Roadmap | The review process | Home
   Heat flow
Heat Flow on Hotspot Swells

Robert N. Harris

College of Oceanic and Atmospheric Sciences, Oregon State University

Two assumptions that commonly guide the interpretation of heat flow data collected on hotspot swells are the choice of a reference model and that values of surface heat flow represent lithospheric heat flow.   In an earlier webpage on this site, Stein & Stein (2005) discuss assumptions related to the choice of a reference model.  Because the expected thermal anomalies are small, anomalous heat flow depends strongly on the reference model.  Stein & Stein (1993) reviewed average heat flow values over hotspot swells relative to their model (GDH1 for global depth and heat flow as a function of crustal age; Stein & Stein, 1992) and concluded that while hotspot swells are anomalously shallow, heat flow is not anomalously high.  Based on the presence of a bathymetric anomaly and lack of a heat flow anomaly, they argued that hot spot swells are dynamically, not thermally, supported.

The second assumption is that heat flow values represent conductive heat transfer through the lithosphere so that these values can be used to infer the thermal state at the base of the lithosphere.  In fact until very recently heat flow surveys over hotspot swells were designed assuming conductive heat transport.  As a consequence heat flow was thought to vary at the long wavelengths associated with a sublithospheric thermal plume.  With this supposition in mind, and the tradeoff between covering large areas with limited ship time, early surveys were designed to consist of about seven or so ~10 x 10 km areas, each with multiple (10-20) heat flow determinations.  Large distances (of the order of 100 km) generally separate these areas and heat flow determinations were averaged together to cancel variability. 

Averages of heat-flow determinations from these widely spaced areas collected for the purpose of measuring the thermal flux from hotspots have been published at four oceanic swells: Hawaii (Von Herzen et al., 1982, 1989), Bermuda (Detrick et al., 1986), Crozet (Courtney & Recq, 1986), and Cape Verde (Courtney & White, 1986).  Harris & McNutt (2007) analyze inter-site variability of individual heat flow determinations.  They show that inter-site variability is generally larger than instrumental noise as indicated by repeat measurements at the same site, and that as the volcanic edifice is approached the scatter of values at each site increases.  Harris & McNutt (2007) argue that this pattern is inconsistent with conductive heat transfer and that the most likely source of perturbation is advective fluid flow.

Discriminating between environments where heat is transferred conductively or advectively requires closely spaced heat flow determinations (1-2 km) collocated with seismic reflection profiles.  Only Hawaii (Harris et al., 2000a) and Reunion (Bonneville et al., 1997) have surveys meeting these requirements.  The Hawaiian survey consists of two profiles, one north of Oahu and one north of Maro Reef (Figure 1).  The Reunion survey also consists of two profiles, both north of Mauritius (Figure 2).  These high-resolution surveys provide evidence for shallow fluid circulation and help to identify its character at distance scales that are badly aliased in the earlier coarse surveys and smoothed out by averaging the results from multiple penetrations at the local sites.  Both surveys cross archipelagic aprons characterized by a thick wedge of volcaniclastic sediments deposited during mass wasting events and which overlie pre-existing sediments and oceanic basement.  Root mean square variations along the Oahu and Maro Reef profiles are 15 and 5 mW m-2, respectively, and along both Reunion heat flow profiles are about 13 mW m-2 (Figures 1 and 2).  These heat flow profiles reveal greater scatter than could plausibly be attributed to lateral or temporal variations in mantle heat flux.  The spectra at both hotspots (Figure 3) are quite similar with most of the signal in the heat flow occurring at wavelengths less than a few hundred kilometers. In each case there is a spectral peak at a wavelength of the order of 10 km suggesting that the source of perturbation is relatively shallow.  Data collected using the coarse survey strategy, at intervals of 100 km or more, have aliased this portion of the heat flow signal.

Figure 1. Thermal data from Hawaii. a) Location of heat flow determinations. White symbols show location of data from South Arch (circles) and Maro Reef (triangles) (Von Herzen et al., 1982, 1989) and thick black lines show data from Harris et al. (2000a). The white triangle colocated with the Maro Reef profile is site 4 of Von Herzen et al., (1989). The extent of the Hawaiian swell is approximated with the 5 km bathymetric contour. Heat flow profile collocated with seismic reflection line for offshore b) Oahu and c) Maro Reef. Heat flow determinations uncorrected (open circles) and corrected for the effects of sedimentation (solid circles). Error bar represents standard deviation. Line drawing of depth-converted seismic reflection profile are shown in the lower portion of each panel (Harris et al., 2000a). Vertical dashed lines show approximate divisions of Von Herzen (2004).

Likely candidates for processes generating the observed variability in the heat flow profiles are reviewed in Harris et al. (2000a, 2000b).  Variability in the Hawaii profiles is greatest in the deepest part of the flexural moat toward the volcanic edifice, and in the Reunion profiles is greatest near bare rock environments (Figures 1 and 2).  Both of these observations seem more consistent with buoyancy driven fluid flow than recent volcanic intrusions.  Harris et al. (2000b) and Harris and McNutt (2007) argue that lateral fluid flow may mask anomalous heat associated with the presence of a thermal plume.

Because the scatter in heat flow determinations is approximately the same magnitude as predicted variations in basal heat flow (e.g. Von Herzen et al., 1989), our ability to resolve anomalous basal heat flux will depend on a better understanding of environmental conditions and non-conductive processes. This understanding will likely require more high-resolution heat flow surveys collocated with seismic reflection lines, observations of fluid flow, and more sophisticated modeling.  Until then, using heat flow data to distinguish between thermal and non-thermal origins of midplate swells may be premature.


Figure 2. Thermal data from Reunion (Bonneville et al., 1997). a) Location of heat flow determinations at Reunion and heat flow determinations along b) profile 1 and c) profile 2. Heat flow determinations uncorrected (open circles) and corrected for the effects of sedimentation (solid circles). Error bar represents standard deviation. Seismic reflection data is also shown.

Figure 3. Spectral analysis of closely spaced heat flow profiles on hotspot swells. Vertical scale shows magnitude of harmonic coefficient and uncertainty. Spectral smoothing was accomplished through windowing four subsections of the original series. Note spectral peaks of the order of 10 km. Along the Oahu profile there is a strong peak at 7 km, along the Maro Reef profile there is a subtle peak at approximately 10 km, and along Reunion profile 1 there is a peak at approximately 50 km.



last updated 4th April, 2007