Compact Reconnaissance Imaging Spectrometer for Mars ... - CiteSeerX

Click Here

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114, E00D11, doi:10.1029/2009JE003348, 2009

for

Full Article

Compact Reconnaissance Imaging Spectrometer for Mars observations of northern Martian latitudes in summer W. M. Calvin,1 L. H. Roach,2 F. P. Seelos,3 K. D. Seelos,3 R. O. Green,4 S. L. Murchie,3 and J. F. Mustard2 Received 29 January 2009; revised 4 August 2009; accepted 21 August 2009; published 2 December 2009.

[1] This paper brings together initial results obtained of the high northern latitudes in

Mars years 28 and 29, between October 2006 and October 2008. These measurements confirm many previous models and shed new light on the nature of polar surface materials, particularly in intermediate-albedo units of the polar layered deposits, many of which are found to be ice-rich. We identify hydrated non ice materials present in many lowalbedo troughs, as well as in the circumpolar erg that was previously associated with gypsum. We identify icy outlier deposits that may be related to subsurface thermophysical properties and permafrost. New observations of the gypsum-rich dune material constrain models for its formation and distribution. Intrinsic properties of ice content and grain size are found to be independent of the albedo of fine layered units and may provide a novel method for stratigraphic identification and correlation. Citation: Calvin, W. M., L. H. Roach, F. P. Seelos, K. D. Seelos, R. O. Green, S. L. Murchie, and J. F. Mustard (2009), Compact Reconnaissance Imaging Spectrometer for Mars observations of northern Martian latitudes in summer, J. Geophys. Res., 114, E00D11, doi:10.1029/2009JE003348.

1. Introduction [2] The north and south polar deposits and circumpolar materials preserve a record of current and recent Martian climate history. The residual ice and subjacent layered units may reflect the last few hundred thousand to few million years. Dune and mantling deposits around the northern polar layered deposits (PLD) span the entire Amazonian period, 3Ga to the present, and multiple sequences of deposition and erosion are recorded. Understanding the interaction between the current climate and residual ice caps will act as a ‘Rosetta stone’ which can be used to interpret the layered deposits in terms of the previous climates which formed them. This modern climate record is intimately tied to the migration pathways of the primary volatile species, CO2 and H2O, and non ice contaminants entrained in the seasonal and residual ice deposits. Recent observations of the permanent and seasonal polar caps have shown the complex dynamics involved in the layers active today and also the extent and consistency of layering in the polar caps themselves. Our observations extend and enhance these recent descriptions of the polar deposits through high-resolution

1 Department of Geological Sciences and Engineering, University of Nevada, Reno, Nevada, USA. 2 Department of Geological Sciences, Brown University, Providence, Rhode Island, USA. 3 Johns Hopkins University Applied Physics Laboratory, Laurel, Maryland, USA. 4 Jet Propulsion Laboratory, Pasadena, California, USA.

Copyright 2009 by the American Geophysical Union. 0148-0227/09/2009JE003348$09.00

compositional information of the various units of the PLD and surrounding terrains. [3] The layers of the northern PLD were recognized in Viking imagery [e.g., Cutts, 1973; Howard et al., 1982] and it was noted that the uppermost high-albedo surfaces were nearly coincident with lower albedo layered units, in contrast to the southern PLD. Viking and telescopic observations demonstrated that the upper, ‘‘bright’’ surface was water ice and the albedo was most consistent with a coarse old water ice surface with a substantial portion of ingrained fines, sand or rocky materials (aka ‘‘dirt’’) [Kieffer et al., 1976; Clark and McCord, 1982; Kieffer, 1990]. Based on the lower albedo of the dark portions of the PLD these were inferred to be largely rocky material [Thomas et al., 1992]. Howard [2000] recognized a wide variety of frost streaks, either bright through the deposition of icy material or dark via removal of a frosty veneer, and these streaks exhibit a wide range of orientation and dynamics. He proposed that the major contribution to the deep troughs was ablation due to katabatic winds and solar insolation patterns, rather than ice cap flow, unlike terrestrial glaciers whose surface and interior morphology is strongly dependent on ice flow and deformation. The circumpolar dune field or ‘erg’ is the largest accumulation of dunes on Mars [Tsoar et al., 1979] and their inertias are lower than other low-albedo dunes [Herkenhoff and Vasavada, 1999]. [4] The Mars Global Surveyor complement of instruments enhanced our understanding of these regions through detailed topographic mapping [Zuber et al., 1998], highresolution imagery [Malin and Edgett, 2001], and temperature evolution [Kieffer and Titus, 2001]. The Mars Observer Laser Altimeter (MOLA) showed the extreme depths of the

E00D11

1 of 15

E00D11

CALVIN ET AL.: CRISM NORTH POLE SUMMER

troughs into the bulk of the PLD deposit, as well as the absolute height profile and upper surface contour [Zuber et al., 1998]. The Mars Observer Camera (MOC) showed numerous fine scale layers and enabled a number of detailed studies of the stratigraphy [e.g., Byrne and Murray, 2002; Edgett et al., 2003; Fishbaugh and Head, 2005; Fishbaugh and Hvidberg, 2006] and the first attempts to find stratigraphic correlations across different regions of the PLD and to identify layer patterns that are coupled to the climate cycles [Laskar et al., 2002; Milkovich and Head, 2005; Levrard et al., 2007]. Attempts to model Viking camera and MOC albedo in terms of ice content concentrated on the brighter units and were limited by the absolute calibration of the cameras [Hale et al., 2005; Benson and James, 2005; Bass et al., 2000]. Estimates of the volume percent of non ice material derived from these studies were consistent with earlier models. The Thermal Emission Spectrometer (TES) albedo and spectral measurements showed seasonal evolution of albedo and the presence of both sustained and variable bright patches [Kieffer and Titus, 2001; Calvin and Titus, 2008]. [5] The major contribution to our understanding of northern polar volatiles from the Odyssey mission was gleaned from the Gamma ray and neutron spectroscopy measurements that showed significant enhancement in hydrogen poleward of 60° latitude, and large concentrations, inferred to be buried permafrost, beyond the visual extent of the PLD [Boynton et al., 2002; Feldman et al., 2004]. For the north polar region, THEMIS data were used to identify a water ice collar on the retreating seasonal cap [Wagstaff et al., 2008]. [6] Spectroscopy in the wavelengths of reflected solar radiation (0.4 to 5 mm, for the coldest polar temperatures on Mars) is ideally suited to the discrimination of water ice from carbon dioxide ice. The OMEGA instrument on Mars Express provided a synoptic view of seasonal changes in the northern polar ices, noting midsummer removal of early fine-grained frosts and the sustained presence of small patches of fine-grained water ice [Langevin et al., 2005a]. In addition, OMEGA discovered gypsum deposits in the north polar erg [Langevin et al., 2005b]. [7] The Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) can amplify these observations and interpretations through high spatial resolution mapping of the polar units and surrounding material. This paper brings together initial results obtained in the early portion of the Mars Reconnaissance Orbiter (MRO) Primary Science Phase (PSP) observations from October 2006 to January 2007, Ls 130° to 180° in Mars Year 28 (MY 28) and selected observations from the following northern summer, July 2008 to October 2008, Ls 95°– 145° in MY 29. Each of the topical areas is the subject of detailed ongoing analysis and future focused reports. The measurements described here confirm many previous models and shed new light on the nature of materials, particularly in intermediate albedo units of the PLD, many of which are found to be ice-rich. We identify icy outlier deposits that may be related to subsurface thermophysical properties (i.e., permafrost). New observations of the gypsum-rich dune material constrain models for its formation and distribution. Intrinsic properties of ice content and grain size are found to be independent of the albedo of fine layered units and may

E00D11

provide a novel method for stratigraphic identification and correlation.

2. Observation Summary [8] CRISM is a visible and near infrared hyperspectral imager with 544 channels from 0.362 to 3.920 mm (6.55 nm spectral sampling). The instrument has two primary data acquisition modes: a gimbaled or targeted hyperspectral mode that acquires 544 channel high spatial resolution data at 20 or 40 m/pixel and a push broom mode that acquires a 72 channel spectral subset at 100 or 200 m/pixel. Gimbaled observation types include Full Resolution Targeted (FRT) and Half-Resolution Long (HRL) or Short (HRS) which differ in spatial resolution and down-track coverage (10 or 20 km). Push broom observation types include Multispectral Survey (MSP, 200 m/pixel) and Multispectral Window (MSW, 100 m/pixel) that can both be acquired with down-track coverage of 45, 180, or 540 km [Murchie et al., 2007]. All CRISM observations have 10 km coverage cross track. The MRO Primary Science Phase (PSP) began at Ls 130° allowing for significant data acquisition at the high northern latitudes in advance of deteriorating atmospheric and illumination conditions. As a result of early PSP science priorities and operational constraints, the majority of the CRISM data acquired at that time were MSP, although a handful of FRT, HRS and HRL observations were acquired at high northern latitudes. 2.1. MSP North Polar Mosaics [9] Early in MRO PSP the spacecraft was restricted to a nadir observing geometry on 50% of the available science orbits to facilitate systematic campaign observations. The CRISM multispectral survey campaign observation strategy is designed to acquire a near-global map using the MSP data acquisition mode. CRISM survey observations early in PSP typically took the form of three minute segments scheduled in series on the dedicated campaign orbits. Multispectral survey data are also acquired on unrestricted orbits on a noninterference basis subject to operational and schedule constraints. The 72 wavelengths in a CRISM multispectral observation consist of a subset of the hyperspectral wavelengths carefully selected to maintain sufficient spectral sampling of key surface and atmospheric absorptions. This allows a robust set of spectral summary parameters to be calculated from multispectral data [Pelkey et al., 2007]. The high mapping survey coverage density near the poles allows for the generation of a nearly complete multispectral survey mosaic for the north polar residual cap and the surrounding terrain (>75° latitude). The north polar mapping mosaics were created from an Ls range of 95 to 180 degrees in two Mars years. Individual mapping strips are converted to apparent I/F using standard pipeline processing, photometrically corrected using a Lambert assumption, and corrected for atmospheric gas absorptions using an empirical transmission spectrum derived from observations over Olympus Mons. Summary parameters are then calculated from the corrected reflectance data. These processed strips and band parameter maps are projected using spacecraft orientation information and then mosaicked using custom registration software [Murchie et al., 2009a]. The spacecraft and instru-

2 of 15

E00D11


E00D11

survey visible and IR false color composites of the portion of Planum Boreum including Olympia Planitia and the surrounding dune field for Ls 130 to 180 in MY28, Figure 2 shows the 1.9 mm band depth determined from MY29 observations.

Figure 1. CRISM multispectral survey (top) VNIR and (bottom) SWIR false color composites. The VNIR mosaic uses channels at 0.71 mm, 0.60 mm, 0.53 mm (RGB), the SWIR uses 2.52 mm, 1.51 mm, and 1.08 mm (RGB). The VNIR color approximates ‘‘true’’ color in the wavelengths selected. In the SWIR composite ice-rich regions appear blue, materials with moderate ice absorptions are white, and ice-poor materials appear yellow to brown. This mosaic used observations from Ls 130° to 180° in MY 28. ment pointing knowledge is excellent, allowing mosaicking with control at the resolution of the data. For consistency, the mapping data are resampled to 256 pixel/deg (231 m/ pixel) spatial resolution. Figure 1 shows multispectral

2.2. Gimbaled Observations [10] Due to the limited observing opportunities early in the MRO primary mission, only a handful of high spatial and full spectral resolution CRISM observations were obtained at the high northern latitudes in the northern summer of MY28. Northward of 75°, there were 28 FRTs, 12 HRL, and 17 HRS observations acquired between 2006 and 272 and 2007 – 30 (year, day of year), or Ls 113.5° to 175.4°. These include 19 FRT, 7 HRL and 10 HRS observations of the residual ice layers and PLD, the remainder are of circumpolar dunes and outlier icy areas. The locations of these targeted observations sample the residual upper ice surfaces, selected areas of troughs, several exposures of basal layers and scarp faces in the PLD, particularly along the Chasma Boreale, and several outlier, ice and non ice deposits. The highest concentration of gypsum identified by OMEGA was also targeted, as well as selected locations along the region of enhanced hydration observed in Olympia Undae [Langevin et al., 2005b; Fishbaugh et al., 2007]. Data were processed using standard pipeline processing to I/F with corrections for atmosphere removal and photometric angles [Murchie et al., 2009a]. Custom software routines allow geographic projection of individual observations for comparison with other mapped data products.

3. Hydrated, Non Ice Materials 3.1. Distribution of Non Ice Hydrated Materials [11] The CRISM MSP data and related analysis products for the high northern latitudes reveal hydrated material within a number of troughs and reentrants into the north polar residual cap, as inferred from the depth and spatial

Figure 2. Band depth at 1.90 mm filtered to remove most areas rich in water ice. Enhanced signatures are discussed in the text. The location of the 6 CRISM targeted observations from MY28 are shown as black polygons with arrows pointing to them over a mosaic created with observations from Ls 95° to 145° in MY29 and the data set is displayed on MOLA topography in the background. 3 of 15


E00D11

E00D11

Figure 3. (left) Subset of select CRISM targeted observations over the region of Olympia Undae with the strongest hydration signature in OMEGA (brown colors). OMEGA hydration signature in color over the Mars Digital Image Mosaic. CRISM observations are from MY 28 (black outline) and MY29 (blue outline). (right) The same CRISM observational footprints over the geologic unit map modified from Tanaka et al. [2005] to include revisions from Tanaka et al. [2008].

coherency of the 1.9 mm absorption band. Figure 2 shows this hydration signature filtered to remove those regions rich in water ice for the area adjacent to Olympia Planitia, as modified from Horgan et al. [2009]. The parameter seeks to isolate the 1.928 mm band depth due to hydrated materials from the broader 2 mm water ice absorption in locations where both water ice and hydrated materials are present. First, a continuum is calculated from the shoulders of the 2 mm water ice absorption and removed to compensate for the negative slope. The hydration signature is measured as the excess 1.928 mm absorption after water ice continuum removal. Theoretically, water ice should have a symmetrical absorption over that wavelength region, so it will not affect the continuum removal or the excess 1.928 mm band depth. In practice, the multispectral CRISM wavelengths available for constructing the continuum are not exactly symmetric around the 2 mm water ice absorption, so a slight component of water ice is included in the resulting hydration parameter. This is most evident around high-albedo ice outliers around 120E and equatorward of 70N. Hydration parameter ¼ 1 CR1928 =CR2119 (

) R1928 CR1928 ¼ ðR2205R1809Þ þ R1809 =ð2:2051:809Þ *ð1:928 1:809Þ

( CR2119 ¼

ðR2205R1809Þ =

R2119 ð2:2051:809Þ *ð2:119

) 1:809Þ

þ R1809

where CR1928 is the continuum-removed reflectance at 1.928 mm, R1928 is the reflectance at 1.928 mm, etc., and a number without a letter prefix refers to the wavelength value. [12] The hydration signature is most commonly associated with low-albedo dunes, both in the vicinity of the Olympia Undae gypsum-bearing sand sea, and in distant polar cap

reentrants. In particular, non ice hydrated material is identified in the Boreales Scopuli adjacent to Olympia Rupes (using nomenclature of Tanaka et al. [2008]), but is common in reentrants and troughs, as also observed in lower resolution using OMEGA data by Horgan et al. [2009]. The detailed analysis of high-resolution hyperspectral CRISM observations of hydrated materials in the vicinity of the north polar residual cap is ongoing and will be the subject of future papers focused on these regions. 3.2. Gypsum [13] Targeted CRISM observations at 21 – 43 m/pixel along Olympia Undae characterize the gypsum spectral signature in the dunes and allow speculation into the location of the gypsum source region. Six targeted observations from MY28 and seven from MY29 covering Olympia Undae and its geologic contacts were analyzed for this study. Figure 2 shows the distribution of CRISM targeted observations from MY28 designed to explore the variation of hydration along the arc of hydration noted by OMEGA [Langevin et al., 2005b]. Figure 3 shows a close-up of the Olympia Undae region of highest hydration signature together with MY28 and MY29 CRISM observations including in this study. In order to clarify the positions and locations of these and other materials we refer to the units as identified by Tanaka et al. [2008]. Using the Tanaka et al. [2008] nomenclature, CRISM observations were concentrated along contacts between ABou and underlying polar units. [14] The distinguishing absorption features of gypsum are a triplet absorption between 1.44 and 1.54 mm, absorptions at 1.75 mm and 1.9 mm, and a doublet absorption at 2.25 and 2.265 mm, all due to combinations and overtones of H2O vibrations and librations [Hunt et al., 1971; Cloutis et al., 2006]. Gypsum was identified in CRISM spectra by a wide 2.2 mm absorption, often with a resolvable doublet, and a drop off after 2.4 mm diagnostic of polyhydrated sulfates. A small absorption at 1.75 mm is also usually present. Atmospheric CO2 absorptions complicate interpreting the 1.9 mm feature.

4 of 15

E00D11


E00D11

Figure 4. Spectra from CRISM observation FRT0000285F of north polar gypsum dunes. (left) Spectra from the dune crest and interdune areas with library spectrum show stronger signatures in dune crests. (right) HiRISE image PSP001432_2610 shows the polygonal crack in substrate exposed in interdune areas; arrow points to dune crest and notes the approximate illumination direction. HiRISE image is 300m across. 3.2.1. Distribution [15] Targeted observations were analyzed to study changes in gypsum signature concentration. They showed a general decrease in the overall strength of the gypsum signature with clockwise distance along Olympia Undae. This trend is not due to differences in the relative areas covered by gypsum-rich dune crests and interdune substrate exposures [Lahtela et al., 2009]. The gypsum signature decrease in a downwind direction from Chasma Boreal is consistent with aeolian transport of gypsum or its sulfur-rich precursor materials. [16] The deepest hydration bands and strongest gypsumrelated absorptions are found in dune crests across Olympia Undae. Dune troughs and sides show weaker but still present gypsum features, while interdune material has the weakest hydration bands (Figure 4) [Murchie et al., 2009b]. The strong absorption bands in the dune crests may indicate coarser grain sizes [Ghrefat et al., 2007], cementation of grains, or more gypsum in the dune crests than the troughs and interdune areas. Grain size modeling indicates 10 mm for interdune material and 10 – 100 mm for dune crest material [Fishbaugh et al., 2007; F. Poulet, personal communication, 2008]. The interdune material’s weak hydration signature indicates it either is covered by a thin layer of gypsum-rich sand grains or hosts small amounts of finegrained gypsum. The unit is light-toned and polygonally fractured on the meter scale (Figure 4) [Murchie et al., 2009b], with sufficient cracks to trap loose gypsum-rich sand grains.

3.2.2. Gypsum Source Region [17] The origin of this gypsum deposit is uncertain, however it must be near or in the dunes, as saltation would quickly break down the soft mineral [Fishbaugh et al., 2007]. Possible sources include the underlying polar basal units, called Planum Boreum 3 (ABb3) and Boreum cavi (ABbc) [Tanaka et al., 2005, 2008], or in situ alteration of pyroxene- and sulfide-rich sand dunes from water derived from polar cap melting [Fishbaugh et al., 2007]. Past work has noted the concentration of sulfate in dune crests rather than interdune substrate exposures, but the original source is uncertain [Roach et al., 2007; Horgan et al., 2009]. We are looking for evidence that ABbc, ABou, or ABb3 are plausible gypsum sources. ABbc consists of light and dark-colored platy layers with cross bedding, interpreted to be alternating basaltic sands and ice-cemented dust layers [e.g., Herkenhoff et al., 2007; Tanaka et al., 2008]. ABou is the Olympia Undae unit consisting of basaltic sands and migrating dunes. ABb3 is a thin, bright, layered unit that overlies ABou and is thought to be composed of water ice and dust [Tanaka et al., 2008]. Both units are exposed at the ice cap edge and underneath or adjacent to the Olympia Undae dunes. [18] The multiple constituents that make up the polar basal units have been mapped at the regional scale [Tanaka et al., 2008], but interpreting those geologic units and their contacts from isolated outcrops in HiRISE and CRISM imagery is challenging. Several units with different compositional and physical properties are exposed beneath and

5 of 15

E00D11


E00D11

Figure 5. False color composite and band depth map at 1.9 mm (BD1900) for HRS0000C074 on the contact between ABb3 and ABou units. Increased hydration, bright in BD1900, are associated with dark dunes. Spectra associated with dunes (ABou-red) and underlying light-toned substrate (ABs or ABrt) (green). Note the 1.9 mm feature is present in both bright and dune units, but is strongest in the dunes. The dunes also show a concavity near 2.2 mm and an absorption near 2.4 mm due to gypsum. between the ABou dunes. They include ABs, ABrt, ABbc, and ABb3. [19] The bright, meter-scale polygonally fractured material exposed in interdune areas shown in Figure 4 we interpret as ABb3, presumably dusty water ice. In that case, the hydration signature would be more due to water ice than gypsum. A more detailed discussion of the spectral character of ABb3 appears in section IV.B.2. Another contact between ABou and ABb3 is seen in HRS0000C074 (Figure 5). Here, ABb3 has strong water ice signatures, and no assessment of its gypsum-bearing potential is possible. The dark dunes in ABou have strong hydration bands and gypsum features at 2.2 and 2.4 mm while the light-toned material underlying the dunes is hydrated but without gypsum features. In Figure 5, the light-toned material exposed between dunes we interpret as either Rupes Tenuis unit (ABrt) or Scandia region unit (ABs), reworked by permafrost activity into an intermediatetoned unit with raised edge polygons a few to 10 m on a side. Tanaka et al. [2008] interpret ABrt as consisting of consolidated aeolian material from weathering ABs, a massive sedimentary unit with evidence for water- and ice-driven mobilization. Figure 6, HiRISE image PSP_010120_2605 shows both the lighter-toned ABb3 and the ABrt or ABs unit which underlie the ABou gypsum-bearing dunes. ABb3 is fractured on the meter to submeter scale, with multiple layers sometimes visible, and is preferentially exposed on the lee side of dunes. These identifications are based on the correspondence of unit descriptions and pictures of Tanaka et al. [2008] and regional mapping [Tanaka et al., 2005] with CRISM imagery. None of the units underlying ABou show gypsum absorption features. [20] Finally, the Planum Boreum cavi unit, ABbc, which has been identified as the dominant source of Olympia Undae sand [Thomas and Weitz, 1989; Byrne and Murray, 2002], was examined to test if it was also a source for the gypsum in the north polar sand sea. CRISM spectra confirm a lack of gypsum in large exposures of ABbc at the edge of the polar cap, but often these observations are dominated by water ice, which will occlude potential hydration signatures.

In smaller outcrops of ABbc that are clearly separated from ice-bearing terrains, there is a slight hydration signature but none of the diagnostic gypsum bands. HRL0000B99F, Figure 7, shows a sharp contact between ABbc and the gypsum-bearing dunes of ABou. Ratio spectra show ABbc is slightly more hydrated than the light-toned ABrt or ABs exposures beneath the gypsum dunes, but neither material has convincing sulfate signatures. Another contact of ABou and ABbc further west from the high gypsum region reveals that both ABou or ABbc have slightly hydrated signatures but no diagnostic gypsum absorptions. Like ABb3, the hydration signature in the dark-toned platy unit ABbc could be due to either a water ice component, a nongypsum hydrated mineral, or small amounts of gypsum. The gypsum signature of the dunes may be stronger than any source region as the gypsum could be concentrated on the surface by capillary wicking and evaporation. [21] In summary, multiple geologic units near ABou have been investigated with CRISM data to determine the source

Figure 6. HiRISE image PSP_010120_2605. Unit relationship to the gypsum signature is described in the text. Illumination if from the top right of the image.

6 of 15

E00D11


Figure 7. CRISM observation HRL0000B99F showing a sharp contact between ABou and ABbc units. Hydration signatures are discussed in the text.

of the gypsum. The units ABs, ABrt, ABbc, and ABb3 all show weak hydration bands but no diagnostic sulfate bands, which suggests the units have a water ice component rather than large amounts of gypsum. However, small amounts of gypsum cannot be ruled out. The polygonal fracturing or platy nature of these units have sufficient roughness to trap loose gypsum-rich sand grains. In that case, the gypsum source would be within the dunes themselves. On the other hand, if one of the basal units is the gypsum source, a postulate is that beneath the gypsum-bearing dunes, the substrate could consist of an eroded friable (gypsum containing) layer over a more resistant layer (gypsum-poor lighter-toned substrate). As current analysis is not conclusive with regard to either the source or formation mechanism for gypsum, continued analysis of high-resolution CRISM and HiRISE data of geologic contacts in the region with the highest gypsum concentration is ongoing.

4. Icy Materials 4.1. Ice Outliers [22] Analysis of late northern summer MSP mosaic data products (e.g., Figure 1) reveals small water ice deposits distributed throughout the northern plains at latitudes quite distant from the residual polar cap and the usual bright outliers between 70 and 80°N. These small outliers range in size from a few hundreds of meters to several kilometers, and are generally associated with the northward facing slopes of crater rims or other elevated landforms. In a few instances the ice deposits are located on the leeward (southeast facing) sides of larger craters, and may indicate the presence of wintertime CO2 frost formation from orographic lifting [Beitia et al., 2008]. The brighter frost sublimes more slowly than the surrounding CO2 ice, ulti-

E00D11

mately forming a late spring cold trap and inducing an accumulation of water ice. 4.1.1. Distribution [23] The spatial resolution and coverage of the MSP data set allowed for consistent detection of water ice deposits at the scale of 500 m and larger. To determine the southernmost extent of the deposits, false color visible and infrared composites of the MSP data set as well as a 1.5mm band depth map were assembled. Here 1° longitude bins were systematically evaluated and the furthest south occurrence (minimum latitude) of water ice was recorded. For comparison to other data sets, a 10° latitude and longitude average was determined about a center latitude of 75°N. Hyperspectral targeted observations were acquired throughout the northern plains resulting from support of Phoenix landing site selection. Although the distribution of hyperspectral observations is sparse for MY28, the higher spatial resolution allows for detection of an order of magnitude smaller ice deposits (50 to 100 m, depending on the observing mode). Figure 8 portrays the mapped minimum latitude of water ice occurrence. For large portions of the mosaic, the minimum latitude follows the outline of the residual polar cap and well-known bright outliers. However, km-scale ice outliers frequently exist well away from the cap to as far south as 67°N. The average latitude is 75.5°N. The longitudinal distribution is also not uniform, with a noticeable increase in minimum latitude around lower albedo regions such as Acidalia Planitia (300– 345°E). Comparison to the GRS dry layer thickness [Boynton et al., 2008; Feldman et al., 2008; Mellon et al., 2008] also shows a strong correlation to the inferred ice table depth, areas with shallow ice also having abundant surface outliers as seen in Figure 8. This suggests that large-scale factors are common controls to both surface and subsurface ice distribution, or that permafrost close to the surface may improve the retention of late season surface ice. 4.1.2. Local Factors Influencing Occurrence [24] A majority of the outlying deposits are located on north facing crater rims or escarpments, but a few appear to extend around to the east and southeast facing side of the crater rim (Figure 9). The prevalence of ice occurrence in association with craters indicates that local slope is a dominant factor controlling the presence of late summer water ice. The deposits that persist on east to southeast facing sides of craters where readily exposed to morning sunlight are less intuitive. These leeward deposits may be due to orographic lifting that causes wintertime deposition of bright fine-grained CO2 frost [Beitia et al., 2008]. The CO2 frost sublimes more slowly than the surrounding CO2 ice and induces a late spring cold trap that favors accumulation of water ice. Figure 9 shows high-resolution examples for deposits adjacent to and within craters. In Figure 9 (left), the orientation suggests a contribution from local winds and the ripple-like morphology of the deposit is suggestive of aeolian influence. In Figure 9 (right), local insolation and topography likely influence the location of sustained ice patches. [25] Outlying late summer water ice deposits are an indication of the extent of environments favorable to interannual persistence of near-surface ice. Since Martian permafrost is presumed to form by vapor diffusion [Mellon et al., 2008] overlying small patches of surface frost may

7 of 15

E00D11


E00D11

Boreum Planum. Given the timing of the start of MRO science phase observations, Ls 130, the early north seasonal evolution of albedo and frost grain size noted by OMEGA and TES over the residual ice deposits [Langevin et al., 2005a; Calvin and Titus, 2008] was finished for Mars year 28. We explored selected observations for variation in ice absorption features, ice content of low-albedo layers, association between albedo and ice band strength, and to explore spectral parameters that may be suggestive of ice content in the lower albedo units. 4.2.1. Overview of Unit Types [27] Simple measures of ice content include band depth of the 1.5 mm absorption feature, slope down from 1.15 to 2.25 mm, and spectral rise from 3.0 to 3.6 mm. Langevin et al. [2007] have noted that icy particles in the atmosphere contribute weakly to these three parameters, but that the spectral shape beyond 3 mm can be quite diagnostic in discriminating ice cloud from icy surfaces. We have performed simple color composites, relative band depths of the 1.5 mm feature and find the upper ice-rich surfaces (ABb4) show subtle changes in ice band strength, and that most of the upper, lower albedo units of the PLD (ABb1– ABb3) are still quite water ice – rich. Figure 10 shows average spectral properties for differing high-albedo (ABb4) and low-albedo (ABb1) units of the PLD, with properties that suggest grain size variation in these surfaces. These spectra were taken from surfaces on Gemini Lingula that remain high albedo throughout the northern summer. The location map in Figure 10 shows that observation FRT00002F46 is roughly 850m lower in absolute elevation than FRT000035B1. The strong features of water ice at 1.04, 1.25, 1.5, 2.0 mm are noted in bright upper surfaces of the PLD. Although the overall albedo is low for water ice, the shape of peaks at 3.05 and 3.65 mm in the 35B1 spectrum (black spectral line) suggest fairly fine ice grain sizes, as the region longward of 3 mm rapidly saturates with large grains [Calvin and Clark, 1991; Calvin et al., 2002; Warren and Brandt, 2008]. The other high-albedo surface

Figure 8. (top) CRISM 1.5 mm band depth map from 65 to 90°N with minimum ice latitudes (white dots). (bottom) GRS dry layer thickness (g/cm2) over MOLA topography with mapped minimum ice latitude. Areas with very high and very low DLT values (Acidalia and the polar cap, respectively) have been masked out.

either provide additional water vapor for underlying permafrost or may prohibit diffusive loss from the underlying icerich soil. Future observations will monitor the evolution of these water ice deposits through time. 4.2. Survey of Polar Layered Deposits [26] Of the 17 FRTs from MY28 that observe the upper surfaces of the PLD, four are predominantly of the upper ice-rich surface, six show fine layering and stratigraphy adjacent to troughs, and seven observe the lowest units of

Figure 9. High-resolution examples of outlier ice deposits. (left) CRISM hyperspectral observation of a southeast facing water ice deposit (appears white; clouds are also apparent). HRL000044AD was acquired at Ls = 180°; image is approximately 11 km across at narrowest. (right) A small fresh crater with water ice deposit (bluish spots, subset). FRT0000332C was acquired at Ls = 142.6°; image is approximately 11 km across at narrowest. In both cases the false color composite uses R = 0.71mm, G = 0.60mm, B = 0.53mm.

8 of 15

E00D11


E00D11

Figure 10. Average full spectral range data from ice and upper unit low-albedo troughs. The location map with arrows shows the regions for the two FRTs used for averages in the plot. FRT000035B1 (black spectrum) is 850m higher than 00002F46 (colored spectra). Bright units in both FRTs are high albedo throughout northern summer. Residual CO2 atmosphere features are seen near 2.0 mm, and values between 2.65 and 2.9 have also been set to zero due to atmospheric interference. The feature near 0.65 mm is a calibration artifact. Spectra have been smoothed using a 5-channel box-car filter. FRT00002F46 acquired at Ls 133 shown below in visible color RGB bands (0.73, 0.60, 0.44 mm), IR color (RGB-2.25, 1.50, 1.25 mm), and 1.5 mm band depth (values range from 2.2 to 6). IR color emphasizes ice-rich material in blue, and the band depth is highest for lightest colors in the false color stretch. Spectra are from high-albedo ice-rich regions (cyan) and the two dark troughs (green and orange) and regions for the spectral extraction are noted on the visible color image as small colored circles. Note the left trough of FRT00002F46 (green spectrum) is more ice-rich than the right (orange spectrum). Additional discussion of spectral features provided in the text. from FRT00002F46 (cyan colored spectrum and also shown with various image color composites) shows stronger ice features shorter than 2.6 mm, but reduced contrast longer than 3 mm, consistent with somewhat larger grain sizes. Color composites highlight ice band strength both in IR color combinations and with the band strength of the 1.5 mm feature. Albedo varies a few percent, along with changes in band depth. For low-albedo surfaces in troughs (ABb2 or 3), the units have nearly identical visible albedos, but have strongly contrasting water ice band strengths (troughs on either side of a small promontory in Figure 10, observation FRT00002F46 at the margins of Gemini Lingula). [28] Figure 10 shows that a wide range of surfaces can have nearly identical visible albedo, yet strongly varying

water ice absorption features. This variation must be due to changes in ice abundance, grain size, or possibly surface distribution of small bright patches as observed in HiRISE imagery [Fishbaugh et al., 2009]. In order to further explore ice content variability in the low-albedo troughs we examined the 1.5 mm band depth (as determined in simple 3 channel ratios) and found that this can consistently track exposed layers, regardless of albedo. As shown in Figure 11, ice content of layers is not correlated to albedo, and may be an independent method for determining stratigraphy or layer ‘‘packets.’’ Figure 11 shows the visible color of FRT000027EC, on the side of the residual cap near Olympia Planitia. The 1.5 mm band depth is calculated and plotted in a density slice for intermediate band strengths. In

9 of 15

E00D11


Figure 11. (a) Visible color for FRT000027EC showing albedo striping in upper PLD units, the data are unprojected. (b) Density slice of the 1.5 mm band depth, displayed over a gray scale of infrared albedo. Only intermediate band strengths are shown and range from browns (low) to colors (moderate). (c) Plot of albedo (horizontal axis) versus 1.5 mm band depth (vertical) for the IR portion of this FRT. Data do not show strong linear trends but clusters that range in albedo and water ice band strength. The red cluster is shown as the red overlay on the IR false color image. (d) False color IR composite using 2.50, 1.50, and 1.02 mm channels, highlight ice-rich, high-albedo surfaces in blue. This section is a subset of the full scene, and the region shown as a black box on Figure 11b. Here consistent ice band depth (red cluster from Figure 11C) occurs over a range of albedos and likely traces intrinsic units on the lower edge of the ice-rich surface (ABb4/ABb3). cluster plots of albedo versus 1.5 mm band depth there are no strong linear trends, yet clusters within the data cloud can isolate units with consistent water content and variable albedo. This is demonstrated by highlighting a coherent set of points in red (Figure 11c) and then showing those same points in red on an IR false color composite (Figure 11d). This suggests that water ice content of low-albedo ‘‘dirty’’ layers can be traced to intrinsic layer properties, rather than surface veneers or small patches of frost on eroded surfaces as seen in HiRISE imagery [Herkenhoff et al., 2007; Fishbaugh et al., 2009]. Continuous horizontal banding in 1.5 mm band depth is also seen to extend over tens of kilometers in the MSP mosaic (Figure 1). Large-scale stratification in water ice content is also observed within the full 3 km vertical stack of the Planum Boreal. The lowest layers with sections just above the what has been called the ‘‘ Basal Unit’’ (BU), identified as ABrt, ABbc [Tanaka et al., 2008] are all more water ice – rich than intermediate sections (ABb1 – ABb3) as described in the next section. Future work will concentrate on determining layer sequences defined by infrared ice absorption feature strength, if these correlate with layers identified based on albedo and morphology, and if there is consistency between

E00D11

these spectral parameterizations and previously noted ‘‘marker’’ beds [Malin and Edgett, 2001; Milkovich and Head, 2006; Fishbaugh et al., 2009]. 4.2.2. Ice Content of the Lower PLD [29] We had previously noted that the lowest sections of the PLD exhibit weak to no water ice signatures based on the 1.5 mm band depth [Murchie, 2007; Seelos et al., 2007]. However, Herkenhoff et al. [2007] found that based on polygonal fractures and mass wasting of blocks that the brighter basal units should be volatile rich. Improved calibration gives us increased confidence in weak ice signatures at 1.5 mm and we find that the 3.0 to 3.6 mm rise is diagnostic of ice in lighter toned basal unit materials where the 1.5 and 2.0 mm features of water ice are not obvious, as shown in Figure 12. Of the 6 FRTs in MY28 that cover the lower units of the PLD initial analysis shows that portions of the basal unit material are ice-cemented, particularly light layered materials adjacent to the upper Basal Unit (BU) scarps (ABrt, ABbc, and/or ABb1–Figure 12, also FRT00002FA6, not shown here). An upturn longward of 3 mm can also be indicative of hydrated minerals rather than water ice [Calvin, 1997; Milliken and Mustard, 2005]. We find that the presence of a distinct 3.7 mm peak, rather than just a spectral rise can reliably distinguish between hydrated minerals and ice content. In hydrated materials, if there is a peak it is also usually at longer wavelengths than those materials that have some water ice. Detailed modeling of the spectral shape longward of 3 mm is needed to confidently identify ice in the darkest materials, but there is evidence that lighter ejecta surrounding craters in the Chasma Boreale floor and lighter units of the differentially eroded basal layers contain some ice (FRT0000287F, not shown). The lower portion of the PLD just above the basal unit (ABbc, and/or ABb1) is found to be relatively ice-rich compared to layers higher up in the sequence, with the uppermost highest albedo surface (ABb4) the most ice-rich (Figures 10 and 11). Though coverage is sparse, this trend is observed in all high-resolution images at widely separated locations. In FRT00002854 (Figure 13) light colored icerich cones are seen on top of the ABrt and ABbc units of the basal unit scarp in contact with these upper ice-rich (ABb1/ ABb2) layers. Russell et al. [2008] also note these features in HiRISE imagery and interpret them as being composed of water ice and related to scarp mass wasting processes. 4.3. Initial Modeling of Spectral Properties [30] In order to examine model fits to the varying spectral properties of the PLD we examined CRISM FRT00002F7F acquired at Ls of 133. This image captures the north central wall of Chasma Boreale and is shown in Figure 14 (curve A). The image contains a portion of the base of Chasma Boreale as well as a section the polar layered deposits including the upper surface. The image shows layering of mixed ice and dust. From this data set a set of average spectra were extracted from the central swath and from the top to the bottom of the image; these spectra included units from ABou, ABb1, ABrt or ABbc, and icy material ABb2. The spatial locations of the extracted spectra were optimally matched between the two CRISM spectrometers. Figure 14 (curve B) shows the calibrated I/F spectra labeled A to I for the different sections from top to bottom. In each case an average of 10 spectra were extracted and averaged. Spectra

10 of 15

E00D11


E00D11

Figure 12. FRT00002854, acquired Ls 114. (top) Visible color RGB bands (0.73, 0.60, 0.44 mm), IR color (2.50, 1.50, and 1.02 mm), and ice emphasis. The ice emphasis is created using the RGB combination of albedo at 1.1 mm, the 1.5 mm band depth, and the spectral rise from 3.3 to 3.6 mm. In the IR color ice-rich units are light blue, and only layered material in the trough wall and small mounds near the base appear to contain water ice. In the ice emphasis image, icy materials are white, cyan and magenta; ice-poor units are blue. Magenta colors clearly highlight spatially coherent units, including the upper most surface and layered material at the bottom of the PLD (right side) and light-toned areas surrounding dunes (top left). (bottom) Corresponding HiRISE color, PSP_0014112_2650. The location of HiRISE is shown as a light rectangle on the visible color CRISM image. The material in the middle of the image is ABs, and the layered material in the bottom right is probably ABrt, ABbc, and/or ABb1.

A, B, C, and E show water ice absorption features at 1.05, 1.25, 1.50, 2.00 mm. Spectrum E was extracted from the fine layer on the surface of the upper unit and is likely a spatial mixture even at full CRISM spatial resolution. [31 ] An atmospheric correction using a modified MODTRAN approach with residual artifact suppression was applied to the I/F data set. MODTRAN was originally designed to model the atmospheric properties of the Earth [e.g., Anderson et al., 2000]. However, the underlying physics of radiative transfer captured in MODTRAN allows it to be configured to model the atmosphere of Mars. In MODTRAN, sensor view and illumination geometry as well as the atmospheric constituent abundances and profiles may be adjusted. By varying the amount of water vapor for a Mars type atmosphere we find the wavelength regions of 1.38, 1.88, and 2.60 mm are most impacted by these low but variable amounts. MODTRAN modeled spectra with varying amounts of carbon dioxide that may be due to path length changes from low to high surface elevation or path length changes due to illumination and observation geometry were also considered. Using this approach we created a MODTRAN atmospheric transmittance look-up table that spans a range of carbon dioxide and water vapor abundances (Figure 15, top). These modeled transmittance spectra were convolved to the CRISM spectral response functions taking into account the instrumental cross-track spectral variation. Figure 15 (middle) shows the atmospherically corrected spectra of the PLD that were derived from this

approach. Following atmospheric correction, weak water ice absorptions are more clearly observable in spectra D, E and F from the upper unit of the PLD. This early result shows good success with a MODTRAN based atmospheric correction approach. Future work will be focused on increased automation of the techniques and on proper inclusion of

Figure 13. Close-up of the right hand side of FRT00002854, from Figure 12, rotated counterclockwise. Visible color (left) RGB bands (0.73, 0.60, 0.44 mm), (right) IR color (1.50, 2.20, and 1.02 mm). Ice-rich materials are light blue in the IR color, though note that the orange/brown layered terrain laying on the Chasma Boreale floor (single arrow left image) is also ice cemented in the ice emphasis image in Figure 12. Note that the ice lenses stratigraphically above the above the floor and lowest layered unit (multiple arrows right image) appear to connect with overlying ice-rich layered materials.

11 of 15

E00D11


E00D11

Figure 14. (left) Unprojected CRISM color composite image FRT00002F7F of the northern edge of Chasma Boreale (RGB, 0.96,0.74,0.51 mm). Dunes to bottom right are ABou, middle is ABb1, layered deposits are probably ABrt or ABbc, and icy material to top right are ABb2. (right) CRISM extracted I/F spectra extending from upper surface of PLD to Chasma Boreale. Averages of 10 spectra are used from approximately the same cross-track column region. atmospheric scattering effects for dust laden atmospheric conditions. [32] To investigate the physical properties of the water ice recorded in these CRISM spectral measurements a radiative transfer grain size forward inversion approach has been used. This approach incorporates the DISORT radiative transfer code [Stamnes et al., 1988] with Mie scattering calculations based upon the spectral complex refractive index of water ice [Wiscombe, 1980]. The combined algorithm is used to model the spectral reflectance signature of an ice surface with specified grain size. At this stage in the model development, the grains of ice are modeled as spheres. To test the forward inversion approach, a set of model spectra spanning a range of grain sizes were calculated. Figure 15 (bottom) shows the best fit between the CRISM spectrum from the upper surface of the PLD and the modeled spectra. The effects of dust are not yet fully integrated in the model, consequently the algorithm was tested only on the range from 0.90 to 2.60 mm. Using a spectral fit of this range, a grain size of 900 mm was estimated for the upper surface of the PLD. This is consistent with late season grain sizes determined for the upper surface and outliers in the range of 700 mm to 1 mm by Langevin et al. [2005a]. A future focus of this effort will be to more completely model the physical properties of the ice and dust mixtures of the PLD and apply the forward inversion approach to a range of CRISM full resolution data sets. The derived ice and dust parameters will be used to investigate the diversity of the PLD and pursue understanding of its history in relation to the Mars polar environment.

5. Discussion, Synthesis, and Conclusions [33] Hydrated non ice material is present in multiple windows in the north polar layered deposits. Some of these hydrated units are associated with dunes, similar to the gypsum signature noted in Olympia Undae, but many

appear related to stratigraphic horizons exposed within troughs. [34] Gypsum is concentrated on dune crests, but a weak hydration signature is also present in some polygonal lighttoned underlying units. Many units of the lower portion of the PLD (ABs, ABrt, and ABbc) and the overlying ABb3 unit are found to have a hydration signature suggestive of water ice. The distribution of gypsum along Olympia Undae is consistent with aeolian reworking. It should also be noted that the massive layered sulfates observed in situ at Meridiani Planum also have weak to no sulfate signatures at CRISM spatial resolution (S. Wiseman et al., Stratigraphic context of hydrated sulfate and phyllosilicate deposits in northern Sinus Meridiani, submitted to Journal of Geophysical Research, 2009). This may suggest that competent substrate units may lack sufficient spatial exposure to be obvious in CRISM, or that massive, textured units reduce surface scattering effects and hence the strength of spectral absorption features. Alternately the polygonal underlying terrain does not contain gypsum, but only traps small grains distributed by the wind, leading to a weak hydration signature. The source of sand in the erg from the Planum Boreum cavi unit and the apparent hydration of several Planum Boreum members (Rupes Tenuis, Planum Boreum cavi, and Planum Borerum unit 2) suggest that the source of gypsum may be related to the formation of the basal unit itself. [35] The distribution of lower latitude outlier ice deposits is found to correlate well with the GRS signature for shallow hydrogen. The Phoenix mission has found ice at shallow depths as predicted prior to landing [Smith, 2008; Mellon et al., 2008] underlying thermophysical properties may help sustain water ice at low latitudes after the retreat of the seasonal CO2 cap. Sun shadowing, local winds, local ground ice lenses and other small scale phenomena will also contribute to the persistence of low latitude outlier ice deposits, as is also seen commonly on Earth.

12 of 15

E00D11


Figure 15. (top) MODTRAN modeled atmospheric transmittance for varying carbon dioxide path length. These spectra have been convolved to the CRISM spectral response functions. (middle) Spectra shown in Figure 14 atmospherically corrected using MODTRAN models. (bottom) Grain size model fit of 900 mm radius with CRISM spectrum C from the upper surface of the PLD.

E00D11

Borealis Formation. CRISM also suggests that the majority of the PLD is ice-rich and that multiple horizons are seen within the vertical layered unit related to variation in ice content or thickness of non ice coating on the upper most surfaces. It has been shown that modest amounts of lowalbedo material coating ice-rich surfaces will dramatically lower the visible albedo but that the infrared ice features will remain strong [Clark, 1981; Kieffer, 1990]. CRISM observations suggest both large scale stratigraphic variation in bulk ice content, as well as localized variation in veneers of either dust or ice coatings. Many low-albedo layers, previously assumed to be ‘‘dirt’’ are found to have strong infrared signatures of water ice. Sections just above the lowest units are found to be more ice-rich. Lighter materials in the lowest elevations of the PLD and adjacent to ice-rich scarps are ice cemented, possibly also ice-rich. A strong underlying ice content may also contribute to interannual variability in the appearance of bright ‘‘frosted’’ patches [Malin and Edgett, 2001; Calvin and Titus, 2008, and references therein]. [37] Previously, stratigraphic units and layering within the PLD have been related to albedo and unique morphological features. Numerous studies have noted that albedo alone is not a good indicator of stratigraphy. We find that ice grain size or content, as determined by the strength of the infrared water ice absorptions, may be an independent identifier of horizontally competent units. The water ice content of ‘‘dirty’’ layers can be mapped and traced to intrinsic properties rather than surface veneers or slopes. This may offer a new approach to stratigraphy of the PLD, and ultimately relating the vertical structure to past climate cycles. [38] Numerous small ice lenses overly the lowest laminated units and are in contact with ice-rich materials above. These may represent slumping of volatile rich blocks, or accumulations of landslide material in the wake of downslope motion as captured by HiRISE [Herkenhoff et al., 2007; Russell et al., 2008]. [39] Preliminary efforts suggest terrestrial models for atmospheric correction and surface physical properties can be readily adapted to Martian conditions. Moderate grain sizes for the upper water ice units are found, consistent with models from OMEGA. This suggests that the lower coherent units of the PLD are sintered by age or compaction, developing a surface lag of non ice material, similar to some glacial ice on earth. [40] The results presented lay the foundation for future studies and analysis that will continue to reveal the nature of the units that have the most direct link to current and recent climate change on Mars. [41] Acknowledgment. This work supported by the Mars Reconnaissance Orbiter CRISM Science Team and Participating Scientist Programs.

References [36] Phillips et al. [2008] have noted that the PLD is largely transparent at the radio frequencies of MARSIS and SHARAD but find multiple interior radar reflection horizons. They suggest 2 – 10% admixture of soil with ice and that some strong reflectors may be possibly as much as 30% dust. Additionally, they see a radar reflection from both the BU under the main lobe and the underlying Vastitas

Anderson, G. P., et al. (2000), MODTRAN4 radiative transfer modeling for remote sensing, Proc. SPIE, 4049, 176 – 183, doi:10.1117/12.410338. Bass, D. S., K. E. Herkenhoff, and D. A. Paige (2000), Variability of Mars’ north polar water ice cap: I. Analysis of Mariner 9 and Viking Orbiter imaging data, Icarus, 144(2), 382 – 396, doi:10.1006/icar.1999.6300. Beitia, E. A., T. N. Titus, K. D. Seelos, F. P. Seelos, A. Colaprete, S. L. Murchie, and P. R. Christensen (2008), Martian H2O ice outliers may result from wintertime near-surface CO2 snows, Lunar Planet. Sci., XXXIX, Abstract 1422.

13 of 15

E00D11


Benson, J. L., and P. B. James (2005), Yearly comparisons of the Martian polar caps: 1999 – 2003 Mars Orbiter Camera observations, Icarus, 174(2), 513 – 523, doi:10.1016/j.icarus.2004.08.025. Boynton, W. V., et al. (2002), Distribution of hydrogen in the near surface of Mars: Evidence for subsurface ice deposits, Science, 297(5578), 81 – 85, doi:10.1126/science.1073722. Boynton, W. V., G. J. Taylor, S. Karunatillake, R. C. Reedy, and J. M. Keller (2008), Elemental abundances determined via the Mars Odyssey GRS, in The Martian Surface: Composition, Mineralogy, and Physical Properties, edited by J. F. Bell, pp. 105 – 124, Cambridge Univ. Press, London. Byrne, S., and B. C. Murray (2002), North polar stratigraphy and the paleoerg of Mars, J. Geophys. Res., 107(E6), 5044, doi:10.1029/ 2001JE001615. Calvin, W. M. (1997), Variation of the 3-mm absorption feature on Mars: Observations over eastern Valles Marineris by the Mariner 6 infrared spectrometer, J. Geophys. Res., 102, 9097 – 9107, doi:10.1029/ 96JE03767. Calvin, W. M., and R. N. Clark (1991), Modeling the reflectance spectrum of Callisto 0.25 to 4.1 mm, Icarus, 89(2), 305 – 317, doi:10.1016/00191035(91)90180-2. Calvin, W. M., and T. N. Titus (2008), Summer season variability of the north residual cap of Mars as observed by the Mars Global Surveyor Thermal Emission Spectrometer (MGS-TES), Planet. Space Sci., 56, 212 – 226, doi:10.1016/j.pss.2007.08.005. Calvin, W. M., M. Milman, and H. H. Kieffer (2002), Reflectance of Antarctica from 3 to 5 mm: Discrimination of surface snow and cloud properties, Ann. Glaciol., 34, 121 – 126. Clark, R. N. (1981), Spectral reflectance of water-mineral mixtures at low temperatures, J. Geophys. Res., 86(B4), 3074 – 3086, doi:10.1029/ JB086iB04p03074. Clark, R. N., and T. B. McCord (1982), Mars residual north polar cap: Earth-based spectroscopic confirmation of water ice as a major constituent and evidence for hydrated minerals, J. Geophys. Res., 87(B1), 367 – 370, doi:10.1029/JB087iB01p00367. Cloutis, E. C., et al. (2006), Detection and discrimination of sulfate minerals using reflectance spectroscopy, Icarus, 184, 121 – 157, doi:10.1016/ j.icarus.2006.04.003. Cutts, J. A. (1973), Nature and origin of layered deposits of the Martian polar regions, J. Geophys. Res., 78, 4231 – 4249, doi:10.1029/ JB078i020p04231. Edgett, K. S., R. M. E. Williams, M. C. Malin, B. A. Cantor, and P. C. Thomas (2003), Mars landscape evolution: Influence of stratigraphy on geomorphology in the north polar region, Geomorphology, 52, 289 – 297. Feldman, W. C., et al. (2004), Global distribution of near-surface hydrogen on Mars, J. Geophys. Res., 109, E09006, doi:10.1029/2003JE002160. Feldman, W. C., M. T. Mellon, O. Gasnault, S. Maurice, and T. H. Prettyman (2008), Volatiles on Mars: Scientific results from the Mars Odyssey neutron spectrometer, in The Martian Surface: Composition, Mineralogy, and Physical Properties, edited by J. F. Bell, pp. 125 – 148, Cambridge Univ. Press, London. Fishbaugh, K. E., and J. W. Head (2005), Origin and characteristics of the Mars north polar basal unit and implications for polar geologic history, Icarus, 174(2), 444 – 474, doi:10.1016/j.icarus.2004.06.021. Fishbaugh, K. E., and C. S. Hvidberg (2006), Martian north polar layered deposits stratigraphy: Implications for accumulation rates and flow, J. Geophys. Res., 111, E06012, doi:10.1029/2005JE002571. Fishbaugh, K. E., F. Poulet, V. Chevrier, Y. Langevin, and J.-P. Bibring (2007), On the origin of gypsum in the Mars north polar region, J. Geophys. Res., 112, E07002, doi:10.1029/2006JE002862. Fishbaugh, K. E., S. Byrne, K. E. Herkenhoff, R. L. Kirk, C. Fortezzo, P. S. Russell, and A. McEwen (2009), Evaluating the meaning of ‘‘layer’’ in the Martian north polar layered deposits and the impact on the climate connection, Icarus, doi:10.1016/j.icarus.2009.04.011, in press. Ghrefat, H. A., P. C. Goodell, B. E. Hubbard, R. P. Langford, and R. E. Aldouri (2007), Modeling grain size variations of aeolian gypsum deposits at White Sands New Mexico, using AVIRIS imagery, Geomophology, 88, 57 – 68, doi:10.1016/j.geomorph.2006.10.013. Hale, A. S., D. S. Bass, and L. K. Tamppari (2005), Monitoring the perennial Martian northern polar cap with MGS MOC, Icarus, 174, 502 – 512, doi:10.1016/j.icarus.2004.10.033. Herkenhoff, K. E., and A. R. Vasavada (1999), Dark material in the polar layered deposits and dunes on Mars, J. Geophys. Res., 104(E7), 16,487 – 16,500, doi:10.1029/1998JE000589. Herkenhoff, K. E., S. Byrne, P. S. Russell, K. E. Fishbaugh, and A. S. McEwen (2007), Meter scale morphology of the north polar region of Mars, Science, 317, 1711 – 1715, doi:10.1126/science.1143544. Horgan, B. H., J. F. Bell III, E. Z. Noe Dobrea, E. A. Cloutis, D. T. Bailey, M. A. Craig, L. H. Roach, and J. F. Mustard (2009), The distribution of

E00D11

hydrated minerals in the north polar region of Mars, J. Geophys. Res., 114, E01005, doi:10.1029/2008JE003187. Howard, A. D. (2000), The role of eolian processes in forming surface features of the Martian polar layered deposits, Icarus, 144, 267 – 288, doi:10.1006/icar.1999.6305. Howard, A. D., J. A. Cutts, and K. R. Blasius (1982), Stratigraphic relationships within Martian polar cap deposits, Icarus, 50, 161 – 215, doi:10.1016/0019-1035(82)90123-3. Hunt, G. R., J. W. Salisbury, and C. J. Lenhoff (1971), Visible and nearinfrared spectra of minerals and rocks: IV. Sulphides and sulfates, Mod. Geol., 3, 1 – 14. Kieffer, H. H. (1990), H2O grain-size and the amount of dust in Mars residual north polar cap, J. Geophys. Res., 95(B2), 1481 – 1493, doi:10.1029/JB095iB02p01481. Kieffer, H. H., and T. N. Titus (2001), TES mapping of Mars’ north seasonal cap, Icarus, 154(1), 162 – 180, doi:10.1006/icar.2001.6670. Kieffer, H. H., S. C. Chase, T. Z. Martin, E. D. Miner, and F. D. Palluconi (1976), Martian north pole summer temperatures: Dirty water ice, Science, 194(4271), 1341 – 1344, doi:10.1126/science.194.4271.1341. Lahtela, H., et al. (2009), Coordinated HiRISE/CRISM observation on gypsum signature in Martian polar dunes, Lunar Planet. Sci., XL, Abstract 2254. Langevin, Y., F. Poulet, J. P. Bibring, and B. Gondet (2005a), Summer evolution of the north polar cap of Mars as observed by OMEGA/Mars express, Science, 307(5715), 1581 – 1584, doi:10.1126/science.1109438. Langevin, Y., F. Poulet, J. P. Bibring, and B. Gondet (2005b), Sulfates in the north polar region of Mars detected by OMEGA/Mars express, Science, 307(5715), 1584 – 1586, doi:10.1126/science.1109091. Langevin, Y., J. P. Bibring, F. Montmessin, F. Forget, M. Vincendon, S. Doute, F. Poulet, and B. Gondet (2007), Observations of the south seasonal cap of Mars during recession in 2004 – 2006 by the OMEGA visible/near-infrared imaging spectrometer on board Mars Express, J. Geophys. Res., 112, E08S12, doi:10.1029/2006JE002841. Laskar, J., B. Levrard, and J. F. Mustard (2002), Orbital forcing of the Martian polar layered deposits, Nature, 419, 375 – 377, doi:10.1038/ nature01066. Levrard, B., F. Forget, F. Montmessin, and J. Laskar (2007), Recent formation and evolution of northern Martian polar layered deposits as inferred from a global climate model, J. Geophys. Res., 112, E06012, doi:10.1029/2006JE002772. Malin, M. C., and K. S. Edgett (2001), Mars Global Surveyor Mars Orbiter Camera: Interplanetary cruise through primary mission, J. Geophys. Res., 106(E10), 23,429 – 23,570, doi:10.1029/2000JE001455. Mellon, M. T., W. V. Boynton, W. C. Feldman, R. E. Arvidson, T. N. Titus, J. L. Bandfield, N. E. Putzig, and H. G. Sizemore (2008), A prelanding assessment of the ice table depth and ground ice characteristics in Martian permafrost at the Phoenix landing site, J. Geophys. Res., 113, E00A25, doi:10.1029/2007JE003067. Milkovich, S. M., and J. W. Head (2005), North polar cap of Mars: Polar layered deposit characterization and identification of a fundamental climate signal, J. Geophys. Res., 110, E01005, doi:10.1029/2004JE002349. Milkovich, S. M., and J. W. Head (2006), Surface textures of Mars’ north polar layered deposits: A framework for interpretation and future exploration, Mars, 2, 21 – 45, doi:10.1555/mars.2006.0003. Milliken, R. E., and J. F. Mustard (2005), Quantifying absolute water content of minerals using near-infrared reflectance spectroscopy, J. Geophys. Res., 110, E12001, doi:10.1029/2005JE002534. Murchie, S. M. (2007), First results from the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM), Lunar Planet. Sci., XXXVIII, Abstract 1472. Murchie, S. M., et al. (2007), Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) on Mars Reconnaissance Orbiter (MRO), J. Geophys. Res., 112, E05S03, doi:10.1029/2006JE002682. Murchie, S. M., et al. (2009a), Compact Reconnaissance Imaging Spectrometer for Mars investigation and data set from the Mars Reconnaissance Orbiter’s primary science phase, J. Geophys. Res., 114, E00D07, doi:10.1029/2009JE003344. Murchie, S. M., et al. (2009b), A synthesis of Martian aqueous mineralogy after one Mars year of observations from the Mars Reconnaissance Orbiter, J. Geophys. Res., 114, E00D06, doi:10.1029/2009JE003342. Pelkey, S. M., et al. (2007), CRISM multispectral summary products: Parameterizing mineral diversity on Mars from reflectance, J. Geophys. Res., 112, E08S14, doi:10.1029/2006JE002831. Phillips, R. J., et al. (2008), Mars north polar deposits: Stratigraphy, age, and geodynamical response, Science, 320(5880), 1182 – 1185, doi:10.1126/science.1157546. Roach, L. H., et al. (2007), CRISM spectral signatures on the north polar gypsum dunes, Lunar Planet. Sci., XXXVIII, Abstract 1970.

14 of 15

E00D11


Russell, P., et al. (2008), Seasonally active frost-dust avalanches on a north polar scarp of Mars captured by HiRISE, Geophys. Res. Lett., 35, L23204, doi:10.1029/2008GL035790. Seelos, F. P., S. L. Murchie, S. M. Pelkey, and K. D. Seelos (2007), CRISM multispectral survey campaign: Status and initial mosaics, Lunar Planet, Sci., XXXVIII, Abstract 2336. Smith, P. (2008), The Phoenix mission explores the Martian arctic, Eos Trans. AGU, 89(53), Fall Meet. Suppl., Abstract P13F – 01. Stamnes, K., S. C. Tsay, W. J. Wiscombe, and K. Jayaweera (1988), Numerically stable algorithm for discrete-ordinate-method radiative transfer in multiple scattering and emitting layered media, Appl. Opt., 27, 2502 – 2509, doi:10.1364/AO.27.002502. Tanaka, K. L., J. A. Rodriguez, J. A. Skinner, M. C. Bourke, C. M. Fortezzo, K. E. Herkenhoff, E. J. Kolb, and C. H. Okubo (2008), North polar region of Mars: Advances in stratigraphy, structure, and erosional modification, Icarus, 196(2), 318 – 358, doi:10.1016/j.icarus.2008.01.021. Tanaka, T. L., J. A. Skinner Jr., and T. M. Hare (2005), Geologic map of the northern plains of Mars, U.S. Geol. Surv. Spec. Invest. Map 2888. Thomas, P., and C. Weitz (1989), Sand dune materials and polar layered deposits on Mars, Icarus, 81, 185 – 215, doi:10.1016/00191035(89)90133-4. Thomas, P., S. Squyres, K. Herkenhoff, A. Howard, and B. Murray (1992), Polar deposits of Mars, in Mars, edited by H. H. Kieffer et al., pp. 767 – 795, Univ. of Ariz. Press, Tucson.

E00D11

Tsoar, H., R. Greeley, and A. R. Peterfreund (1979), Mars: The north polar sand sea and related wind patterns, J. Geophys. Res., 84, 8167 – 8180, doi:10.1029/JB084iB14p08167. Wagstaff, K. L., T. N. Titus, A. B. Ivanov, R. Castano, and J. L. Bandfield (2008), Observations of the north polar water ice annulus on Mars using THEMIS and TES, Planet. Space Sci., 56, 256 – 265, doi:10.1016/ j.pss.2007.08.008. Warren, S. G., and R. E. Brandt (2008), Optical constants of ice from the ultraviolet to the microwave: A revised compilation, J. Geophys. Res., 113, D14220, doi:10.1029/2007JD009744. Wiscombe, W. J. (1980), Improved Mie scattering algorithms, Appl. Opt., 19(9), 1505 – 1509, doi:10.1364/AO.19.001505. Zuber, M. T., et al. (1998), Observations of the north polar region of Mars from the Mars Orbiter Laser Altimeter, Science, 282, 2053 – 2060, doi:10.1126/science.282.5396.2053.

W. M. Calvin, Department of Geological Sciences and Engineering, University of Nevada, MS 172, Reno, NV 89557, USA. ([email protected]) R. O. Green, Jet Propulsion Laboratory, Pasadena, CA 91109, USA. S. L. Murchie, F. P. Seelos, and K. D. Seelos, Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723, USA. J. F. Mustard and L. H. Roach, Department of Geological Sciences, Brown University, Providence, RI 02912, USA.

15 of 15