationswith a thin plate, transient lower boundary condition, andrapid sedimentation are compared to results based on a thickplate with constant basal heat flux. This test provides aconservative estimate of the impact of lithospheric cooling,

because there is less opportunity for the thermal influence ofsedimentation to reach the lower plate boundary if the platethickens as sediments are deposited.[41] We ran these tests using reference heat fluxes appro-

priate for 0.5–5 Ma seafloor and sediment accumulationrates of 100 and 500 m/Ma. The difference in seafloor heatflux between the two sets of simulations, one with a thinplate in which heat flux at the base decreased with time, andone with a thick plate and constant heat flux, was <1%.

4. Application to the Eastern Flank of theJuan de Fuca Ridge Near 48�N4.1. Regional Setting and Problem

[42] The eastern flank of the Juan de Fuca Ridge near48�N has been the focus of numerous surface-ship, drilling,and submersible programs exploring the nature, drivingforces, and influences of hydrothermal circulation withinyoung (�3.6 Ma) seafloor (Figure 12a) [Davis et al., 1989,1996, 1992a, 1997a, 1999, 1997b, 1992b; Fisher et al.,2005; Mottl et al., 1998; Thompson et al., 1995; Wheat andMottl, 1994; Wheat et al., 1997; Zuehlsdorff et al., 2005].This region has also yielded many observations that providea basis for analytical and numerical models of coupled fluid,heat, and/or solute transport [Davis et al., 1999, 1997c;Elderfield et al., 1999; Fisher et al., 2003a; Giambalvo etal., 2000; Hutnak et al., 2006; Rudnicki et al., 2001; Spinelliand Fisher, 2004; Stein and Fisher, 2003]. Because of itsproximity to the North American continental margin, youngbasement rocks have been rapidly buried by Pleistoceneturbidites and hemipelagic mud, causing a transition in thethermal state of the upper crust. To the west, near theactively spreading Juan de Fuca Ridge, sediment cover isthin and patchy and basement is exposed over broad areas.On this part of the ridge flank, cool seawater recharges andcirculates rapidly through the crust, resulting in measuredseafloor heat flux values <20% of predictions from litho-spheric models.[43] Sediments generally thicken and become more con-

tinuous to the east, and basement temperatures and seafloorheat flux increase as upper basement becomes hydrologi-cally isolated from the overlying ocean [e.g., Davis et al.,1992a, 1999]. There are several basement outcrops locatedon 3.5–3.6 Ma seafloor in the vicinity of Ocean DrillingProgram (ODP) sites 1026 and 1027 (Figure 12a). Located�100 km east of the active spreading center, these basementoutcrops provide entry and exit points for hydrothermalfluids [Davis et al., 1992a; Fisher et al., 2003a; Hutnak etal., 2006; Mottl et al., 1998]. Although these outcrops arehydrothermally important locally, they do not appear toremove significant quantities of lithospheric heat on aregional basis. However, recent studies of seafloor heat fluxon 3.5–3.6 Ma seafloor in this area suggest that someprocess may suppress the measured heat flux by as muchas 15–20% on average across a wide swath of seafloorextending �50 km north and south of sites 1026 and 1027[Hutnak et al., 2006; Zuehlsdorff et al., 2005], independentof the cooling effects of sedimentation [Davis et al., 1999].In this section, we reevaluate environmental corrections thatmay be needed for heat flux measurements from this area,including the influence of conductive thermal reboundfollowing the cessation of efficient advective heat extraction


11 of 19

B12101

from basement, in an attempt to reconcile observations andlithospheric cooling models.

4.2. Regional Sedimentation

[44] The sedimentation history of this area is well knownon the basis of sediments recovered from ODP sites 1026and 1027 (Figure 12b), the former being located above a

buried basement high (sediment thickness is 265 m) and thelatter being located above a buried basement low (sedimentthickness is 598 m) [Davis et al., 1999, 1997b]. The currentdepths of various biostratigraphic markers provide an indi-cation of sediment accumulation rates, and the sedimentsections were unloaded by backstripping [e.g., Watts andRyan, 1976] to derive deposition rate histories for both sites.Sedimentation corrections to seafloor heat flux are calcu-lated using both deposition and accumulation rates (includ-ing variations in both of ±5%), in order to evaluateuncertainties in sedimentation corrections (Figure 12b).[45] Sedimentation at both sites was relatively slow until

several tens to hundreds of meters of material had beendeposited:�17–24 m/Ma at Site 1026, and�68–125 m/Maat Site 1027 (lower and higher values in both ranges indicatingaccumulation and deposition rates, respectively). Sedimenta-tion rates increased abruptly during the last 0.8–1.3 Ma; theincrease occurred�0.5Ma later at Site 1026 than at Site 1027,presumably followingsubsidenceof thebasement ridgebelowthe base level of turbidite deposition [e.g.,Davis et al., 1992b;Underwood et al., 2005].[46] Sedimentation alone would have lowered seafloor

heat flux by 2% at Site 1026, and by 5–8% at Site 1027,during the first (slower) phase of sedimentation as the twosites accumulated 50–250 m of sediment (Figure 13).Continued, rapid sedimentation would have suppressedseafloor heat flux by 10–12% at Site 1026, and by 15–18% at Site 1027, by the time these sites received theircomplete sediment sections. The sedimentation correctioncurves for the two sites overlap (±1–2%) when the sedi-ment thickness reaches 265 m, and differ by <4% forthinner sediments (Figure 13b).

4.3. Thermal Rebound Following Cessation ofRegional Advective Heat Extraction

[47] Parametric tests completed earlier demonstrate that ahigh Nu value in basement, resulting from vigorous localconvection, will influence the sedimentation correction atthe 2–4% level for sediment thicknesses of the order of250–600 m. Additional simulations were run to evaluatethe influence of conductive thermal rebound in basement,following cessation of advective heat extraction. It is likelythat there were many more basement outcrops exposed inthis region prior to the deposition of the last 150–200 m ofsediment at Site 1026, and the last 400 m of sediment at Site1027, on the basis of the number of buried basement highs

Figure 12. (a) Location map shows the Juan de Fuca Platein the northeast Pacific subducting below the NorthAmerican Plate, with a star marking the location of OceanDrilling Project (ODP) sites 1026 and 1027 (2.2 km apart).(b) Biostratigraphic ages are determined from sediments andmagnetic basement ages at ODP sites 1026 (squares) and1027 (circles) on the eastern flank of the Juan de FucaRidge [Davis et al., 1999; Shipboard Scientific Party, 1997;Underwood et al., 2005]. Closed symbols and thick solidlines correspond to accumulation rates, whereas opensymbols and thin dashed lines correspond to depositionrates. Lines bracket the four nominal accumulation anddeposition histories by ±5%, in an effort to assessuncertainties in sedimentation corrections.


12 of 19

B12101

close to the seafloor that have been mapped thus far withlimited regional seismic data [Davis et al., 1992a; Hutnak etal., 2006; Zuehlsdorff et al., 2005]. Prior to 0.1–0.5 Maago, basement in this area would have lost heat mainly byadvection, with hydrothermal fluids circulating rapidlybetween recharging and discharging outcrops.[48] For modeling purposes, we assume that prior to the

burial of critical basement outcrops, regional hydrothermalcirculation removed lithospheric heat from the Site 1026/1027 area with 80% efficiency, consistent with the percent-age of heat removed advectively closer to the ridge in thisregion and with that removed from other sites wherebasement outcrops penetrating ridge-flank sediments arecommon [e.g., Davis et al., 1992a; Fisher et al., 2003b;Green et al., 1981; Villinger et al., 2002]. We representvigorous local hydrothermal heat redistribution by assigningNu = 1000 for the basement aquifer, consistent with currentobservations and models suggesting that regional uppermost

basement is largely isothermal despite large variations inbasement topography [Davis et al., 1999, 1997c; Hutnak etal., 2006; Spinelli and Fisher, 2004]. We evaluate the extentof hydrothermal rebound that is expected to have beencompleted by assuming that regional advective extractionof lithospheric heat ended 0.1–0.5 Ma, prior to the last 25–165 m of sediment accumulation (Figure 14). If the cessa-tion of regional advective extraction is more recent, whichcertainly is possible, thermal rebound will be less complete.[49] Cessation of advective heat extraction at 0.5 Ma

would have allowed thermal rebound at Site 1026 to becomplete at present, making seafloor heat flux �14% lowerthan the lithospheric value as a result of sedimentation alone(Figure 14). In contrast, rebound of the thicker sedimentsection at Site 1027 would have resulted in the seafloor heatflux being �4–6% lower than can be explained by sedi-mentation alone. Cessation of advective heat extractionwithin the last 0.1–0.2 Ma at both sites would result insuppression of seafloor heat flux by up to 25% more thancan be explained by sedimentation alone (Figure 14). Thismagnitude of heat flux suppression is consistent withobservations from this area, suggesting that heat flux is15–20% lower than can be accounted for by sedimentation[Zuehlsdorff et al., 2005; Hutnak et al., 2006].[50] The range of parameters analyzed in these models is

not exhaustive, and there is no direct evidence concerningthe time at which significant, regional advective heatextraction in this area ended. However, changing the initial

Figure 13. (a) Calculated seafloor heat flux fractionversus time is shown for the range of deposition andaccumulation rates appropriate for ODP sites 1026 and 1027(Figure 12), based on physical properties from these sites[Davis et al., 1999; Shipboard Scientific Party, 1997], andboundary conditions and other properties described in thetext. Thick dashed curves indicate values used in subse-quent analyses, intended to be typical of the valuesgenerated, given uncertainties in the sedimentation historiesof these sites. (b) Calculated seafloor heat flux fractionversus sediment thickness is shown for the range ofdeposition and accumulation rates appropriate for ODPsites 1026 and 1027. The two dashed curves used insubsequent analyses are in good agreement for sedimentthicknesses >250 m and differ by <4% for thinnersediments.

Figure 14. Calculated seafloor heat flux fraction versustime for ODP sites 1026 (thick curves) and 1027 (thincurves), assuming an initial 80% efficiency in regionaladvective extraction of lithospheric heat by hydrothermalcirculation, followed by cessation at 0.5, 0.2, and 0.1 Ma.Stippled band indicates the magnitude of likely sedimenta-tion corrections appropriate for these sites based on earliercalculations (Figure 13), whereas hatched band indicatesthe mean of >300 filtered heat flux observations from 3.5–3.6 Ma seafloor on the eastern flank of the Juan de FucaRidge [Hutnak et al., 2006; Zuehlsdorff et al., 2005], whichare lower than lithospheric (following sedimentation correc-tion) by 15–20%. Observational data are most consistentwith a cessation of regional heat extraction during the last0.1–0.2 Ma, at a time when sedimentation in this area wasextremely rapid and many basement outcrops were buried.


13 of 19

B12101

efficiency of advective heat extraction will impact theresidual suppression by a few percent at most, because theinitial phase of rebound is most abrupt and the later phase ismuch more gradual (Figures 8, 9, and 14). In addition, if theadvective extraction of lithospheric heat did not end abruptlybut instead was reduced gradually, then there would be aneven greater fraction of rebound remaining (Figure 9).[51] There is a global implication to these calculations:

conductive rebound following the end of regional, advectiveheat extraction is more rapid above buried basement highswhere sediments are thinner. Unlike local redistribution ofheat by vigorous convection in basement, which moves heatlaterally but results in no regional bias (i.e., the excess heat

flux above a buried basement high balances the deficit in theadjacent troughs), different rates of hydrothermal reboundabove buried basement highs and lows will produce asystematic regional bias. Seafloor heat flux will tend to belower than lithospheric, on average, for millions of yearsfollowing cessation of hydrothermal heat extraction, eventhough conditions are fully conductive. One cannot neces-sarily average the high and low values found above buriedbasement highs and lows to determine the local lithosphericheat flux, nor is it sufficient to determine it from measure-ments at sites that are currently far from the nearest outcrop:these areas may have been much closer to outcrops in thepast, and in any case, efficient advective heat extraction ispossible where sediments average 400–500 m in thicknessand basement outcrops are separated by 20–50 km [Fisheret al., 2003b; Hutnak et al., 2007]. Because thermalrebound is slower through thick sediments, using data fromheavily sedimented sites above flat basement to assesslithospheric heat flux may be misleading. Unless the historyof hydrothermal activity in an area over the last several tensof million years is known, seafloor thermal data mayindicate a regional or background heat flux below the truelithospheric value as a result of incomplete conductiverebound.

5. Application to the Eastern Flank of the EastPacific Rise Near 8–10�N5.1. Regional Setting

[52] The Cocos Plate seaward of the Nicoya Peninsula, onthe Pacific Ocean side of Costa Rica, is moderate-aged (18–24 Ma) seafloor within which there are two distinct heattransport regimes (Figure 15a) [Fisher et al., 2003b; Hutnaket al., 2007]. Seafloor in the northwestern part of this areawas constructed along the fast spreading East Pacific Rise(EPR) to the west, whereas seafloor in the southeastern partof this area was constructed at the medium-spreading

Figure 15. Regional map and calculated seafloor heat fluxfraction versus time and sediment thickness for an area ofthe seafloor experiencing sediment accumulation at 10 or50 m/Ma, consistent with typical global rates. The heat fluxplots continue to 100–180 Ma, based on the observationthat even the large seamounts and other basement outcropspresent in this field area would eventually become buried,even given the low sedimentation rates, were the plate notsubducted. (a) Location map shows East Pacific Rise(EPR)– and Cocos-Nazca Ridge (CNR)–generated seafloorsubducting along the Middle America Trench (MAT)seaward of Costa Rica, with a star making the location ofODP Site 1039. Circles and ovals show outcrop locationsand approximate diameters. Thick dashed lines correspondto a thermal boundary between areas with anomalously lowheat flux and those with elevated heat flux [Fisher et al.,2003b; Hutnak et al., 2007]. (b) Sediment accumulation of10 m/Ma and cessation of advective heat extraction areshown (initially with 80% efficiency) at 20, 50, and 100 Ma(200, 500, and 1000 m of sediment). (c) Sedimentaccumulation of 50 m/Ma and cessation of advective heatextraction are shown (initially with 80% efficiency) at 5, 10,20, and 40 Ma (200, 500, 1000, and 2000 m of sediment).


14 of 19

B12101

Cocos-Nazca Ridge (CNR) to the south [Barckhausen et al.,2001; Hey, 1977]. A triple junction trace and fracture zonecollectively define a plate ‘‘suture’’ between crust generatedat the EPR and that formed at the CNR.[53] The Cocos Plate is currently being subducted along

the Middle America Trench, leading to volcanic and seismicactivity on, near, and below the Nicoya Peninsula [e.g.,Newman et al., 2002]. The thermal state of the plate hasimportant influences on many subduction-related processes[e.g., Harris and Wang, 2002; Hyndman and Wang, 1993;Moore and Saffer, 2001; Peacock and Wang, 1999; Scholz,1990]. Values of seafloor heat flux from CNR-generatedseafloor are highly scatted but are on average consistentwith lithospheric cooling models (95–120 mW/m2). Incontrast, values of seafloor heat flux from EPR-generatedseafloor are typically much lower, generally 10–30% oflithospheric [Fisher et al., 2003b; Hutnak et al., 2007;Langseth and Silver, 1996; Vacquier et al., 1967; VonHerzen and Uyeda, 1963].[54] Seamounts and other basement outcrops are common

on EPR-generated seafloor where the heat flux is sup-pressed across a large area, and provide permeable path-ways for hydrothermal fluids to bypass the regionally thick(�400–500 m), low-permeability sediments and efficientlyextract lithospheric heat. It is useful to consider both thehistory of sedimentation in this area, and to project currentconditions into the future to assess what would happen at asite such as this one if it were not subducted, but wasinstead allowed to evolve along a more typical trajectory.Many of the seamounts in this area are 3–5 km in diameterand rise 1000–2000 m above the surrounding top ofbasement [Fisher et al., 2003b; Hutnak et al., 2007; Wessel,2001], but eventually even outcrops as large as these wouldbe blanketed with sediments that would restrict entry andexit of hydrothermal fluids, initiating conductive thermalrebound in basement. We can use models of future (hypo-thetical) thermal conditions at this site, if the plate were notsubducted, to gain insight into processes that may occurwithin much of the seafloor when widely distributed base-ment outcrops become buried.

5.2. Regional Sedimentation

[55] ODP Site 1039 is located within the trench on theincoming plate and provides a detailed history of sedimen-tation on this ridge flank. The upper �70 m of the coredinterval was deposited rapidly because the area was close tothe continental margin [Shipboard Scientific Party, 1997].However, margin sediments do not extend beyond thetrench and have no influence on the majority of seafloorheat flux measurements on the incoming plate in this region[Fisher et al., 2003b; Hutnak et al., 2007]. Lower in thesection, pelagic and hemipelagic deposits indicate accumu-lation rates more typical of global averages for thesesediment types [e.g., Spinelli et al., 2004b]. For modelingpurposes we explore the implications of sedimentation onthe EPR-generated side of the Cocos Plate at accumulationrates of 10 and 50 m/Ma, using physical properties versusdepth relations derived from the pelagic and hemipelagicintervals recovered from ODP Site 1039.[56] Sediment accumulation at 10 m/Ma reduces the

seafloor heat flux by 3–4% relative to lithospheric after20 Ma, and by 7–8% at a plate age of 180 Ma, the age of

some of the oldest remaining seafloor (Figure 15b). Asediment accumulation rate of 50 m/Ma would reduce theseafloor heat flux by 10–11% after 20 Ma, and 20–21%after 100 Ma (Figure 15c). These calculations illustrate thatsedimentation corrections can be important for assessing thelithospheric heat flux across much of the seafloor, even forrelatively low sediment accumulation rates, particularly ifsedimentation continues for tens of millions of years.

5.3. Thermal Rebound Following Cessation ofRegional Advective Heat Extraction

[57] We evaluated the thermal influence of the cessationof regional advective heat extraction once the sedimentthickness reaches 200, 500, 1000, and 2000 m. This rangeof sediment thicknesses includes the median height of the�15,000 seamounts apparent globally from satellite gra-vimetry [Wessel, 2001], and likely includes most seamountsand similar outcrops since the vast majority are too small tobe mapped by satellite. The initial efficiency of regionaladvective heat extraction is 80%, as seen today on EPR-generated seafloor of the Cocos Plate, and modeled cessa-tion occurs at plate ages of 5–100 Ma, depending onsediment thickness and accumulation rate. This rangeincludes the mean global age of 65 Ma at which observedseafloor heat flux values tend to fall, on average, onstandard lithospheric cooling curves [Lister, 1977; Parsonsand Sclater, 1977; Stein and Stein, 1992].[58] Thick sediment layers that accumulate on moderate

to old seafloor, even when deposited slowly, can mask theconductive thermal rebound of the sediments and underly-ing basement following the cessation of regional advectiveheat extraction. For example, when advective extractionceases following the accumulation of 500 m of sediment at10 m/Ma, conductive thermal rebound remains incompleteat the 10% level after 1.3 Ma, and incomplete at the 5%level after 4.2 Ma (Figure 15b). Sedimentation accounts foran additional 4–6% of heat flux suppression in these cases.The influence of conductive thermal rebound is greaterwhen sediments are thicker, requiring 3.6 Ma to recoverto within 10% of the sediment-suppressed value when thereis 1000 m of sediment, and 10.6 Ma to recover to within5%. In the extreme case of cessation of advective heatextraction when sediment thickness reaches 2000 m at anaccumulation rate of 50 m/Ma (as might have been the casefor the EPR-generated Cocos Plate were it not subducted,given the size of seamounts in that area), recovery of theseafloor heat flux to the 10% level requires 6.3 Ma, andrecovery to the 5% level requires 15.1 Ma.[59] Two important conclusions should be drawn from

these analyses. First, even relatively low sediment accumu-lation rates will suppress seafloor heat flux by 5–10% ormore if sedimentation continues as the plate ages, leading tosignificant regional biases in estimates of lithospheric heatflux if data are not corrected. In addition, regionally thicksediments and moderate to old plate ages do not mean thatresearchers can ignore the potential influence of pasthydrothermal activity on seafloor heat flux. The influenceof regional advective heat extraction on seafloor heat fluxmeasurements can continue for 10 Ma or more, dependingon the thickness of sediments and the timing of thermalrebound. All simulations using sedimentation rates that aremost typical of the global seafloor include instantaneous


15 of 19

B12101

cessation of regional advective heat extraction. If cessationinstead occurs slowly, the time required for conductiverebound will be commensurately longer (Figure 9).

6. Summary and Conclusions

[60] We have developed a one-dimensional thermal mod-el of heat flux through an accumulating sediment layer thatallows temporal variability of boundary conditions, repre-sents several scales of hydrothermal and transient conduc-tive processes within basement below sediments, andincludes the thermal influence of upward fluid seepage.We compare results of this model to analytical and othernumerical solutions, conduct sensitivity analyses to quantifythe importance of various physical parameters and processes(Table 2), and apply the model to assess thermal conditionsand seafloor heat flux in several settings.[61] Sensitivity analyses using the new model indicate

that selection of basement thermal conductivities within areasonable range may influence seafloor heat flux correc-tions at the 2–4% level. The thickness of the hydrothermalaquifer in upper basement and the vigor of local mixing inthis aquifer influence the suppression of seafloor heat fluxdue to sedimentation at the 1–6% level. Vertical fluidseepage through accumulating sediments tends to reducethe suppression of seafloor heat flux associated with sedi-mentation, raising measured values by up to 8% for seepagerates �2 mm/a. Higher seepage rates would have a greaterinfluence on seafloor heat flux, but are difficult to maintaingiven typical sediment properties and seepage drivingforces on ridge flanks.[62] Additional calculations explore the rates and influ-

ence of conductive thermal rebound following the cessationof hydrothermal circulation on seafloor heat flux measure-

ments. New models suggest that conductive rebound maylast up to several tens of millions of years, considerablylonger than inferred from previous analytical calculations,mainly because earlier analyses were based on having aconstant heat flux lower boundary immediately below thesediments. New models place the lower heat flux boundaryat the bottom of a lithospheric plate. This configurationallows the upper part of the plate to act as a thermalcapacitor, delaying conductive rebound and lowering sea-floor heat flux. Calculations illustrate that conductiverebound occurs more slowly below thicker sediment layers,and could impart a regional bias leading to an underestimateof lithospheric heat flux.[63] Models were run to assess heat flux measurements

on the eastern flank of the Juan de Fuca Ridge. Earlierstudies have suggested that the surface heat flux across awide area of 3.5–3.6 Ma seafloor in this area may be lowerthan lithospheric predictions by 15–20%, even after cor-recting for (rapid) sedimentation. New models are consis-tent with earlier calculations suggesting sedimentationcorrections of 10–18%. The remaining 15–20% deficitbetween sediment-corrected observations and lithosphericpredictions can be explained by ongoing, conductive ther-mal rebound following the cessation of hydrothermal heatextraction, provided that this thermal transition began in thelast 0.1–0.2 Ma. Many more basement outcrops wereexposed in this area during the late Pleistocene; much ofthe lithospheric heat would have been extracted advectivelyuntil the spacing between outcrops became too great tosustain dominantly advective cooling of the crust.[64] Models were also run to assess environmental cor-

rections to heat flux values from seafloor of the Cocos Plategenerated on the East Pacific Rise, seaward of the MiddleAmerica Trench in the eastern Pacific Ocean. In this area,large, widely distributed basement outcrops allow hydro-thermal fluids to enter and exit the seafloor easily, and fluidflow between these outcrops extracts 70–90% of thelithospheric heat. Sedimentation rates in this area weremuch lower than those on the eastern flank of the Juan deFuca Ridge, but they are still sufficient to suppress heat fluxby 3–11% at a current plate age of �20 Ma, and would leadto suppression of 7–20% when the crust is 180 Ma old (ifthe crust were not subducted first). In addition, conductivethermal rebound (to sedimentation-only thermal conditions)following the cessation of regional advective heat extractionwould require 1–6 Ma at the 10% significance level, and5–15 Ma at the 5% level, depending on the sedimentthickness and accumulation rate during and followingcessation.[65] These calculations show that it may be difficult to

assess lithospheric heat flux from many seafloor sites,including sites located far from outcrops where basementis deeply buried by sediments and conditions are dominant-ly conductive at present. Unless the sedimentation andhydrothermal histories of a site are known for the lastseveral to several tens of Ma, significant biases in regionalheat flux values can lead to misinterpretation of heattransport processes. This will continue to make efforts todistinguish between competing thermal models of the oce-anic lithosphere difficult, and makes it challenging forresearchers to assess the extent and magnitude of localand regional seafloor heat flux anomalies.

Table 2. Relative Importance of Sediment and Crustal Parameters

to the Magnitude of Environmental Corrections to Seafloor Heat

Flux Valuesa

Parameter Range Tested Influence, %

Sedimentation,b m/Ma 10–500 2–32Basement thermalconductivity,c W/m-K

2.0–2.9 2–4

Aquifer thickness andthermal homogeneity,d m

100–1000Nu = 2–104

1–6

Vertical fluid seepage,e �2 mm/a 8Lithospheric thickening,f 5–17 km <1Conductive thermal rebound,g 0.2–100 Ma 10

aPrimary parametric tests were run with sediment accumulation rates of100–500 m/Ma. Site-specific analyses used other parameters. All influencevalues are relative to a lithospheric reference.

bSedimentation corrections depend on rate history of deposition,compaction trends, sediment and basement properties, and other parametersexplored in this study.

cValue applies to conductive layer below hydrothermal aquifer.dVertical thermal homogeneity is used, determined through application of

a high-Nu approximation, as described in the text.eUpward seepage tends to increase seafloor heat flux, in contrast to all

other processes modeled in this study. Additional simulations were run withconstant overpressure and a seepage rate that varies with time as sedimentsaccumulate.

fLithospheric thickening was represented by deepening of the 1200�Cisotherm.

gTime interval reported is that required to achieve recovery to within10% of lithospheric value (Figure 10). For short time periods, reboundcould be more or less than 10%, depending on site-specific parameters.


16 of 19

B12101

[66] Acknowledgments. Funding in support of this research wasprovided by grants from the National Science Foundation (OCE–000001892, OCE-0326699, and OCE-0550713). This research also useddata provided by the Ocean Drilling Program (ODP) and the IntegratedOcean Drilling Program (IODP) and funding from the Joint OceanographicInstitutions (JOI), Inc., projects T301A7 and T301B7. The ODP and IODPare sponsored by the U.S. National Science Foundation (NSF) andparticipating countries under the management of JOI. Support for M. H.was also provided by a Schlanger Ocean Drilling Fellowship, part of theNSF-sponsored U.S. Science Support Program (USSSP) to ODP and IODP.This manuscript benefited greatly from careful reviews by R. N. Harris andan anonymous reviewer, as well as from discussions with R. P. Von Herzenand E. A. Silver.

ReferencesAbbott, D., W. Menke, M. Hobart, and R. Anderson (1981), Evidence forexcess pore pressures in southwest Indian Ocean sediments, J. Geophys.Res., 86, 1813–1827.

Anderson, R., and M. Hobart (1976), The relation between heat flow,sediment thickness, and age in the eastern Pacific, J. Geophys. Res.,81, 2968–2989.

Anderson, R. N., M. A. Hobart, and M. G. Langseth (1979), Geothermalconvection through oceanic crust and sediments in the Indian Ocean,Science, 204, 828–832.

Barckhausen, U., C. Renaro, R. von Huene, S. C. Cande, and H. A. Roeser(2001), Revised tectonic boundaries in the Cocos Plate off Costa Rica:Implications for the segmentation of the convergent margin and for platetectonic models, J. Geophys. Res., 106, 19,207–19,220.

Bartetzko, A., P. Pezard, D. Goldberg, Y.-F. Sun, and K. Becker (2001),Volcanic stratigraphy of DSDP/ODP Hole 395A: An interpretation usingwell-logging data, Mar. Geophys. Res., 22, 111–127.

Becker, K., and A. Fisher (2000), Permeability of upper oceanic basementon the eastern flank of the Endeavor Ridge determined with drill-stringpacker experiments, J. Geophys. Res., 105, 897–912.

Becker, K., M. Langseth, R. P. Von Herzen, and R. Anderson (1983), Deepcrustal geothermal measurements, Hole 504B, Costa Rica Rift, J. Geo-phys. Res., 88, 3447–3457.

Becker, K., M. Langseth, R. Anderson, and M. Hobart (1985), Deep crustalgeothermal measurements, Hole 504B, Costa Rica Rift, legs 69, 70, 83,and 92, Initial Rep. Deep Sea Drill. Proj., 83, 405–418.

Benfield, A. E. (1949), The effect of uplift and denudation on undergroundtemperatures, J. Appl. Phys., 20, 66–70.

Berger, W. H. (1978), Pelagic sedimentation, pelagic sediments, in Ency-clopedia of Sedimentology, edited by R. W. Fairbridge and J. Bourgeois,pp. 544–558, Dowden, Hutchinson, and Ross, New York.

Bredehoeft, J. D., and I. S. Papadopulos (1965), Rates of vertical ground-water movement estimated from the Earth’s thermal profile,Water Resour.Res., 1, 325–328.

Bryant, W. R., and R. H. Bennett (1988), Origin, physical, and mineralo-gical nature of red clays: The Pacific-Ocean basin as a model, Geo Mar.Lett., 8(4), 189–249.

Busch, W. H., P. R. Castillo, P. A. Floyd, and G. Cameron (1992), Effects ofalteration on physical properties of basalts from the Pigafetta and EastMariana basins, Proc. Ocean Drill. Program Sci. Results, 129, 485–500.

Carslaw, H. S., and J. C. Jaeger (1959), The Conduction of Heat in Solids,Oxford Univ. Press, Oxford, U. K.

Conrad, C. P., and C. Lithgow-Bertelloni (2006), Influence of continentalroots and asthenosphere on plate-mantle coupling, Geophys. Res. Lett.,33, L05312, doi:10.1029/2005GL025621.

Davis, E. E., and K. Becker (2002), Observations of natural-state fluidpressures and temperatures in young oceanic crust and inferences regard-ing hydrothermal circulation, Earth Planet. Sci. Lett., 204, 231–248.

Davis, E. E., and K. Becker (2004), Observations of temperature and pres-sure: Constraints on ocean crustal hydrologic state, properties, and flow,in Hydrogeology of the Oceanic Lithosphere, edited by E. E. Davis andH. Elderfield, pp. 225–271, Cambridge Univ. Press, Cambridge, U. K.

Davis, E. E., and C. R. B. Lister (1974), Fundamentals of ridge cresttopography, Earth Planet. Sci. Lett., 21, 405–413.

Davis, E. E., and C. R. B. Lister (1977), Heat flow measured over the Juande Fuca Ridge: Evidence for widespread hydrothermal circulation in ahighly heat-transportive crust, J. Geophys. Res., 82, 4845–4860.

Davis, E. E., and H. Villinger (1992), Tectonic and thermal structure of theMiddle Valley sedimented rift, northern Juan de Fuca Ridge, Proc. OceanDrill. Program Initial Rep., 139, 9–41.

Davis, E. E., D. S. Chapman, C. Forster, and H. Villinger (1989), Heat-flowvariations correlated with buried basement topography on the Juan deFuca Ridge flank, Nature, 342, 533–537.

Davis, E. E., et al. (1992a), FlankFlux: An experiment to study the nature ofhydrothermal circulation in young oceanic crust, Can. J. Earth Sci.,29(5), 925–952.

Davis, E. E., et al. (1992b), Proceedings of the Ocean Drilling Program,Initial Reports, vol. 139, 1026 pp., Ocean Drill. Program, College Station,Tex.

Davis, E. E., D. S. Chapman, and C. B. Forster (1996), Observationsconcerning the vigor of hydrothermal circulation in young volcanic crust,J. Geophys. Res., 101, 2927–2942.

Davis, E. E., D. S. Chapman, H. Villinger, S. Robinson, J. Grigel,A. Rosenberger, and D. Pribnow (1997a), Seafloor heat flow on theeastern flank of the Juan de Fuca Ridge: Data from the ‘‘FlankFlux’’studies through 1995, Proc. Ocean Drill. Program Initial Rep., 168,23–33.

Davis, E. E., A. T. Fisher, and J. Firth (1997b), Proceedings of the OceanDrilling Program, Initial Reports, vol. 168, 470 pp., Ocean Drill. Pro-gram, College Station, Tex.

Davis, E. E., K.Wang, J. He, D. S. Chapman, H.Villinger, andA.Rosenberger(1997c), An unequivocal case for high Nusselt-number hydrothermal con-vection in sediment-buried igneous oceanic crust, Earth Planet. Sci. Lett.,146, 137–150.

Davis, E. E., D. S. Chapman, K. Wang, H. Villinger, A. T. Fisher, S. W.Robinson, J. Grigel, D. Pribnow, J. Stein, and K. Becker (1999), Regionalheat flow variations across the sedimented Juan de Fuca Ridge easternflank: Constraints on lithospheric cooling and lateral hydrothermal heattransport, J. Geophys. Res., 104, 17,675–17,688.

Elderfield, H., C. G. Wheat, M. J. Mottl, C. Monnin, and B. Spiro (1999),Fluid and geochemical transport through oceanic crust: A transect acrossthe eastern flank of the Juan de Fuca Ridge, Earth Planet. Sci. Lett., 172,151–165.

Emiliani, C., and J. D. Milliman (1966), Deep-sea sediments and theirgeological record, Earth Sci. Rev., 1, 105–132.

Fisher, A. T. (1998), Permeability within basaltic oceanic crust, Rev. Geo-phys., 36(2), 143–182.

Fisher, A. (2003), Geophysical constraints on hydrothermal circulation:Observations and models, in Energy and Mass Transfer in SubmarineHydrothermal Systems, edited by P. Halbach, V. Tunnicliffe, and J. Hein,pp. 29–52, Dahlem Univ. Press, Berlin.

Fisher, A. (2004), Rates and patterns of fluid circulation, in Hydrogeologyof the Oceanic Lithosphere, edited by E. E. Davis and H. Elderfield, pp.339–377, Cambridge Univ. Press, Cambridge, U. K.

Fisher, A., and K. Becker (2000), Channelized fluid flow in oceanic crustreconciles heat-flow and permeability data, Nature, 403, 71–74.

Fisher, A. T., K. Becker, T. N. Narasimhan, M. G. Langseth, and M. J.Mottl (1990), Passive, off-axis convection on the southern flank of theCosta Rica Rift, J. Geophys. Res., 95, 9343–9370.

Fisher, A. T., et al. (2003a), Hydrothermal recharge and discharge across50 km guided by seamounts on a young ridge flank, Nature, 421, 618–621.

Fisher, A. T., C. A. Stein, R. N. Harris, K. Wang, E. A. Silver, M. Pfender,M. Hutnak, A. Cherkaoui, R. Bodzin, and H. Villinger (2003b), Abruptthermal transition reveals hydrothermal boundary and role of seamountswithin the Cocos Plate, Geophys. Res. Lett., 30(11), 1550, doi:10.1029/2002GL016766.

Fisher, A. T., T. Urabe, and A. Klaus (2005), Proceedings of the IntegratedOcean Drilling Program, vol. 301, doi:10.2204/iodp.proc.301.2005,Integrated Ocean Drill. Program, College Station, Tex.

Giambalvo, E. R. (2001), Factors controlling fluxes of fluid, heat, andsolutes from sedimented, ridge-flank hydrothermal systems, Ph.D. thesis,Univ. of Calif., Santa Cruz.

Giambalvo, E., A. T. Fisher, L. Darty, J. T. Martin, and R. P. Lowell (2000),Origin of elevated sediment permeability in a hydrothermal seepage zone,eastern flank of Juan de Fuca Ridge, and implications for transport offluid and heat, J. Geophys. Res., 105, 913–928.

Green, K. E., R. P. Von Herzen, and D. L. Williams (1981), The Galapagosspreading center at 86�W: A detailed geothermal field study, J. Geophys.Res., 86, 979–986.

Harrington, P. K. (1985), Formation of pockmarks by pore-water escape,Geo Mar. Lett., 5(3), 193–197.

Harris, R. N., and K. Wang (2002), Thermal models of the Middle AmericaTrench at the Nicoya Peninsula, Costa Rica, Geophys. Res. Lett., 29(21),2010, doi:10.1029/2002GL015406.

Harris, R. N., A. T. Fisher, and D. Chapman (2004), Fluid flow throughseamounts and implications for global mass fluxes, Geology, 32(8), 725–728, doi:10.1130/G20387.1.

Hay, W. W., J. L. Sloan, and C. N. Wold (1988), Mass age distribution andcomposition of sediments on the ocean-floor and the global rate of sedi-ment subduction, J. Geophys. Res., 93, 14,933–14,940.

Hey, R. N. (1977), Tectonic evolution of the Cocos-Nazca spreading center,Geol. Soc. Am. Bull., 88, 1404–1420.

Hobart, M., M. Langseth, and R. N. Anderson (1985), A geothermal andgeophysical survey of the south flank of the Costa Rica Rift, sites 504and 505, Initial Rep. Deep Sea Drill. Proj., 137/140, 379–404.


17 of 19

B12101

Hutchison, I. (1985), The effects of sedimentation and compaction onoceanic heat flow, Geophys. J. R. Astron. Soc., 82, 439–459.

Hutnak, M., A. T. Fisher, L. Zuehlsdorff, V. Speiss, P. H. Stauffer, and C. W.Gable (2006), Hydrothermal recharge and discharge guided by basementoutcrops on 0.7 – 3.6 Ma seafloor east of the Juan de Fuca Ridge:Observations and numerical models, Geochem. Geophys. Geosyst., 7,Q07O02, doi:10.1029/2006GC001242.

Hutnak, M., et al. (2007), The thermal state of 18–24 Ma upper lithospheresubducting below the Nicoya Peninsula, northern Costa Rica margin,in Interplate Subduction Zone Seismogenesis, edited by T. Dixon andC. Moore, pp. 86–122, Columbia Univ. Press, New York.

Hyndman, R. D., and K. Wang (1993), Thermal constraints on the zone ofmajor thrust earthquake failure: The Cascadia subduction zone, J. Geo-phys. Res., 98, 2039–2060.

Jarrard, R. D., and C. Broglia (1991), Geophysical properties of oceaniccrust at sites 768 and 770, Proc. Ocean Drill. Program Sci. Results, 124,75–90.

Johnson, H. P., K. Becker, and R. P. V. Herzen (1993), Near-axis heat flowmeasurements on the northern Juan de Fuca Ridge: Implications for fluidcirculation in oceanic crust, Geophys. Res. Lett., 20(17), 1875–1878.

Karato, S. (1983), Physical properties of basalts from Deep Sea DrillingProject Hole 504B, Initial Rep. Deep Sea Drill. Proj., 69, 687–695.

Karato, S., and K. Becker (1983), Physical properties of sediments from theGalapagos region and their implications for hydrothermal circulation,Initial Rep. Deep Sea Drill. Proj., 111, 355–368.

Karato, S., R. Wilkens, and M. Langseth (1983), Shipboard physical-properties measurements of basalts from the Costa Rica Rift, DeepSea Drilling Project legs 69 and 70, Initial Rep. Deep Sea Drill. Proj.,69, 675–681.

Kastner, M., H. Elderfield, and J. B. Martin (1991), Fluids in convergentmargins: What do we know about their composition, origin, role indiagenesis and importance for oceanic chemical fluxes, Philos. Trans.R. Soc., Ser. A, 335(1638), 243–259.

Kimura, G., E. Silver, and P. Blum (1997), Proceedings of the OceanDrilling Program, Initial Reports, vol. 170, 458 pp., Ocean Drill. Pro-gram, College Station, Tex.

Langseth, M. G., and B. Herman (1981), Heat transfer in the oceanic crustof the Brazil Basin, J. Geophys. Res., 86, 10,805–10,819.

Langseth, M. G., and E. A. Silver (1996), The Nicoya convergent margin:A region of exceptionally low heat flow, Geophys. Res. Lett., 23(8), 891–894.

Langseth,M. G., R. D. Hyndman, K. Becker, S. H. Hickman, andM. Salisbury(1984), The hydrogeological regime of isolated sediment ponds in mid-oceanic ridges, Initial Rep. Deep Sea Drill. Proj., 78B, 825–837.

Langseth, M. G., M. J. Mottl, M. A. Hobart, and A. T. Fisher (1988), Thedistribution of geothermal and geochemical gradients near Site 501/504:Implications for hydrothermal circulation in the oceanic crust, Proc.Ocean Drill. Program Initial Rep., 111, 23–32.

Langseth, M. G., K. Becker, R. P. Von Herzen, and P. Schultheiss (1992),Heat and fluid flux through sediment on the western flank of the Mid-Atlantic Ridge: A hydrogeological study of North Pond, Geophys. Res.Lett., 19, 517–520.

Lawver, L., and D. Williams (1979), Heat flow in the central Gulf ofCalifornia, J. Geophys. Res., 84, 3465–3478.

Lister, C. R. B. (1977), Estimators for heat flow and deep rock propertiesbased on boundary layer theory, Tectonophysics, 41, 157–171.

Lonsdale, P., and K. Becker (1985), Hydrothermal plumes, hot springs, andconductive heat flow in the southern trough of Guaymas Basin, EarthPlanet. Sci. Lett., 73, 211–225.

Lucazeau, F., and S. Le Douaran (1985), The blanketing effect of sedimentsin basins formed by extension: A numerical model. Application to theGulf of Lion and Viking gragen, Earth Planet. Sci. Lett., 74, 92–102.

Mayer, L. (1981), Erosional troughs in deep-sea carbonate and their rela-tionship to basement structure, Math. Geol., 39, 59–80.

Moore, J. C., and D. Saffer (2001), Updip limit of the seismogenic zonebeneath the accretionary prism of southwest Japan: An effect of diage-netic to low-grade metamorphic processes and increasing effective stress,Geology, 29, 183–186.

Moore, T. C., Jr., N. C. Mitchell, M. Lyle, J. Backman, and H. Palike(2007), Hydrothermal pits in the biogenic sediments of the equatorialPacific Ocean, Geochem. Geophys. Geosyst., 8, Q03015, doi:10.1029/2006GC001501.

Morin, R. H., A. E. Hess, and K. Becker (1992), In situ measurements offluid flow in DSDP holes 395A and 534A: Results from the dianautprogram, Geophys. Res. Lett., 19, 509–512.

Mottl, M. J. (1989), Hydrothermal convection, reaction and diffusion insediments on the Costa Rica Rift flank, pore water evidence from ODPsites 677 and 678, Proc. Ocean Drill. Program Sci. Results, 111, 195–214.

Mottl, M. J., and C. G. Wheat (1994), Hydrothermal circulation throughmid-ocean ridge flanks: Fluxes of heat and magnesium, Geochim. Cos-mochim. Acta, 58, 2225–2237.

Mottl, M. J., et al. (1998), Warm springs discovered on 3.5 Ma oceaniccrust, eastern flank of the Juan de Fuca Ridge, Geology, 26, 51–54.

Newman, A. V., S. Y. Schwartz, H. R. DeShon, J. M. Protti, V. Gonzalez,and L. Dorman (2002), Along-strike variability in the seismogenic zonebelow Nicoya Peninsula,Costa Rica, Geophys. Res. Lett., 29(20), 1977,doi:10.1029/2002GL015409.

Parker, R. L., and D. W. Oldenberg (1973), Thermal model of ocean ridges,Nature Phys. Sci., 242, 137–139.

Parsons, B., and J. G. Sclater (1977), An analysis of the variation of oceanfloor bathymetry and heat flow with age, J. Geophys. Res., 82, 803–829.

Peacock, S. M., and K. Wang (1999), Seismic consequences of warmversus cool subduction metamorphism: Examples for southwest andnortheast Japan, Science, 286, 937–939.

Pezard, P., and R. N. Anderson (1989), Morphology and alteration of theupper oceanic crust from in-situ electrical experiments in DSDP/ODPHole 504B, Proc. Ocean Drill. Program Sci. Results, 111, 133–146.

Rohr, K. (1994), Increase of seismic velocities in upper oceanic crust andhydrothermal circulation in the Juan de Fuca plate, Geophys. Res. Lett.,21, 2163–2166.

Rohr, K., U. Schmidt, C. Lowe, and B. Milkereit (1994), Multichannelseismic reflection data across Endeavour segment, Juan de Fuca Ridge,Geol. Surv. of Can., Sydney, B. C.

Rosenberg, N., A. Fisher, and J. Stein (2000), Large-scale lateral heat andfluid transport in the seafloor: Revisiting the well-mixed aquifer model,Earth Planet. Sci. Lett., 182, 93–101.

Rudnicki, M. D., H. Elderfield, and M. J. Mottl (2001), Pore fluid advectionand reaction in sediments of the eastern flank, Juan de Fuca Ridge, 48�N,Earth Planet. Sci. Lett., 187, 173–189.

Ruppel, C., and M. Kinoshita (2000), Fluid, methane, and energy flux in anactive margin gas hydrate province, offshore Costa Rica, Earth Planet.Sci. Lett., 179, 153–165.

Sakai, H., T. Gamo, E. S. Kim, M. Tsutsumi, T. Tanaka, J. Ishibashi,H. Wakita, M. Yamano, and T. Oomori (1990), Venting of carbon-dioxiderich fluid and hydrate formation in mid-Okinawa trough backarc basin,Science, 248, 1093–1096.

Scholz, C. H. (1990), Geophysics: Earthquakes as chaos, Nature, 348,197–198.

Shipboard Scientific Party (1997), Proceedings of the Ocean DrillingProgram, Initial Reports, vol. 168, edited by G. Kimura, E. Silver,and P. Blum, Ocean Drill. Program, College Station, Tex.

Shipboard Scientific Party (2005), Proceedings of the Integrated OceanDrilling Program, Initial Reports, vol. 301, doi:10.2204/iodp.proc.301.101.2005, Ocean Drill. Program, College Station, Tex.

Spinelli, G. A., and A. Fisher (2004), Hydrothermal circulation withintopographically rough basaltic basement on the Juan de Fuca Ridge flank,Geochem. Geophys. Geosyst., 5, Q02001, doi:10.1029/2003GC000616.

Spinelli, G. A., and D. Saffer (2004), Along-strike variations in underthrustsediment dewatering on the Nicoya margin, Costa Rica related to theupdip limit of seismicity, Geophys. Res. Lett. , 31 , L04613,doi:10.1029/2003GL018863.

Spinelli, G. A., L. Zuhlsdorff, A. T. Fisher, C. G. Wheat, M. Mottl, V. Spiess,and E. G. Giambalvo (2004a), Hydrothermal seepage patterns above aburied basement ridge, eastern flank of the Juan de Fuca Ridge, J. Geophys.Res., 109, B01102, doi:10.1029/2003JB002476.

Spinelli, G. A., E. G. Giambalvo, and A. T. Fisher (2004b), Sedimentpermeability, distribution, and influence on fluxes in oceanic basement,in Hydrogeology of the Oceanic Lithosphere, edited by E. E. Davis andH. Elderfield, pp. 151–188, Cambridge Univ. Press, Cambridge, U. K.

Stein, C., and S. Stein (1992), A model for the global variation in oceanicdepth and heat flow with lithospheric age, Nature, 359, 123–129.

Stein, J. S., and A. T. Fisher (2001), Lateral hydrothermal circulation beneaththe eastern flank of the Juan de Fuca Ridge: Thermal, chemical and mod-eling constraints,, Eos Trans. AGU, 82(47), Fall Meet. Suppl., AbstractS21A-0560.

Stein, J. S., and A. T. Fisher (2003), Observations and models of lateralhydrothermal circulation on a young ridge flank: Numerical evaluation ofthermal and chemical constraints, Geochem. Geophys. Geosyst., 4(3),1026, doi:10.1029/2002GC000415.

Thompson, R. E., E. Davis, and B. J. Burd (1995), Hydrothermal ventingand geothermal heating in Cascadia Basin, J. Geophys. Res., 100, 6121–6141.

Turcotte, D. L., and G. Schubert (1982), Geodynamics, 450 pp., JohnWiley, New York.

Underwood, M., K. D. Hoke, A. T. Fisher, E. G. Giambalvo, E. E. Davis, andL. Zuhlsdorff (2005), Provenance, stratigraphic architecture, and hydro-


18 of 19

B12101

geologic effects of turbidites in northwestern Cascadia Basin, PacificOcean, J. Sediment. Res., 75(1), 149–164.

Vacquier, V., J. G. Sclater, and C. E. Corry (1967), Studies in the thermalstate of the Earth, the 21st paper: Heat-flow, eastern Pacific, Bull. Earth-quake Res. Inst. Univ. Tokyo, 45, 375–393.

Villinger, H., I. Grevemeyer, N. Kaul, J. Hauschild, and M. Pfender (2002),Hydrothermal heat flux through aged oceanic crust: Where does the heatescape?, Earth Planet. Sci. Lett., 202(1), 159–170.

Von Herzen, R. P. (2004), Geothermal evidence for continuing hydrother-mal circulation in older (>60 Ma) ocean crust, in Hydrogeology of theOceanic Lithosphere, edited by E. E. Davis and H. Elderfield, pp. 414–450, Cambridge Univ. Press, Cambridge, U. K.

Von Herzen, R. P., and S. Uyeda (1963), Heat flow through the easternPacific Ocean floor, J. Geophys. Res., 68, 4219–4250.

Wang, K., and E. E. Davis (1992), Thermal effects of marine sedimentationin hydrothermally active areas, Geophys. J. Int., 110, 70–78.

Watts, A. B., and W. B. F. Ryan (1976), Flexure of lithosphere and con-tinental-margin basins, Tectonophysics, 36, 25–44.

Watts, A. B., and S. Zhong (2000), Observations of flexure and the rheol-ogy of oceanic lithosphere, Geophys. J. Int., 142(3), 855–875.

Wessel, P. (2001), Global distribution of seamounts inferred from griddedGeosat/ERS-1 altimetry, J. Geophys. Res., 106, 19,431–19,442.

Wheat, C. G. (1990), Fluid circulation and diagenesis in an off-axis hydro-thermal system: The Mariana Mounds, Ph.D. thesis, Univ. of Wash.,Seattle.

Wheat, C. G., and R. McDuff (1995), Mapping the fluid flow of the Mari-ana Mounds ridge-flank hydrothermal system: Pore water chemical tra-cers, J. Geophys. Res., 100, 8115–8131.

Wheat, C. G., and M. J. Mottl (1994), Hydrothermal circulation, Juan deFuca Ridge eastern flank: Factors controlling basement water composi-tion, J. Geophys. Res., 99, 3067–3080.

Wheat, C. G., and M. Mottl (2004), Geochemical fluxes through mid-oceanridge flanks, in Hydrogeology of the Oceanic Lithosphere, edited byE. E. Davis and H. Elderfield, pp. 627–658, Cambridge Univ. Press,Cambridge, U. K.

Wheat, C. G., M. J. Mottl, E. T. Baker, and R. A. Feely (1997), Chemicalplumes from low-temperature hydrothermal venting on the eastern flankof the Juan de Fuca Ridge, J. Geophys. Res., 102, 15,433–15,446.

Wheat, C. G., M. Mottl, A. Fisher, D. KadKo, E. Davis, and E. Baker(2004), Heat flow through a basaltic outcrop on a sedimented youngridge flank, Geochem. Geophys. Geosyst., 5, Q12006, doi:10.1029/2004GC000700.

Wilson, D. S., et al. (2005), Drilling to gabbro in intact ocean crust, Science,312, 1016–1020, doi:10.1126/science.1126090.

Zuehlsdorff, L., M. Hutnak, A. Fisher, V. Speiss, E. Davis, M. Nedimovic,S. Carbotte, H. Villinger, and K. Becker (2005), Data report: Site surveysrelated to IODP Expedition 301: ImageFlux (SO149) and RetroFlux(TN116) Expeditions and earlier studies, Proc. Integrated Ocean Drill.,301, doi:10.2204/iodp.proc.301.102.2005.

��A. T. Fisher and M. Hutnak, Department of Earth and Planetary Sciences,

University of California Santa Cruz, Santa Cruz, CA 95064, USA.([email protected])


19 of 19

B12101

Date post:	11-Jun-2020
Category:	Documents
Upload:	others
View:	3 times
Download:	0 times

Influence of sedimentation, local and regional ...afisher/CVpubs/pubs/Hutnak... · rapid...

Documents