GEOPHYSICAL RESEARCH LETTERS, VOL. 22, NO. 17, PAGES 2433-2436, SEPTEMBER 1, 1995 CometShoemaker-Levy 9' Fragmentand Progenitor Impact Energy Toshiko Takata* and Thomas J. Ahrens+ LindhurstLaboratory of ExperimentalGeophysics, Division of Geological& PlanetarySciences, California Institute of Technology, Pasadena Alan W. Harris JetPropulsion Laboratory, Pasadena, California Abstract. Initial observational data from the impact of fragment. Previous smoothedparticle hydrodynamic (SPH) fragmentsof Comet Shoemaker-Levy9 (SL9) are compared simulations [Ahrens et al., 1994; Takata et al., 1994] for the with smoothedparticle hydrodynamic(SPH) calculationsto impact of solid ice fragments demonstratedthat the impact determine their pre-impact diameters and the equivalent energy depositedin the Jovian atmosphereresulted in plumes diameter of the SL9 progenitor. Diameters (solid ice) of which rise to more than severaltens of scaleheights. 2.0-1-0.1,2.0-20.05, 2.1+0.04 and 1.9+0.05 km for fragmentsA, The plumesobserved by Hammel et al. [ 1995] for fragments E, G1, and W are obtainedfrom impact-inducedplume heights A, E, G1, and W reached maximum heights of 2966+ 370, from the Hubble SpaceTelescope (HST) data. Applying these 2916+170(lower bound),3346+170 and 2666+170km (lower values to scale apparentdiameters for the balance of 18 SL bound), respectively. The nearly constant height of these fragmentsin Weaver et al.'s [1995] catalogof 22 objectsyields plumes (Fig. 1) is surprising as much larger variations in a SL9 progenitordiameter of 5.0-2_1.8km. This correspondsto fragmentsizes are inferredboth from the pre-impactphotometric totalimpact energy of 1.2(+1.8 - 0.8)x 1030erg. Such an data and in the variability of the area of opaque material energeticevent occurson Jupiterand Earth at leastevery 4,900 depositedafter impact of the SL9 fragmentson the cloud tops. +4,700,-2,700, and -0.5 x 108years, respectively. Here we used 0 km at the 1 bar datum. Interpolation of maximum plume heightsbetween our SPH simulations[Takata Introduction et al., 1994] for a 0.4 and 2 km diameterbolide yields peak The most spectacularand energetic planetary event ever heightsof-500 and -3000 km, respectively(Fig. 1) from: witnessedby humankindwas the recentimpact of fragmentsof Comet Shoemaker-Levy9 (SL9) on Jupiterduring the period of E (ergs)=3.1 x 1025 H (kin) (1) 16 to 21 July 1994. Energy of SL9 fragments is the most important parameter required for understandingthe impact where E is impact energy and H is plume height above the 480 phenomenaand effectson the Jovianatmosphere. Although the km level. We assumein Eq. I that the plume has a ballistic- velocity of the fragments(60 kin/s) was well determinedfrom type behavior for which the maximum height achieved is knowledgeof the orbit [Yeomans and Chodos, 1993], actual proportional to the kinetic energy per unit mass. This fragmentsizes and densitywere uncertain. Numericalmodels of assumptionneeds further testing. Scotti and Melosh [1993] and Asphaugand Benz [1994], which Sincefor a densityof l g/cm3 andimpact velocity of 60 arein part dependenton materialproperties of SL9, wereused to km/sec,E (erg) = 9.4x 1027D 3 (km)3, it followsthat reproducethe orbital positionof the line of fragmentsafter the break-upand indicatedthat the diameterof the parentbody was D (kin)= 0.15(H'(km) - 459)1/3 (2) < 2 km (diameter)and the total expectedimpact energywas - 8 x 1028erg. In contrast,photometric measurements (withthe 3500 HST) yielded an estimated(maximum) 7.4 km diameterparent body [Weaveret al., 1995] or total impactenergy (solid ice) of 4 3000 x 1030•.. erg. Thus there was an uncert ax 'nty of a factorof -50 for 2500 the total impactenergy. 2000 Plume Height and and Impact Energy 1500 The height achievedby the SL9 fragmentimpact-induced shock-heatedgas "plume" rising up in the inhomogeneous 1000 Jovianatmosphere is stronglydependent on impactenergy and to a degree,studied, in part, by Crawford et al. [1995] and 500 0.4km ] Bosloughet al. [1995] on the effectivediameter of the impacting *Presentaddress: Geological Institute, U.of Tokyo, Hongo, Bunkyo- 0 500 1000 1500 2000 ku, Tokyo, Japan Time after Impact (s) +Correspondent Figure 1. Plumeheight obtained from SPH calculationscompared Copyright1995 by the AmericanGeophysical Union. to HST [Hamreel et al., 1995] observations. Three-dimensional SPH calculationsgive plume height for impact of D = 0.4 and 2 Papernumber 95GL02237 km diameter at 60 km/sec on Jupiter. Here, 0 km height 0094-8534/95/95GL-02237503.00 representsthe 1 bar pressurelevel. 2433 2434 TAKATAET AL.: COMETSL9: ENERGY Table 1. Calculateddiameter of 22 SL9 fragments.Diameters in firstcolumn are from thepre-impact photometric data [Weaver et al., 1995]. Fragment,A, G1, E andW d'mmetersobtained from comparisonof plumeheights of our SPH simulationsand the HST observations[Hammel et al., 1995]. Lowerand upper limit valuescome from variancein plumeheights observed from fragmentsand that derivedfrom SPH calculationwhich is 3.1 x 1025erg/km above height of 479 km. Fragment UpperLimit Bestestimate LowerLimit UpperLimit ID diameter(kin) diameter (km) diameter(km) diameter(kin) fromWeaver presentwork presentwork presentwork et al. [ 1995] A* 1.4 2.01 1.91 2.1i B 1.68 1.12 0.36 1.63 C 2.10 1.40 0.45 2.04 D 1.4 0.94 0.30 1.36 E 2.8 2.00 1.95 2.04 F 2.1 1.40 0.45 2.04 G2 0.81 0.54 0.17 0.79 G1 4.06 2.11 2.07 2.15 H 3.08 2.06 0.66 2.98 K 3.78 2.53 0.81 3.66 L 3.50 2.34 0.75 3.39 N 1.4 0,94 0.30 1.36 P2 1.96 1.31 0.42 1.90 P1 0.854 0.57 0.18 0.83 Q2 3.08 2.06 0.66 2.98 Q1 4.06 2.71 0.87 3.93 R 2.52 1.68 0.54 2.44 S 2.94 1.96 0.63 2.85 T 0.644 0.43 0.14 0.62 U 0.91 0.61 0.20 0.88 V 1.36 0.91 0.29 1.32 W 2.38 !.93 1.88 1.98 Total -22 7.43 4.98 3.20 6.79 fragments, progenitor diameter(kin) *Boldface: Indicatesplume heights reported by Hammelet al. [1995]. whereH' is altitudemeasured from the 1 bar level. Applying of the 4 scaling parameters calculated from Eq. 3, the thisto observedplume heights gives DSP H = 2.0-+0.1,2.0-•.05, uncertaintiesare dominated by theuncertainties in Sa. 2.1+0.04, and 1.9_+0.5km for A, E, G1, andW. Upperlimits on Theplume gas ejected above the stratosphere follows a ~101 the diameterof each fragment,DW, are availablefrom Weaver minuteparabolic trajectory. Re-impactat a maximumradius of et al. [1995] optical data. We assumethat the squaresof these ~4h from the ejection point, where h is the maximumplume upperlimits are proportionalto the observedbrightness of each height,yields a radial rangeof the ejectaof ~12,000 km for the fragment. This assumption would be closely valid if the G 1 impact.This agreeswith theradius (10,000 - 14,000kin) for' brightnessis proportionalto disc area of each fragment. For the crescent-shapeddark region observed after impact of each of fragmentsA, E, G1, and W, then, we can calculatean fragmentG1 [Hamreel et al., 1995]. However, the reasonsthe areascaling par •ameter. • other impacts,for which plumes achievedsimilar heights,did not produceas large a radius dark regions,is not understood. Sa= (DSPH/Dw)2 (3) Recently,Boslough et al. [1995] have suggesteda modelof this process. which representsour best estimate of the ratio of the true area We compareour SPH calculationsto the HST data for the A, (inferred from the SPH simulationsand the observedplume E, !31, andW plume heights(Fig. 1) to determinethe diameter heights)to the effectivephotometric area; we obtainS a = 2.3, of A, E, G1, and W fragments,and then to scaleWeaver's et al. 0.53, 0.28, and 0.69, respectively,with an averagevalue of 0.94 [1995] catalog(Table 1). This yields5.0 -+ 1.8 km for an (ice) 4-0.90. Thus, the diameter of each the 18 fragmentswhose cometprogenitor diameter. The correspondingequivalent plumeswere not observedby H ,an)meletal. [1995], but whose enemyof a solidice projectile at60 km/sec is 1.2(1.8, - 0.8)x brightness was obtained by Weaver et al. [1995], can be 103•'ergs. estimated to lie between 0.2 and 1.36 times D W , with •.94 = 0.97DW asthe best estimate (column 3of Table Detectionof Deep AtmosphericConstituents 1). Additional observational(HST) uncertaintiesfor A, E, G 1, Priorto the SL9 impacton Jupiter,it waswidely believed that and W are reflected in Table 1 (columns 4 and 5). Since the theseimpacts wo•d exhumeH20 frombelow its presumed3 to balanceof the fragmentdiameters were scaledwith the average 5 barpressure level clouddeck and possibly H20 from deeper'm TAKATA ET AL.: COMET SL9: ENERGY 2435 the planet. Both finite difference[Boslough et al., 1994; Mac Infrared obscrvaQons[Orton et al., 1995] indicate that the Low and Zahnle, 1994; Zahnle and MacLow, 1994] and our temperatureof the impact sites, from a structuremodeled as [Ahrenset al., 1994; Takataet al., 1994] SPH calculationsagree 0.25-gin particles,for the G, L and Q featuresat the 1 to 200 that-8x 1028 erg (2 Pan diameter) impactors penetrate Jupiter mbarpressure levels did not changeappreciably after the impact, to pressuresof-30 bar or depthsof 170 Pan below the 1 bar and hence, it is claimed that SL9 fragments may not have level. The HST team detectedabundant absorption features penetratedto these pressurelevels. However, although the correspondingto NH 2, S2, CS2, and H2S and emissionfrom atmosphericplume radii observed exceed 2 x 103km for many Mg, Mg+, CS andFe [Nollet al., 1995].