John A. Goff University of Texas Institute for Geophysics • Austin, Texas USA Walter H.F. Smith and Karen M. Marks NOAA Laboratory for Satellite Altimetry • Silver Spring, Maryland USA Oceanography • Vol. 17 • No. 1/2004 24 Special Issue—Bathymetry from Space The Contributions of Abyssal Hill Morphology and Noise to Altimetric Gravity Fabric (e.g., Smith, 1998; Goff and Smith, 2003) and will be shown here, geological information is contained in the small-scale fabric of altimetric gravity data. This fabric, which has the appearance of the skin of an orange peel, or the surface of a football, exists everywhere in the altimetry data, filling the otherwise uninteresting reaches between larger-scaled, recognizable features such as mid-ocean ridges, fracture zones and seamounts. Although noise is an important factor, small-scale gravity fabric, and rms roughness in partic- ular, varies regionally in ways that are suggestive of a tectonic origin; that is, related to mid-ocean ridge spreading rates or ridge-axis morphology at the time of seafloor creation. Such variations must, in some way, be related to the small-scale (i.e., < ~20 km full-wave- length) morphology of the seafloor. In this study we examine the gravity roughness characteristics in two regions, the South Atlantic Ocean (Figure 1) and southeast Pacific Ocean (Figure 2), where changes in spreading rate over time have been observed in the magnetic anomaly record. We employ a statistical characterization method developed by Goff and Smith (2003) that provides estimates of root-means-square (rms) roughness, characteristic length and width, and strike azimuth to identify fabric. Our samples avoid major physiographic features, such as fracture zones, so that our expectation is that the measured gravity fabric is related to the primary fabric of the seafloor: abyssal hills, which constitute the most widespread geomor- phology on Earth. Abyssal hills are related to the nature of the mid-ocean ridges at which they form (Goff, 1991; Goff, 1992; Goff et al., 1995; Goff et al., 1997). The seafloor roughness represented by abyssal hills is also of interest to the physical oceanographic community because of the critical role it plays in ocean circulation and mixing (e.g., Mauritzen et al., 2002). Through synthetic experiments we demonstrate theoretically the relationship between gravity fabric and abyssal hills at a variety of sizes and shapes. We explore as well the effect of noise on our ability Introduction Earth’s deep seafloor is, for the most part, a vast, unexplored terrain. Only a miniscule fraction has ever been observed directly, using deep submersibles and remotely operated vehicles. Most of our understanding of seafloor physiography is instead derived from remotely sensed data, such as sonar or satellite altime- try. Sonar data must be collected by ships or underwa- ter vehicles, which is an expensive and time-consum- ing process. Technological advances over the past two decades have enabled sonar devices to collect detailed “swath” coverage (Figures 1b and 2b), but only a few percent of the seafloor has been mapped in this fash- ion, typically in areas of prominent seafloor structures such as mid ocean ridges, fracture zones and trenches. More generally, sonar coverage of the ocean floor is limited to profile coverage along disparate ship tracks, well concentrated in shipping lanes and very sparse elsewhere. Over the past decade or so, satellite altimetry data have been employed to fill in the gaps left by incom- plete sonar coverage. The marine geoid and the gravi- ty field derived from it are dominated by the seafloor topography signal at scales less than a few hundreds of kilometers (Smith and Sandwell, 1997; Smith, 1998). Major seafloor features are easily detected in the alti- metric gravity data (Figures 1a and 2a), and their bathymetry can be predicted based on geophysical for- mulations constrained by existing bathymetric cover- age. However, this predictive ability is severely limited at smaller scales. The altimetric gravity field is, in essence, a filtered and somewhat noisy expression of bathymetry (Smith, 1998). Depending on water depth, that filter is ~18-20 km wide (full-wavelength); we can- not expect a one-to-one correspondence of gravity and seafloor features below this scale. Furthermore, noise would be translated in this predictive scheme into large bathymetric artifacts. Hence, little effort has gone into trying to interpret small-scale features in the grav- ity field. However, as has been demonstrated previously This article has been published in Oceanography, Volume 17, Number 1, a quarterly journal of The Oceanography Society. Copyright 2003 by The Oceanography Society. All rights reserved.Reproduction of any portion of this article by photocopy machine, reposting, or other means without prior authorization of The Oceanography Society is strictly prohibited. Send all correspondence to: info@tos.org, or 5912 LeMay Road, Rockville, MD 20851-2326, USA.