Reports/ / 30 May 2013 / Page 1/ 10.1126/science.1235768High-resolution gra...
/ / 30 May 2013 / Page 2/ 10.1126/science.1235768proximately 15 km thick at...
/ / 30 May 2013 / Page 3/ 10.1126/science.12357685. G. A. Neumann, M. T. Zu...
/ / 30 May 2013 / Page 4/ 10.1126/science.123576828 January 2013; accepted ...
/ / 30 May 2013 / Page 5/ 10.1126/science.1235768Fig. 1. Free-air gravity a...
/ / 30 May 2013 / Page 6/ 10.1126/science.1235768Fig. 2. Vertical cross sec...
/ / 30 May 2013 / Page 7/ 10.1126/science.1235768Fig. 3. Vertical displacem...
Upcoming SlideShare
Loading in …5

The origin of_lunar_mascon_basins


Published on

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

The origin of_lunar_mascon_basins

  1. 1. Reports/ / 30 May 2013 / Page 1/ 10.1126/science.1235768High-resolution gravity data obtained from NASA’s dual GravityRecovery and Interior Laboratory (GRAIL) spacecraft now provide un-precedented high-resolution measurements of the gravity anomaliesassociated with lunar impact basins (1). These gravity anomalies are themost striking and consistent features of the Moon’s large-scale gravityfield. Positive gravity anomalies in basins partially filled with mare bas-alt, such as Humorum (Fig. 1), have been known since 1968, when lunarmass concentrations or “mascons” were discovered (2). Mascons havesubsequently been identified in association with impact basins on Mars(3) and Mercury (4). Previous analysis of lunar gravity and topographydata indicated that at least nine such mare basins possess central positiveanomalies exceeding that attributable to lava emplacement alone (5).This result is confirmed by GRAIL observations over basins that lackbasaltic infilling, such as Freundlich-Sharanov (Fig. 1), which are alsocharacterized by a central positive free-air gravity anomaly surroundedby a concentric gravity low. These positive anomalies indicate an excessof subsurface mass beyond that required for isostatic (mass) balance—a“superisostatic” state. Mascon formation seems ubiquitous in lunar ba-sins, whether mare-filled or not, despite their formation by impacts (aprocess of mass removal that leaves a topographic low, which normallyimplies a negative gravity anomaly), making mascons one of the oldestpuzzles of lunar geophysics. Their elucidation is one of the goals of theGRAIL mission.The gravity anomaly structure of lunar mascon basins was previous-ly attributed to mantle rebound during collapse of the transient cratercavity (5, 6). This process requires a lithosphere beneath the basin capa-ble of supporting a superisostatic load immediately after impact, a pro-posal that conflicts with the expectation that post-impact temperatureswere sufficiently high to melt both crustal and mantle rocks (7). Alterna-tively, it was proposed (8) that mascons are created by flexural uplift ofa thickened annulus of subisostatic (a deficiency of the subsurface massrequired for isostasy) crust surroundingthe basin, concomitantly lifting thebasin interior as it cooled and the un-derlying lithosphere became stronger.This alternative model emphasizes theannulus of anomalously low gravita-tional acceleration surrounding all mas-cons (Fig. 1) (1, 9, 10), a featurepreviously attributed to thickened crust(5, 6) or perhaps brecciation of the crustduring impact. Many mascons alsoexhibit an annulus of positive gravita-tional acceleration surrounding theannulus of negative gravity anomaly, sothe gravity structure of most lunar ba-sins resembles a bulls-eye target (Fig.1).The role of uplift in the formationof mascon basins has been difficult totest because little is known about themechanical state of basins immediatelyafter cavity collapse. Here we coupleGRAIL gravity and lunar topographydata from the Lunar Orbiter Laser Al-timeter (LOLA) (11) with numericalmodeling to show that the gravityanomaly pattern of a mascon is thenatural consequence of impact craterexcavation in the warm Moon, followedby post-impact isostatic adjustment (12)during cooling and contraction (13) of avoluminous melt pool. In mare-filledbasins this stage in basin evolution was followed by emplacement ofmare-basalt lavas and associated subsidence and lithospheric flexure.We used the axisymmetric iSALE hydrocode (14–16) to simulate theprocess of crater excavation and collapse. Our models used a typicallunar impact velocity of 15 km/s (17) and a two-layer target simulating agabbroic lunar crust (density = 2550 kg/m3; 19) and a dunite mantle(3200 kg/m3). Our objective was to simulate the cratering process thatled to the Freundlich-Sharanov and Humorum basins, which are locatedin areas where the crustal thickness is 40 and 25 km, respectively, asinferred from GRAIL and LOLA observations (19). We sought a combi-nation of impactor diameter and lunar thermal gradient that yielded anannulus of thickened crust at a radius of ~200 km, consistent with theannulus of negative free-air gravity anomaly around those basins.The dependence of material strength on temperature and pressurehas the most marked effect on the formation of large impact basins (20).With little certainty regarding the temperature–depth profile of the earlyMoon or the diameter of the impactor, we considered impactor diametersranging from 30 to 80 km and three possible shallow thermal gradients(21), 10, 20, and 30 K/km, from a 300 K surface. To avoid melted mate-rial in the mantle, the thermal profile was assumed to follow that for asubsolidus convective regime (0.05 K/km adiabat) at temperatures above1300 K. We found that impact at vertical incidence of a 50-km-diameterimpactor in conjunction with a 30 K/km initial thermal gradient bestmatched the extent of the annular gravity low and led to an increase incrustal thickness of 10–15 km at a radial distance of 200–260 km fromboth basin centers (Fig. 2), despite the differences in initial crustal thick-ness (22).A crucial aspect of the model is the formation of the subisostatic col-lar of thickened crust surrounding the deep central pool of melted mantlerock. The crust is thickened as the impact ejects crustal material onto thecool, strong preexisting crust. The ejected material forms a wedge ap-The Origin of Lunar Mascon BasinsH. J. Melosh,1,2* Andrew M. Freed,1Brandon C. Johnson,2David M.Blair,1Jeffrey C. Andrews-Hanna,3Gregory A. Neumann,4Roger J.Phillips,5David E. Smith,6Sean C. Solomon,7,8Mark A. Wieczorek,9Maria T. Zuber61Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, 550 Stadium Mall Drive,West Lafayette, IN 47907, USA.2Department of Physics, Purdue University, 525 Northwestern Avenue,West Lafayette, IN 47907, USA.3Department of Geophysics, Colorado School of Mines, 1500 IllinoisStreet, Golden, CO 80401–1887, USA.4Solar System Exploration Division, NASA Goddard Space FlightCenter, Greenbelt, MD 20771, USA.5Planetary Science Directorate, Southwest Research Institute,Boulder, CO 80302, USA.6Department of Earth, Atmospheric and Planetary Sciences, MassachusettsInstitute of Technology, Cambridge, MA 02139–4307, USA.7Department of Terrestrial Magnetism,Carnegie Institution of Washington, Washington, DC 20015, USA.8Lamont-Doherty Earth Observatory,Columbia University, Palisades, NY 10964, USA.9Institut de Physique du Globe de Paris, Sorbonne ParisCité, Université Paris Diderot, 75205 Paris Cedex 13, France.*Corresponding author. E-mail: jmelosh@purdue.eduHigh-resolution gravity data from the GRAIL spacecraft have clarified the origin oflunar mascons. Free-air gravity anomalies over lunar impact basins display bulls-eye patterns consisting of a central positive (mascon) anomaly, a surroundingnegative collar, and a positive outer annulus. We show that this pattern results fromimpact basin excavation and collapse followed by isostatic adjustment and coolingand contraction of a voluminous melt pool. We employed a hydrocode to simulatethe impact and a self-consistent finite-element model to simulate the subsequentviscoelastic relaxation and cooling. The primary parameters controlling the modeledgravity signatures of mascon basins are the impactor energy, the lunar thermalgradient at the time of impact, the crustal thickness, and the extent of volcanic fill.onJune2,2013www.sciencemag.orgDownloadedfrom
  2. 2. / / 30 May 2013 / Page 2/ 10.1126/science.1235768proximately 15 km thick at its inner edge that thins with increasing dis-tance from the center. The preexisting crust is drawn downward and intothe transient crater cavity due to a combination of loading by ejecta andinward flow of the underlying mantle, deforming it into a subisostaticconfiguration. This arrangement is maintained by the frictional strengthof the cool (but thoroughly shattered) crust, as well as by the viscoelasticmantle that requires time to relax. It is the subsequent relaxation of themantle that leads to a later isostatic adjustment. The result is a thick,low-density crustal collar around the central hot melt pool that is initiallyprevented from mechanically rebounding from its disequilibrium state.The higher thermal gradient of 30 K/km, somewhat counter-intuitively,yields a thicker subisostatic crustal collar than the thermal gradients of10 and 20 K/km. This difference occurs because the weaker mantle as-sociated with a higher thermal gradient flows more readily during thecollapse of the transient crater, exerting less inward drag on the crustalcollar, which consequently experiences less stretching and thinning.Calculations suggest that the impact into relatively thin crust at Hu-morum basin fully exposed mantle material in the central region of thebasin (Fig. 2B), whereas a ~15-km-thick cap of crustal material flowedover the central region of the Freundlich-Sharanov basin (Fig. 2A). Thiscrustal cap material was warm weak lower crustal material that migratedto the basin center during crater collapse (Fig. S1). At the end of thecrater collapse process, the basins (defined by their negative topography)were 4–5 km deep out to 150 km from the basin center, with shallownegative topography continuing to a radial distance of 350–400 km,approximately twice the excavation radius. A substantial melt pool, de-fined as mantle at temperatures above 1500 K, developed in both basins.This melt pool extended out to ~150 km from the basin center and tomore than 100 km depth (Fig. 2).To model the subsequent evolution of the basins, we used the finiteelement code Abaqus (23, 24). We developed axisymmetric models ofthe Humorum and Freundlich-Sharanov basins from the hydrocode out-put, adjusting the thermal structure of the melt to account for rapid post-impact convection and thermal homogenization of the melt pool. Thedensity of solid and liquid silicate material was calculated from the bulkcomposition of the silicate Moon (22, 25).Our models (Figs. 1, C and D) show that the depressed basin topog-raphy, the thickened crustal collar, and the lower density of heated mate-rial combine to create a substantial negative free-air gravity anomaly atthe basin centers (22). The post-impact free-air anomaly is slightly posi-tive outside of the basin owing to ejecta supported by the cool, strongcrust and mantle. The overall shape of the modeled post-impact free-airgravity anomaly is similar to that observed but is much more negative,suggesting that the general pattern of the observed gravity anomaly is theresult of the impact, but that subsequent evolution of the basin drove thecentral anomalies positive.As the impact-heated mantle beneath the basin cooled, the pressuregradient from its exterior to its interior drove viscoelastic flow towardthe basin center, uplifting the basin floor. The inner basin (where thecentral mascon develops) cannot rise above isostatic equilibrium solelydue to forces from its own subisostatic state. However, mechanical cou-pling between the inner and outer basin—where the collar of thickenedcrust was also rising isostatically—provided additional lift to the innerbasin floor, enabling it to achieve a superisostatic state. This mechanicalcoupling is achieved if the lithosphere above the melt pool thickenedsufficiently as it cooled. In the case of the Freundlich-Sharanov basin,the 15-km-thick layer of cool crust provided an initial (if thin) litho-sphere from the beginning, which thickened as the underlying mantlecooled. For Humorum basin, the melt pool reached the surface and thusthere was initially no lithosphere, although one developed during cool-ing. Our calculations show that if the viscosity of the mantle outside themelt pool is consistent with dry dunite, its viscoelastic strength woulddelay isostatic uplift of the basin floor such that lithospheres sufficientfor development of a mascon develop over the melt pools in both basins.In addition to these isostatic forces, cooling increases the density of themelt through contraction; given a strong lithosphere that hinders thesinking of this higher-density material, this process further increases thegravity anomaly at the basin center. The net effect is that isostatic upliftof the surrounding depressed surface topography and crustal collar,combined with cooling and contraction of the melt pool, create the cen-tral positive free-air anomaly. The flexural strength that enables the innerbasin to rise into a superisostatic state prevented the outer basin fromfully rising to isostatic equilibrium, leaving the observed ring of negativefree-air anomaly that surrounds the inner basin.Isostatic uplift raised the surface topography of the Freundlich-Sharanov basin by ~2 km at the center of the basin (Fig. 3A). Theseeffects place the final basin depth at just over 4 km, consistent withLOLA elevation measurements (11, 22). For the Humorum basin, theinner basin was calculated to rise ~3 km (Fig. 3B). This uplift distribu-tion would have left the Humorum basin ~4 km deep prior to mare fill.Infilling of a 3-km-thick mare unit and associated subsidence brings thefloor depth of the Humorum basin to just over 1.5 km deep, modestlydeeper then the 1 km depth measured by LOLA (22).The free-air gravity anomalies of both basins increased markedly af-ter crater collapse as a result of cooling and isostatic uplift. The free-airanomaly of the Freundlich-Sharanov basin is predicted to have risen to apositive 80 mGal in the inner basin and -200 mGal in the outer basinabove the thickened crust, in excellent agreement with GRAIL observa-tions (1) (red line in Fig. 1C). Furthermore, the model predicts an outerannulus of positive anomalies, also in agreement with observations. Asimilar post-impact increase in the free-air anomaly is observed in ourmodel of Humorum basin (red line in Fig. 1D), although this gravityanomaly cannot be verified because the Humorum basin was subse-quently partially filled with mare basalt. Our results support the infer-ence that lunar basins possess a positive gravity anomaly in excess of themare load (5). As a final step in our analysis, we emplaced a mare unit 3km thick and 150 km in radius (tapered to zero thickness over the outer-most 50 km in radial distance), within the Humorum basin. The additionof the mare increases the mascon at the center of the Humorum basin to320 mGal (blue line in Fig, 1D), matching GRAIL measurements (1).This basin evolution scenario depends primarily on the energy of theimpactor, the thermal gradient of the Moon at the time of the impact, andthe thickness of the crust. A high thermal gradient enables weaker man-tle to flow more readily during the collapse of the transient crater, result-ing in less inward motion and thinning of the crust. In contrast tohydrocode parameters that control crater excavation and collapse, suchas the energy of the impactor and the initial thermal gradient, the closematch of our predicted free-air gravity anomalies to those observed byGRAIL is not a product of finding a special combination of finite-element model parameters associated with isostatic uplift and cooling.These processes are controlled by the evolution of the density and vis-cosity structure in the model, which follow from the mineralogy of thelunar crust and mantle and the evolution of temperature as the regionconductively cools.References and Notes1. M. T. Zuber et al., Gravity field of the Moon from the GravityRecovery and Interior Laboratory (GRAIL) mission. Science 339,668 (2013). doi:10.1126/science.1231507 Medline2. P. M. Muller, W. L. Sjogren, Mascons: lunar mass concentrations.Science 161, 680 (1968). doi:10.1126/science.161.3842.680 Medline3. D. E. Smith et al., An improved gravity model for Mars: GoddardMars model 1. J. Geophys. Res. 98, (E11), 20,871 (1993).doi:10.1029/93JE018394. D. E. Smith et al., Gravity field and internal structure of Mercuryfrom MESSENGER. Science 336, 214 (2012).doi:10.1126/science.1218809 MedlineonJune2,2013www.sciencemag.orgDownloadedfrom
  3. 3. / / 30 May 2013 / Page 3/ 10.1126/science.12357685. G. A. Neumann, M. T. Zuber, D. E. Smith, F. G. Lemoine, Globalstructure and signature of major basins. J. Geophys. Res. 101, (E7),16,841 (1994). doi:10.1029/96JE012466. M. A. Wieczorek, R. J. Phillips, Lunar multiring basins and thecratering process. Icarus 139, 246 (1999).doi:10.1006/icar.1999.61027. E. Pierazzo, H. J. Melosh, Melt production in oblique impacts. Icarus145, 252 (2000). doi:10.1006/icar.1999.63328. J. C. Andrews-Hanna, The origin of non-mare mascon gravityanomalies on the Moon. Lunar Planet. Sci. 43, 2804 (2012).9. W. L. Sjogren, R. N. Wimberly, W. R. Wollenhaupt, Lunar gravityvia the Apollo 15 and 16 subsatellites. Moon 9, 115 (1974).doi:10.1007/BF0056539810. M. T. Zuber, D. E. Smith, F. G. Lemoine, G. A. Neumann, Theshape and internal structure of the moon from the clementinemission. Science 266, 1839 (1994).doi:10.1126/science.266.5192.1839 Medline11. D. E. Smith et al., Initial observations from the Lunar Orbiter LaserAltimeter (LOLA). Geophys. Res. Lett. 37, L18204 (2010).doi:10.1029/2010GL04375112. “Isostatic adjustment” as used here is the process by which thestresses imparted in a non-isostatic crust–mantle volume are relievedas they drive density boundaries toward mass balance (isostasy). Thelevel of isostasy achieved depends on viscosity-controlled flow andalso on the finite strength of the system as characterized bylithospheric flexure. This “isostatic adjustment” includes the uplift ofthe basin center to a super-isostatic position as a result of its flexuralcoupling to the sub-isostatic annulus.13. H. J. Melosh, D. M. Blair, A. M. Freed, Origin of superisostaticgravity anomalies in lunar basins. Lunar Planet. Sci. 43, 2596 (2012).14. A. A. Amsden, H. M. Ruppel, C. W. Hirt, SALE: A simplified ALEcomputer program for fluid flow at all speeds. LANL Rep. LA-8095,101 pp., Los Alamos Natl. Lab., Los Alamos, N. M. (1980).15. G. S. Collins, H. J. Melosh, B. A. Ivanov, Modeling damage anddeformation in impact simulations. Meteorit. Planet. Sci. 39, 217(2004). doi:10.1111/j.1945-5100.2004.tb00337.x16. K. W[UNKNOWN ENTITY &udie;]nnemann, G. S. Collins, H. J.Melosh, A strain-based porosity model for use in hydrocodesimulations of impacts and implications for transient crater growth inporous targets. Icarus 180, 514 (2006).doi:10.1016/j.icarus.2005.10.01317. The precise value of the impact velocity is not critical for thiscomputation because a lower impact velocity can be compensated bya larger impactor, and vice versa. The impact velocity distribution onthe Moon is strongly skewed toward high velocities, with a mode at10 km/s and a median about 15 km/s (18).18. M. Le Feuvre, M. A. Wieczorek, Nonuniform cratering of the Moonand a revised crater chronology of the inner Solar System. Icarus214, 1 (2011). doi:10.1016/j.icarus.2011.03.01019. M. A. Wieczorek et al., The crust of the Moon as seen by GRAIL.Science 339, 671 (2013). doi:10.1126/science.1231530 Medline20. B. A. Ivanov, H. J. Melosh, E. Pierazzo, E. Basin-forming impacts:Reconnaissance modeling. in Large Meteorite Impacts and PlanetaryEvolution IV, W. U. Reimold, R. L. Gibson, Eds. (Special Paper 465,Geological Society of America, Boulder, Colo., 2010), pp. 29–49.21. G. Schubert, D. L. Turcotte, P. Olson, Mantle Convection in theEarth and Planets (Cambridge Univ. Press, Cambridge, 2001), pp.940 pp.22. More detailed descriptions of these models and methods areavailable as supplementary materials on Science Online.23. A. M. Freed, S. C. Solomon, T. R. Watters, R. J. Phillips, M. T.Zuber, Could Pantheon Fossae be the result of the Apollodoruscrater-forming impact within the Caloris basin, Mercury? EarthPlanet. Sci. Lett. 285, 320 (2009). doi:10.1016/j.epsl.2009.02.03824. A. M. Freed et al., On the origin of graben and ridges within andnear volcanically buried craters and basins in Mercury’s northernplains. J. Geophys. Res. 117, E00L06 (2012).doi:10.1029/2012JE00411925. S. R. Taylor, Planetary Science: A Lunar Perspective (Lunar andPlanetary Institute, Houston, TX, 1982), p. 401.26. J. F. Brotchie, Flexure of a liquid-filled spherical shell in a radialgravity field. Mod. Geol. 3, 15 (1971).27. W. Benz, A. G. W. Cameron, H. J. Melosh, The origin of the moonand the single-impact hypothesis III. Icarus 81, 113 (1989).doi:10.1016/0019-1035(89)90129-2 Medline28. K. Wünnemann, G. S. Collins G. R. Osinski, Numerical modeling ofimpact melt production in porous rocks. Earth Planet. Sci. Lett. 269,530 (2008). doi:10.1016/j.epsl.2008.03.00729. R. W. K. Potter, G. S. Collins, W. S. Kiefer, P. J. McGovern, D. A.Kring, Constraining the size of the South Pole-Aitken basin impact.Icarus 220, 730 (2012). doi:10.1016/j.icarus.2012.05.03230. E. Azmon, “The melting of gabbro up to 45 kilobars.” NASATechnical Report NASA-CR-86627 (1967).31. R. M. Stesky, W. F. Brace, D. K. Riley, P. Y. Robin, Friction infaulted rock at high temperature and pressure. Tectonophysics 23,177 (1974). doi:10.1016/0040-1951(74)90119-X32. M. Shimada, A. Cho, H. Yukutake, Fracture strength of dry silicaterocks at high confining pressures and activity of acoustic emission.Tectonophysics 96, 159 (1983). doi:10.1016/0040-1951(83)90248-233. T. M. Davison, G. S. Collins, F. J. Ciesla, Numerical modeling ofheating in porous planetesimal collisions. Icarus 208, 468 (2010).doi:10.1016/j.icarus.2010.01.03434. I. A. H. Ismail, S. A. F. Murrell, The effect of confining pressure onstress-drop in compressive rock fracture. Tectonophysics 175, 237(1990). doi:10.1016/0040-1951(90)90140-435. D. McKenzie, M. Bickle, The volume and composition of meltgenerated by extension of the lithosphere. J. Petrol. 29, 625 (1988).doi:10.1093/petrology/29.3.62536. R. W. K. Potter, Numerical modeling of basin-scale impact craterformation, Ph.D. thesis, Imperial College London, England (2012).37. F. J. Spera, Physical properties of magma. in Encyclopedia ofVolcanoes, H. Sigurdsson, Ed. (Academic Press, San Diego, CA,2000), pp. 171–190.38. E. Ito, E. Takahashi, Melting of peridotite at uppermost lower-mantleconditions. Nature 328, 514 (1987). doi:10.1038/328514a039. A. R. McBirney, Igneous Petrology (Jones and Bartlett, Boston, ed.2, 1993), p. 27.40. A. Khan, J. Maclennan, S. R. Taylor, J. A. D. Connolly, Are theEarth and the Moon compositionally alike? Inferences on lunarcomposition and implications for lunar origin and evolution fromgeophysical modeling. J. Geophys. Res. 111, (E5), E05005 (2006).doi:10.1029/2005JE00260841. E. Knittle, Static compression measurements of equations of state. inMineral Physics and Crystallography, T. J. Ahrens, Ed. (ReferenceShelf 2, American Geophysical Union, Washington, DC, 1995), pp.98–142.42. Y. Fei, Thermal expansion. Mineral Physics and Crystallography, T.J. Ahrens, Ed. (Reference Shelf 2, American GeophysicalUnion,Washington, DC, 1995), pp. 29–44.43. E. P. Turtle, H. J. Melosh, Stress and flexural modeling of theMartian lithospheric response to Alba Patera. Icarus 126, 197 (1997).doi:10.1006/icar.1996.5638Acknowledgments: The GRAIL mission is supported by NASA’sDiscovery Program and is performed under contract to theMassachusetts Institute of Technology and the Jet PropulsionLaboratory. The Lunar Reconnaissance Orbiter Lunar Orbiter LaserAltimeter (LOLA) investigation is supported by the NASA ScienceMission Directorate under contract to the NASA Goddard SpaceFlight Center and Massachusetts Institute of Technology. Data fromthe GRAIL and LOLA missions have been deposited in theGeosciences Node of NASA’s Planetary Data System.Supplementary TextFigs. S1 to S6Tables S1 to S4References (26–43)onJune2,2013www.sciencemag.orgDownloadedfrom
  4. 4. / / 30 May 2013 / Page 4/ 10.1126/science.123576828 January 2013; accepted 16 May 2013Published online 30 May 201310.1126/science.1235768onJune2,2013www.sciencemag.orgDownloadedfrom
  5. 5. / / 30 May 2013 / Page 5/ 10.1126/science.1235768Fig. 1. Free-air gravity anomalies over (A) the mare-free Freundlich-Sharanov basin (radius to the center of the free-airgravity low: 215 km) and (B) the mare-filled Humorum basin (radius to the center of the annular free-air gravity low: 215km) from GRAIL observations (1). (C and D) Comparison of observed and calculated free-air gravity anomalies for theFreundlich-Sharanov and Humorum basins, respectively. The observed anomalies and associated one-standard-deviationranges were derived from averages of the data within concentric rings at different radial distances. The black linesrepresent the predicted gravity anomaly just after impact and transient cavity collapse, from the hydrocode calculation. Thered lines represent the predicted anomaly after uplift following isostatic response and cooling, appropriate for comparisonto the Freundlich-Sharanov data. The blue line in (D) represents the predicted gravity anomaly after mare emplacement inthe Humorum basin and is appropriate for comparison to data from that basin.onJune2,2013www.sciencemag.orgDownloadedfrom
  6. 6. / / 30 May 2013 / Page 6/ 10.1126/science.1235768Fig. 2. Vertical cross section of crust and mantle geometry and thermal structure after crater collapse (2hours after impact) for the (A) Freundlich-Sharanov basin (40-km-thick original crust) and (B) Humorum basin(25-km-thick original crust), according to the hydrocode calculation.onJune2,2013www.sciencemag.orgDownloadedfrom
  7. 7. / / 30 May 2013 / Page 7/ 10.1126/science.1235768Fig. 3. Vertical displacement calculated by the finite element model relative to the initial post-crater-collapseconfiguration predicted by the hydrocode for the unfilled (A) Freundlich-Sharanov basin (B) Humorum basin.The deformation is exaggerated by a factor of 10.onJune2,2013www.sciencemag.orgDownloadedfrom