skip navigation link


David T. Sandwell - Scripps Institution of Oceanography
Walter H. F. Smith - Geosciences Laboratory, NOAA -

view large global gravity gif image.

The surface of the ocean bulges outward and inward mimicking the topography of the ocean floor. The bumps, too small to be seen, can be measured by a radar altimeter aboard a satellite. Over the past year, data collected by the European Space Agency ERS-1 altimeter along with recently declassified data from the US Navy Geosat altimeter have provided detailed measurements of sea surface height over the oceans. These data provide the first view of the ocean floor structures in many remote areas of the Earth. For scientific applications, the Geosat and ERS-1 altimeter data are comparable in value to the radar altimeter data recently collected by the Magellan spacecraft during its systematic mapping of Venus.

Read More About:

  1. Introduction
  2. The Method of Satellite Altimetry
  3. Explanation of the Gravity Anomaly
  4. Applications of Satellite Altimetry
  5. More Information About this Article


The geologic and topographic structures of the ocean floor primarily reflect plate tectonic activity that has occurred over the past 150 million years of the 4.5 billion year age of the Earth. Seafloor geology is far simpler than the geology of the continents because erosion rates are lower and also because the continents have suffered multiple collisions associated the opening and closing of ocean basins (Wilson Cycle). Despite their youth and geologic simplicity, most of this deep seafloor has remained poorly understood because it is masked by 3-5 km of seawater. For example, the Pacific-Antarctic rise, which has an area about equal to South America, is a broad rise of the ocean floor caused by sea floor spreading between two major tectonic plates (see Poster southeast of New Zealand). To the west of the ridge lies the Louisville seamount chain which is a chain of large undersea volcanoes having a length equal to the distance between New York and Los Angeles. These features are unfamiliar because they were discovered less than 20 years ago. The Louisville seamount chain was first detected in 1972 using depth soundings collected along random ship crossings of the South Pacific. Six years later the full extent of this chain was revealed by a radar altimeter aboard the Seasat (NASA) spacecraft. Recently, high density data collected by the Geosat (US Navy) and ERS-1 (European Space Agency) spacecraft data show the Pacific-Antarctic Rise and the Louisville Ridge in unprecedented detail. In an age when we are mapping the surfaces of Venus and Mars, it is difficult to believe that so little is known about our own planet.

The reason that the ocean floor, especially the southern hemisphere oceans, is so poorly charted is that electromagnetic waves cannot penetrate the deep ocean (3-5 km = 2-3 mi). Instead, depths are commonly measured by timing the two-way travel time of an acoustic pulse. However because research vessels travel quite slowly (6m/s = 12 knots) it would take approximately 125 years to chart the ocean basins using the latest swath-mapping tools. To date, only a small fraction of the sea floor has been charted by ships.

Fortunately, such a major mapping program is largely unnecessary because the ocean surface has broad bumps and dips which mimic the topography of the ocean floor. These bumps and dips can be mapped using a very accurate radar altimeter mounted on a satellite. In this brief report we attempt to answer some basic questions related to satellite measurements of the ocean basins. What causes the surface of the ocean to bulge outward and inward mimicking the topography of the ocean floor? How big are these bumps? How can they be measured in the presence of waves and tides? What are some of the non-military applications of these data? What has been discovered from the new Geosat and ERS-1 data?

Satellite Altimetry

According to the laws of physics, the surface of the ocean is an "equipotential surface" of the earth's gravity field. (Lets ignore waves, winds, tides and currents for the moment.) Basically this means that if one could place balls everywhere on the surface of the ocean, none of the balls would roll down hill because they are all on the same "level". To a first approximation, this equipotential surface of the earth is a sphere. However because the earth is rotating, the equipotential ocean surface is more nearly matched by an ellipsoid of revolution where the polar diameter is 43 km less than the equatorial diameter. While this ellipsoidal shape fits the earth remarkably well, the actual ocean surface deviates by up to 100 meters from this ideal ellipsoid. These bumps and dips in the ocean surface are caused by minute variations in the earth's gravitational field. For example the extra gravitational attraction due to a massive mountain on the ocean floor attracts water toward it causing a local bump in the ocean surface; a typical undersea volcano is 2000 m tall and has a radius of about 20 km. This bump cannot be seen with the naked eye because the slope of the ocean surface is very low.

These tiny bumps and dips in the geoid height can be measured using a very accurate radar mounted on a satellite (Figure). For example, the Geosat satellite was launched by the US Navy in 1985 to map the geoid height at a horizontal resolution of 10-15 km (6 - 10 mi) and a vertical resolution of 0.03 m (1 in). Geosat was placed in a nearly polar orbit to obtain high latitude coverage (+- 72 deg latitude). The Geosat altimeter orbits the earth 14.3 times per day resulting in an ocean track speed of about 7 km per second (4 mi/sec). The earth rotates beneath the fixed plane of the satellite orbit, so over a period of 1.5 years, the satellite maps the topography of the surface of the earth with an ground track spacing of about 6 km (4 mi).

Two very precise distance measurements must be made in order to establish the topography of the ocean surface to an accuracy of 0.03 m (1 in) (Figure). First, the height of the satellite above the ellipsoid h* is measured by tracking the satellite from a globally-distributed network of lasers and/or doppler stations. The trajectory and height of the satellite are further refined by using orbit dynamic calculations. Second, the height of the satellite above the closest ocean surface h is measured with a microwave radar operating in a pulse-limited mode on a carrier frequency of 13 GHz. (The ocean surface is a good reflector at this frequency.) The radar illuminates a rather large spot on the ocean surface about 45 km (28 mi) in diameter. A smaller effective footprint (1-5 km in diameter = 0.6 - 3 mi)) is achieved by forming a sharp radar pulse and accurately recording its 2-way travel time. The footprint of the pulse must be large enough to average out the local irregularities in the surface due to ocean waves. The spherical wavefront of the pulse ensures that the altitude is measured to the closest ocean surface. A high repetition rate (1000 pulses per second) is used to improve the signal to noise ratio, especially when the ocean surface is rough. Corrections to the travel time of the pulse are made for ionospheric and atmospheric delays and known tidal corrections are applied as well. The difference between the height above the ellipsoid and the altitude above the ocean surface is approximately equal to the geoid height N = h* - h.

view large satellite altimeter gif image.

Gravity Anomaly

As the spacecraft orbits the earth it collects a continuous profile of geoid height across an ocean basin. Profiles from many satellites, collected over many years, are combined to make high resolution images. The Poster shows gravity anomaly derived from geoid height measurements from 4.5 years of Geosat measurements and 2 years of ERS-1 measurements. We have developed a new method to convert these raw geoid height measurements, which have a variety of accuracies, track spacings and data densities, into images (or grids) of gravity anomaly. This conversion is done to enhance the small-scale features of the seafloor. Moreover, after the conversion, the satellite-derived gravity measurements can be compared and combined with gravity anomaly measurements made by ships. The algorithms of the conversion are based on laws of physics, geometry and statistics. Since the data sets are large, diverse, and contaminated with errors, many sophisticated computer operations are required. The ultimate test of the accuracy of our methods is through comparisons with shipboard gravity measurements. Our latest grids show agreement with ship data at a level of 5 milligal (mgal). One mgal is about one millionth the normal pull of gravity (9.8 m/s2). Typical variations in the pull of gravity are 20 milligal although over the deep ocean trenches they exceed 300 mgal.

view large gif image of trackline locations.

Applications of Satellite Altimetry


The Geosat data were collected by the US Navy to fulfill their navigational and mapping requirements. Consider measuring accelerations in a moving submarine or aircraft in order to determine your position as a function of time. (Of course your starting position and velocity must also be known.) If the windows of your vehicle are closed, a true acceleration cannot be distinguished from a variation in the pull of gravity. Thus the gravity data are needed for correction of inertial navigation/guidance systems. The military applications are obvious and provided the rationale for the 80 million dollar cost of the Geosat mission as well as the classification of these data, especially during the cold war when nuclear submarines were more active than they are today. On the commercial side, Honeywell Inc. is using these data to update their inertial navigation systems in commercial aircraft. In particular, when this correction is not applied, they have found large navigational errors along Pacific Ocean flight paths which follow the major ocean trenches.

Prediction of Seafloor Depth

We are using these dense satellite altimeter measurements in combination with sparse measurements of seafloor depth to construct a uniform resolution map of the seafloor topography. These maps do not have sufficient accuracy and resolution to be used to assess navigational hazards but they are useful for such diverse applications as locating the obstructions/constrictions to the major ocean currents and locating shallow seamounts where fish and lobster are abundant.

view concept gif image.

view large gif image of predicted topography.

On a broad scale the topography of the ocean floor reflects the cooling and subsidence of the plates as they move away from the spreading center. While this process is fairly well understood, there are interruptions in this normal subsidence caused by mantle plumes and other types of solid-state convection in the mantle of the Earth that are current topics of research. As the seafloor ages it also becomes covered by a slow rain of sediments. The analysis of the gravity data along with measured can be used to map the thickness of the sedimentary layers.

Related Site - Predicted and Measured Seafloor Topography

Planning Shipboard Surveys

The satellite-derived gravity grids reveal all of the major structures of the ocean floor having widths greater than 10-15 km (6-9 mi). This resolution matches the total swath width of the much higher multibeam mapping system on a ship (100 m resolution) so the gravity maps are the perfect reconnaissance tool for planning the more detailed shipboard surveys. Scientists aboard research vessels use the gravity grids along with other measurements to optimize their survey strategy; in many cases this is done in real time. The cost to operate a research vessel is typically $20,000 per day so these gravity data have become an essential item.

Plate Tectonics

These satellite altimeter data provide an important and definitive confirmation of the theory of plate tectonics. Indeed, almost everything apparent in the marine gravity field was created by the formation and motion of the plates. The Indian Ocean Triple junction (27 deg S latitude, 70 deg E longitude) is a textbook example of seafloor spreading. Spreading ridges are characterized by an orthogonal pattern of ridges and transform faults. The scar produced in the active transform valley is carried by seafloor spreading out onto older seafloor leaving evidence of the past plate motions. At this Indian Ocean site, three spreading ridges intersect forming a triple junction as described by plate tectonic theory. The theory predicts that the ridges would intersect at 120° angles if the three ridges were spreading at exactly the same rate. In this case, one can measure the intersection angles and infer the relative spreading rates of each ridge.

view large gif image of a triple junction.

Plates are created at spreading ridges and destroyed (subducted) at the deep ocean trenches. All of the major ocean trenches are evident in the gravity map as linear troughs. The deep ocean basins away from the trenches are characterized by fracture zone gravity signatures inherited at the spreading ridge axis. This pattern is sometimes overprinted by linear volcanic chains which are believed to be formed as the plate moves over a stationary mantle plume. The hot plume head melts the mantle rocks which erupt on the surface as a hot spot. Because all of these major features are evident in the gravity maps, the geologic history of the ocean basins can now be established in great detail.

view large gif image of western Pacific Ocean.

Undersea Volcanoes

The global gravity grids reveal all volcanoes on the seafloor greater than about 1000 m tall. Approximately 1/2 of these volcanoes were not charted previously. One of the more important aspects of these new data will be to locate all of these volcanoes and identify spatial patterns that may help determine how they formed. Many volcanoes appear in chains, perhaps associated with mantle plumes, there are many more that do not fit this simple model. Moreover, numerous undersea volcanoes are long linear ridges with aspect ratios of 20 or more. These features suggest that the plates are not exactly ridged as predicted by the simple plate tectonic theory. Using these data we are exploring the internal deformations of the plates, especially outboard of trenches where the forces generated by the slab-pull force of the subducted plates is greatest.

view large gif image of undersea volcanoes.

Petroleum Exploration

All of the major petroleum exploration companies use satellite altimeter gravity data from Geosat and ERS-1 to locate offshore sedimentary basins in remote areas. This information is combined with other reconnaissance survey information to determine where to collect or purchase multi-channel seismic survey data. Currently, the regions of most intense exploration interest are the continental shelves of Australia and the former Soviet Union; recently companies have expressed interest in the Caspian Sea. Developments in offshore drilling technology now make it economical to recover oil from continental slope areas in water once thought prohibitively deep.

While we are not directly involved in this activity, we fill data requests from many companies including UNOCAL. Dr. Mark Odegard of UNOCAL Inc. says "We routinely use satellite gravity data in any exploration effort in the oceans outside of the Gulf of Mexico. We consider the current data quality to be better than the standard regional type of survey that was run in the previous twenty five years or so. . . . If the conventional gravity data were collected over the continental shelves of the World, where we have interest in exploration, the cost by my estimate, would be in the range of $200-$400 million dollars, maybe more. This obviously will not be done, but we are beginning to collect high-resolution (shipboard) data over selected targets outside the US The other companies that use the satellite data, that I know of, are: Exxon, Mobil, and Texaco. UNOCAL has been a recognized leader in potential fields in the oil industry, so we have probably been quicker to utilize the data than many companies. I would suspect BP, Total, and AGIP are foreign companies that probably utilize the data."

view large image of Gulf.

Lithospheric Structure

There are numerous other scientific applications that cannot be described in a short report. One of the traditional uses of marine gravity measurements is to estimate the thickness of the elastic portion of the tectonic plates. When a volcano forms on the ocean floor it provides a large downward load on the plate causing it to deform. This deformation is appears in the gravity field as a donut-shaped gravity low surrounding the gravity high associated with the volcano itself. By measuring the amplitude and width of the gravity low and relating this to the size of the volcano as measured my a ship with an echo sounder, one can establish the thickness and strength of the elastic plate. The new satellite-derived gravity data enable researchers to perform this type of analysis everywhere in the oceans. Thus scientists can now probe the outermost part of the earth using these and other methods.

For more information see:

  • McGraw-Hill Yearbook of Science and Technology, p. 178-180, 1995.
  • Sandwell, D. T., Geophysical Applications of Satellite Altimetry, Reviews of Geophysics Supplement, p. 132-137, 1990.