R EPORTS 23. M. H. Carr, Icaurus 56, 476 (1983). 29. E. R. D. Scott, A. N. Krot, A. Yamaguchi, Meteoritics 35. We thank K. Manser for assistance with U-Pb analy- 24. ࿜࿜࿜࿜ Water on Mars (Oxford Univ. Press, xx, 1996), 33, 709 (1998). ses and D. C. Golden for assistance with P analyses. p. 229. 30. G. Turner, S. F. Knot, R. D. Ash, J. D. Gilmour, We also thank three anonymous reviewers for com- 25. W. K. Hartmann et al., Basaltic Volcanism on the Geochim. Cosmochim. Acta 61, 3835 (1997). ments on the manuscript. This work was supported Terrestrial Planets (Pergamon, New York, 1981), pp. 31. D. D. Bogard and D. H. Garrison, Meteoritics 34, 451 by NASA RTOP numbers 344-35-05-02 and 344-31- 1049Ð1127. (1999). 00-05 and was completed while L.E.B. held a National 26. G. Neukum and D. U. Wise, Science, 194, 1381 (1976) 32. D. S. McKay et al., Science 273, 924 (1996). Research Council (Johnson Space Center) Research 27. H. Y. McSween Jr. and R. P. Harvey, Int. Geol. Rev. 40, 33. D. York, Can. J. Phys. 44, 1079 (1966). Associateship. 774 (1998). 34. M. Tatsumoto, R. J. Knight, C. J. Alle`gre, Science 180, 28. P. H. Warren, J. Geophys. Res. 103, 16759 (1998). 1279 (1973). 21 June 1999; 17 August 1999 Marineris, and Isidis impact basin are resolved The Gravity Field of Mars: as individual gravitational anomalies in the areoid. Of particular note is the apparent circu- Results from Mars Global larity of the Tharsis areoid feature, in contrast to the complex topography of the region (12). The central portion of the Tharsis areoid encom- Surveyor passes the three Tharsis Montes in a single large David E. Smith,1* William L. Sjogren,2 G. Leonard Tyler,3 anomaly and isolates Olympus Mons to the Georges Balmino,4 Frank G. Lemoine,1 Alex S. Konopliv2 northwest. The Alba Patera volcanic construct also appears as a subtle feature separated from Observations of the gravity field of Mars reveal a planet that has responded the main Tharsis areoid feature. The separation differently in its northern and southern hemispheres to major impacts and of Olympus and Alba from the main dome of volcanic processes. The rough, elevated southern hemisphere has a relatively Tharsis is similar to the topographic separation featureless gravitational signature indicating a state of near-isostatic compen- (12) and indicates that these prominent volcanic sation, whereas the smooth, low northern plains display a wider range of shields have distinctive source regions in the gravitational anomalies that indicates a thinner but stronger surface layer than mantle that may explain the topography. in the south. The northern hemisphere shows evidence for buried impact basins, The zonal variation of gravity anomalies although none large enough to explain the hemispheric elevation difference. (Fig. 3) shows differences between the southern The gravitational potential signature of Tharsis is approximately axisymmetric and northern hemispheres: the lack of gravity and contains the Tharsis Montes but not the Olympus Mons or Alba Patera anomalies below major topographic features volcanoes. The gravity signature of Valles Marineris extends into Chryse and over most of the southern hemisphere, as well provides an estimate of material removed by early fluvial activity. as substantial anomalies that lack topographic expression in the northern hemisphere. The The gravity field of Mars reflects internal and itational power (Fig. 2) is indicative of the smooth character of the gravity indicates that external processes over several billion years, ability of the planet’s lithosphere to support the topography in the southern latitudes is iso- similar to the moon’s. However, the magni- large stresses associated with surface and sub- statically compensated much as are highland tude of the variations on Mars indicates stress surface loads (11). The thick lithosphere on terrains on the moon (14). differences of about six times those in the Mars is a consequence of the more rapid loss of The range of gravity variation from south Moon. The radio science investigation on the accretional and radiogenic heat from the mar- to north (Fig. 3A) increases to as large as Mars Global Surveyor (MGS) mission (1) has tian interior as compared with Earth’s. A strik- ϳ160 mgal. This suggests a latitude-depen- developed global high-resolution gravitation- ing feature of the field is the limited range of dent variation in compensation that may be al field models (Fig. 1) for Mars from track- anomalies (Ϯϳ100 mgal) over a large fraction associated with Mars’ pole-to-pole 0.036° ing data (Table 1) (2, 3) and provides new of the planet. Anomalies with substantial am- slope in topography (12) that might be ex- insight into the manner in which Mars has plitudes are limited to the Tharsis, Isidis, and plained by a systematic variation of crustal evolved through time in response to major Elysium regions in the eastern and western thickness with latitude. The longitudinal vari- impacts, surficial geological processes, and hemispheres. A broad high of ϳ100 mgal in the ation of gravity anomalies between latitudes internal dynamics. midlatitude eastern hemisphere surrounds the 20°N to 75°N (Fig. 3B), and 20°S to 75°S Tracking of the Mariner-9 and Viking-1 and Hellas basin and appears to be associated with (Fig. 3C) also emphasizes the difference be- -2 orbiters had limited coverage and lower ac- material excavated from this structure (12). tween the two hemispheres. The uniform curacy data (2, 4–6) than MGS. The largest Mars’ gravitational potential, or areoid gravity field over most of the southern hemi- free-air gravity anomalies on Mars (Fig. 1A) (13) (Fig. 1B), displays two hemispheric- sphere in comparison with the much higher associated with Olympus Mons and the Tharsis scale quasi-circular features with a dynamic amplitude variation seen in the north is the Montes (Ascraeus, Pavonis, and Arsia volca- range of more than 2 km, an order of magni- antithesis of the topography, which is smooth noes) exceed 3000 mgal (3), more than an order tude greater than that observed for Earth. The in the north and rugged in the south. This of magnitude greater than those on Earth (9) for Tharsis Montes, Olympus Mons, Valles points to a quantitative difference in the evo- similar wavelengths (10). The substantial grav- Table 1. Summary of spacecraft tracking data. 1Laboratory for Terrestrial Physics, National Aeronau- tics and Space Administration (NASA) Goddard Space Periapsis Inclination Flight Center, Greenbelt, MD 20771, USA. 2Jet Pro- Spacecraft Tracking height (km) (degrees) pulsion Laboratory, Pasadena, CA 91109, USA. 3Cen- ter for Radio Astronomy, Stanford University, Stan- Mariner 9 1600 64 S-band Doppler ford, CA 94035Ð4055, USA. 4Group de Recherches de Geodesie Spatiales, Toulouse, France. Viking 1 300, 1500 39, 55 S-band Doppler Viking 2 300, 800, 1500 55, 75, 80 S-band Doppler *To whom correspondence should be addressed: MGS 380, 263, 170 93 X-band Doppler, range [email protected] 94 1 OCTOBER 1999 VOL 286 SCIENCE www.sciencemag.org R EPORTS lution of the northern and southern hemi- A possible explanation for the high-lati- is able to withstand buoyancy forces from the spheres. The gravity data suggest a thin, tude northern hemisphere gravity anomalies, mantle. strong lithosphere in the north and a thick, adjacent to and remote from the residual ice After Isidis, the largest impact-associated weak lithosphere in the south. Crustal mag- cap, is that they represent moderate-diameter gravity anomaly (Table 2) is Utopia (Fig. netization measurements (15) suggest that the (100 km) impact basins buried beneath the 5A), which has been interpreted as an ancient crust is older in the south than in the north, resurfaced northern hemisphere (18). The basin on the basis of surface geology (24). which implies that the south had a longer mass excesses implied by these positive Topographic data indicate that the basin is a period of time available to achieve isostasy. anomalies may represent a combination of quasi-circular depression ϳ1500 km in diam- Gravity over the north polar region (Fig. 4A) volcanic and sedimentary fill within the basin eter (25), a factor of about 2 greater than reveals several positive anomalies that have no cavity and thinning of the northern hemi- originally proposed. Utopia’s gravity anoma- obvious correlation with topography (16). A sphere crust beneath the basin (19) that have ly is diffuse, occupying an area of ϳ107 km2 combination of ice and crustal material has been not relaxed to an isostatic state (20). with no clear center. The Utopia structure is proposed (17) to account for anomalies situated Impact basins in Mars’ southern hemi- buried beneath the northern hemisphere re- in the immediate vicinity of the north polar sphere show primarily negative annular surfacing, but the size of the depression and layered terrains. In contrast, the south polar re- anomalies with a small central positive anom- gravity anomaly suggest that the original ba- gion (Fig. 4B) shows a positive anomaly of aly. Such signatures are characteristic of sev- sin could have been of a size comparable to ϳ200 mgal immediately over the pole, which eral basins on the moon (14, 21) that lack that of Hellas (Fig. 5B) (16). However, these could represent the load associated with the mare fill. The central anomaly probably rep- two massive structures appear in complete permanent ice cap. The lack of a comparable resents mantle uplift, and the negative ring contrast gravitationally. Both appear to have anomaly over the northern cap could indicate around it may represent a combination of readjusted isostatically, but Utopia was sub- that the southern cap is younger and has not yet uncompensated crustal thickening produced had sufficient time to adjust isostatically, or that during the impact (22) and lithospheric flex- the southern layered deposits contain a larger ure in response to the loading (23).