EarthRef.org Reference Database (ERR) -- McNutt & Bonneville 2000

Young hotspot volcanoes within plate interiors are frequently surrounded by smooth, broad regions of shallow seafloor termed midplate swells. These swells are typically hundreds of kilometers wide and can be more than a kilometer in elevation. The most frequently invoked explanation for these swells is that they represent the thermal and dynamic surface uplift from rising mantle plumes. Here we argue that buoyancy of a volcanic unit underplating the Marquesan volcanoes just below the Moho, as imaged by a seismic refraction experiment, is capable of producing much, if not all, of what has been previously interpreted as a thermal swell in both the bathymetry and the geoid. The shallow compensation depth previously calculated for the swell based on geoid observations is thus expected given that the compensation resides at 20-km depth. The volcanic unit underplating the islands would also be expected to have a thermal effect, but its contribution to swell uplift is never more than 10% of the chemical contribution. The predicted increase in heat flow, on the other hand, is very large, but the thermal transient rapidly decays in the first 2 million years after emplacement. If similar large volumes of underplated material underlie other hotspot volcanoes, problems in explaining the lack of heat flow anomalies at later times and negligible rates of swell subsidence can be alleviated. Although volcanic underplating is unlikely to explain the entire swell signature for particularly large and wide swells, the results from the Marquesas suggest that at least some portion of hotspot swells may not be thermal in origin. After removing the geoid signature from the Marquesas volcanoes and swell, what remains in the geoid are stripes of amplitude 1 m and wavelength 650 km trending normal to the fracture zone. The anomalies are presumably relict from when this crust was created, but their amplitude is too large to be caused by variations in isostatically compensated crustal thickness.