JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 91, NO. B8, PAGES 8325-8340, JULY 10, 1986 Seasat-DerivedGravity Over the Musicians Seamounts ADAM P. FREEDMAN AND BARRY PARSONS Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technolo•Ty, Cambrid•Te The two-dimensional gravity field over the Musicians seamount province in the Pacific Ocean has been derived from Seasataltimetry. Geoid maps were producedby fitting a minimum curvature surface to the sea surface height measurements. As a check on the quality of this interpolation method, we also gridded the data using weighted grid point averages. Fourier transformsof both the geoid and geoid gradient were used to determine the gravity field. We have compared gravity maps produced these different ways in order to identify the problemsinvolved in pushing Seasatdata to the limits of its spatial resolution and accuracy.Our minimum-curvature interpolation schemewas determined to be the more accurateand cost effective mapping method, while gravity obtained by transforming the geoid produced more reliable gravity maps. The bathymetry of this area was usedto predict the gravity field by filtering the bathymetry under the assumptionthat it is regionally compensated by a thin elastic plate. Gravity fieldspredictedfor a variety of effective elasticthicknesses were compared to the Seasat-derived gravity, particularly in areaswith good track coverage. The derived gravity tends to favor a thin plate with an effective elasticthickness of •5 km, though the east-west ridges in the south display a smaller signal more consistent with Airy compensation. This variation may be indicative of early fracturing of the lithospherein the south, or it may be a manifestation of the age difference and early thermal structure acrossthe Murray fracture zone, which separates the seamount province into northern and southern sections. Neighboringseamounts with different flexural signatures, particularly in the south,may indicate that volcanism occurred in the same locationover an extended period of time. INTRODUCTION Many studies have shown that gravity and bathymetry in the oceans are highly correlated for short- and medium- wavelength features [e.g., McNutt, 1979; Watts and Daly, 1981]. The details of this correlation can yield information about how topographic features are compensated and hence about the thermal structureand tectonichistory of the region being studied. The gravity-bathymetry relationship can be characterized through the use of a response function, or ad- mittance. The admittance Z is a function of wave number k, and is definedby z(t0 = (1) where k---2rr/,•, ,• is the wavelength of the bathymetric and gravimetric features,and G(k) and B(k) are the Fourier trans- forms of the gravity and bathymetry, respectively. In many studies, G and B determined from observations are used to compute Z, which is then comparedto theoretical admittance models [Lewis and Dotman, 1970' McKenzie and Bowin, 1976' Banks et al., 1977' Watts, 1978' McNutt, 1979; Ribe and Watts, 1982; Ribe, 1982]. Alternatively, as in this study, a theoretical admittance is used as a filter in conjunction with a known value of/• or C,to predict the other quantity and to compare it to observations [Watts, 1978, 1979; McNutt, 1979; Watts et al., 1980' Dixon et al., 1983' Watts and Ribe, 1984]. The radar altimeter on board the Seasat satellite obtained uniform coverage, high-resolution sea surface height infor- mation over most of the oceansurface during a 70-day period from July through October 1978 [Lame and Born, 1982]. The shape of the sea surface closely approximates the marine geoid. Other factors that affect sea surface height, such as tides, currents, and atmospheric and meteorologicaleffects, can be removedfrom the data or have small magnitudes (less Copyright 1986by the American Geophysical Union. Paper number 5B5854. 0148-0227/86/005 B- 5854 $05.00 than 50 cm) compared to many features of interest in the geoid (with magnitudes > 1 m) [Wunsch and Gaposchkin, 1980]. Global geoid maps are easily made using Seasatdata [e.g., Parke and Dixon, 1982]. The combinationof thesegeoid maps and comprehensive bathymetricmaps,such as the digi- tal global bathymetric data set SYNBAPS [SyntheticBathy- metricProfiling System, 1979], facilitates admittancestudies of features whose gravity could neverbefore be studied in detail. Most previous admittance studies have concentrated on linear featuressuchas mid-oceanridges,aseismic ridges,frac- ture zones, and hot-spot-related island chains, where one- dimensional modeling is adequate. In the case of individual seamounts or seamountprovinces with randomly located sea- mounts, however, two-dimensionalanalysisis required [Ribe, 1982; Watts and Ribe, 1984]. Only a few truly two- dimensional comparisons of the short- and intermediate- wavelength(4 < 500 km) marine gravity or geoid field with the bathymetry have been made [Watts et al., 1975; McNutt, 1979; Cazenaveet al., 1980; Watts and Ribe, 1984]. These previous studies have focused on the methods of compensation of bathymetricfeatures. The oceanic lithosphere tendsto act as an elastic plate overlyinga fluid asthenosphere. Crustal loads formed on old, thick lithosphere are supported by elastic stresses producedby the bending of the plate. The plate distributes the compensating masses over a large area. Directly over the load, therefore, the gravity anomaly is large. For crustal loads formed on young seafloor, the effective elas- tic plate thickness is smalland the lithosphere is weak. Elastic stresses play a smallpart in the supportof the load. Supportis instead providedby buoyancy forces due to the compensating masses located beneath the load. Hence directly over the load, the gravity anomalyis small.The actual behaviorof the plate depends on the horizontal dimensions of the crustal load as well as the plate thickness, so the ratio of gravity to bathy- metry as a function of wavelength, i.e., the admittance,can indicate the effective elastic thickness of the lithosphere. By relating the elasticthickness to the age of the lithosphere at the time of loading, Watts et al. [1980] and others have shown how the elastic thickness is a function of time and increases with age. 8325