REPORTS a broad range of temperatures (5) (table S1). Diviner Lunar Radiometer Daytime brightness temperatures (Fig. 1B) show that the LCROSS impact site in Cabeus crater Observations of the LCROSS Impact is in persistent shadow, with a nearly isother- mal surface at ~ 40 K just before impact (6). Figure 1 shows that the LCROSS impact oc- Paul O. Hayne,1* Benjamin T. Greenhagen,2 Marc C. Foote,2 Matthew A. Siegler,1 curred near the center of one of the coldest re- Ashwin R. Vasavada,2 David A. Paige1 gions of Cabeus crater. Thermal models (2)place the LCROSS impact site’s annual average tem- The Lunar Reconnaissance Orbiter (LRO) Diviner instrument detected a thermal emission signature perature at 37 K, in the 99.7th percentile (by 90 seconds after the Lunar Crater Observation and Sensing Satellite (LCROSS) Centaur impact and on area) for the region poleward of 80°S. two subsequent orbits. The impact heated a region of 30 to 200 square meters to at least 950 kelvin, The LRO orbit was modified to put the providing a sustained heat source for the sublimation of up to ~300 kilograms of water ice during closest approach (slant range 78 km) at 90 s after the 4 minutes of LCROSS post-impact observations. Diviner visible observations constrain the mass of impact of the launch vehicle’s spent Centaur up- the sunlit ejecta column to be ~10−6 to 10−5 kilograms per square meter, which is consistent with per stage, in order to better observe the LCROSS LCROSS estimates used to derive the relative abundance of the ice within the regolith. impact while protecting the spacecraft from de- bris. Also, the spacecraft was rolled 81.1° to allow n 9 October 2009, the Diviner Lunar cooling during a planetary impact in a region the Lyman Alpha Mapping Project (LAMP) to Radiometer onboard the Lunar Recon- where volatiles may be cold-trapped (3, 4). We view the lunar limb above Cabeus crater. Diviner naissance Orbiter (LRO) observed a con- report on the effects of volatiles on the temper- O 1 Department of Earth and Space Sciences, University of trolled impact of the Lunar Crater Observation ature behavior observed by Diviner and the 2 1 California, Los Angeles, CA 90095, USA. Jet Propulsion Lab- Downloaded from and Sensing Satellite (LCROSS) ( ) into one of LCROSS Shepherding Spacecraft (SSc). oratory, California Institute of Technology, Pasadena, CA the coldest places on the Moon (2). Diviner’smul- Diviner is a push-broom visible and infrared 91109, USA. tispectral thermal infrared measurements of this radiometer with nine channels spanning wave- *To whom correspondence should be addressed. E-mail: event provide insight into energy dissipation and lengths from 0.3 to 400 mm, with sensitivity to [email protected] http://science.sciencemag.org/ on July 18, 2017 Fig. 1. (A) Pre- and post-impact images (T–2h, T+90s) of the LCROSS (6). The colors for the solar channel scale linearly with radiance from impact site. (B) Diviner 50- to 100-mm pre-impact brightness tem- 1.2 × 10−2 to 4.3 × 10−2 Wm−2 sr−1; the infrared frames are linear perature map from nadir data, 1 to 15 October 2009. The width of with brightness temperature, with ranges of 85 to 105 K (13 to 23 mm), eachinsetframe(whitebox)isabout15km.TheLCROSSimpactsite 60 to 90 K (25 to 41 mm), 55 to 80 K (50 to 100 mm), and 65 to 80 K (x) had an estimated surface temperature of ~40 K just before impact (100 to 400 mm). www.sciencemag.org SCIENCE VOL 330 22 OCTOBER 2010 477 REPORTS has independent azimuth and elevation actua- pact (11). A simple model using two temperature Cooling of the impact zone occurred primar- tors, enabling targeted observations despite this components accurately reproduces the Diviner ily by surface radiation to space, sublimation of unusual spacecraft attitude. Diviner targeted the data, captures the primary features of the spec- ice, and (to a lesser degree) downward conduc- impact site for eight orbits, separated by about tral energy distribution (figs. S2 and S4), and tion at the base of the hot layer. We simulated 2 hours each, the third orbit occurring about 90 s is generally consistent with SSc mid-infrared these processes with a one-dimensional heat dif- after Centaur impact (T+90s). Diviner’sviewof images of the impact site shortly after impact fusion model, including the conductivity effects the impact site was oblique, with an emission (12). In this model, dissipation generates a high- of ice as well as sublimation and vapor diffusion angle of ~ 48° and pixel size of ~400 m. temperature component within the crater (diam- (14). An unknown fraction of the ~7 × 109 J All seven of Diviner’s infrared channels mea- eter ~25 m) and a larger region with a somewhat total impact energy is irreversibly converted to sured thermal emission from the Centaur impact lower temperature represents an ejecta blanket heat, although an upper limit of 20% is plau- crater during the T+90s observations (Fig. 1A). (diameter ~150 m). sible (15, 16). We investigated the cooling be- Also, the more sensitive of the two solar channels We simulated radiance values for Diviner’s havior of the hot inner region (30 to 200 m2) observed both thermal emission from the sur- spectral channels (table S1) by convolving each for different values of Eh, the fraction of the face and scattered sunlight from the ejecta cloud. channel’s spectral response with the area-weighted total impact energy contributing to heating this About 2 hours later, Diviner’s three longest- blackbody spectral radiance for each of the two zone (Fig. 2). Only cooling curves with Eh > wavelength channels again measured a thermal temperature components, with a constant back- 3% and initial temperature > 950 K are consist- signal from the subpixel impact site. At 4 hours ground radiance filling the remainder of the field ent with the T+90s observations; upper limits on after impact (T+4h), only the channel spanning of view. Diviner’s high-sensitivity solar channel these quantities are not well constrained (6). Ice wavelengths of 25 to 41 mm measured a signal cannot detect temperatures below ~ 400 K for within the regolith causes more rapid cooling by above the noise level. No signal due to the an area less than 103 m2 (~1% pixel fill factor). loss of latent and sensible heat, as well as more LCROSS Centaur impact was observed on later Therefore, the clear thermal emission signal in this efficient downward conduction. If the regolith orbits. The complete set of measurements is channel places a strong lower limit on the tem- pores were filled with ice (50% porosity), we find Downloaded from summarized in table S1. This report focuses only perature of the hot component at T+90s. Using that the heat energy fraction Eh must be >12%, on the Centaur impact, because emission from the the 8-mm channels’ T+90s radiances, we find a otherwise cooling is too rapid to match the T+90s þ50 E SSc impact (which occurred after the T+90s best-fit hot component temperature of 600−90 K, Diviner data. Conversely, if h =5%,icemass þ140 2 6 13 observations and with much lower energy) has with an area of 60−30 m ( , ). fractions must be <4%. Thus, an independent not yet been found in the data. Diviner’s longer-wavelength infrared channels constraint on Eh would effectively place limits on In addition to thermal emission from the constrain the average ejecta blanket temperature the initial ice content in the impact zone. http://science.sciencemag.org/ þ50 : þ10 4 2 ’ Centaur impact site, we attribute an enhanced to be 110−20 K, with an area of 8 0−6 ×10 m Diviner s estimate for the area of the hot signal in Diviner’s solar channel, evident over atT+90s.Between~0.5and1kmawayfromthe component, 30 to 200 m2, is somewhat less than a broad (~140 km2) region, to scattered sunlight impact, brightness temperatures are elevated the area of the Centaur’s impact crater, ~500 to from the ejecta plume. Using the radiance of the slightly from the pre-impact background; this is 700 m2 (12), which indicates that some portion scattered sunlight, we calculate a total normal attributed to emission by warm grains in the of this region cooled below the detection limit optical depth of 2.0 × 10−5 to 8.1 × 10−4 over sunlit ejecta cloud. About 2 hours after the Centaur of ~300 K (for an area of 500 m2)oftheDi- this region at T+90s (7, 8). Given a typical lunar impact, the hot region had cooled below the viner 8-mmchannelsinlessthan90s.Thisis soil grain size distribution (9), the total column ~450 K detection threshold (area 60 m2)ofthe consistent with SSc NIR2 images of the Centaur mass is 1.0 × 10−6 to 4.1 × 10−5 kg m−2.As- 8-mm channels. A c2 minimization procedure crater about 4 min after impact, which reveal a 2 suming an even distribution over the 140-km using the observed emission in two channels dark inner region, perhaps due to more rapid on July 18, 2017 region, the total mass above the sunlight hori- (25 to 41 mm and 50 to 100 mm) failed to cooling of icy material at depth (12). An in- þ3200 10 zon of ~800 m is 2600−1400 kg ( ). This range is yield a reliable best-fit temperature for the hot crease in ice content of just a few percent by consistent with those derived from the LCROSS component; however, we estimate a mean tem- mass over the excavation depth (~1 to 3 m) could spectrometer data (1), although some material perature of ~100 K for the combined crater– result in more than 100 K of cooling within the may have been outside Diviner’s field of view.