Global sea floor topography from satellite altimetry and ship depth soundings

Science

Volumn 277, Issue 5334, 1997, Pages 1956-1962

  Smith, Walter H F ^a   Sandwell, David T ^b  

^a NOAA EARTH SYSTEM RESEARCH LABORATORY   (United States)

^b UNIVERSITY OF CALIFORNIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

BATHYMETRY; BOTTOM TOPOGRAPHY; ERS-1; GEOSAT; GRAVITY; SATELLITE ALTIMETRY; SEA FLOOR TOPOGRAPHY; SEAFLOOR MAPPING;

ARTICLE; CALCULATION; MATHEMATICAL ANALYSIS; MATHEMATICAL MODEL; PACIFIC OCEAN; PREDICTION; PRIORITY JOURNAL; SEA; SEDIMENT; SPACE FLIGHT; TELECOMMUNICATION; TOPOGRAPHY;

EID: 0030842741     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.277.5334.1956     Document Type: Article

References (67)

1. R. G. Lueck and T. M. Mudge, Science 276 1831 (1997).
2. J. A. Koslow, Am. Sci. 85, 168 (1997).
3. W. H. F. Smith, J. Geophys. Res. 98, 9591 (1993).
4. W. B. F. Ryan, W. Menke, D. Caress, W. F. Haxby, RIDGE Multibeam Synthesis Project (Lamont Doherty Earth Observatory, Columbia University, Palisades, NY, 1995).
5. Medea Committee, Scientific Utility of Naval Environmental Data (Mitre, McLean, VA, 1995).
6. Canadian Hydrographic Office, General Bathymetric Chart of the Oceans (GEBCO) (Hydrographic Chart Distribution Office, Ottawa, Canada, ed. 5, 1981). The fifth edition was the last global synthesis; newer contours in regional areas are included on the GEBCO Digital Atlas CD-ROM, available from the British Oceanographic Data Centre.
7. National Geophysical Data Center, ETOPO-5 Bathymetry/Topography Data (Data Announcement 88-MGG-02, National Oceanic and Atmospheric Administration, U.S. Department of Commerce, Boulder, CO, 1988). The ETOPO-5 data combine various sources of land elevation data with a Digital Bathymetric Data Base at 5 arc min produced in the early 1980s by the U.S. Naval Oceanographic Office from hand-drawn contour charts. The U.S. Navy spent 176 person-years contouring and gridding its charts (5). Errors in this data set are discussed in (3).
8. National Geophysical Data Center, Worldwide Marine Geophysical Data TRKDAS CD-ROM version 3.1 (Data Announcement 95-MGG-02, National Oceanic and Atmospheric Administration, U.S. Department of Commerce, Boulder, CO, 1995). As the World Data Center A for Marine Geology and Geophysics, the NGDC is the world's primary archive for ship sounding data.
9. P. Wessel and A. B. Watts, J. Geophys. Res. 93, 393 (1988).
10. P. Wessel and W. H. F. Smith, Eos 72, 441 (1991).
11. D. T. Sandwell and W. H. F. Smith, J. Geophys. Res. 102, 10039 (1997).
12. In March 1995, ERS-1 completed its mapping, and in July 1995, all of the data collected by Geosat were declassified. The ERS-1 data cover latitudes between ±81.5° along orbital tracks spaced 8 km apart at their widest point (on the equator); the Geosat data cover latitudes between ±72° on tracks 5 km apart at the equator. We combined and processed these two data sets to form a global marine gravity grid between latitudes of ±72° (11).
13. SIO has a reserved archive of Indian Ocean surveys by the R/V Marion Dufresne. M. Munschy and R. Schlich granted us use of these data, which filled an important gap in coverage.
14. S. Cande has assembled a database with quality control optimized for studies of magnetic anomalies; it includes data originally archived at LDEO plus additional cruise data in the Southern Ocean that are not available elsewhere.
15. The most reliable soundings come from the LDEO-derived databases because over time many researchers have inspected these data, corrected errors, and digitized older analog soundings. One version of the LDEO data was quality-controlled for gravity studies by Wessel (9) and for bathymetry by Smith (3). A slightly different version was assembled by Cande (14). The NGDC archive (8) has the most complete geographic coverage but has been through minimal quality control; data from it were used only where the LDEO-derived data sets lacked coverage. The SIO database was searched for cruise data that were not in the other three databases (13). Data from the four sources were merged into a common format at SIO with the use of the database and quality-control methods developed at LDEO by Wessel and Smith (3, 9, 10). Where multibeam surveys exist, we used only the center beam to match the format of the more common single-beam data.
16. New surveys, particularly along the Pacific-Antarctic and Southeast Indian ridges, went to previously unexplored areas and were largely guided by the satellite gravity observations; thus, they covered some of the newly discovered high-relief features in the ocean basins.
17. Our gravity (11) and bathymetry grids have cell dimensions of 2 min in longitude and cosine (latitude) by 2 min in latitude (Mercator projection) so that cells are equidimensional but vary in size with latitude, being 3.7 km at the equator and 1.1 km at 72°. A total of 68,428,800 cells cover the area between ±72° latitude. The gravity solution combines 4.5 years of Geosat data and 2 years of ERS-1 data. The higher accuracy profiles come from averages of 16 ERS-1 repeat cycles along its 35-day repeat orbit and up to 66 Geosat repeat cycles along its 17-day repeat orbit. The higher spatial density profiles (12) come from the 1.5-year Geosat Geodetic Mission and the 1-year ERS-1 Geodetic Phases I and II.
18. K. M. Marks, Geophys. Res. Lett. 23, 2069 (1996).
19. We used the 30-arc-sec global land topography "GTOPO30" archived by the U.S. Geological Survey at http://edcwww.cr.usgs.gov/landdaac/gtopo30/ gtopo30.html. Shorelines north of 60°S were taken from P. Wessel and W. H. F. Smith [J. Geophys. Res. 101, 8741 (1996)]; these have 50-m to 5-km resolution. Shorelines south of 60°S were taken from the Scientific Committee on Antarctic Research [Antarctic Digital Database CD-ROM (British Antarctic Survey, Cambridge, UK, 1993)]; these come from charts at scales between 1:200,000 and 1:5,000,000.
20. We used the 30-arc-sec global land topography "GTOPO30" archived by the U.S. Geological Survey at http://edcwww.cr.usgs.gov/landdaac/gtopo30/ gtopo30.html. Shorelines north of 60°S were taken from P. Wessel and W. H. F. Smith [J. Geophys. Res. 101, 8741 (1996)]; these have 50-m to 5-km resolution. Shorelines south of 60°S were taken from the Scientific Committee on Antarctic Research [Antarctic Digital Database CD-ROM (British Antarctic Survey, Cambridge, UK, 1993)]; these come from charts at scales between 1:200,000 and 1:5,000,000.
21. T. H. Dixon, M. Naraghi, M. K. McNutt, S. M. Smith, J. Geophys. Res. 88, 1563 (1983).
22. N. M. Ribe, Geophys. J. R. Astron. Soc. 70, 273 (1982).
23. R. L. Parker, ibid. 31, 447 (1973).
24. W. H. F. Smith and D. T. Sandwell, J. Geophys. Res. 99, 21803 (1994).
25. G. Ramillien and A. Cazenave, J. Geodyn. 23, 129 (1997).
26. See also N. Baudry and S. Calmant, Geophys. Res. Lett. 18, 1143 (1991); W.-Y. Jung and P. R. Vogt, Tectonophysics 210, 235 (1992); L. Sichoix and A. Bonneville, Geophys. Res. Lett. 23, 2469 (1996); M. McNutt and A. Bonneville, Endeavour 20, 157 (1996).
27. See also N. Baudry and S. Calmant, Geophys. Res. Lett. 18, 1143 (1991); W.-Y. Jung and P. R. Vogt, Tectonophysics 210, 235 (1992); L. Sichoix and A. Bonneville, Geophys. Res. Lett. 23, 2469 (1996); M. McNutt and A. Bonneville, Endeavour 20, 157 (1996).
28. See also N. Baudry and S. Calmant, Geophys. Res. Lett. 18, 1143 (1991); W.-Y. Jung and P. R. Vogt, Tectonophysics 210, 235 (1992); L. Sichoix and A. Bonneville, Geophys. Res. Lett. 23, 2469 (1996); M. McNutt and A. Bonneville, Endeavour 20, 157 (1996).
29. See also N. Baudry and S. Calmant, Geophys. Res. Lett. 18, 1143 (1991); W.-Y. Jung and P. R. Vogt, Tectonophysics 210, 235 (1992); L. Sichoix and A. Bonneville, Geophys. Res. Lett. 23, 2469 (1996); M. McNutt and A. Bonneville, Endeavour 20, 157 (1996).
30. This choice is a compromise between including more gravity information in areas of widely spaced ship soundings and minimizing the errors from not knowing the spatial variations in elastic thickness. Justification of this choice and derivation of the Gaussian filter in terms of the uncertainty in elastic thickness are given in (23).
31. W. H. F. Smith and P. Wessel, Geophysics 55, 293 (1990).
32. We started with the original sounding data rather than ETOPO-5, combined gravity data from two satellites, allowed for regional variation in sediment thickness, did not make assumptions about the elastic thickness of the lithosphere, and forced the solution to agree with measured depths and coastlines.
33. W. H. F. Smith and D. T. Sandwell, Measured and Estimated Sea Floor Topography (Publication RP-1, World Data Center A for Marine Geology and Geophysics, Boulder, CO, 1997). This publication is a 34″ by 53″ printed map available from the NGDC and SIO. The digital grid (version 6.2) is available by anonymous ftp at topex.ucsd.edu; digital images can be downloaded from http://topex.ucsd.edu/marine_topo/ mar_topo.html.
34. One of us (D.T.S.) participated in this 1997 expedition, which was led by M. Maia, Chief Scientist. Depths used in this comparison are center beam values from a Simrad 12D (Kongsberg Simrad AS, Horton, Norway) multibeam system.
35. J. Mammerickx, Earth Planet. Sci. Lett. 113, 293 (1992).
36. C. Devey et al., Cruise Report SO-100: The Foundation Seamount Chain (Kiel University, Kiel, Germany, 1995).
37. The Foundation Seamounts area was poorly surveyed by ships until a 1995 expedition of the R/V Sonne (32). In 1991, Mammerickx (31), by visual inspection of gravity anomalies apparent in the widely spaced satellite altimeter profiles then available, inferred the existence of this 1600-km-long volcanic chain, which she named the Foundation Seamounts. When dense Geosat altimeter data south of 30°S were declassified in 1992, we refined the maps of this area using our estimation method (23). Our map guided the R/V Sonne expedition (32), which found 44 previously unsurveyed volcanoes with heights ranging from 1500 to 4000 m; eleven of these volcanoes come to within 500 m of the ocean surface. In 1997, the R/V Atalante surveyed another part of this area (30), generally more than 160 km away from the Sonne survey. We made an initial depth estimation using the Sonne data and the older sparse data, but not the Atalante data, in order to use the latter as an independent "blind" test of our method.
38. H. W. Menard and S. M. Smith, J. Geophys. Res. 71, 4305 (1966).
39. Menard and Smith (34) compared their solution with one obtained in 1921 and noted the same phenomenon: the newer solution found more area at shallower depths because the ability to detect seamounts had improved in the interim. Our comparison with their results was not strictly valid because they included the Arctic Ocean, whereas we worked within ±72° latitude. However, the Arctic Ocean is only 3% of the entire ocean area, and it is shallower on average; thus, adding the Arctic to our data would only increase the shallowing of our solution over theirs.
40. D. M. Müller, W. R. Roest, J.-Y. Royer, L. M. Gahagan, J. G. Sclater, J. Geophys. Res. 102, 3211 (1997).
41. D. L. Divins gave us a sediment thickness grid that he is preparing for NGDC distribution.
42. Areas where sediment thickness exceeded 2 km (continental slopes and rises) were excluded because the sediment correction is complicated in areas of thick sediment; see S. Le Douaran and B. Parsons, J. Geophys. Res. 87, 4715 (1982). The Bauer Microplate was excluded because the age model (36) may be wrong in this area. The Scotia Plate and the Marianas and Philippine back-arc basins were excluded because their depths are known to be anomalous. The following ridges, rises, banks, and plateaus were excluded because they are known to have anomalous crustal thickness: Agulhas, Azores, Chagos-Laccadive, Conrad, Del Caño, Galapagos, Hess, Iceland, Kerguelen, Line Islands, Mid-Pacific Mountains, Naturaliste, Nazca, Ninety-East, Ontong-Java, Rio Grande, Seychelles, Shatsky, Tuamotu, and Walvis.
43. M. L. Renkin and J. G. Sclater, J. Geophys. Res. 93, 2919 (1988).
44. B. Parsons, ibid. 87, 289 (1982).
45. The age grid (36) was interpolated by assuming a constant spreading rate between identifiable magnetic anomalies, hence the steplike pattern.
46. J. D. Hays and W. C. Pitman III, Nature 246, 18 (1973); W. C. Pitman III, Geol. Soc. Am. Bull. 89, 1389 (1978).
47. J. D. Hays and W. C. Pitman III, Nature 246, 18 (1973); W. C. Pitman III, Geol. Soc. Am. Bull. 89, 1389 (1978).
48. W. H. Berger and E. L. Winterer, Int. Assoc. Sedimentol. Spec. Publ. 1, 11 (1974).
49. 1/4 measures the dispersion of the distribution. The mode is that value of x that occurs most frequently; here, it is the center of the 50-m interval of depth having the most area.
50. B. Parsons and J. Sclater, J. Geophys. Res. 82, 803 (1977).
51. D. L. Turcotte and E. R. Oxburgh, J. Fluid Mech. 28, 29 (1967); E. E. Davis and C. R. B. Lister, Earth Planet. Sci. Lett. 21, 405 (1974).
52. D. L. Turcotte and E. R. Oxburgh, J. Fluid Mech. 28, 29 (1967); E. E. Davis and C. R. B. Lister, Earth Planet. Sci. Lett. 21, 405 (1974).
53. C. A. Stein and S. Stein, Nature 359, 123 (1992).
54. The hypsometric curve predicted by a plate model was found by treating the distribution of area with age as ΔA/ΔAt, a discrete approximation of the rate of change of cumulative area with age, treating the hypsometric curve similarly as ΔA/Δz, and relating the two through dz/dt, the derivative of the subsidence predicted by the plate model. This method was given in (40), but that paper assumed a linear function for ΔA/ΔAt, whereas we used the actual data (gray area of Fig. 5).
55. J. C. Marty and A. Cazenave, Earth Planet. Sci. Lett. 94, 301 (1989).
56. S. Nagihara, C. R. B. Lister, J. G. Sclater, ibid. 139, 91 (1996).
57. H. W. Menard and M. K. McNutt, J. Geophys. Res. 87, 8570 (1982); M. K. McNutt, ibid. 89, 11180 (1984); W. H. F. Smith, H. Staudigel, A. B. Watts, M. S. Pringle, ibid. 94, 10501 (1989).
58. H. W. Menard and M. K. McNutt, J. Geophys. Res. 87, 8570 (1982); M. K. McNutt, ibid. 89, 11180 (1984); W. H. F. Smith, H. Staudigel, A. B. Watts, M. S. Pringle, ibid. 94, 10501 (1989).
59. H. W. Menard and M. K. McNutt, J. Geophys. Res. 87, 8570 (1982); M. K. McNutt, ibid. 89, 11180 (1984); W. H. F. Smith, H. Staudigel, A. B. Watts, M. S. Pringle, ibid. 94, 10501 (1989).
60. S. T. Crough, Geophys. J. R. Astron. Soc. 55, 451 (1978).
61. R. L. Heestand and S. T. Crough, J. Geophys. Res. 86, 6107 (1981).
62. S. T. Crough, Tectonophysics 61, 321 (1979).
63. S. Calmant and A. Cazenave, Nature 328, 236 (1987); M. K. McNutt and K. Fischer, Geophys. Monogr. Am. Geophys. Union 43, 25 (1987).
64. S. Calmant and A. Cazenave, Nature 328, 236 (1987); M. K. McNutt and K. Fischer, Geophys. Monogr. Am. Geophys. Union 43, 25 (1987).
65. R. L. Larson, Sci. Am. 272, 82 (February 1995).
66. If the thermal diffusivity and thermal expansivity are constant, then depth is proportional to the integral of temperature over the lithosphere (45, 46). Any reheating mechanism that resets the effective thermal age, as sensed by depth, will do; several of these mechanisms are discussed in (50).
67. This work was supported primarily by NSF Ocean Sciences grant OCE95-21518, with some additional support from NASA Solid Earth and Natural Hazards grant NAGW3035. We thank the many scientists who provided us with soundings from their recent cruises.