Kerber et al. (2011) - Brown University Planetary Geosciences

Primitivo Vieira | Download | HTML Embed
  • Nov 28, 2011
  • Views: 17
  • Page(s): 15
  • Size: 3.53 MB
  • Report



1 Planetary and Space Science 59 (2011) 18951909 Contents lists available at ScienceDirect Planetary and Space Science journal homepage: The global distribution of pyroclastic deposits on Mercury: The view from MESSENGER ybys 13 Laura Kerber a,n, James W. Head a, David T. Blewett b, Sean C. Solomon c, Lionel Wilson d, Scott L. Murchie b, Mark S. Robinson e, Brett W. Denevi e, Deborah L. Domingue f a Department of Geological Sciences, Brown University, 324 Brook Street, Box 1846, Providence, RI 02912, USA b The Johns Hopkins University Applied Physics Laboratory, 11100 Johns Hopkins Road, Laurel, MD 20723, USA c Department of Terrestrial Magnetism, Carnegie Institution of Washington, 5241 Broad Branch Road, N.W., Washington, DC 20015, USA d Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, UK e School of Earth and Space Exploration, Arizona State University, Tempe, AZ 85251, USA f Planetary Science Institute, 177 E. Fort Lowell, Suite 106, Tucson, AZ 85719, USA a r t i c l e i n f o abstract Article history: We present a global survey of candidate pyroclastic deposits on Mercury, derived from images obtained Received 14 July 2010 during MESSENGER ybys 13 that provided near-global coverage at resolutions between 5 and Received in revised form 0.5 km/pixel. Thirty-ve deposits were identied and characterized and are located principally on the 28 February 2011 oors of craters, along rims of craters, and along the edge of the Caloris basin. Deposits are commonly Accepted 29 March 2011 centered on rimless, often irregularly shaped pits, mostly between 5 and 45 km in diameter. The Available online 8 April 2011 deposits identied are generally similar in morphology and absolute reectance to lunar pyroclastic Keywords: deposits. Spectrally the deposits appear brighter and redder than background Mercury terrain. On the Mercury basis of the available coverage, the candidate pyroclastic deposits appear to be essentially globally Pyroclastic distributed. The diameters of the deposits, when mapped to lunar gravity conditions, are larger than Volcanism their lunar counterparts, implying that more abundant volatiles were present during the typical Volcanoes Volatile eruptive process than on the Moon. These observations indicate that if these deposits resulted from Pits hawaiian-style eruptions, the volatile contents required would be between 1600 and 16,000 ppm CO or an equivalent value of H2O, CO2, SO2, or H2S (for a more oxidizing interior), or N2, S2, CS2, S2Cl, Cl, Cl2, or COS (for a more reducing interior). These abundances are much greater than those predicted by existing models for Mercurys formation. An apparent lack of small deposits, compared with the Moon, may be due to resolution effects, a topic that can be further assessed during the orbital phase of the MESSENGER mission. These results provide a framework within which orbital observations by MESSENGER and the future BepiColombo mission can be analyzed. & 2011 Elsevier Ltd. All rights reserved. 1. Introduction bodies such as the Moon and Mercury, eruptions range from gasless, effusive eruptions to explosive hawaiian, strombolian, or During the rst yby of the planet Mercury by the MErcury vulcanian eruptions, depending on the gas content and rise rate of Surface, Space ENvironment, GEochemistry, and Ranging (MES- the magma body and gas bubbles. Pyroclastic and other volcanic SENGER) spacecraft (Solomon et al., 2008), several deposits were deposits thus represent important sources of information about discovered that were hypothesized to be composed of pyro- planetary structure, composition (including volatile content), clastic material (Head et al., 2008, 2009; Murchie et al., 2008; stress state, and thermal history (Wilson, 2009). Compositional Robinson et al., 2008; Blewett et al., 2009a; Kerber et al., 2009). information, derived from visible to near-infrared spectroscopic Pyroclastic deposits are explosive volcanic eruption products remote sensing, as well as X-ray, neutron, and gamma-ray formed by the fragmentation and upward propulsion of magma spectrometry, can be used to characterize a surface deposit as particles driven by the expansion of volatile species released from well as to make inferences about the parent magma from which it rising bodies of magma (e.g., Wilson and Head, 1981). A wide was derived and the depth of its generation. The presence of range of pyroclastic eruption types and deposit morphologies is volcanism on the surface of Mercury implies conditions sufcient anticipated during planetary explosive eruptions. On airless to generate melts within the planet, combined with a planetary stress state conducive to allowing the magma generated to propagate to the surface (Wilson and Head, 1981, 2008). Precise n Corresponding author. Tel.: 1 14018633485. global mapping of volcanic features can demonstrate trends in the E-mail address: [email protected] (L. Kerber). thermal history of the planet through time and space and can aid 0032-0633/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.pss.2011.03.020

2 1896 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 in assessing the amount and distribution of volatile and heat- to nearby units (Robinson et al., 2008; Blewett et al., 2009a; producing elements within the interior (e.g., Solomon and Kerber et al., 2009). Many of the deposits are associated Chaiken, 1976; Wilson, 2009). with irregularly shaped vent-like depressions (Head et al., 2008, The discovery of pyroclastic deposits on the surface of Mercury 2009; Murchie et al., 2008; Robinson et al., 2008; Kerber et al., has already provided an important constraint on the interior 2009). volatile budget of the planet (Kerber et al., 2009), which had Subsequent ybys provided many more observations from previously been hypothesized to be extremely volatile poor (e.g., which additional pyroclastic deposits could be identied. From Boynton et al., 2007). The presence of pyroclastic and other images obtained during the second yby, Blewett et al. (2009b) volcanic deposits has also revealed that the stress state of the identied two additional pyroclastic deposits, one within the crust of Mercury has not, on a global basis, been sufciently crater Mistral and another within a crater modied by Antoniadi compressive throughout its history to prohibit the propagation of Dorsum, and Denevi et al. (2009) mentioned a possible pyroclastic dikes to the surface (Wilson and Head, 2008). In this work we deposit in Praxiteles crater. With the completion of the third provide a detailed documentation of what we interpret to be yby, it is possible now to identify, describe, and map pyroclastic pyroclastic deposits identied in the course of the three MESSEN- deposits on a global basis. GER ybys of Mercury and an analysis of their morphologies, physical properties, and distribution. This analysis is used to 2.1. Identication of pyroclastic deposits provide a framework for future study of pyroclastic deposits by MESSENGER after its insertion into orbit about Mercury We have identied candidate pyroclastic deposits globally on (scheduled for March 18, 2011) and by the later BepiColombo Mercury on the basis of spectral character, morphology, and mission, which is expected to begin data collection in 2020 surface texture inferred from a combination of WAC 11-band (Benkhoff et al., 2010), by identifying key areas for targeting, color image mosaics (with bands centered at 430, 480, 560, 630, outlining major science questions and objectives, and summariz- 700, 750, 830, 900, 950, 1000, and 1020 nm wavelength, and a ing techniques and analyses successfully used to study lunar resolution of 5 km/pixel), and NAC high-resolution image pyroclastic deposits. Given the recent and projected inux of data mosaics (centered at 750 nm, with a resolution of 500 m/pixel) from both Mercury and the Moon, the extremely rich opportunity (Hawkins et al., 2007; Robinson et al., 2008; Becker et al., 2009; to study pyroclastic deposits on the two bodies in concert is also Domingue et al., 2010). The image mosaics were calibrated to discussed. irradiance/solar ux (I/F), photometrically adjusted to the stan- dard bidirectional geometry of 301 solar incidence and 01 emis- sion angle (Robinson et al., 2008; Domingue et al., 2010), and 2. Characteristics of pyroclastic deposits on Mercury analyzed using ENVI, an image visualization software package. Representative spectra were selected from the pyroclastic deposit Possible pyroclastic deposits on Mercury were rst identied studied in detail by Kerber et al. (2009), here termed RS-03 (after in Mariner 10 data. Rava and Hapke (1987) identied, on the oor Red Spot 3, the designation given to the feature by Blewett et al., of the crater Lermontov, a candidate pyroclastic deposit charac- 2009a); Caloris interior plains material; bright crater-ll material; terized by diffuse borders, high reectance, and relatively red dark crater material; and plains materials exterior to Caloris. color (i.e., displaying a more steeply sloped reectance spectrum A linear spectral unmixing (end-members summing to unity from visible to near-infrared wavelengths). MESSENGER has using a weight of 4) was then performed to highlight the units provided additional evidence to support the inference that this that had spectra that are most similar to the spectrum of RS-03. deposit is indeed pyroclastic in nature. Several other diffuse, low- One representation of the result of the spectral unmixing is albedo, relatively blue deposits were suggested to be either shown in Fig. 1a and b, for which red (R) is pyroclastic material, ballistically emplaced ejecta deposits or pyroclastic deposits green (G) is bright crater-ll material, and blue (B) exterior Caloris (Robinson and Lucey, 1997). New, higher-resolution data and plains. The locations of the end-member spectra are indicated multi-band color information acquired by the Mercury Dual in Fig. 1a. Imaging System (MDIS) narrow-angle camera (NAC) and wide- Several other spectral unmixing procedures were performed angle camera (WAC) (Hawkins et al., 2007) have suggested that with other types of deposits chosen as the end-members. These several of these low-albedo, relatively blue deposits (northwest of results were compared to RGB color composite images obtained Lermontov crater and within Homer basin) are most likely of with several combinations of WAC bands (e.g., 430, 750, and impact origin (Blewett et al., 2009b). 1000 nm). The areas consistently identied as similar to RS-03 During the rst MESSENGER yby, ve additional candidate were targeted for further analysis. For regions where Mariner 10 pyroclastic deposits were identied, mostly immediately inside images had a resolution or viewing angle that was preferable to the southern rim of the 1550-km-diameter Caloris impact basin that provided by MESSENGER NAC data, these images provided (Head et al., 2008; Murchie et al., 2008; Robinson et al., 2008; supplementary information for morphological identication. Blewett et al., 2009a; Kerber et al., 2009). These deposits have Areas with high incidence angles or non-ideal viewing geometry diffuse boundaries and are generally bright and red with respect appear around the edges of the spectral classication composite Fig. 1. (a) Spectral classication results (resolution 5 km/pixel), based on the calibrated WAC color mosaic (Hawkins et al., 2007; Robinson et al., 2008; Becker et al., 2009; Domingue et al., 2010). Representative end-member spectra were selected from (1) RS-03 (the pyroclastic deposit studied by Kerber et al., 2009), (2) plains material interior to the Caloris basin, (3) bright crater-ll material, (4) dark crater material, and (5) plains material exterior to Caloris (end-member locations indicated). A linear spectral unmixing was performed to identify the units that had spectra most similar to that of RS-03. Three end-member abundance images are presented here as an RGB composite (R: pyroclastic material; G: bright crater-ll material; B: plains exterior to Caloris). (b) Locations of candidate pyroclastic deposits. The deposits are generally named after the crater in which they are found. In cases where there is more than one deposit in a crater, the location of the deposit in the crater is added at the end (e.g., Praxiteles NE is a deposit in the northeastern part of Praxiteles crater). In cases where the deposit is not within a crater but there was a named crater nearby, the deposit name indicates the named crater, with the direction of the deposit from the crater added at the beginning (e.g., NE Rachmaninoff is a deposit located to the northeast of Rachmaninoff basin). Deposits associated with an unnamed crater are designated as such and numbered. (c) Composite image showing the maximum incidence angles for areas imaged by MESSENGER. Areas in red were imaged at high Sun (low incidence angle), and areas in blue at low Sun (high incidence angle). Most of the pyroclastic deposits that were identied appear in the areas between these two extremes, because sufcient color and morphologic indicators were both present. Future searches should be directed toward the reddest and bluest areas shown here, as these are the areas where deposits were most likely to have been missed.

3 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1897 (Fig. 1a) as blue, purple, and pink. Although low-incidence-angle high incidence angle (blue on the gure) would be difcult to (high Sun) data are preferable for spectroscopic analysis, high- identify spectroscopically, whereas deposits located in the areas incidence-angle (low Sun) data are preferable for morphologic of low incidence angle (red on the gure) would be difcult to analysis and identication of potential volcanic vents. A map of identify morphologically. As can be seen in Fig. 1c, most of the the maximum incidence angles for images collected by MESSEN- pyroclastic deposits that were identied are located in the regions GER is shown in Fig. 1c. Pyroclastic deposits located in the areas of between these extremes.

4 1898 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 Specic criteria used to identify possible pyroclastic deposits data are not yet available. The Beckett crater deposit is an included the presence of an irregular central depression, an example of a feature with good lighting for morphological albedo anomaly with diffuse boundaries, and a distinct spectral analysis but poor lighting geometry for spectral analysis, whereas signature similar to that of the previously identied pyroclastic the Melville, Hemingway, and RS-04 deposits are examples of deposit RS-03. These criteria are similar to those used to identify features with good lighting geometry for spectral analysis but lunar pyroclastic deposits, though lunar deposits are also distin- poor lighting for morphological analysis (Fig. 2). The additional guished by their low albedo compared with surrounding terrain candidate pyroclastic deposits identied from images obtained (Pieters et al., 1974; Weitz et al., 1998; Gaddis et al., 2003), during the second and third ybys are generally located on the whereas mercurian pyroclastic deposits tend to have a higher oors of impact craters, though some are located along crater albedo than surrounding terrain. peak rings, and one (Melville) is located just outside the rim of a On the basis of these criteria, a total of 35 candidate pyroclas- crater (Table 1). tic deposits have been identied on Mercury, including 19 newly Note added in proof: Four additional pyroclastic deposits were documented deposits and multiple distinct deposits at some of identied after this manuscript was submitted. These four fea- the previously identied sites (Fig. 2, Table 1). Most of the tures, which bring the total number of candidate pyroclastic pyroclastic deposits so identied t all three criteria, though deposits to 39, are included in Figs. 1 and 2 and Table 1 (labeled several large deposits lack a discernible central vent, and several Penta and unnamed crater 5ac, after the areas in which they are deposits are located in the areas for which sufcient WAC color found) but are not included in the analyses that follow. Fig. 2. Candidate pyroclastic deposits identied from images obtained during the three MESSENGER ybys. Higher-resolution images were taken from a global mosaic of Mercury combining MDIS NAC data with images from Mariner 10 (Becker et al., 2009) and overlaid with color images from a global WAC mosaic (Domingue et al., 2010) (R: 1000 nm, G: 750 nm, B: 430 nm). Numbers correspond to the position of each deposit in Table 1 (with 1 the largest and 35 the smallest). The WAC data were overlaid on the NAC data using an Adobe Photoshop blending lter that highlights WAC color variations while preserving NAC contrast and texture, effectively reducing the appearance of WAC pixels. Most deposits have diffuse edges and a distinguishable color anomaly and are centered on an irregular depression. (a) Some large deposits such as Hemingway (3) and Hesiod a (4) do not have clear vents. Deposit RS-05 has several possible vent structures. RS-02 (located in the crater Navoi, 6) lies in an area with poor viewing geometry (Fig. 1c). (b) To Ngoc Van (10), Scarlatti (12), and Gibran (15) were identied as pit craters by Gillis-Davis et al. (2009). The To Ngoc Van deposit (10) lies in an area imaged at poor viewing geometry (Fig. 1). (c) Rachmaninoff SE (17), RS-04b and c (20 and 23), and RS-01 (located in the crater Moody, 25) lack discernible vents. RS-04 is located in an area of high illumination angle, making morphological features more difcult to discern (Fig. 1c). (d) Deposits Beckett (34), Glinka (29), and unnamed crater 4 (not measured) lie at the edges of the WAC image data in the areas of high illumination angle.

5 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1899 Fig. 2. (continued) 2.2. Pits and pit craters craters on the basis of their non-circular shape, rimless margins, lack of an ejecta deposit, and lack of association with secondary Irregular pits interpreted to be vents have been identied in just crater chains. Deposits without irregular central pits often have over half of the identied deposits. Pits are distinguished from impact a rough interior that may consist of several vents (designated

6 1900 Table 1 Candidate pyroclastic deposits on Mercury. No. Site Latitude Longitude Area Measured Adjusted Size Geologic Pit Pit Pit Spectral (east) (km2) radius (km) lunar radius (km) class setting length width morphology signature 1 NE Rachmaninoff 35.8 63.8 19,466 71 164 Vlarge Cratered highlands 31 16 Oval Yes 2 RS-05 24.3 179 10,414 54 125 Vlarge Edge of Caloris basin Uncertain Yes 3 Hemingway 17.6 2.7 8920 54 124 Vlarge Crater oor Uncertain Yes 4 Hesiod a 57.2 31.7 4950 30 70 Vlarge Crater edge Not identied Yes 5 RS-03 22.3 146.2 4063 24 56 Vlarge Edge of Caloris basin 17 12 Rounded arcuate oval Yes 6 RS-02 58.8 160.6 3951 23 53 Vlarge Edge of Caloris basin 12 5 Pinched oval(s) Yes 7 Lermontov NE 15.8 48.2 3806 33 76 Vlarge Crater oor 12 5 Highly irregular Yes L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 8 Raphael 21 74.4 3240 30 70 Vlarge Basin oor Not identied Yes 9 Lermontov SW 15.1 49.1 3174 31 72 Vlarge Crater oor 11 10.5 Highly irregular Yes 10 To Ngoc Van 52.6 111.8 2924 22 51 Vlarge Crater oor 31 13 Rounded arcuate oval Yes 11 Praxiteles NE 26.9 59.1 2594 26 60 Vlarge Crater oor 12 9 Quasi-circular Yes 12 Scarlatti 41.3 100.7 2490 22 51 Vlarge Crater ring 43 14 Arcuate pinched oval Yes 13 Hesiod c 53.2 30.9 2233 21 49 Vlarge Crater edge Not identied Yes 14 Hesiod b 55 30 2079 19 45 Vlarge Crater edge Not identied Yes 15 Gibran 35.8 111.3 1852 22 52 Vlarge Crater oor 32.2 28 Circular No 16 Geddes 27.2 29.5 1654 21 51 Vlarge Crater oor 25.8 17 Rounded oval Yes 17 Rachmaninoff SE 26.2 59.8 1389 20 50 Vlarge Basin oor Not identied Yes 18 Mistral SE 4.2 54.2 1245 19 46 Vlarge Crater oor Uncertain Yes 19 Praxiteles SW 26 60.3 1210 18 42 Vlarge Crater ring 7 5 Rounded oval Yes 20 RS-04b 16.7 156.9 1196 20 47 Vlarge Edge of Caloris basin Not identied Yes 21 RS-04a 14.1 159.2 32 19 43 Vlarge Edge of Caloris basin Not identied Yes 22 NE Derzhavin 48.4 33.7 1111 15 38 Vlarge Crater oor 17 6 Rounded oval Yes 23 RS-04c 14.2 162.1 1048 19 43 Vlarge Edge of Caloris basin Not identied Yes 24 Hesiod e 51.5 27.9 1021 14 32 Vlarge Crater edge Not identied Yes 25 RS-01 13.3 144.9 1008 18 43 Vlarge Crater oor Not identied Yes 26 Unnamed crater 1 22 67.5 921 17 40 Large Crater oor 27 8 Arcuate Yes 27 RS-04d 15 164 911 17 40 Large Edge of Caloris basin 7 4 Circular (crater-like) Yes 28 RS-03 SW 21.7 145.4 875 19 43 Large Edge of Caloris basin Uncertain Marginal 29 Glinka 14.9 112.4 846 16 38 Large Crater oor 25 8 Arcuate Yes 30 Melville 26.2 4.2 746 17 41 Large Crater edge Uncertain Yes 31 RS-04e 15.4 165.3 502 12 28 Large Edge of Caloris basin 7 6 Circular (crater-like) Yes 32 Hesiod d 52.2 28.6 453 9 20 Large Crater edge Not identied Yes 33 Mistral NW 5.4 55.2 421 11 35 Large Crater oor 17 7 Pinched oval Marginal 34 Beckett 40 111.2 408 10 22 Large Crater oor 38 9 Arcuate Yes 35 RS-03 SE 19.4 150.2 317 7 26 Medium Edge of Caloris basin 8 4 Pinched oval No Unnamed crater 4 0.5 161.9 Crater oor Yes Pentas 3.45 50.4 Crater ring Yes Unnamed crater 5a 52.1 139.4 Crater oor Yes Unnamed crater 5b 55.1 142.5 Crater oor Yes Unnamed crater 5c 56.2 143.8 Intercrater area Yes

7 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1901 uncertain in Table 1). Others have no clear evidence for a vent-like deposit Sulpicius Gallus (Lucchitta and Schmitt, 1974; Gaddis et al., structure (designated not identied in Table 1). The shapes of the 2003). The pits appear to be morphologically similar, although the pit pits range from oval to highly irregularly shaped to arcuate. Two pits, in Lermontov is approximately twice as large as the lunar pit. In centered in deposits RS-04d and e, are more circular and crater-like, general, while both lunar and mercurian pits have irregular shapes and could possibly be impact craters that merely expose underlying and sloping interiors, the mercurian pits tend to be wider and larger. pyroclastic material rather than true pyroclastic vents. The size of the Pit craters and the pit-oor craters that host them were deposits in these cases is comparable to that of ejecta deposits from dened and mapped by Gillis-Davis et al. (2009) on the basis of nearby craters of a similar size. The long axis of the pits ranges from images from the rst MESSENGER yby. These authors found 6 to 43 km, and the short axis of the pits ranges from 3 to 28 km evidence to support an endogenic hypothesis for pit-crater (Table 1). Fig. 3 compares the pit associated with the Lermontov NE formation. Specically, they argued that pit craters formed by pyroclastic deposit with that associated with the lunar pyroclastic either explosive volcanic activity from dikes in the near subsur- face or withdrawal of magma from a near-surface reservoir followed by subsequent reservoir roof collapse. Because of the seeming paucity of pyroclastic-like deposits surrounding the pit craters and the association of pit craters with nearby smooth plains, Gillis-Davis et al. (2009) favored the latter hypothesis. Given the morphological similarities between the pit craters and the irregular pits discussed above, we reassessed the possibility that these pits are related to pyroclastic activity. Of the seven pit craters identied by Gillis-Davis et al. (2009), three [located in unnamed crater 3 (now named Geddes), Scarlatti, and Glinka; see Figs. 1 and 2] show a color signature similar to that of pyroclastic material elsewhere on Mercury. One of the examples given by Gillis-Davis et al. (unnamed crater 1 in that work, now named To Ngoc Van) has a diffuse deposit with a bright albedo and a color anomaly that may be similar to pyroclastic material elsewhere on the planet, but the data quality at that location makes this interpretation inconclusive (see Figs. 1 and 2b). Three of the pit craters identied by Gillis-Davis et al. (2009) do not show spectral evidence for pyroclastic activity [located in Beckett, unnamed crater 4, and newly named Gibran (previously unnamed crater 2)]. However, these features are all located in the areas where either WAC viewing geometry or illumination angle were not ideal, so the lack of a pyroclastic signature need not preclude the possibility that they are pyroclastic in origin (Figs. 1, 2b and d). For example, the pit in Gibran crater is associated with a diffuse deposit with a high albedo, thus meeting two of the three criteria discussed above (Fig. 2b). We conclude that the majority of the pit craters identied by Gillis-Davis et al. (2009) meet the three criteria expected of pyroclastic deposits. We chose to include the remain- ing three pit craters in Table 1 on the basis of the morphological similarity of their pits to those associated with pyroclastic depos- its and the presence in some cases of a diffuse, high-albedo anomaly. For these deposits, measurements of their areal extents are less certain (Table 1). It is intended that their inclusion will motivate further data collection and analysis, allowing a more denitive conclusion to be made regarding the relation of pit craters to pyroclastic deposits. 2.3. Global distribution The sites of candidate pyroclastic deposits on Mercury are shown in map view in Fig. 4. Their locations on the oors of craters and along the edges of basins are similar to those seen on the Moon (Fig. 4), where approximately half of the deposits are found on crater oors and most of the remainder are located at the edge of mare deposits near large basins (Gaddis et al., 2003). However, the global distribution of the pyroclastic deposits discerned so far suggests that pyroclastic deposits on Mercury are more widely distributed than those on the Moon. Thus far, pyroclastic deposits have been identied in most areas of the planet where good coverage is available, with less certain detec- Fig. 3. Comparison of the central pits of the Lermontov NE deposit on Mercury tions in the areas where data coverage, phase angle, and viewing (NAC mosaic overlain by a WAC color composite) and the Sulpicius Gallus Formation on the Moon (Apollo 17 Hasselblad color photograph AS17-149- geometry make positive identication more difcult. Addition- 22880). Although the morphologies of the pits are similar, the pit in Lermontov ally, whereas the largest pyroclastic deposits on the Moon are is about twice as large (11 km in its long dimension) as the lunar pit (5.5 km). often located on the edges of mare-lled basins, large pyroclastic

8 1902 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 Fig. 4. Global distribution of candidate pyroclastic deposits on Mercury compared with those on the Moon (from Gaddis et al., 2003) using the same schematic representation. Deposits on each body are numbered in order of decreasing size (1 is the largest). The mercurian deposits are referenced by number in Table 1. The lunar deposits are referenced by number in Table 1 of Gaddis et al. (2003). Whereas lunar pyroclastic deposits tend to cluster near mare deposits, mercurian pyroclastic deposits are more widely distributed. There is a clustering of deposits at the edges of the Caloris basin, which is partially lled with volcanic material (Head et al., 2008; Murchie et al., 2008). The background image of the Moon (Gaddis et al., 2003) is the global 750-nm Clementine mosaic. The background image for Mercury is the 750-nm MESSENGER NAC mosaic from ybys 1, 2, and 3 (Becker et al., 2009). Some clustering in the Mercury distribution is likely to be a consequence of variations in viewing angle and incidence angle affecting the identication of pyroclastic features, as depicted in Fig. 1. deposits on Mercury occur in a variety of geographical locations was not obvious) to the edge of the deposit (Table 1). For this reason, and are often located on crater oors. Unlike many pyroclastic size rankings according to area are not always the same as size deposits on the Moon (Gaddis et al., 2003), pyroclastic deposits on rankings according to average radius, depending on the irregularity Mercury do not appear to be associated with oor-fractured of the deposit and the location of the features interpreted to be vents craters, though they can be associated with crater peak rings. (Table 1). Because the edges of pyroclastic deposits are commonly The distribution of pyroclastic deposits on Mercury may suggest diffuse, radius measurements are approximate and can be uncertain a more even distribution of heat-producing elements or interior by up to several kilometers. Deposit edges were determined using a volatiles, or a more uniform crustal thickness, each of which might combination of albedo variations in the 500 m/pixel NAC mono- allow magma to form and propagate to the surface with a more chrome image mosaic supplemented by 5 km/pixel WAC color widespread distribution than seen on the Moon. Alternatively, boundaries. Due to the lower resolution of WAC images, NAC albedo the different spatial distribution of basins may exert a control on variations were usually more precise, except for areas where light- the distribution of pyroclastic deposits on the Moon and Mercury. ing or viewing geometry was not ideal. Some of the large, broad, The apparently more widespread and more global distribution of diffuse deposits lacked a coherent shape or a visible vent. The plains on Mercury relative to the Moon (e.g., Head et al., 2008; measured radii are least accurate for these deposits, as they could be Denevi et al., 2009) may also be associated with this difference. composed of several overlapping deposits from different vents. Because of poor image quality in the vicinity of unnamed crater 4 2.4. Physical properties (Fig. 2d), this deposit was not measured. Radii calculated from the area (under the assumption that each deposit is circular in extent) The area and the average radius were measured for each result in radii equal to (when the vent is nearly circular) or larger candidate pyroclastic deposit. NAC and WAC image mosaics than (when the vent is highly irregular) radii directly measured (Becker et al., 2009) were placed onto an azimuthal-equidistant and averaged from the images. Features were classied as projection centered on the deposit being measured. Areas were very large (100149,000 km2), large (4011000 km2), med- calculated using ArcGIS, a geographic information system software ium (201400 km2), small (101200 km2), and very small package, by dening the edges of each feature. Radii were deter- (1100 km2) according to the size classication used by Gaddis mined by averaging 12 transects from the center of the candidate et al. (2003) for the Moon (shown in Fig. 4). The measured radii were vent (or the approximate center of the deposit in cases where a vent tabulated, and a graphical representation of these results is

9 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1903 Fig. 5. (a) Graphical representation of the measured radii of the 35 candidate mercurian pyroclastic deposits (red) compared with those of the 76 identied lunar pyroclastic deposits (blue, from Gaddis et al., 2003). The Moon has many more pyroclastic deposits in the small size range than are seen on Mercury. The comparative absence of these features on Mercury is likely an effect of the resolution and coverage of the current data, and it is expected that additional smaller pyroclastic deposits will be discovered during the orbital phase of the MESSENGER mission and the future BepiColombo mission to Mercury. (b) The equivalent comparative sorting but with the radii of pyroclastic deposits on Mercury scaled to lunar gravitational acceleration. Because the pyroclastic beads that make up pyroclastic deposits are propelled in near- ballistic trajectories from the vent, and their range is inversely proportional to gravitational acceleration, pyroclastic beads on Mercury and the Moon that were propelled with the same velocity of ejection would reach different ranges. Scaling the deposit radii in this way allows for direct comparison of the magma volatile contents implied by the deposit dimensions. compared with a similar plot for lunar deposits in Fig. 5a. Area lunar conditions by recalling that the range of a ballistically measurements of Gaddis et al. (2003) were used to calculate emplaced object is inversely proportional to gravitational accel- approximate radii (under the assumption of circular deposits) for eration g. According to the laws of projectile motion, the hor- this gure. izontal range, X, of a particle in a vacuum is Although the radii of mercurian pyroclastic deposits appear to X v0 t sin y, be broadly similar to those of the lunar pyroclastic deposits, they are not directly comparable. Two eruptions of identical energies where v0 is the initial eruption velocity from the vent; y is the on each of the bodies would result in deposits with markedly ejection angle (measured from the zenith), and different radii because of the effect of the different surface gravitational accelerations on the ballistic emplacement of the 2v0 cos y t , pyroclasts. The radii for the mercurian deposits can be scaled to g

10 1904 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 yielding (for an angle of 451, chosen to determine the minimum Mercury than on the Moon (Kerber et al., 2009). Pyroclasts energy needed to emplace the particles) deposited in this way would be more likely to crystallize as opposed to being quenched as glasses. 2v20 sin ycos y v2 sin 2y v2 1 X 0 0p : On the Moon, crystallized pyroclastic beads are darker than g g g g quenched glasses because of the crystallization of opaque miner- In this way we may scale a deposit radius measured on Mercury als such as ilmenite, which also tends to make the deposit to what it would be on the Moon for the same initial ejection velocity relatively blue (Pieters et al., 1974). Lunar pyroclastic deposits XMoon gMercury 3:7m=s2 that are relatively bright and red relative to other lunar pyroclas- ) XMoon 2:3XMercury : tic deposits appear to be lower in iron (Lucey et al., 1995). The XMercury gMoon 1:6m=s2 bright and red character of the pyroclastic deposits on Mercury From this scaling, it can be seen that the candidate pyroclastic may thus indicate that they are relatively low in iron compared deposits thus far identied on Mercury required markedly higher with surrounding terrain (which may already be low in iron). eruption velocities (and hence more energy) to achieve their Although pyroclastic deposits on Mercury may be more likely to observed radial extents than their lunar counterparts (Fig. 5b). have crystallized than their lunar counterparts, they apparently If mercurian deposit radii calculated from area measurements did not crystallize a large amount of opaque minerals. The lack are used instead of directly measured radii, this difference is of a 1000-nm ferrous iron band in any of the pyroclast spectra enhanced. supports the conclusion that the deposits are low in ferrous iron The surface area of Mercury is approximately two times and is consistent with the general paucity of iron in silicates at the greater than that of the Moon, suggesting that, all else being surface of Mercury (Blewett et al., 2002, 2009a; Warell and equal, two times as many pyroclastic deposits would be expected Blewett, 2004; Robinson et al., 2008; Denevi et al., 2009). An on Mercury. Only about half as many pyroclastic deposits have absence of opaque minerals could reect a paucity of titanium, if been identied on Mercury to date, however, as have been the most common opaque mineral was ilmenite. identied on the Moon. This difference may mean that pyroclastic Reectance spectra of mercurian pyroclastic deposits can be volcanism was less common on Mercury or that pyroclastic broadly compared with reectance spectra derived from ve- deposits were more commonly buried by other types of deposits band Clementine data for lunar pyroclasts (Gaddis et al., 2003) by on Mercury than on the Moon. However, many of the deposits removing the WAC bands that are not shared between the two documented on the Moon are small or very small deposits. No instruments (the 415 nm Clementine and 430 nm WAC bands small or very small deposits have yet been found on Mercury are left to illustrate the visible continuum; Fig. 6). The reectance (shown in Fig. 4, where the deposits are displayed with symbols values of the mercurian pyroclastic deposits fall within the corresponding to the same size classes). This absence of small general range reported for the lunar deposits (Fig. 6). Thus, pyroclastic deposits on Mercury could be an effect of resolution, whereas mercurian pyroclastic deposits appear brighter than and as the resolution of data for Mercury improves it is expected surrounding terrain, and lunar deposits appear dark compared that additional smaller pyroclastic deposits may be discovered. with their surroundings, this difference in contrasts appears to Milkovich et al. (2002) estimated from analysis of lunar volcanic have more to do with the relative albedos of the surrounding features at different resolutions that small volcanic domes required terrain than with the pyroclastic deposits themselves. Further both low Sun and resolutions between 100 and 500 m/pixel in order compositional analysis of pyroclastic deposits will be possible to be identiable in images. The current NAC mosaic has a spatial following the detailed photometric and scattered-light calibration resolution of approximately 500 m/pixel, and that for the of the WAC multi-spectral image data (e.g., Domingue et al., WAC mosaic is closer to 5 km/pixel (Becker et al., 2009). The 2010). Deposits of particular interest are those that appear to orbital phase of the MESSENGER mission will generate higher- resemble other pyroclastic deposits but are located in the areas resolution data, including NAC image mosaics with an average for which the spectral data are uncertain (e.g., NE Rachmaninoff, resolution of better than 250 m/pixel and targeted images with Rachmaninoff SE, Beckett, unnamed crater 4, and Gibran). resolutions of 25 m/pixel. WAC color data will be available with resolutions of 1.1 km/pixel, and targeted images will have resolu- tions of approximately 300 m/pixel, meaning that smaller volca- 3. Implications of deposit dimensions for interior volatile nic features on Mercury should be resolvable (Milkovich et al., contents 2002). The presence of volatile elements deep within the interior of 2.5. Color and albedo Mercury sufcient to drive pyroclastic eruptions has implications for the planets mode of formation and subsequent evolution (e.g., Multi-band spectral reectance data from the MDIS WAC Kerber et al., 2009). It had long been thought that Mercurys interior instrument obtained during the MESSENGER ybys allow for would be volatile-poor, since it is likely to have accreted in a hot broad correlations between spectrally similar units. However, part of the solar nebula (Wetherill, 1994). In addition, the large core- limited data taken at a variety of phase angles and viewing to-mantle ratio of Mercury has been hypothesized to be due to some geometries make detailed compositional analysis of individual type of later heating episode, either by the nebula (Cameron, 1985; spectra difcult, especially for smaller pyroclastic deposits and Fegley and Cameron, 1987) or by a giant impact (Wetherill, 1988; those located at high latitudes or along the limb of the planet. Benz et al., 1988, 2007). Both types of heating event would have Pyroclastic deposits on Mercury generally appear to be brighter further devolatilized the planets interior. However, the presence of and redder (that is, displaying higher reectance and a more pyroclastic volcanism on Mercury suggests that the devolatilization steeply inclined visible to infrared reectance slope) than average of the planet was not complete, or that the initial volatile budget of background terrain (Robinson et al., 2008; Blewett et al., 2009a; the planet was greater than previously hypothesized, perhaps due to Kerber et al., 2009). On Mercury, volcanic re fountains are the incorporation during accretion of planetesimals formed over a expected to have a greater optical density than re fountains on large range of solar distances, excursions of the semi-major axis of the Moon. Together with shorter ight times due to the greater Mercurys orbit in the early stages of accretion (Wetherill, 1988), or gravitational acceleration of Mercury, this effect would result in a bombardment by volatile-rich embryos from the outer solar system larger percentage of pyroclasts falling to the surface warm on (Morbidelli et al., 2000).

11 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1905 Fig. 6. Comparison of reectance spectra of pyroclastic deposits on Mercury and the Moon. The spectrum of each deposit represents an average of 5 individually selected pixels taken from outside each candidate vent. All MDIS spectra (collected during the rst and second Mercury ybys) were corrected for phase angle and calibrated to absolute reectance. The Hillier et al. (1999) correction was applied to the Clementine lunar spectra. Bands that were not shared between the two data sets were not included, except for the 415-nm band of Clementine and the 430-nm band of MDIS in order to illustrate the visible continuum. Although detailed spectral and albedo comparisons must await further collection and calibration of instrument data, it appears from current information that the albedos of mercurian pyroclastic deposits fall within the range of albedos for lunar pyroclastic deposits. The ferrous iron absorption near 1000 nm seen in the spectra of lunar pyroclastic deposits is absent in spectra of the deposits on Mercury, consistent with the overall paucity of ferrous iron in silicate minerals on Mercury compared with the Moon (Blewett et al., 2002, 2009a; Warell and Blewett, 2004; Robinson et al., 2008). The recent documentation of water and other volatiles in lunar to v20/g, as described above, and thus is also directly proportional pyroclastic glasses (Saal et al., 2008) and water in lunar minerals to the kinetic energy of the particle at the time of eruption, (McCubbin et al., 2010) suggest that lunar formation models (1/2)mv20, where m is particle mass. The kinetic energy at the time predicting near-complete devolatilization may bear revisiting. These of eruption is, to a good approximation, directly proportional to ndings need not imply, however, that water incorporated during the released magma gas fraction by mass, f (Wilson, 1980), so f/g accretion was the only, or even the major, volatile responsible for is proportional to X. Wilson and Head (1981) determined that explosive volcanism on the Moon. It has been suggested, for approximately 500 ppm of CO (equivalent to an eruption speed of instance, that lunar pyroclastic deposits could have been fueled by 90 m/s) would be needed to emplace pyroclasts to a distance of an oxidation reaction producing CO from elemental carbon through 5 km on the Moon. From the relationship above, it would take the reduction of Cr2O3, TiO2, or FeO (e.g., Sato, 1976; Fogel and 2.3 times the amount of a particular volatile species to emplace Rutherford, 1995; Nicholis and Rutherford, 2005). pyroclasts to the same distance on Mercury. Zolotov (2011) explored this process for Mercury and sug- In this way, we converted each measured deposit radius into gested that several other species (S, Cl, and N) could survive the equivalent proportion of volatiles needed to eject a pyroclastic devolatilization events because they are stable in their solid, particle to this distance on Mercury. The results, displayed in reduced forms. The inferred dry, reducing conditions of the Fig. 7 and Table 2, are shown in ppm CO because CO is a volatile mantle of Mercury would result in volatile species that are species that could be produced on Mercury under a variety of distinct from those commonly encountered on Earth, perhaps conditions, and the values can be easily compared with values including N2, CO, S2, CS2, S2Cl, Cl, Cl2, and COS, rather than H2O, discussed for the Moon, where CO is considered a likely volatile CO2, SO2, H2S, and HCl (Zolotov, 2011). The dominant volatiles (Nicholis and Rutherford, 2005). It is more likely in practice that expected to drive pyroclastic eruptions would depend on the an eruption on Mercury would be driven by a combination of initial composition of the accreted planet (e.g., elemental carbon volatile species. The required amount of different volatile species may not be as abundant in planetesimals that formed in the inner can be readily calculated, as the energy available from the parts of the solar nebula), the redox state of the mantle, and the expansion of a gas is inversely proportional to its molecular temperature and pressure conditions encountered during the rise weight. These calculations hold for any type of explosive eruption, of the erupting magma (which affect how the volatile compo- as they address only the amount of energy imparted to the nents partition into the gas phase) (Zolotov, 2011). entrained pyroclasts as they exit the vent and are not dependent Analysis following the rst MESSENGER yby (Kerber et al., on the manner or timing of degassing. However, depending on the 2009) indicated that in order to produce a deposit the size of RS-03 type of eruption, volatiles could have passively degassed from the (the fth largest in Table 1; Fig. 2a), approximately 5550 parts per magma, causing a depletion in volatiles needed for an explosive million (ppm) of CO (or an equivalent amount of another volatile) eruption, or they could have become concentrated during the dike would be required. This calculation can be made because the propagation and eruption processes, yielding calculated abun- horizontal range, X, of any ballistic particle is directly proportional dances that are greater than their original abundances in the melt.

12 1906 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 Fig. 7. Magmatic content of CO needed to propel pyroclasts to the measured radius for each deposit, under the assumption that CO was the only driving volatile. Results for other volatiles are shown in Table 2. Hemingway and RS-05 both lack a discernible vent, meaning that the calculations for these deposits may yield an erroneously large result. NE Rachmaninoff has an irregular central depression and fairly well-dened edges, but the deposits large radius would require an extremely high volatile content. This result suggests that the NE Rachmaninoff deposits may not be completely composed of pyroclastic material, that there are additional vents within the area of the deposit, or that volatiles were concentrated below the surface prior to eruption (as in vulcanian eruptions); images at higher resolution will allow these possibilities to be evaluated. It is possible, especially in the case of vulcanian eruptions, to interior volatiles or solid phases of C, S, N, or Cl present in the two concentrate a volatile-rich magmatic foam at the tip of a propa- bodies when they formed. gating dike or below a plug blocking the vent mouth. In such a situation, magma containing a moderate amount of volatiles can eventually lead to an energetic eruption composed of mostly gas 4. Future analyses and fragmented foam (Wilson, 1980; Wilson and Head, 1981; Fagents and Wilson, 1993). Mercury and the Moon are similar in that they are both small, The majority of the newly recognized pyroclastic deposits on airless bodies with generally ancient silicate surfaces. Both bodies Mercury are smaller in size than the original deposit analyzed by have similar dominant surface processes: impact cratering, vol- Kerber et al. (2009). Of the four deposits that are larger than canism, structural deformation, and space weathering (e.g., RS-03 in areal extent, only one (NE Rachmaninoff; Fig. 1a) has Hiesinger and Head, 2006). However, Mercury differs from the well-dened edges and a prominent central irregular depression. Moon in its surface area ( 2 times that of the Moon), the radius However, with a radius almost three times that of RS-03, the of its core (at least 5 times that of the Moon), the composition proportion of volatiles needed to emplace pyroclasts to that of its crust (silicates with very low FeO content), and the distance is almost three times as great. For comparison, measure- distribution and expression of its volcanic output. ments made from eruptions of Kilauea volcano in Hawaii implied The Moon, because it is closer to Earth, better studied, and volatile abundances in the hotspot mantle source of 3000 ppm sampled, provides an excellent framework for an increased H2O, 6500 ppm CO2, and 1300 ppm S (Gerlach, 1986). If the understanding of Mercury. Conversely, Mercury, lacking the NE Rachmaninoff deposit was formed through a re-fountaining, complicating factors of the Moons proximity to (and likely origin hawaiian-like explosive eruption, where calculated volatile con- from) the Earth (e.g., Hartmann and Davis, 1975; Cameron and tents would be similar to those found in the mantle source, the Ward, 1976; Benz et al., 1986), provides a good context through deposit dimensions would imply volatile contents in the source which we may better understand the Moon. New data currently region of up to 11,000 ppm H2O, 26,000 ppm CO2, 13,000 ppm being acquired by missions to the Moon will provide a wealth of SO2, or a combination of these or other volatiles (see Zolotov, information about that body. The concurrent exploration of 2011). If the volcanic gas was created through the oxidation of Mercury will provide a rich context that will lead to a substantial carbon, nitrogen, or similar species, a somewhat oxidizing crust and synergistic increase in understanding of both bodies. would be required in order to supply oxygen for the process The MESSENGER mission and the upcoming BepiColombo (Zolotov, 2011). It is possible that these high volatile abundances mission provide an opportunity to begin to understand Mercurys could be achieved through concentration of gas in a vulcanian surface processes in ways that were rst possible on the Moon eruption, as discussed above. The NE Rachmaninoff deposit almost a half-century ago. Several techniques may be used to currently lies at the edge of the usable MESSENGER WAC color better understand pyroclastic deposits, each contributing to pro- data and will be an important target during the orbital phase of gress on the outstanding issues discussed above. First, MESSEN- the mission. GER NAC images have made it possible to view large portions of The presence of large pyroclastic deposits on Mercury and the the planet in much greater detail than was previously available. implied amount of volatile species needed to create them sug- This improved resolution allows for the careful inventory and gests that Mercury may be more volatile-rich than previously characterization of the morphologies of possible pyroclastic thought. The observation that there appears to be a greater deposits, including their shapes and dimensions and character- proportion of large pyroclastic deposits on Mercury than there istics of their central vents (e.g., Figs. 2 and 5, Table 1). On the is on the Moon may yield clues to the relative abundances of Moon, morphological studies have allowed the classication of

13 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1907 Table 2 Table 3 Magmatic abundances, in ppm, of selected volatiles required to emplace pyro- Anticipated data sets that will further the understanding of pyroclastic deposits on clastic material to the maximum radial extent of each deposit. Abundances for Mercury. some of the volatiles suggested as likely by Zolotov (2011) can be expressed in a similar way by scaling the calculated CO value by the ratio of the molecular Instruments Mission Possible insights weight of each volatile to that of CO, as follows: N2 CO (identical molecular weights), S2 2.3 (CO), CS 2.7 (CO), SCl 3.6 (CO), Cl 1.3 (CO), Cl 2.5 (CO), MDIS NAC MESSENGER Morphology, albedo, geological context COS 2.1 (CO). Most of these abundances would therefore be high, near values for MDIS WAC MESSENGER Fe, Ti, and opaque mineral SO2 at 2.4 (CO). characterization, estimates of crystallization extent Site CO H2O CO2 SO2 H2S MASCS MESSENGER Mineral identication SIMBIO-SYS BepiColombo Morphology, albedo Rachmaninoff NE 16,000 11,000 26,000 37,000 21,000 MERTIS BepiColombo Mineral identication, thermal properties RS-05 12,000 8000 20,000 29,000 16,000 XRS MESSENGER Al/Si ratio, Al/Mg ratio, Fe and Ca Hemingway 12,000 8000 19,000 28,000 16,000 detections during solar ares Lermontov NE 7600 4900 12,000 17,000 9800 MIXS BepiColombo Al/Si ratio, Al/Mg ratio Lermontov SE 7200 4700 11,000 17,000 9300 Raphael 7000 4500 11,000 16,000 9000 Hesiod 7000 4500 11,000 16,000 8900 Ages are not currently available for the pyroclastic deposits on Praxiteles NE 6000 3800 9400 14,000 7700 RS-03 5600 3600 8700 13,000 7000 Mercury. Whereas samples of lunar pyroclastic deposits have RS-02 5300 3400 8400 12,000 6800 been returned to the Earth where they can be dated with To Ngoc Van 5100 3300 8000 12,000 6600 laboratory techniques, mercurian deposits must be dated using Gibran 5100 3300 8000 12,000 6600 a combination of crater-retention age dating (better for larger, Scarlatti 5100 3300 8000 12,000 6500 thicker, and more coherent deposits), spectral maturity, and Geddes 5000 3200 7800 11,000 6400 Hesiod c 4900 3100 7700 11,000 6300 stratigraphic relationships. In the areas where Caloris impact- Rachmaninoff SW 4700 3000 7300 11,000 6000 related features are visible, mapping conducted after the Mariner RS-04b 4600 3000 7200 10,000 5900 10 ybys placed many of the craters hosting pyroclastic deposits Hesiod b 4500 2900 7000 10,000 5700 at ages concurrent with or directly after the Caloris impact event Mistral SE 4300 2800 6800 9900 5600 RS-04a 4300 2800 6800 9900 5500 (e.g., Guest and Greeley, 1983; Spudis and Prosser, 1984). Further RS-04c 4300 2800 6700 9800 5500 dating of associated host craters and nearby volcanic units during RS-03 SW 4300 2800 6700 9800 5500 the orbital phase of MESSENGER will help constrain the ages of RS-01 4200 2700 6600 9600 5400 the deposits to the several recognized periods in Mercurys Praxiteles SW 4100 2700 6500 9400 5300 history. Melville 4000 2600 6300 9200 5200 Unnamed crater 1 4000 2600 6300 9200 5100 X-ray uorescence instruments own on the Apollo 15 and 16 RS-04d 3800 2400 6000 8800 4900 missions documented the correlation between lunar dark man- Glinka 3800 2400 6000 8700 4900 tling deposits and very high Mg/Al ratios (Schonfeld and Bielefeld, NE Derzhavin 3500 2200 5400 7900 4400 1978). The later European Space Agency mission Small Missions Hesiod e 3200 2100 5000 7400 4200 Mistral NW 2600 1700 4100 6000 3300 for Advanced Research in Technology (SMART-1) continued map- Beckett 2200 1400 3500 5000 2800 ping the Moon in X-rays, specically targeting lunar pyroclastic Hesiod d 2000 1300 3200 4700 2600 deposits as a possible source of oxygen for future industrial use RS-03 SE 1600 1000 2400 3600 2000 (Dunkin et al., 2003). The MESSENGER X-Ray Spectrometer (XRS) instrument will be able to search for correlations between high Mg/Al ratios and the mercurian pyroclastic deposits (Schlemm features into different volcanological regimes (Wilson and Head, et al., 2007). During solar ares, the XRS will also be able to 1981) that depend on the availability of magma, the abundance of measure the distribution of Fe, Ti, and Ca (Schlemm et al., 2007). volatiles, and the stress state of the lithosphere (Head and Wilson, These XRS measurements, together with those taken by MASCS 1992). Similar analyses have begun on Mercury (Wilson and and MDIS, will serve to better characterize the petrological Head, 2008) and will be greatly facilitated by the high-resolution evolution of Mercurys source magmas, especially in comparison image data to be collected by MESSENGER and BepiColombo. with those of the Moon (Dunkin et al., 2003). The best XRS In addition to morphological analysis, MESSENGERs multi-spec- resolution, over Mercurys northern hemisphere ( 40 km per tral MDIS WAC and hyper-spectral Mercury Atmospheric and Sur- pixel), will be sufciently high to resolve the larger pyroclastic face Composition Spectrometer (MASCS) instrument (McClintock deposits (Schlemm et al., 2007). BepiColombos Mercury Imaging and Lankton, 2007) make compositional analysis of pyroclastic X-ray Spectrometer (MIXS) will gather X-ray data complementary deposits possible. The BepiColombo Mercury Planetary Orbiter to those of MESSENGER given that MPO will y in a different orbit (MPO) will be equipped with the Spectrometers and Imagers for from that of MESSENGER and during a time of different levels of MPO BepiColombo Integrated Observatory System (SIMBIO-SYS) solar X-ray activity (Fraser et al., 2010). A summary of anticipated and the Mercury Thermal Infrared Spectrometer (MERTIS) instru- data resources is shown in Table 3. ments that will view the planet in visible, infrared, and thermal wavelengths (Flamini et al., 2010; Hiesinger et al., 2010). For the Moon, the Apollo samples have enabled in-depth characterization of 5. Discussion and conclusions pyroclastic bead composition and distribution, which, combined with Clementine multi-spectral imagery, has allowed for the classi- The discovery of pyroclastic deposits on the surface of Mercury cation of pyroclastic deposits by bead color, iron content, and (Rava and Hapke, 1987; Head et al., 2008, 2009; Murchie et al., titanium content (Delano, 1986; Gaddis et al., 2003). Newly acquired 2008; Robinson et al., 2008; Blewett et al., 2009a; Kerber et al., spectral and imaging data (e.g., from Chandrayaan 1, Kaguya, and 2009) yielded several important insights. Pyroclastic deposits on Lunar Reconnaissance Orbiter instruments) will enable even more Mercury were found to be similar to those on the Moon in that detailed classication of pyroclastic deposits in order to determine both are mantling deposits with diffuse edges often distributed the key parameters that lead to the differences between them, around an approximately central, irregularly shaped depression including source magma and style of volcanologic emplacement. (Fig. 3).

14 1908 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 Although the particular mercurian deposit (RS-03) analyzed by imaging, be separable into several deposits erupted from different Kerber et al. (2009) appears bright compared with the planetary vents; and that subtle morphologic and compositional differences average, in contrast to lunar pyroclastic deposits that appear dark among pyroclastic deposits on Mercury will provide information compared with the lunar average, the RS-03 deposit has reec- about their inferred eruption mechanisms, source regions, ages, and tance values in the general range of pyroclastic deposits on the subsequent space weathering or impact gardening that has altered Moon. It was determined by Kerber et al. (2009) that the them from their pristine states. Validation of each of these predic- magmatic volatile abundances needed to form RS-03 were com- tions, while informative in itself, will provide important contextual parable to those found in terrestrial oceanic basaltic magmas. information on a variety of broader questions on Mercurys interior The identication and characterization of pyroclastic deposits structure, composition (including volatile budget), stress state, and across the surface of Mercury in this paper has yielded further thermal history. Equally important will be the synergy created insight. First, it was found that there are many deposits similar to through analysis and comparison of new data being gathered for RS-03, including four deposits that are larger in areal extent and pyroclastic deposits on the Moon. eight deposits that are larger in average radius than RS-03. Three of these deposits have more than twice the radius of RS-03 (Table 1, Fig. 5). Pyroclastic deposits on Mercury are found to be Acknowledgments systematically larger (after accounting for the difference in gravitational accelerations between bodies) than those found on The authors are grateful for helpful suggestions from two the Moon (Fig. 5), with implied volatile contents in some cases anonymous reviewers. The contribution of D.T.B. is made possible many times greater than those found in terrestrial oceanic basalts by NASAs MESSENGER Participating Scientist program. The (Table 2, Fig. 7). These volatile species must be both present in the MESSENGER project is supported by the NASA Discovery program melt and subject to appropriate conditions to allow them to under contracts NASW-00002 to the Carnegie Institution of partition into the gas phase. Washington and NAS5-97271 to the Johns Hopkins University Second, pyroclastic deposits on Mercury are widely distributed Applied Physics Laboratory. The incidence-angle map in Fig. 1c across the surface and are found mostly on the oors of craters was kindly provided by Caleb I. Fassett. We also thank L. M. Garcia (Fig. 4). Whereas lunar pyroclastic deposits are often found for helpful discussions. associated with fractures in crater oors, mercurian pyroclastic deposits are often associated with crater central peaks and peak- References ring structures (Fig. 2, Table 1). Many lunar deposits are asso- ciated with mare-lled basins (Gaddis et al., 2003), and several Becker, K.J., Robinson, M.S., Becker, T.L., Weller, L.A., Turner, S., Nguyen, L., Selby, mercurian pyroclastic deposits are seen at the edge of the C., Denevi, B.W., Murchie, S.L., McNutt, R.L., Solomon S.C., 2009. Near global volcanic plains (Murchie et al., 2008) interior to the Caloris basin. mosaic of Mercury, Eos Trans. AGU 90 (Fall Meet. Suppl., abstract P21A-1189). The widespread distribution of pyroclastic deposits on Mercury Benkhoff, J., van Casteren, J., Hayakawa, H., Fujimoto, M., Laakso, H., Novara, M., Ferri, P., Middleton, H.R., Ziethe, R., 2010. BepiColombocomprehensive may indicate a more uniform distribution of heat-producing exploration of Mercury: mission overview and science goals. Planet. Space elements or interior volatiles or a more uniform crustal thickness Sci. 58, 220. on Mercury than on the Moon. Benz, W., Slattery, W.L., Cameron, A.G.W., 1986. The origin of the Moon and the single-impact hypothesis. Icarus 66, 515535. Other outcomes of this study include the nding that pyroclastic Benz, W., Slattery, W.L., Cameron, A.G.W., 1988. Collisional stripping of Mercurys deposits on Mercury generally have larger vents than deposits on mantle. Icarus 74, 516528. the Moon, perhaps reecting a higher explosivity (Fig. 2, Table 1). Benz, W., Anic, A., Horner, J., Whitby, J.A., 2007. The origin of Mercury. Space Sci. Rev. 132, 189202. Mercurian pyroclastic deposits lack the ubiquitous 1000-nm ferrous Blewett, D.T., Hawke, B.R., Lucey, P.G., 2002. Lunar pure anorthosite as a spectral iron band seen in lunar pyroclastic deposits, indicating that pyro- analog for Mercury. Meteorit. Planet. Sci. 37, 12451254. clastic deposits, like other surface materials on Mercury, have little Blewett, D.T., Robinson, M.S., Denevi, B.W., Gillis-Davis, J.J., Head, J.W., Solomon, S.C., Holsclaw, G.M., McClintock, W.E., 2009a. Multispectral images of Mercury ferrous iron in silicate phases (Fig. 6) (Blewett et al., 2002, 2009a; from the rst MESSENGER yby: analysis of global and regional color trends. Warell and Blewett, 2004; Robinson et al., 2008; Kerber et al., 2009). Earth Planet. Sci. Lett. 285, 263271. It was also found that mercurian pyroclastic deposits exhibit a range Blewett, D.T., Kerber, L., Head, J.W., Denevi, B.W., Robinson, M.S., Murchie, S.L., Gillis-Davis, J.J., Solomon, S.C., 2009b. Mercury pyroclastics: color, morphol- of albedos, but all values appear to be within the range of albedos ogy, and volatile content. Lunar Planet. Sci. Conf. 40 (abstract 1793). for lunar pyroclastic deposits (Fig. 6). Boynton, W.V., Sprague, A.L., Solomon, S.C., Starr, R.D., Evans, L.G., Feldman, W.C., Further study of Mercurys pyroclastic deposits will benet Trombka, J.I., Rhodes, E.A., 2007. MESSENGER and the chemistry of Mercurys from a variety of anticipated data sets, as indicated in Table 3, surface. Space Sci. Rev. 131, 85104. Cameron, A.G.W., 1985. The partial volatilization of Mercury. Icarus 64, 285294. including morphological analysis, dating, and correlation Cameron, A.G.W., Ward, W.R., 1976. The origin of the Moon. Lunar Sci. 7, 120122. with other volcanological features from images obtained with Delano, J.W., 1986. Pristine lunar glasses: criteria, data, and implications. Proc. MESSENGERs MDIS NAC (in conjunction with existing Mariner 10 Lunar Planet. Sci. Conf. 16, J. Geophys. Res. 91, D201D213. Denevi, B.W., Robinson, M.S., Solomon, S.C., Murchie, S.L., Blewett, D.T., Domingue, images and future BepiColombo images); compositional analysis D.L., McCoy, T.J., Ernst, C.M., Head, J.W., Watters, T.R., Chabot, N.L., 2009. The provided by MESSENGERs MDIS WAC, MASCS, and XRS; further evolution of Mercurys crust: a global perspective from MESSENGER. Science modeling of the ascent and eruption of magma-lled dikes (e.g., 324, 613618. Domingue, D.L., Vilas, F., Holsclaw, G.M., Warell, J., Izenberg, N.R., Murchie, S.L., Wilson and Head, 2008), combining the stress elds suggested by Denevi, B.W., Blewett, D.T., McClintock, W.E., Anderson, B.J., Sarantos, M., tectonic features such as wrinkle ridges and lobate scarps (e.g., 2010. Whole-disk spectrophotometric properties of Mercury: synthesis of Watters et al., 2009) with the eruption energies calculated here; MESSENGER and ground-based observations. Icarus 209, 101124. Dunkin, S.K., Grande, M., Casanova, I., Fernandes, V., Heather, D.J., Kellett, B., and theoretical comparisons with recent lunar sample laboratory Muinonen, K., Russell, S.S., Browning, R., Waltham, N., Parker, D., Kent, B., ndings (Saal et al., 2008), including the volatile contents and Perry, C.H., Swinyard, B., Perry, A., Feraday, J., Howe, C., Phillips, K., McBride, G., types of coatings that would be predicted for the particles Huovelin, J., Muhli, P., Hakala, P.J., Vilhu, O., Thomas, N., Hughes, D., Alleyne, H., Grady, M., Lundin, R., Barabash, S., Baker, D., Clark, P.E., Murray, C.D., Guest, J., forming pyroclastic deposits on Mercury. dUston, L.C., Maurice, S., Foing, B., Christou, A., Owen, C., Charles, P., Predictions that can be made from the analysis of data collected Laukkanen, J., Koskinen, H., Kato, M., Sipila, K., Nenonen, S., Holmstrom, M., so far by MESSENGER include the likelihood that many small Bhandari, N., Elphic, R., Lawrence, D., 2003. Scientic rationale for the D-CIXS pyroclastic deposits, currently below the limit of detection, will be X-ray spectrometer on board ESAs SMART-1 mission to the Moon. Planet. Space Sci. 51, 435442. discovered when higher-resolution images are available; the possi- Fagents, S.A., Wilson, L., 1993. Explosive volcanic eruptionsVII. The ranges of bility that larger, regional deposits will, with higher-resolution pyroclasts ejected in transient volcanic explosions. Geophys. J. Int. 113, 359370.

15 L. Kerber et al. / Planetary and Space Science 59 (2011) 18951909 1909 Fegley, B., Cameron, A.G.W., 1987. A vaporization model for Fe/Si fractionation in McCubbin, F.M., Steele, A., Hauri, E.H., Nekvasil, H., Yamashita, S., Hemley, R.J., the Mercury protoplanet. Earth Planet. Sci. Lett. 82, 207222. 2010. Nominally hydrous magmatism on the Moon. Proc. Natl. Acad. Sci. 107, Flamini, E., Capaccioni, F., Colangeli, L., Cremonese, G., Doressoundiram, A., Josset, 1122311228. J.L., Langevin, Y., Debei, S., Capria, M.T., De Sanctis, M.C., Marinangeli, L., Milkovich, S.M., Head, J.W., Wilson, L., 2002. Identication of mercurian volca- Massironi, M., Mazzotta Epifani, E., Naletto, G., Palumbo, P., Eng, P., Roig, J.F., nism: resolution effects and implications for MESSENGER. Meteorit. Planet. Caporali, A., Da Deppo, V., Erard, S., Federico, C., Forni, O., Sgavetti, M., Sci. 37, 12091222. Filacchione, G., Giacomini, L., Marra, G., Martellato, E., Zusi, M., Cosi, M., Morbidelli, A., Chambers, J., Lunine, J.I., Petit, J.M., Robert, F., Valsecchi, G.B., Cyr, Bettanini, C., Calamai, L., Zaccariotto, M., Tommasi, L., Dami, M., Ficai Veltroni, K.E., 2000. Source regions and timescales for the delivery of water to the Earth. J., Poulet, F., Hello, Y., The SIMBIO-SYS Team, 2010. SIMBIO-SYS: the spectro- Meteorit. Planet. Sci. 35, 13091320. meter and imagers integrated observatory system for the BepiColombo Murchie, S.L., Watters, T.R., Robinson, M.S., Head, J.W., Strom, R.G., Chapman, C.R., planetary orbiter. Planet. Space Sci. 58, 125143. Solomon, S.C., McClintock, W.E., Prockter, L.M., Domingue, D.L., Blewett, D.T., Fogel, R.A., Rutherford, M.J., 1995. Magmatic volatiles in primitive lunar glasses: I. 2008. Geology of the Caloris basin, Mercury: a view from MESSENGER. Science FTIR and EPMA analyses of Apollo 15 green and yellow glasses and revision of 321, 7376. Nicholis, M.G., Rutherford, M.J., 2005. Pressure dependence of graphiteCO phase the volatile assisted re-fountain theory. Geochim. Cosmochim. Acta 59, equilibria and its role in lunar volcanism. Lunar Planet. Sci. 36 (abstract 1726). 201215. Pieters, C.M., McCord, T.B., Charette, M.P., Adams, J.B., 1974. Lunar surface: Fraser, G.W., Carpenter, J.D., Rothery, D.A., Pearson, J.F., Martindale, A., Huovelin, J., identication of the dark mantling material in the Apollo 17 soil samples. Treis, J., Anand, M., Anttila, M., Ashcroft, M., Benkoff, J., Bland, P., Bowyerg, A., Science 183, 11911194. Bradley, A., Bridges, J., Brown, C., Bulloch, C., Bunce, E.J., Christensen, U., Evans, Rava, B., Hapke, B., 1987. An analysis of the Mariner 10 color ratio map of Mercury. M., Fairbend, R., Feasey, M., Giannini, F., Hermann, S., Hesse, M., Hilchenbach, Icarus 71, 397429. M., Jordenm, T., Joy, K., Kaipiainen, M., Kitchingman, I., Lechner, P., Lutz, G., Robinson, M.S., Lucey, P.G., 1997. Recalibrated Mariner 10 color mosaics: implica- Malkki, A., Muinonen, K., Naranen, J., Portin, P., Prydderch, M., San Juan, J., tions for Mercurian volcanism. Science 275, 197200. Sclater, E., Schyns, E., Stevenson, T.J., Struder, L., Syrjasuo, M., Talboys, D., Robinson, M.S., Murchie, S.L., Blewett, D.T., Domingue, D.L., Hawkins III, S.E., Head, Thomas, P., Whitford, C., Whitehead, S., 2010. The Mercury imaging X-ray J.W., Holsclaw, G.M., McClintock, W.E., McCoy, T.J., McNutt Jr., R.L., Prockter, spectrometer (MIXS) on BepiColombo. Planet. Space Sci. 58, 7995. L.M., Solomon, S.C., Watters, T.R., 2008. Reectance and color variations on Gaddis, L.R., Staid, M.I., Tyburczy, J.A., Hawke, B.R., Petro, N.E., 2003. Compositional Mercury: regolith processes and compositional heterogeneity. Science 321, analyses of lunar pyroclastic deposits. Icarus 161, 262280. 6669. Gerlach, T.M., 1986. Exsolution of H2O, CO2, and S during eruptive episodes at Saal, A.E., Hauri, E.H., Cascio, M.L., Van Orman, J., Rutherford, M.C., Cooper, R.F., Kilauea Volcano, Hawaii. J. Geophys. Res. 91, 1217712185. 2008. The volatile content of the lunar volcanic glasses: evidence for the Gillis-Davis, J.J., Blewett, D.T., Gaskell, R.W., Denevi, B.W., Robinson, M.S., Strom, presence of water in the Moons interior. Nature 454, 192195. R.G., Solomon, S.C., Sprague, A.L., 2009. Pit-oor craters on Mercury: evidence Sato, M., 1976. Oxygen fugacity and other thermochemical parameters of Apollo of near-surface igneous activity. Earth Planet. Sci. Lett. 285, 243250. 17 high-Ti basalts and their implications on the reduction mechanism. Proc. Guest, J.E., Greeley, R., 1983. Geologic Map of the Shakespeare Quadrangle Lunar Sci. Conf. 7, 13231344. of Mercury. Map I-1408, Misc. Investigations Ser., U.S. Geological Survey, Schlemm II, C.E., Starr, R.D., Ho, G.C., Bechtold, K.E., Hamilton, S.A., Boldt, J.D., Denver, CO. Boynton, W.V., Bradley, W., Fraeman, M.E., Gold, R.E., Goldsten, J.O., Hayes, J.R., Hartmann, W.K., Davis, D.R., 1975. Satellite-sized planetesimals and lunar origin. Jaskulek, S.E., Rossano, E., Rumph, R.A., Schaefer, E.D., Strohbehn, K., Shelton, Icarus 24, 504515. R.G., Thompson, R.E., Trombka, J.I., Williams, B.D., 2007. The X-Ray Spectro- Hawkins III, S.E., Boldt, J.D., Darlington, E.H., Espiritu, R., Gold, R.E., Gotwols, B., meter on the MESSENGER spacecraft. Space Sci. Rev. 131, 393415. Grey, M.P., Hash, C.D., Hayes, J.R., Jaskulek, S.E., Kardian Jr., C.J., Keller, M.R., Schonfeld, E., Bielefeld, M.J., 1978. Correlation of dark mantle deposits with high Malaret, E.R., Murchie, S.L., Murphy, P.K., Peacock, K., Prockter, L.M., Reiter, Mg/Al ratios. Proc. Lunar Planet. Sci. Conf. 9, 30373048. R.A., Robinson, M.S., Schaefer, E.D., Shelton, R.G., Sterner II, R.E., Taylor, H.W., Solomon, S.C., Chaiken, J., 1976. Thermal expansion and thermal stress in the Watters, T.R., Williams, B.D., 2007. The Mercury Dual Imaging System on the Moon and terrestrial planets: clues to early thermal history. Proc. Lunar Sci. MESSENGER spacecraft. Space Sci. Rev. 131, 247338. Conf. 7, 32293243. Head, J.W., Wilson, L., 1992. Lunar mare volcanism: stratigraphy, eruption Solomon, S.C., McNutt Jr., R.L., Watters, T.R., Lawrence, D.J., Feldman, W.C., Head, conditions, and the evolution of secondary crusts. Geochim. Cosmochim. Acta J.W., Krimigis, S.M., Murchie, S.L., Phillips, R.J., Slavin, J.A., Zuber, M.T., 2008. 56, 21552175. Return to Mercury: a global perspective on MESSENGERs rst Mercury yby. Head, J.W., Murchie, S.L., Prockter, L.M., Robinson, M.S., Solomon, S.C., Strom, R.G., Science 321, 5962. Chapman, C.R., Watters, T.R., McClintock, W.E., Blewett, D.T., Gillis-Davis, J.J., Spudis, P.D., Prosser, J.G., 1984. Geologic Map of the Michelangelo Quadrangle of 2008. Volcanism on Mercury: evidence from the rst MESSENGER yby. Mercury. Map I-1659, Misc. Investigations Ser., U.S. Geological Survey, Denver, Science 321, 6972. CO. Head, J.W., Murchie, S.L., Prockter, L.M., Solomon, S.C., Chapman, C.R., Strom, R.G., Warell, J., Blewett, D.T., 2004. Properties of the Hermean regolith: V. New optical reectance spectra, comparison with lunar anorthosites, and mineralogical Watters, T.R., Blewett, D.T., Gillis-Davis, J.J., Fassett, C.I., Dickson, J.L., Morgan, modeling. Icarus 168, 257276. G.A., Kerber, L., 2009. Volcanism on Mercury: evidence from the rst Watters, T.R., Solomon, S.C., Robinson, M.S., Head, J.W., Andre, S.L., Hauck, S.A., MESSENGER yby for extrusive and explosive activity and the volcanic origin Murchie, S.L., 2009. The tectonics of Mercury: the view after MESSENGERs of plains. Earth Planet. Sci. Lett. 285, 227242. rst yby. Earth Planet. Space Lett. 285, 283296. Hiesinger, H., Head, J.W., 2006. New views of lunar geoscience: an introduction Weitz, C.M., Head, J.W., Pieters, C.M., 1998. Lunar and regional dark mantle and overview. Rev. Mineral. Geochem. 60, 181. deposits: geologic, multispectral, and modeling studies. J. Geophys. Res. 103, Hiesinger, H., Helbert, J., MERTIS Co-I Team, 2010. The Mercury Radiometer and 2272522759. Thermal Infrared Spectrometer (MERTIS) for the BepiColombo mission. Planet. Wetherill, G.W., 1988. Accumulation of Mercury from planetesimals. In: Vilas, F., Space Sci. 58, 144165. Chapman, C.R., Matthews, M.S. (Eds.), Mercury. University of Arizona Press, Hillier, J.K., Buratti, B.J., Hill, K., 1999. Multispectral photometry of the Moon and Tucson, AZ, pp. 670691. absolute calibration of the Clementine UV/Vis camera. Icarus 141, 205225. Wetherill, G.W., 1994. Provenance of the terrestrial planets. Geochim. Comochim. Kerber, L., Head, J.W., Solomon, S.C., Murchie, S.L., Blewett, D.T., Wilson, L., 2009. Acta 58, 45134520. Explosive volcanic eruptions on Mercury: eruption conditions, magma volatile Wilson, L., 1980. Relationships between pressures, volatile content and ejecta content, and implications for interior volatile abundances. Earth Planet. Sci. velocity in three types of volcanic eruptions. J. Volcanol. Geotherm. Res. 8, Lett. 285, 263271. 297313. Lucey, P.G., Taylor, G.J., Malaret, E., 1995. Abundance and distribution of iron on Wilson, L., 2009. Volcanism in the solar system. Nat. Geosci. 2, 389397. the Moon. Science 268, 11501153. Wilson, L., Head, J.W., 1981. Ascent and eruption of basaltic magma on the Earth Lucchitta, B.K., Schmitt, H.H., 1974. Orange material in the Sulpicius Gallus and Moon. J. Geophys. Res. 86, 29713001. Formation at the southwestern edge of Mare Serenitatis. Proc. Lunar. Sci. Wilson, L., Head, J.W., 2008. Volcanism on Mercury: a new model for the history of Conf. 5, 223234. magma ascent and eruption. Geophys. Res. Lett. 35, L23205. doi:10.1029/ McClintock, W.E., Lankton, M.R., 2007. The Mercury Atmospheric and Surface 2008GL035620. Composition Spectrometer for the MESSENGER mission. Space Sci. Rev. 131, Zolotov, M. Yu., 2011. On the chemistry of mantle and magmatic volatiles on 481521. Mercury. Icarus 212, 2441. doi:10.1016/j.icarus.2010.12.014.

Load More