Potential Effects of Hydrothermal Circulation and Magmatism on Heatﬂow at Hotspot Swells

Home , Gillian Foulger, Hydrothermal circulation

The Geological Society of America Special Paper 430 2007

Potential effects of hydrothermal circulation and magmatism on heatﬂow at hotspot swells

Carol A. Stein* Department of Earth and Environmental Sciences, University of Illinois at Chicago, 845 West Taylor Street, Chicago, Illinois 60607-7059, USA Richard P. Von Herzen* Department of Geology and Geophysics, Woods Hole Oceanographic Institution, Clark 244B, MS#22, Woods Hole, Massachusetts 02543, USA, and Department of Earth Sciences, University of California at Santa Cruz, Earth and Marine Science Building, Santa Cruz, California 95064-1077, USA

ABSTRACT

The lack of high heatflow values at hotspots has been interpreted as showing that the mechanism forming the associated swells is not reheating of the lower half of oceanic lithosphere. Alternatively, it has recently been proposed that the hotspot surface heatflow signature is obscured by fluid circulation. We re-examine closely spaced heatflow measurements near the Hawaii, Réunion, Crozet, Cape Verde, and Bermuda hotspots. We conclude that hydrothermal circulation may redistribute heat near the swell axes, but it does not mask a large and spatially broad heatflow anomaly. There may, however, be heatflow perturbations associated with the cooling of igneous intrusions emplaced during hotspot formation. Although such effects may raise heatflow at a few sites, the small heatflow anomalies indicate that the mechanisms producing hotspots do not significantly perturb the thermal state of the lithosphere.

Keywords: heatﬂow, swell, hotspot, mantle plume

INTRODUCTION anomalies at hotspots (e.g., Crough, 1978). A consequence of this mechanism is unusually high surface heatflow, which After the recognition that hotspots, regions of excess vol- should be measurable by established methodologies. The first canism and higher elevation compared to surrounding areas, re- detailed study of the Hawaiian hotspot found high mean heat- sult from the rise of material through the mantle to the Earth’s flow (Von Herzen et al., 1982), but subsequent studies showed surface beneath both oceans and continents, marine geothermal that the anomaly was overestimated because the reference measurements were among the first geophysical techniques model predicted significantly lower heatflow than observed in used to assess their nature. Initially it was proposed that signif- unperturbed older lithosphere (Von Herzen et al., 1989; Stein icant reheating of the bottom half of the lithosphere over a dis- and Stein, 1993). The conclusion was that lithospheric reheat- tance of ~300 km in diameter was required to match the depth ing is not a major factor in hotspot formation. Thus, possible

*E-mails: Stein, [email protected]; Von Herzen, [email protected].

Stein, C.A., and Von Herzen, R.P., 2007, Potential effects of hydrothermal circulation and magmatism on heatﬂow at hotspot swells, in Foulger, G.R., and Jurdy, D.M., eds., Plates, plumes, and planetary processes: Geological Society of America Special Paper 430, p. 261Ð274, doi: 10.1130/2007.2430(13). For permission to copy, contact [email protected]. ©2007 The Geological Society of America. All rights reserved. 261 262 Stein and Von Herzen

mechanisms producing the swells must be consistent with small on the answer to this question, different mechanisms have been or essentially zero heatflow anomalies, such as those expected proposed for hotspot formation. for dynamic models (e.g., Ribe and Christensen, 1994). Many mechanisms have been suggested to explain hotspot However, it has been recently proposed (Harris et al., formation. An important observation to model is the anom- 2000a,b) that hydrothermal circulation may obscure the signal alously shallow bathymetry and its subsequent subsidence (e.g., from the processes forming hotspots. Harris et al. (2000a, DeLaughter et al., 2005). For example, the reheating model re- p. 21,368) suggested that “fluid flow has the potential to obscure quires significant upwelling of mantle material that rapidly thins basal heatflow patterns associated with the Hawaiian hot spot,” the bottom half of the overlying plate (Crough, 1978, 1983). For which would “not bode well for capturing the form and magni- the dynamic plume model that incorporates a large viscosity tude of this signature.” Most hydrothermal circulation occurs in contrast between the plume and overlying lithosphere, most re- young oceanic crust (0 to ca. 65 Ma), resulting in lower mea- heating is at the base of the lithosphere (e.g., Liu and Chase, sured heatflow compared to conductive cooling models of the 1989; Ribe and Christensen, 1994). The compositional buoyancy lithosphere. Heatflow transferred by hydrothermal circulation model predicts shallower depths from relatively low-density in older oceanic lithosphere is thought to be much less (e.g., magmas emplaced either within the lithosphere or near its base Stein and Stein, 1994; Von Herzen, 2004). However, as sug- (e.g., Robinson, 1988; Sleep, 1994; Phipps Morgan et al., 1995). gested above, if the overall rough basement and islands and Different amounts of reheating and excess heatflow with time seamounts associated with hotspots cause fluid flow to occur since hotspot formation are expected from these different mech- both in the igneous basement and perhaps in the relatively per- anisms. Hence, heatflow data may help to discriminate among meable volcaniclastic sediments, then “capturing the maximum them. basal heat flux may be confounded by fluid flow” (Harris et al., For the Hawaiian chain, on the relatively fast-moving Pa- 2000a, p. 21,368). Evidence on the magnitude and extent of cific plate, matching the observed uplift and subsequent subsi- hydrothermal heat transfer may be provided by the detailed dence by the reheating mechanism requires that the bottom third heatflow at the five hotspots on crust older than ~65 Ma. to half of the oceanic lithosphere is rapidly reheated to as- In this article, we first examine heatflow anomalies near thenospheric temperatures over a zone 300 km wide centered on these hotspots. The anomalies are computed relative to the the axis (Von Herzen et al., 1982). At the time reheating begins, Global Depth and Heatflow reference model (Stein and Stein, the surface heatflow anomaly should be zero, but as the plate 1992) and with crustal ages from Mueller et al. (1997). Second, moves away from the upwelling, the heating of the lower litho- we examine which sites may be affected by hydrothermal cir- sphere propagates upward by conduction. The predicted anom- culation and thus mask higher heatflow. Last, we consider heat- aly gradually increases to a maximum of ~25 mW mÐ2 ~15Ð20 flow perturbations associated with the cooling of igneous struc- m.y. after reheating and then gradually decreases. The heatflow tures that result from the hotspot processes, including the igneous anomaly should be a maximum at the axis of the chain and de- material emplaced on the surface of pre-existing oceanic crust, crease to zero over ~600 km normal to the axis (Fig. 1A). and intruded sills at subcrustal depths. Although anomalously high heatflow was initially reported for the Hawaiian chain (Von Herzen et al., 1982), consistent with ESTIMATED HEATFLOW significant lithospheric reheating, subsequent data and analysis ANOMALIES FROM HOTSPOTS (Von Herzen et al., 1989; Stein and Stein, 1993) showed that the expected Gaussian distribution of anomalous heat flux orthogo- Although the depth in the Earth from which hotspots orig- nal to the axis did not exist, and many of the apparent anomalies inate is still debated (e.g., Sleep, 2003; Anderson, 2005; Foul- along axis result from comparing the data to reference thermal ger, this volume; Garnero et al., this volume), a geophysical models that underestimate heatflow elsewhere. Thus, the impli- challenge is to separate the effects of mechanisms forming cation is that heatflow data do not support swell formation by hotspots from shallower lithospheric processes, including the extensive lithospheric heating. “normal” lithospheric cooling as the plate ages. Lithospheric in- The small heatflow anomaly and absence of the pattern ex- teractions with the upward flow of magma, the absolute veloc- pected for lithospheric reheating at Hawaii and other marine ity of the plate, and near-surface physical processes (including hotspots (Réunion [Bonneville et al., 1997], Bermuda [Detrick excess volcanism and hydrothermal circulation) complicate the et al., 1986], Cape Verde [Courtney and White, 1986], and Crozet task of understanding the deeper mantle processes leading to the [Courtney and Recq, 1986]) imply that uplift may result from formation of hotspots. Also, because heat is largely transferred the dynamic effects of rising plumes. Their thermal effects are by conduction through the lithosphere, excess heat loss associ- concentrated at the base of the lithosphere and hence would raise ated with hotspot formation at a given location may occur long surface heatflow at most slightly, given that tens of millions after the hotspot upwelling has ceased. Although hotspots are of years are required for the conduction of heat to the surface. associated with anomalously shallow bathymetry, it is unclear For example, Ribe and Christensen (1994) predicted heatflow whether their surface heatflow is anomalously high. Depending anomalies for Hawaii of less than 1 mW mÐ2 (Fig. 1B). The heat- Potential effects of hydrothermal circulation and magmatism on heatflow 263

Figure 1. Predicted heatflow anomalies from reheating and dynamic plume models. (A) The reheating model predicts heatflow anomalies up to ~25 mW mÐ2 (Von Herzen et al., 1982) in contrast to (B) a dynamic plume model of less than 1 mW mÐ2 (Ribe and Christensen, 1994). The location of heatflow measurements for Hawaii from Von Herzen et al. (1982, 1989) and Harris et al. (2000a) are shown in panel A. In panel B, the predicted surface heatflow from an upwelling plume (location indicated by the solid semicir- cle) beneath a moving plate is shown for up to 2400 km downstream of where the plume has passed (horizontal axis) and for up to 1600 km perpendicular (vertical axis) from the path of the plume. F.Z. —fracture zone.

flow anomaly associated with compositional buoyancy depends hotspot frame of ~104 mm/year (Gripp and Gordon, 2002). Heat- on whether the magma is near the base of the lithosphere (e.g., flow measurements on ca. 84- to ca. 118-Ma crust show neither Sleep, 1994) or at relatively shallow depth in the lithosphere the expected anomalously high heatflow near the axis parallel to (e.g., Phipps Morgan et al., 1995). the relative motion of the hotspot track nor a decrease from the center of the axis of the swell toward the edges (Figs. 3A,B and DATA 4A). The anomalies are ~5Ð10 mW mÐ2 higher than expected for similar aged lithosphere (Von Herzen et al., 1989; Stein and Of the five oceanic hotspots (Hawaii, Réunion, Cape Verde, Stein, 1992). These anomalies may reflect lithosphere, before Bermuda, and Crozet) with closely spaced “pogo”1 heatflow perturbation by the hotspot processes, having a somewhat measurements (Fig. 2), Hawaii is the best studied (Von Herzen higher heatflow than typical, but still within the observed range et al., 1982, 1989; Harris et al., 2000a). The ca. 81-Ma hotspot of variability for its age (DeLaughter et al., 2005). track (e.g., Keller et al., 1995; Clouard and Bonneville, 2005; Réunion is the location of most recent volcanism for a ca. Norton, this volume) presently has an absolute velocity in the 65-Ma hotspot track (e.g., Bonneville et al., 1997) with a present-day ~13 mm/year absolute velocity in the hotspot reference frame (Gripp and Gordon, 2002). The heatflow measurements 1For “pogo” measurements, the equipment is initially lowered to the seafloor, were made in the form of long (~200 and ~400 km) parallel pro- and then after heatflow is determined at a location it is raised slightly above the files ~100 km apart extending westward from the Mascarene seafloor and towed to other nearby sites, allowing for closely spaced measurements typically for ~24 hours, before the equipment is returned to the ship. Ridge (perpendicular to the hotspot track) on crust affected by Hence, it is similar to jumping with a pogo stick. hotspot volcanism ~15Ð20 m.y. ago (Fig. 2B). The crustal age is 264 Stein and Von Herzen

Figure 2. Bathymetry and sites with closely spaced heatflow measurements for (A) Hawaii, (B) Réunion, (C) Crozet, (D) Cape Verde, and (E) Bermuda. Von Herzen (2004) examined heatflow and bathymetry to infer whether the sites have hydrothermal circulation (red; diamond), probably have hydrothermal circulation (orange; upward-pointing tri- angle), probably have conductive heat transfer (green; downward-pointing tri- angle), or have conductive heat transfer (blue; square). Sites without a determi- nation are shown as black circles. Most sites have conductive heat transfer. The associated numbers and letters next to the heatflow locations are the site names from the published heatflow papers. For Crozet, the upper-right inset is an en- largement of the region with the islands indicated by the small box. The eastern islands (Po—Île de la Possession; E—Île de l’Est) are separated by the Indivat Basin (Ind. B.) from the western islands (Pi—Île des Pingouins; C—Île aux Co- chons; and A—Îles des Apotres).

ca. 65 Ma west of the Mahanoro fracture zone. To the east, the pected increase in the heatflow anomaly toward the axis of the crustal ages from Mueller et al. (1997) are ~10Ð15 m.y. younger hotspot from a lithospheric reheating mechanism is not ob- than those used by Bonneville et al. (1997; derived from Dy- served (Fig. 3C,D). ment, 1991) to analyze their heatflow measurements. With the Unlike the Hawaiian and Réunion hotspots with clear and revised younger crustal ages, many measured heatflow values long volcanic signatures, volcanism for Crozet on the Antarctic to the east of the Mahanoro fracture zone, which were initially plate started at ca. 8 Ma and lacks a clear track (Fig. 2C). The interpreted as greater than the expected average by Bonneville shallow Crozet Bank appears to have been emplaced on old, me- et al. (1997), are now about that expected, thus suggesting no chanically thick crust different from the adjacent Del Cano Rise, heatflow anomaly. The heatflow anomaly is largest west of formed on or near the Southwest Indian ridge before anomaly the Mahanoro fracture zone (~6 mW mÐ2 for the northern pro- 24 time (Goslin and Diament, 1987; Recq et al., 1998). The file, although it may be affected by hydrothermal circulation) Crozet Bank’s eastern islands (ële de la Possession and ële de and decreases toward the Mascarene Ridge. Hence, the ex- l’Est) are separated by the Indivat basin from the western islands Potential effects of hydrothermal circulation and magmatism on heatflow 265

Figure 3. Heatflow anomaly with distance for profiles taken perpendicular to a hotspot track for Hawaii and Réunion (see locations from Figure 2A,B). Al- though the heatflow may be ~5 mW mÐ2 greater than predicted by the Global Depth and Heatflow reference model, the heatflow anomaly does not show the expected Gaussian-type pattern from significant lithospheric reheating with a large maximum at the axis (0 km). Sim- ilarly, few sites show evidence of hydrothermal circulation (Von Herzen, 2004). Symbols as for Figure 2. Horizontal bar at left in panels A and B represents mean heatflow anomaly on incoming Pacific plate before reaching hotspot. F.Z.— fracture zone.

(ële des Pingouins, ële aux Cochons, and ëles des Apotres). The the volcanism, have heatflow approximately equal to that ex- eastern islands are usually thought to be the oldest, with the old- pected for the crustal age. It is difficult to interpret one isolated est dated volcanism at ca. 8 Ma, but detailed paleomagnetic measurement, M3, at the base of the shallowing volcanic con- work indicates that volcanism also occurred from 5 to 0.5 Ma struct of the archipelago, where the heatflow is ~35 mW mÐ2 on the eastern island of Possession (Camps et al., 2001). Also, greater than expected. This heatflow value is discussed later in the western island of Apotres has volcanism dated as old as 5.5 the article. Ma (Giret et al., 2003). Both western and eastern islands have Two additional heatflow surveys are on swells perturbing some Holocene volcanism. The relatively few heatflow mea- old crust. Cape Verde is on the African plate (Courtney and surements for the Crozet hotspot are all on ca. 68-Ma crust ex- White, 1986) and Bermuda (Detrick et al., 1986) is on the North tending southeast from the eastern islands (Courtney and Recq, American plate (Fig. 2D,E). Some characteristics of Cape Verde 1986) and presumably into crust unperturbed by the hotspot are similar to those of Crozet. Volcanism producing the islands activity. Of the three reliable multipenetration sites, M4, with is relatively recent on ca. 125-Ma crust. The easternmost islands the shallowest seafloor, has a heatflow anomaly of ~14 mW mÐ2 formed ca. 10Ð20 Ma, with the western ones younger than 8 Ma (Fig. 4B). The other two sites, M5 and M6, farther away from (Courtney and White, 1986). The absolute velocity in the hotspot 266 Stein and Von Herzen

Figure 4. Heatflow anomaly with distance for profiles taken parallel to a hotspot track for (A) Hawaii. (B) Crozet, (C) Cape Verde, and (D) Bermuda have no clear hotspot track, so distances for the sites are measured from the local geoid maximum in the approximate direction of the heatflow sites. Symbols as for Figure 2. Horizontal bar at left in panel A represents mean heatflow anomaly on incoming Pacific plate before reaching hotspot. GDH1—Global Depth and Heatflow reference model.

reference frame for Cape Verde is somewhat higher (~21 within a few hundred kilometers of the maximum geoid high mm/year) than for Crozet (~6 mm/year) (Gripp and Gordon, (Fig. 4C). 2002). As with Crozet, the heatflow sites are located on a pro- Bermuda (Fig. 2E), on the North American plate (with an file of approximately similar age crust extending from near the absolute velocity of ~38 mm/year in the hotspot reference frame; volcanic islands into crust unperturbed by the hotspot activity. Gripp and Gordon, 2002), is a shallow region without volcan- Of the seven sites with closely spaced measurements (Fig. 2D), ism for ~33 m.y. Bermuda may be either a “one-shot” hotspot site A was considered unreliable because of the difficulty pene- or perhaps part of a track with a very large spatial gap in vol- trating the seafloor with the heatflow probe, the lack of measured canism (e.g., Vogt and Jung, this volume). Compared to the other thermal conductivity, and likelihood of bottom erosion to the four hotspot regions, the “pogo” measurements at Bermuda (De- seafloor (Courtney and White, 1986). We have included the val- trick et al., 1986) have a similarly small maximum anomaly (~8 ues here, using an estimated thermal conductivity of 1 W/(m K). mW mÐ2). Also, a heatflow value determined from a drill hole The heatflow anomaly is relatively high (~9 mW mÐ2) at site C, on the island of Bermuda (Hyndman et al., 1974) has a small (~5 Potential effects of hydrothermal circulation and magmatism on heatflow 267

mW mÐ2) anomaly. The heatflow data do not display the broad transfer. Forty-four of the fifty-eight pogo surveys are on or near anomaly pattern expected for significant thermal rejuvenation of the swells discussed in this article (Fig. 2). For the surveys as- the lithosphere (Fig. 4D), although the mean heatflow values are sociated with hotspots, Von Herzen (2004) suggested that twenty- systematically distributed with distance from Bermuda. four (~55%) have conductive heatflow, eight (~18%) probably have conductive heatflow, eight (~18%) have convection, and HYDROTHERMAL CIRCULATION four (~9%) probably have convection. About 27% of the surveys near the swells were interpreted “Fluid flow has the potential to obscure basal heatflow pat- as having or probably having convection. Most are within ~200 terns associated with the Hawaiian hot spot” (Harris et al., 2000a, km of the swell axes (Figs. 3 and 4). These sites include five that p. 21,368). Hydrothermal circulation is assumed to transfer rel- are close to the Hawaiian axis (site #4, two sections of the Maro atively little heat in oceanic lithosphere older than ca. 65 Ma Reef profile, and two sections of the Oahu profile; Fig. 2A), the (e.g., Stein and Stein, 1994; Von Herzen, 2004), about the min- two closest sites to Bermuda (sites #5 and #6), and site C from imum crustal age in these surveys. However, even in old crust, Cape Verde, ~40 km from the geoid high and closest to the a few areas have been interpreted as having hydrothermal cir- youngest region of volcanism. Conversely, most heatflow sites culation (e.g., Noel and Hounslow, 1988; Fisher and Von Herzen, off axis are apparently not affected by hydrothermal circulation. 2005). Also, areas where isolated igneous basement peaks pen- Thus, it is doubtful that hydrothermal circulation masks a high etrate or are only shallowly buried by the sediment cover are and wide heatflow anomaly associated with lithospheric reheat- thought to be the main flow path for fluids between the over- ing mechanism for the formation of the swells, but it may occur lying ocean and the crystalline crust (e.g., Harris et al., 2004). near the swell axes. This effect has been observed on 3.5-Ma crust on the Juan de Fuca plate (Mottl et al., 1998) and inferred from heatflow mea- VOLCANISM sured on 83-Ma crust of the northwestern Atlantic (Embley et al., 1983). Studies suggest that pore water circulation removes heat Although magmatic activity is associated with the for- from the surrounding area, extending a few to tens of kilometers mation of hotspot swells, its thermal effects are typically not in- distance from (near-) outcrops of basement rock (e.g., Stein and cluded in heat flux estimates. Crustal thickness at hotspots is Stein, 1997; Fisher et al., 2003). significantly greater than for normal oceanic crust (e.g., White An additional factor for hotspots is that if hydrothermal cir- et al., 1992), although the conditions producing the melt have culation removes large amounts of heat near seamounts, signif- large uncertainties (e.g., Sallarès and Calahorrano, this volume). icant pore water flow may also occur within the relatively highly The emplacement of the additional hot material at relatively permeable volcaniclastic sediments. The permeabilities of sedi- shallow depths should have a thermal signature. Although some ment obtained for the Canary Islands seafloor by Urgeles et al. excess crust is associated with the near-surface volcanism con- (1999) are as high as ~10−16 m2. However, the mostly thick sed- structing the islands, some hotspots have a deeper sill structure iments at the hotspots we discuss should prevent any significant intruded near the base of oceanic crust. Such sills are deduced vertical flow of water (e.g., Spinelli et al., 2004) and thus any from seismic data for Hawaii (ten Brink and Brocher, 1987) significant advective heat transfer. Slow horizontal pore water (Fig. 5A,B) and the Marquesas (Caress et al., 1995; McNutt and velocities may occur, given this sediment permeability, but Bonneville, 2000), but others do not appear to have obvious un- modeling of flow in oceanic crystalline basement suggests that derplating (e.g., Crozet; Recq et al., 1998). Next, we examine flow sufficient to transport significant amounts of heat requires the heat flux associated with the cooling of the island/seamount higher permeabilities, at a minimum roughly three to five orders chain and the underplated sill for Hawaii, and the implications of magnitude larger (Fisher et al., 2003; Stein and Fisher, 2003; for the other heatflow surveys on hotspots. Hutnak et al., 2006; A. Fisher, personal commun., 2006). For the sediments, for some models, Harris et al. (2000b, p. 21,382) Total Heat Contribution “found that horizontal permeabilities as large as 10-11 m2 were needed to produce significant perturbations to the surface heat- The contribution of the cooling magma to the geothermal flow.” budget in part depends on its production with time. We estimate To test this hypothesis of significant hydrothermal circula- the production rate at the Hawaiian hotspot from the magma tion, we use the analysis of fifty-eight high-quality closely cross-section geometry (Fig. 5B) and the absolute plate veloc- spaced heatflow measurements (“pogo” surveys; Von Herzen, ity of ~100 mm/year. Assuming that erosion of Oahu has re- 2004). This approach used the correlation of heatflow variations duced its initial volume by a factor ranging between 0.3 and 0.5, with the shape of the basaltic basement and sediment cover to the mean magma volume generation/unit time (V) is ~3Ð6 m3/s, infer whether areas have hydrothermal circulation. Von Herzen comparable to other estimates of the rate of Hawaiian magma- (2004) classified the sites in four categories: (1) having hydro- tism (Swanson, 1972; Shaw, 1973; Wadge, 1980). thermal circulation, (2) perhaps having hydrothermal circula- The total thermal power (P) from the magma is its latent tion, (3) uncertain if all conductive, and (4) conductive heat heat from solidification (PL) plus the cooling (PT). For the οο A 158 156 W

A’ ο 22

Oahu

ο A 20 N Hawaii 0 200 km

B AA’ 0

infil volcanic 4 edifice pre-existing 8 oceanic crust km

sill 16

20 -3000 300 km

C -2 120

100 Hawaii Marquesas 80

40 MARO REEF 20 OAHU

HEAT FLOW ANOMALY (mW m ) ANOMALY FLOW HEAT 0 0 5 10 15 20 AGE SINCE SILL EMPLACEMENT (Ma) Potential effects of hydrothermal circulation and magmatism on heatﬂow 269

“island” region, the total thermal power is given in equation 1 TABLE 1. PHYSICAL PROPERTIES ASSUMED and values for the parameters from Table 1: FOR THE COOLING MAGMA

Symbol Term Value P = PL + PT = Vρ (L + c ΔT) = 1Ð2 × 104 MW. (1) ρ Density 3000 kg/m3 c Heat capacity 800 J/kg-K Additional heat comes from the deeper sill that has an av- k Thermal conductivity 2.2 W/m-K erage thickness of 2.5 km. We estimate that the volume of the κ Thermal diffusivity 8 × 10–7 m2/s sill is about half that of the edifice of the islands before erosion. L Latent heat of melting 3 × 105 J/kg Δ Although the emplacement of the underplated material may T Temperature drop after solidification 1000 K consist of thinner sills that average to a lower effective initial temperature, for our calculations, we assume a single sill initially at the solidus temperature (~1000 ¡C above the intruded to 7 and 9.5 km, respectively, relative to the seafloor depths rock), to maximize its thermal effects. Thus, including the con- away from the edifice. The surface heatflow for the deep sill is tribution of both the “island” and the sill, the total power is 1.5 zero at emplacement and then rapidly increases to a maximum Ð 3 × 104 MW (50% greater than that estimated in equation 1). of ~75 mW m−2 at ~0.5 m.y. after emplacement, but subse- Another way of examining this issue is to calculate the power quently rapidly decreases with increasing age since emplace- released from the cooling for a volume of magma, 1-m2 area ment. This model predicts excess heatflow of ~10 mW m−2 at with a thickness of 2.5 km. The total power, 8.25 × 1012 J, for ca. 3.5 Ma (near Oahu) but only ~2 mW mÐ2 at ca. 15 Ma (near the square-meter area divided by ~10 m.y. of cooling produces Maro Reef). However, the observed heatflow anomalies (mea- an average heatflow of ~26 mW m−2 or three times that if the sured less than expected from GDH1; Stein and Stein, 1992) “island” and eroded material are included. along the Oahu and Maro Reef profiles (Fig. 6) to ~150 km from the islands are about the same value and similar to other meas- The Sill Model urements for the Hawaiian chain (~5Ð10 mW m−2). After correcting the heatflow values at Oahu and Maro Reef The cooling of a magmatic underplated region should have for thermal blanketing caused by deposition of cold sediment a longer-term effect compared to the near-surface volcanism. At (e.g., Von Herzen and Uyeda, 1963; Hutchison, 1985), the data mid-ocean ridges, heatflow rapidly decreases from cooling lava are consistent with the suggestion that additional heat is derived (Johnson and Hutnak, 1997), and heat is quickly dissipated by from an intrusive body. The “corrected” heatflow (Harris et al., seawater. Similarly, on land, atmospheric cooling dissipates heat 2000a) at Oahu is anomalous by ~24 mW m−2 and at Maro after an eruption. The surface heatflow from the underplated by ~12 mW m−2, approximately the expected difference in heat- area with time is calculated assuming 1-D cooling of an infi- flow at these locations modeled by conductive cooling of the sill nitely extending horizontal sill, appropriate because of the large intrusion. However, the amount of the correction depends ratio of width to thickness, such as observed near Oahu, Hawaii. greatly on such factors as the sedimentation rate, thermal con- The analytical solution for the surface heat flux (q) with time (t) ductivity of the sediments, and assumed boundary conditions at is given in Von Herzen et al. (1989, equation 5; initially obtained the sediment/igneous crust interface. Most oceanic sites have from Carslaw and Jaeger, 1959) as: sedimentation rates that are too low to require corrections, and for those with somewhat thicker sediment, the corrections are q(t) = [kΔT/(πκt)1/2]{exp[Ða2/(4κt)] Ð exp[Ðb2/(4κt)]}, (2) typically small (a few percent). In contrast, for the measurement sites closest to Oahu and Maro Reef, ~2 km of sediment have where a and b are the depths of the upper and lower sill surfaces, been deposited in the moat formed by the “loading” of the is- respectively and the other parameters are from Table 1. lands on the crust. The calculated correction is large if the heat- The deep sill structure found by ten Brink and Brocher flow at the sediment-basement interface varies with normal con- (1987) is ~200 km wide with an average thickness of 2.5 km, ductive cooling of oceanic lithosphere, as assumed by Harris and we presume it was formed by intruded magma associated et al. (2000a). However, the correction is somewhat smaller if with the hotspot. The heat flux over time is illustrated in Figure vigorous hydrothermal circulation occurs within the top of the 5C, for values of the top (a) and the bottom (b) of the sill equal igneous oceanic crust (Wang and Davis, 1992). Additional uncertainties for the calculation are introduced if significant hydrothermal circulation occurs within permeable near-surface vol- Figure 5. Sill intrusions and heatflow anomalies. (A) Location of caniclastic sediments. Thus, for Oahu, a case can be made for or seismic study across Oahu shown in panel B. (B) Cross-section with against additional heat from igneous intrusions, depending on volcanic sill and edifice above (from ten Brink and Brocher, 1987). assumptions. (C) Heatflow anomaly expected for the observed sills (with an average thickness of 2.5 km) under Oahu and (with an average thickness of Magmatic activity may explain the higher heatflow at two 5 km) under the Marquesas. Heatflow for Hawaii is a maximum at sites near Crozet and perhaps one at Cape Verde. For Crozet, at ca. 0.75 Ma and decreases with increasing age. station M3 the heatflow anomaly is ~35 mW m−2, ~50 km from 270 Stein and Von Herzen

Figure 6. (A) Heatflow anomalies, (B) measured and sediment-corrected heatflow, and (C) depth-converted seismic reflection profiles for (left) Oahu and (right) Maro Reef. The top panels show the heatflow anomalies for the measured (solid circles) and sediment-corrected (purple tri- angles) heatflow. Blue circles—no hydrothermal circulation inferred; red circles—hydrothermal circulation inferred; black circles—no determi- nation of circulation. The middle panel shows the measured (open circles) and sediment-corrected (solid circles) heatflow values. The location of these Hawaiian measurements are shown in Figure 2A. The middle and lower panels are from Harris et al. (2000a, Figs. 10 and 11).

the Holocene volcanism on Île de l’Est; and at M4, the heatflow nificant lithospheric reheating does not occur, and the upwelling anomaly is ~15 mW m−2, ~100 km distant (Fig. 4B). The inter- plume of material must have a substantially smaller thermal im- pretation at M3 is problematic, because only one reliable mea- print than the suggested 300-km diameter (e.g., Crough, 1978). surement was made, but the site is clearly on the slope of the volcanic structure. At Cape Verde, station C, the ~10-mWmÐ2 heat- Issues with Hydrothermal Circulation flow anomaly is within ~100 km of Sao Vincent, an island with Holocene volcanism (Fig. 2D). It seems possible that hydrothermal circulation may redistribute some heat within ~200 km of the swell axis. However, it DISCUSSION AND CONCLUSIONS does not mask a large and spatially broad heatflow anomaly expected from reheating of a significant portion of lower litho- Reanalysis of closely spaced heatflow surveys near five sphere (e.g., Crough, 1978). hotspots confirms that the heatflow anomalies are quite small, Hydrothermal circulation occurs at the two sites nearest to generally less than ~10 mW m−2, and do not have the Gaussian- Bermuda’s volcanic edifice. Unlike the four other hotspots with shaped distribution of values extending ~600 km away from the present-day volcanism, the most recent volcanic activity here is swell axis expected from the lithospheric reheating models. This dated at ca. 33 Ma (Vogt and Jung, this volume). Bermuda does finding accords with previous studies (Von Herzen et al., 1989; not appear to be associated with low mantle seismic velocities, Stein and Stein, 1993; DeLaughter et al., 2005) showing that sig- as inferred at the other four active hotspots (e.g., Zhao, 2004; 60 80 100 120 140 160 180 80 80 )

2 Swell

70 Non-Swell 70 Global Average Heat Flow (Stein, 2003) 60 60

Figure 7. Heatﬂow data with crustal age 50 50 for the swell (solid symbols) and non- swell (open symbols) “pogo” surveys

Average Heat Flow (mW/m analyzed by Von Herzen (2004). (A) The average heatflow from each survey with 40 40 age is similar to the observed global 60 80 100 120 140 160 180 average heatflow (Stein, 2003). (B) The standard deviations of the pogo surveys AGE (Myr) are significantly less than the global av- 60 80 100 120 140 160 180 erage. Colors as for Figure 2. Two sites from Von Herzen (2004) are not in- 30 30 cluded in this figure. One located near a

) seamount has extremely high and vari- 2 Approximate Global able heatflow (presumably due to fluid flow) and the other, in the Gulf of Mex- Standard Deviation ico, has a large uncertainty in crustal age. 20 20

10 10 Standard Deviation (mW/m

0 0 60 80 100 120 140 160 180 272 Stein and Von Herzen

Dziewonski, 2005; Ritsema, 2005). The three sites with the high- the Oahu and Maro Reef profiles (Fig. 6). These sites are on the est heatflow anomalies are located near the base of the volcanic relatively thinly sedimented flexed arches that have been asso- edifice (site #4), and the others (sites #5 and #6) are higher on ciated with relatively recent and extensive seafloor volcanism the structure (Fig. 2C). The first is interpreted as having (Clague et al., 2002). Similarly, the highest heatflow for Crozet conductive heat transfer but the other two as perhaps having con- (M3; Fig. 2C) is on the volcanic edifice near recent volcanism. vection (Von Herzen, 2004). Hence, some higher than expected Additional work on hotspots would be useful. As far as we heatflow near the volcanic edifices may result from convection, know, no pogo heatflow measurements have been carried out on rather than from volcanic processes. the Marquesas. Seismic data from the Marquesas hotspot (Caress et al., 1995, Fig. 4) indicate a similarly shaped, but substantially Heatflow and Absolute Plate Velocity larger, subcrustal intrusion compared to Hawaii, with ~10 km maximum thickness at the center, thinning to ~3Ð4 km at 150 km The maximum heatflow anomalies for all five hotspots are distance to either side. Given this intrusion and the recent age of about the same, yet the duration of the hotspot tracks and the ab- the chain (less than 6 Ma; Desonie et al., 1993), the Marquesas solute plate velocities and age are quite different. Maximum may be an excellent location for future study (Fig. 5C). heatflow anomalies for the two long hotspot tracks of Hawaii and Réunion are approximately the same as observed at the vol- canically active, more concentrated, large volcanic highs of ACKNOWLEDGMENTS Crozet and Cape Verde (Fig. 2). Despite the large range of absolute plate velocities for these five hotspots (6–104 mm/year) We thank Andy Fisher and Seth Stein for important discussions. there is no clear trend of larger heatflow anomalies with de- Constructive reviews by Gillian Foulger, Rob Harris, Donna creasing absolute plate velocity, as might be expected if volcanic Jurdy, and Will Sager significantly improved the article. Also, we activity occurred over a longer period of time in one region. thank Gillian Foulger for organizing the stimulating Chapman Conference, The Great Plume Debate, and inviting this contribution. The GMT software package (Wessel and Smith, 1998) Pogo Surveys and the Global Heatflow Data was used to make some of the figures in this article. We have examined how the heatflow on and off swells varies with lithospheric age and compares to the global data set REFERENCES CITED (Fig. 7A). Swell sites are those sites on the regions of shallow depths associated with the hotspots discussed in this article. Off- Anderson, D.L., 2005. Scoring hotspots: The plume and plate paradigms, in swell sites include those away from the shallow depth regions Plates, plumes, and paradigms, Foulger, G.R., et al., eds., Plates, plumes, and others discussed in Von Herzen (2004). On average, pogo and paradigms: Boulder, Colorado, Geological Society of America Special surveys have heatflow values similar to global averages for Paper 388, p. 31Ð54, doi: 10.1130/2005.2388(04). Bonneville, A., Von Herzen, R.P., and Lucazeau, F., 1997, Heatflow over oceanic lithosphere. Réunion hot spot track; Additional evidence for thermal rejuvenation of Overall, the standard deviations for the pogo sites (Fig. 7B) oceanic lithosphere: Journal of Geophysical Research, v. 102, p. 22,731Ð are much lower than for widely spaced measurements on simi- 22,747, doi: 10.1029/97JB00952. larly aged crust (Stein and Stein, 1994). There is a slight de- Camps, P., Henry, B., Prevot, M., and Faynot, L., 2001, Geomagnetic paleo- crease in the standard deviations for the pogo surveys with secular variation recorded in Plio-Pleistocene volcanic rocks from Posses- sion Island (Crozet Archipelago, southern Indian Ocean): Journal of Geo- increasing age. These observations suggest that although hy- physical Research, v. 106, p. 1961Ð1971, doi: 10.1029/2000JB900370. drothermal circulation transfers relatively little heat in old Caress, D.W., McNutt, M.K., Detrick, R.S., and Mutter, J.C., 1995, Seismic oceans, its effect is still slightly decreasing with age, even in imaging of hotspot-related crustal underplating beneath the Marquesas Mesozoic crust. Von Herzen (2004) reached a similar conclu- Islands: Nature, v. 373, p. 600Ð603, doi: 10.1038/373600a0. sion based on the decreasing percentage with increasing crustal Carslaw, H.S., and Jaeger, J.C., 1959. Conduction of heat in solids: Oxford, Oxford University Press, 510 p. age of pogo survey sites on old oceanic lithosphere that were in- Clague, D.A., Uto, K., Satsake, K., and Davis, A.S., 2002. Eruption style and terpreted to have hydrothermal circulation. flow emplacement in the submarine North Arch volcanic field, Hawaii, in Takahashi, E., et al., eds., Hawaiian volcanoes: Deep underwater per- Future Work spectives: Washington, D.C., American Geophysical Union Geophysical Monograph 128, p. 65Ð84. Clouard, V., and Bonneville, A., 2005. Age of seamounts, islands, and plateaus The obvious difficulty in interpreting the data for swells is on the Pacific plate, in Foulger, G.R., et al., eds., Plates, plumes, and para- the low density of heatflow measurements. For example, al- digms: Boulder, Colorado, Geological Society of America Special Paper though the Hawaiian Swell has the most extensive geothermal, 388, p. 71Ð90, doi: 10.1130/2005.2388(06). geological, and geophysical data compared to other hotspots, it Courtney, R.C., and Recq, M., 1986, Anomalous heatflow near the Crozet Plateau and mantle convection: Earth and Planetary Science Letters, v. 79, is still difficult to determine the extent and thermal effects of p. 373Ð384, doi: 10.1016/0012-821X(86)90193-7. hotspot magmatism. It is interesting to note that a high measured Courtney, R.C., and White, R.S., 1986, Anomalous heatflow and geoid across heatflow anomaly for Hawaii occurs at the northeastern ends of the Cape Verde Rise: Evidence for dynamic support from a thermal plume Potential effects of hydrothermal circulation and magmatism on heatflow 273

in the mantle: Geophysical Journal of the Royal Astronomical Society, Harris, R.N., Fisher, A.T., and Chapman, D., 2004, Seamounts induce large fluid v. 87, p. 815Ð867. fluxes: Geology, v. 32, p. 725Ð728, doi: 10.1130/G20387.1. Crough, S.T., 1978, Thermal origin of mid-plate hot-spot swells: Geophysical Hutchison, I., 1985, The effects of sedimentation and compaction on oceanic Journal of the Royal Astronomical Society, v. 355, p. 451Ð469. heatflow: Geophysical Journal of the Royal Astronomical Society, v. 82, Crough, S.T., 1983, Hotspot swells: Annual Review of Earth and Planetary Sci- p. 439Ð459. ences, v. 11, p. 165Ð193, doi: 10.1146/annurev.ea.11.050183.001121. Hutnak, M., Fisher, A.T., Zuhlsdorff, L., Spiess, V., Stauffer, P., and Gable, C.W., DeLaughter, J.E., Stein, C.A., and Stein, S., 2005. Hotspots: A view from the 2006, Hydrothermal recharge and discharge guided by basement outcrops swells, in Foulger, G.R., et al., eds., Plates, plumes, and paradigms: Boul- on 0.7–3.6 Ma seafloor east of the Juan de Fuca Ridge: Observations and der, Colorado, Geological Society of America Special Paper 388, p. 257Ð numerical models: Geochemistry, Geophysics, Geosystems, v. 7, doi: 278, doi: 10.1130/2005.2388(16). 10.1029/2006GC001242. Desonie, D.L., Duncan, R.A., and Natland, J.H., 1993, Temporal and geochem- Hyndman, R.D., Muecke, G.K., and Aumento, F., 1974, Deep Drill 1972. Heat- ical variability of volcanic products of the Marquesas hotspot: Journal of flow and heat production in Bermuda: Canadian Journal of Earth Sciences, Geophysical Research, v. 98, p. 17,649Ð17,665. v. 11, p. 809Ð818. Detrick, R.S., Von Herzen, R.P., Parsons, B., Sandwell, D., and Dougherty, M., Johnson, P., and Hutnak, M., 1997, Conductive heat loss in recent eruptions at 1986, Heatflow observations on the Bermuda Rise and thermal models of mid-ocean ridges: Geophysical Research Letters, v. 24, p. 3089Ð3092, doi: midplate swells: Journal of Geophysical Research, v. 91, p. 3701Ð3723. 10.1029/97GL02998. Dyment, J., 1991. Structure et évolution de la lithosphère océanique dans l’océan Keller, R.A., Duncan, R.A., and Fisk, M.R., 1995. Geochemistry and 40Ar/39Ar Indien: Apport des anomalies magnétiques [Ph.D. thesis]: Strasbourg, geochronology of basalts from ODP leg 145 (North Pacific transect), in Rea, France, University of Strasbourg, 374 p. D.K., et al., eds., Proceedings of the Ocean Drilling Program, Scientific Re- Dziewonski, A.M., 2005. The robust aspects of global seismic tomography, in sults (vol. 145): College Station, Texas, Ocean Drilling Program, p. 333Ð344. Foulger, G.R., et al., eds., Plates, plumes, and paradigms: Boulder, Col- Liu, M., and Chase, C.G., 1989, Evolution of midplate hotspot swells: Numer- orado, Geological Society of America Special Paper 388, p. 147Ð154, doi: ical solutions: Journal of Geophysical Research, v. 94, p. 5571Ð5584. 10.1130/2005.2388(09). Mottl, M.J., Wheat, G., Baker, E., Becker, N., Davis, E., Feely, A.R., Grehan, Embley, R.W., Hobart, M.A., Anderson, R.N., and Abbott, D., 1983, Anomalous A., Kadko, D., Lilley, M., Massoth, G., Moyer, C., and Sansone, F., 1998, heatflow in the northwest Atlantic: A case for continued hydrothermal cir- Warm springs discovered on 3.5 Ma oceanic crust, eastern flank of the Juan culation in 80-m.y. crust: Journal of Geophysical Research, v. 88, p. 1067Ð de Fuca Ridge: Geology, v. 26, p. 51Ð54, doi: 10.1130/0091-7613(1998) 1074. 026<0051:WSDOMO>2.3.CO;2. Fisher, A.T., Davis, E.E., Hutnak, M., Spiess, V., Zühlsdorff, L., Cherkaoui, A., McNutt, M., and Bonneville, A., 2000. A shallow, chemical origin for the Mar- Christiansen, L., Edwards, K.M., Macdonald, R., Villinger, H., Mottl, M., quesas Swell: Geochemistry, Geophysics, Geosystems, v. 1, doi: 10.1029/ Wheat, C.G., and Becker, K., 2003, Hydrothermal circulation across 50 km 1999GC00028. on a young ridge flank: The role of seamounts in guiding recharge and dis- Mueller, R.D., Roest, W.R., Royer, J.-Y., Gahagan, L.M., and Sclater, J.G., 1997. charge at a crustal scale: Nature, v. 421, p. 618Ð621, doi: 10.1038/nature Digital isochrons of the world’s ocean floor: Journal of Geophysical Re- 01352. search, v. 102, p. 3211Ð3214 (ftp://ftp.es.usyd.edu.au/pub/agegrid/, version Fisher, A.T., and Von Herzen, R.P., 2005, Models of hydrothermal circulation 1.6). within 106 Ma seafloor: Constraints on the vigor of fluid circulation and Noel, M., and Hounslow, M.W., 1988, Heatflow evidence for hydrothermal crustal properties below the Madeira Abyssal Plain: Geochemistry, Geo- convection in Cretaceous crust of the Madiera Abyssal Plain: Earth and physics, Geosystems, v. 6, doi: 10.1029/2005GC001013. Planetary Science Letters, v. 90, p. 77Ð86, doi: 10.1016/0012-821X(88) Foulger, G.R., 2007 (this volume), The “plate” model for the genesis of melting 90113-6. anomalies, in Foulger, G.R., and Jurdy, D.M., eds., Plates, plumes, and Norton, I.O., 2007 (this volume), Speculations on Cretaceous tectonic history planetary processes: Boulder, Colorado, Geological Society of America of the northwest Pacific and a tectonic origin for the Hawaii hotspot, in Special Paper 430, doi: 10.1130/2007.2430(01). Foulger, G.R., and Jurdy, D.M., eds., Plates, plumes, and planetary processes: Garnero, E.J., Lay, T., and McNamara, A., 2007 (this volume), Implications of Boulder, Colorado, Geological Society of America Special Paper 430, doi: lower-mantle structural heterogeneity for existence and nature of whole- 10.1130/2007.2430(22). mantle plumes, in Foulger, G.R., and Jurdy, D.M., eds., Plates, plumes, and Phipps Morgan, J., Morgan, W.J., and Price, E., 1995. Hotspot melting gener- planetary processes: Boulder, Colorado, Geological Society of America ates both hotspot volcanism and a hotspot swell?: Journal of Geophysical Special Paper 430, doi: 10.1130/2007.2430(05). Research, v. 100, p. 8045Ð8062. Giret, A., Weis, D., Zhou, X., Cottin, J.-Y., and Tourpin, S., 2003. Géologie des Recq, M., Gosslin, J., Charvis, P., and Operto, S., 1998, Small-scale crustal vari- îles Crozet (Geology of the Crozet Islands): Géologues, v. 137, Union ability within an intraplate structure: The Crozet Bank (southern Indian Française des Géologues, Paris, 15–23. Ocean): Geophysical Journal International, v. 134, p. 145Ð156, doi: 10.1046/ Goslin, J., and Diament, M., 1987, Mechanical and thermal isostatic response j.1365-246x.1998.00530.x. of the Del Cano Rise and Crozet Bank (southern Indian Ocean) from al- Ribe, N.M., and Christensen, U.R., 1994, Three-dimensional modeling of plume- timetry data: Earth and Planetary Science Letters, v. 84, p. 285Ð294, doi: lithosphere interaction: Journal of Geophysical Research, v. 99, p. 669Ð 10.1016/0012-821X(87)90093-8. 682, doi: 10.1029/93JB02386. Gripp, A.E., and Gordon, R.G., 2002, Young tracks of hotspots and current plate Ritsema, J., 2005. Global seismic structure maps, in Foulger, G.R., et al., eds., velocities: Geophysical Journal International, v. 150, p. 321Ð361, doi: Plates, plumes, and paradigms: Boulder, Colorado, Geological Society of 10.1046/j.1365-246X.2002.01627.x. America Special Paper 388, p. 11Ð18, doi: 10.1130/2005.2388(02). Harris, R.N., Von Herzen, R.P., McNutt, M.K., Garven, G., and Jordahl, K., Robinson, E.M., 1988, The topographic and gravitational expression of density 2000a, Submarine hydrogeology of the Hawaiian archipelagic apron 1. anomalies due to melt extraction in the uppermost oceanic mantle: Earth Heatflow patterns north of Oahu and Maro Reef: Journal of Geophysical and Planetary Science Letters, v. 90, p. 221Ð228, doi: 10.1016/0012-821X Research, v. 105, p. 21,353Ð21,369, doi: 10.1029/2000JB900165. (88)90102-1. Harris, R.N., Garven, G., Georgen, J., McNutt, M.K., Christiansen, L., and Von Sallarès, V., and Calahorrano, A., 2007 (this volume), Geophysical characteri- Herzen, R.P., 2000b, Submarine hydrogeology of the Hawaiian archipe- zation of mantle melting anomalies: A crustal view, in Foulger, G.R., and lagic apron 2. Numerical simulations of coupled heat transport and fluid Jurdy, D.M., eds., Plates, plumes, and planetary processes: Boulder, Col- flow: Journal of Geophysical Research, v. 105, p. 21,371Ð21,385, doi: orado, Geological Society of America Special Paper 430, doi: 10.1130/ 10.1029/2000JB900164. 2007.2430(25). 274 Stein and Von Herzen

Shaw, H.R., 1973, Mantle convection and volcanic periodicity in the Pacific: Islands: Journal of Volcanology and Geothermal Research, v. 94, p. 305Ð321, Evidence from Hawaii: Geological Society of America Bulletin, v. 84, doi: 10.1016/S0377-0273(99)00109-2. p. 1505Ð1526, doi: , doi: 10.1130/0016-7606(1973)84<1505:MCAVPI> Vogt, P.R., and Jung, W.-Y., 2007 (this volume), Origin of the Bermuda volca- 2.0.CO;2. noes and Bermuda Rise: History, observations, models, and puzzles, in Foul- Sleep, N.H., 1994, Lithospheric thinning by midplate mantle plumes and the ther- ger, G.R., and Jurdy, D.M., eds., Plates, plumes, and planetary processes: mal history of hot plume material ponded at sublithospheric depths: Journal Boulder, Colorado, Geological Society of America Special Paper 430, doi: of Geophysical Research, v. 99, p. 9327Ð9343, doi: 10.1029/94JB00240. 10.1130/2007.2430(27). Sleep, N.H., 2003, Mantle plumes?: Astronomy & Geophysics, v. 44, p. 1.11Ð Von Herzen, R.P., 2004. Geothermal evidence for continuing hydrothermal cir- 1.13, doi: 10.1046/j.1468-4004.2003.44111.x. culation in older ( 60 m.y.) ocean crust, in Davis, E.E., and Elderfield, E.H., Spinelli, G.A., Giambalvo, E.R., and Fisher, A.T., 2004. Sediment permeability, eds., Hydrogeology of the oceanic lithosphere: Cambridge, Cambridge distribution, and influences on fluxes in oceanic basement, in Davis, E.E., University Press, p. 414Ð447. and Elderfield, E.H., eds., Hydrogeology of the oceanic lithosphere: Cam- Von Herzen, R.P., and Uyeda, S., 1963, Heatflow through the eastern Pacific bridge, Cambridge University Press, p. 151Ð188. Ocean floor: Journal of Geophysical Research, v. 68, p. 4219Ð4250. Stein, C.A., 2003, Heatflow and flexure at subduction zones: Geophysical Re- Von Herzen, R.P., Detrick, R.S., Crough, S.T., Epp, D., and Fehn, U., 1982, search Letters, v. 30, doi: 10.1029/2003GL018478. Thermal origin of the Hawaiian swell: Heatflow evidence and thermal Stein, C.A., and Stein, S., 1992, Amodel for the global variation in oceanic depth models: Journal of Geophysical Research, v. 87, p. 6711Ð6723. and heatflow with lithospheric age: Nature, v. 359, p. 123Ð129, doi: 10.1038/ Von Herzen, R.P., Cordery, M.J., Detrick, R.S., and Fang, C., 1989, Heatflow 359123a0. and the thermal origin of hotspot swells: The Hawaiian Swell revisited: Stein, C.A., and Stein, S., 1993, Constraints on Pacific midplate swells from Journal of Geophysical Research, v. 94, p. 13,783Ð13,799. global depth-age and heatflow-age models, in Pringle, M., et al., eds., The Wadge, G., 1980, Output rate of magma from active central volcanoes: Nature, Mesozoic Pacific: Washington, D.C., American Geophysical Union Geo- v. 288, p. 253Ð255, doi: 10.1038/288253a0. physical Monograph 77, p. 53Ð76. Wang, K., and Davis, E.E., 1992, Thermal effects of marine sedimentation in Stein, C.A., and Stein, S., 1994, Constraints on hydrothermal heat flux through hydrothermally active areas: Geophysical Journal International, v. 110, the oceanic lithosphere from global heatflow: Journal of Geophysical Re- p. 70Ð78. search, v. 99, p. 3081Ð3095, doi: 10.1029/93JB02222. Wessel, P., and Smith, W.H.F., 1998, New, improved version of generic map- Stein, C.A., and Stein, S., 1997, Estimation of lateral hydrothermal flow dis- ping tools released: Eos (Transactions, American Geophysical Union), tances from spatial variations in heatflow: Geophysical Research Letters, v. 79, p. 579, doi: 10.1029/98EO00426. v. 24, p. 2323Ð2326, doi: 10.1029/97GL02319. White, R.S., McKenzie, D., and O’Nions, R.K., 1992, Oceanic crustal thickness Stein, J.S., and Fisher, A.T., 2003. Observations and models of lateral hydro- from seismic measurements and rare earth element inversions: Journal of thermal circulation on a young ridge flank: Numerical evaluation of ther- Geophysical Research, v. 97, p. 19,683Ð19,715. mal and chemical constraints: Geochemistry, Geophysics, Geosystems, Zhao, D., 2004, Global tomographic images of mantle plumes and subducting v. 4, doi: 10.1029/2002GC000415. slabs: Insight into deep Earth dynamics: Physics of the Earth and Planetary Swanson, D.A., 1972, Magma supply rate of Kilauea volcano, 1952Ð1971: Sci- Interiors, v. 146, p. 3Ð34, doi: 10.1016/j.pepi.2003.07.032. ence, v. 175, p. 169Ð170, doi: 10.1126/science.175.4018.169. ten Brink, U.S., and Brocher, T.M., 1987, Multichannel seismic evidence for a subcrustal intrusive complex under Oahu and a model for Hawaiian volcanism: Journal of Geophysical Research, v. 92, p. 13,687Ð13,707. Urgeles, R., Canals, M., Roberts, J., and SNV “Las Palmas” Shipboard Party, 1999, Fluid flow from pore pressure measurements off La Palma, Canary MANUSCRIPT ACCEPTED BY THE SOCIETY 31 JANUARY 2007