Classification and analysis of candidate impact crater ... · Classiﬁcation and analysis of...

Icarus 260 (2015) 346–367

Contents lists available at ScienceDirect

Icarus

journal homepage: www.elsevier .com/ locate/ icarus

Classification and analysis of candidate impact crater-hostedclosed-basin lakes on Mars

http://dx.doi.org/10.1016/j.icarus.2015.07.0260019-1035/� 2015 Elsevier Inc. All rights reserved.

⇑ Corresponding author. Fax: +1 (401) 863 3978.E-mail address: [email protected] (T.A. Goudge).

Timothy A. Goudge a,⇑, Kelsey L. Aureli a, James W. Head a, Caleb I. Fassett b, John F. Mustard a

a Department of Earth, Environmental and Planetary Sciences, Brown University, 324 Brook St., Box 1846, Providence, RI 02912, USAb Department of Astronomy, Mount Holyoke College, South Hadley, MA 01075, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 26 March 2015Revised 26 June 2015Accepted 20 July 2015Available online 27 July 2015

Keywords:Mars, surfaceGeological processesMars

We present a new catalog of 205 candidate closed-basin lakes contained within impact craters across thesurface of Mars. These basins have an inlet valley that incises the crater rim and flows into the basin butno visible outlet valley, and are considered candidate closed-basin lakes; the presence of a valley flowinginto a basin does not necessitate the formation of a standing body of water. The major geomorphic dis-tinction within our catalog of candidate paleolakes is the length of the inlet valley(s), with two majorclasses – basins with long (>20 km) inlet valleys (30 basins), and basins with short (<20 km) inlet valleys(175 basins). We identify 55 basins that contain sedimentary fan deposits at the mouths of their inlet val-leys, of which nine are fed by long inlet valleys and 46 are fed by short inlet valleys. Analysis of the min-eralogy of these fan deposits suggests that they are primarily composed of detrital material. Additionally,we find no evidence for widespread evaporite deposit formation within our catalog of candidateclosed-basin lakes, which we conclude is indicative of a general transience for any lakes that did formwithin these basins. Morphometric characteristics for our catalog indicate that as an upper limit, thesebasins represent a volume of water equivalent to a �1.2 m global equivalent layer (GEL) of water spreadevenly across the martian surface; this is a small fraction of the modern water ice reservoir on Mars. Ourcatalog offers a broader context within which results from the Mars Science Laboratory Curiosity rovercan be interpreted, as Gale crater is a candidate closed-basin lake contained within our catalog. Gale isalso one of 12 closed-basin lakes fed by both long and short inlet valleys, and so in situ analyses byCuriosity can shed light on the relative importance of these two types of inlets for any lacustrine activitywithin the basin.

� 2015 Elsevier Inc. All rights reserved.

1. Introduction

The surface of Mars displays abundant evidence for both fluvialactivity (e.g., Pieri, 1980; Carr, 1987; Goldspiel and Squyres, 1991;Cabrol and Grin, 1999, 2001, 2010; Howard et al., 2005; Irwin et al.,2005; Fassett and Head, 2008a,b; Hynek et al., 2010) and impactcratering (e.g., Hartmann, 1966; Pike, 1971; Barlow, 1988; Stromet al., 1992; Robbins and Hynek, 2012a,b), two of the most impor-tant geomorphic processes that have operated to shape the mar-tian surface. An interesting and fundamental part of the martianlandscape where these two processes readily interact is in the for-mation of paleolake basins, which are commonly contained withinimpact craters (e.g., Goldspiel and Squyres, 1991; Cabrol and Grin,1999, 2001, 2010; Fassett and Head, 2008b). Impact craters onMars provide the dominant mechanism by which basin topography

has been created (Cabrol and Grin, 2010), which is notably differ-ent from Earth, where the majority of lake basins are formed fromtectonics and glaciation (Wetzel, 2001; Cohen, 2003).

Impact crater-contained paleolakes on Mars have been recog-nized for several decades, including in many key contributionsfrom studies of Viking data (e.g., Goldspiel and Squyres, 1991;Forsythe and Zimbelman, 1995; Forsythe and Blackwelder, 1998;Cabrol and Grin, 1999, 2001). An important characteristic of pale-olakes, on both Earth and Mars, is their hydrologic setting, andmartian paleolake basins are commonly classified as eitheropen-basin or closed-basin lakes (Cabrol and Grin, 1999, 2001;Fassett and Head, 2008b). Open-basin lakes have an outlet valleythat drains the basin, which requires that the basin filled suffi-ciently with water to overtop and breach (Fassett and Head,2008b). Open-basin lakes on Mars typically have at least one inletvalley (Fassett and Head, 2008b); however, Warner et al. (2010)mapped a series of hydrologically open basins in Xanthe Terraand Arabia Terra with outlet valleys but no inlet valleys, suggesting

http://crossmark.crossref.org/dialog/?doi=10.1016/j.icarus.2015.07.026&domain=pdf

http://dx.doi.org/10.1016/j.icarus.2015.07.026

mailto:[email protected]

http://dx.doi.org/10.1016/j.icarus.2015.07.026

http://www.sciencedirect.com/science/journal/00191035

http://www.elsevier.com/locate/icarus

T.A. Goudge et al. / Icarus 260 (2015) 346–367 347

open-basin lakes may also have formed without the incision ofinlet valleys that remain preserved (e.g., via groundwater influxor inlet valleys that have not remained observable).

Closed-basin lakes are basins that may have ponded water, sup-plied either by an inlet valley or groundwater influx, but that haveno visible outlet valley. It is more difficult to definitively say that ahydrologically closed basin once contained a standing body ofwater – water flowing into a closed topographic depression doesnot require the formation of a lake. Therefore, any identifiedclosed-basin lake can be referred to only as a candidate paleolake;however, for simplicity in the remainder of the paper, we will referto candidate closed-basin lakes simply as closed-basin lakes.

Yet, the incision of inlet valleys associated with closed-basinlakes indicates some level of fluvial activity, and these featurescan provide insight into the evolution of the martian hydrologiccycle. However, recent catalogs of paleolakes on Mars have focusedon open-basin lakes (e.g., Fassett and Head, 2008b; Goudge et al.,2012), and there has not been a global catalog of closed-basin lakescompiled with post-Viking data (Cabrol and Grin, 1999).Additionally, Gale crater, the site of exploration for the MarsScience Laboratory (MSL) Curiosity rover, is a closed-basin lake(Cabrol et al., 1999; Irwin et al., 2005). Broader studies ofclosed-basin lakes on Mars can therefore provide context for theinterpretation of MSL results indicative of paleolake activity (e.g.,Grotzinger et al., 2014). In this contribution we present a new cat-alog of 205 closed-basin lakes contained within impact cratersacross the surface of Mars, along with morphologic, morphometric,and mineralogic details of these basins and associated sedimentarydeposits.

2. Previous catalogs of closed-basin lakes on Mars

The most recent global catalog of closed-basin lakes on Marswas compiled by Cabrol and Grin (1999), who used VikingOrbiter data to map 179 paleolake basins on Mars contained withinimpact craters. These basins were classified into three categories:closed, open, and lake-chains. The latter two types are largelyincorporated in the catalog of Fassett and Head (2008b), and com-pose approximately two thirds of the Cabrol and Grin (1999) cata-log (119/179 of their basins). The remaining 60 basins are classifiedas closed-basin lakes by Cabrol and Grin (1999), suggesting thathydrologically closed paleolakes are less abundant than hydrolog-ically open paleolakes on Mars.

Following the pioneering work by Cabrol and Grin (1999),Fassett and Head (2008b) used higher-resolution image data com-bined with topographic data to define a catalog of 210 open-basinlakes, almost doubling the number originally identified.Additionally, Irwin et al. (2005) used post-Viking data to show thatmany of the paleolake basins identified by Cabrol and Grin (1999)are not reliably classified given modern data. Of particular issue forthe study of closed-basin lakes is the lack of reliable global-scaletopography available during the analysis of Cabrol and Grin(1999). Without topography, it is difficult to say from image dataalone whether a valley that breaches an impact crater basin is aninlet or an outlet. Here we build on the earlier work of Cabroland Grin (1999) by conducting a new global survey ofclosed-basin lakes using the wealth of image and topographic datacollected for the surface of Mars in the last �15 years.

3. Methods

3.1. Identification of closed-basin lakes

We performed a grid-based search for closed-basin lakes from�60� to 60�N across all longitudes. High latitude regions (i.e.,

<�60� and >60�N) were not included in our search to avoid areasthat have been heavily modified by ice-related processes in Mars’recent past (e.g., Head et al., 2003). The initial search was com-pleted using gridded topography at �463 m/pixel from the MarsOrbiter Laser Altimeter (MOLA) instrument (Smith et al., 2001)onboard the Mars Global Surveyor spacecraft, and the�100 m/pixel global daytime infrared mosaic (Edwards et al.,2011) from the Thermal Emission Imaging System (THEMIS)instrument (Christensen et al., 2004) onboard the Mars Odysseyspacecraft. These data were analyzed in ESRI’s ArcMap geographicinformation system (GIS) software. Grid cells with side lengths of�100–250 km (varied with latitude) were systematically searchedfor impact craters that display topographic and/or morphologicindications of inlet valleys breaching the crater rim.

Each potential basin identified from this initial search was fur-ther examined using higher resolution image data, including�6 m/pixel data from the Context Camera (CTX) instrument(Malin et al., 2007) onboard the Mars Reconnaissance Orbiter(MRO) spacecraft and <50 m/pixel data from the High ResolutionStereo Camera (HRSC) instrument (Neukum et al., 2004) onboardthe Mars Express spacecraft. Where available, these data were alsosupplemented with topography from stereo-derived digital eleva-tion models (DEMs) from HRSC images (Neukum et al., 2004;Gwinner et al., 2010).

To be added to the catalog of closed-basin lakes, each basinneeded to satisfy the following requirements: (1) the basin isdefined by an impact crater, or multiple coalesced impact craters,(2) the basin has one, or more, identifiable valleys that breachthe crater rim and flow into the basin, and (3) the basin has no vis-ible valleys that breach the crater rim and drain away from thebasin (i.e., outlet valleys). The requirement of a visible inlet valleyinherently limits our catalog to basins involving some level of flu-vial incision, excluding any paleolakes that may have been formedpredominantly from groundwater influx (e.g., Warner et al., 2010;Wray et al., 2011; Michalski et al., 2013). Identifying candidategroundwater-fed closed-basin lakes in a global survey, however,is exceedingly difficult given the limited morphologic evidence ofhydrologic activity that is typically associated with such basins(e.g., Wray et al., 2011; Michalski et al., 2013). Therefore, thesebasins are not included in our catalog, although they may repre-sent an important component of the paleolake record on Mars(e.g., Warner et al., 2010; Wray et al., 2011; Michalski et al., 2013).

The restriction of our search to basins defined by impact cratersis different from the Fassett and Head (2008b) catalog ofopen-basin lakes, which also includes broader, inter-crater depres-sions. However, since fluvial valleys will always flow down slope,the terminus of most valleys on Mars will represent the inlet to atopographic depression that is defined by a closed contour at theelevation of the valley terminus. This represents an unrealisticnumber of potential closed-basins with an inlet valley, and to avoidthis complication we have excluded inter-crater basins from oursearch. In addition to inter-crater depressions, basins created bytectonic features associated with Valles Marineris are alsoexcluded from our catalog. While it has been hypothesized thatmany of these basins contained lakes in the past (e.g., Lucchittaet al., 1994; Di Achille et al., 2006a, 2007; Harrison andChapman, 2008, 2010; Lucchitta, 2010; Warner et al., 2013),impact cratering is the primary mechanism for basin formationon Mars (Cabrol and Grin, 2010), and so we have limited the scopeof our search to only crater-contained paleolakes.

Examples of closed-basin lakes included in our catalog areshown in Figs. 1 and 2. Identified basins were assigned one of threedegrees of confidence for identification, which is primarily basedon a qualitative confidence in the morphologic identification ofan inlet valley and whether that inlet is likely to have been formedthrough fluvial incision. From most to least confident, the assigned

Fig. 1. Examples of closed-basin lakes with short inlet valleys (white arrows). Blueoutlines indicate maximum lake level elevation contour. (A) Closed-basin lake at18.1�N,�23.0�E (basin number 57; Table 1). Mosaic of CTX images G01_018662_1981and P12_005740_1974. North is up. (B) Closed-basin lake at 20.4�N, �20.9�E (basinnumber 56; Table 1). Mosaic of CTX images B02_010381_1995, B02_010526_2014, andB05_011449_2011. North is up. (C) Closed-basin lake at 19.0�N, 4.9�E (basin number60; Table 1). Mosaic of CTX images B05_011514_1985, D13_032322_1986,P15_006952_1975 and P13_006240_1974 and HRSC nadir image h6251_0000. Northis down. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

348 T.A. Goudge et al. / Icarus 260 (2015) 346–367

classes are: certain, probable, and possible. Basins with inlet val-leys that clearly incise the host-crater rim and are readily identifi-able as predominantly fluvial features in image and topographicdata were assigned a confidence of ‘certain’. Basins with partiallyeroded inlet valleys that may be slightly modified by non-fluvialprocesses were assigned a confidence of ‘probable’. Basins withheavily eroded inlet valleys that may have been significantly mod-ified by non-fluvial processes were assigned a confidence of‘possible’.

The length of the inlet valley(s) of each closed-basin lake in thecatalog was analyzed as a primary morphometric attribute. Twomajor classes were defined based on an empirical threshold valueof 20 km: (1) basins with inlet valleys <20 km in length (e.g., Fig. 1),and (2) basins with inlet valleys >20 km in length (e.g., Fig. 2).Basins with both long and short inlet valleys were classified as hav-ing long inlet valleys. Although inlet length is a readily quantifiedmorphometric parameter, the ultimate control on this parameter islikely to be related to a number of competing factors such as dura-tion and magnitude of fluvial erosion events, valley slope, regionaltopography, and local lithology.

We also documented whether the basin had an identifiable sed-imentary fan deposit at the mouth of the inlet valley feeding thebasin (e.g., Fig. 3). Here we use the term fan or fan deposit as anon-genetic, descriptive term for a fan-shaped deposited hypothe-sized to be sedimentary in origin, as opposed to genetic terms suchas alluvial fan or delta deposit. Fan identifications were assignedone of three degradation classes. Class 1 fans are the leastdegraded, often showing distinct preserved layering (e.g., Fig. 3).Class 2 fans are moderately degraded, showing observable signsof erosion and/or burial by younger deposits, although not to anysignificant degree. Class 3 fans are the most degraded, showingsigns of significant amounts of erosion and/or burial by youngerdeposits. In addition to fan deposits, evidence for shorelines and/orterraces potentially related to lacustrine activity (e.g., Forsythe andZimbelman, 1995; Cabrol and Grin, 1999, 2001; Ori et al., 2000;Banfield et al., 2015) were searched for in our analysis.

MOLA gridded topography was used to find the maximum ele-vation that the basin could be filled before overtopping by examin-ing contours of increasing elevation at 5 m intervals and findingthe highest contour that is completely contained within the basin.This contour level was defined as the maximum possible extent ofthe paleolake within the basin – if the basin filled any further, itwould breach and form an outlet valley. The maximum paleolakeextent is based on modern MOLA topography, and so is likely tobe an underestimate of the true maximum extent due to anydegradation of the crater rim that has occurred since the periodof inlet valley incision. Using this maximum lake level, we calcu-lated the surface area of the maximum lake extent and the currentvolume contained within the basin below the maximum lake level.The diameter of the host impact crater was measured for basinsdefined by a single, non-elliptical impact crater.

3.2. Mineralogy of sedimentary deposits within closed-basin lakes

We studied the mineralogy of sedimentary fan deposits (e.g.,Fig. 3) identified within our catalog of closed-basin lakes usingspectral reflectance data from the Compact ReconnaissanceImaging Spectrometer for Mars (CRISM) instrument (Murchieet al., 2007) onboard the MRO spacecraft. CRISM is a visible tonear-infrared (VNIR) imaging spectrometer that measures the sig-nal of reflected light from the martian surface from �0.36 to3.9 lm. We analyzed all documented sedimentary fan depositscovered by targeted CRISM image data, which have full(�18 m/pixel; FRT images) or half (�36 m/pixel; HRL and HRSimages) spatial resolution and full (�6.5 nm/channel) spectral res-olution (Murchie et al., 2007).

Fig. 2. Examples of closed-basin lakes with long inlet valleys (white arrows). Blue outlines indicate maximum lake level elevation contour. (A) Closed-basin lake at 5.3�N,�58.6�E (basin number 7; Table 1). Mosaic of CTX images B11_013824_1846, B17_016369_1853, P13_005926_1849, B17_016158_1882, P20_008866_1844, andP03_002247_1847 overlain on the THEMIS � 100 m/pixel global daytime infrared mosaic. North is down. (B) Closed-basin lake at �3.3�N, 88.3�E (basin number 8; Table 1).Mosaic of CTX image B18_016812_1755 and HRSC nadir image h7396_0001. North is down. (C) Closed-basin lake at 8.2�N, �49.3�E (basin number 73; Table 1). Mosaic of CTXimages P21_009314_1892, B05_011727_1881, G02_019019_1879, P16_007165_1883, P18_008022_1894, and P19_008312_1883 overlain on the THEMIS �100 m/pixelglobal daytime infrared mosaic. North is to the right. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


Calibrated CRISM I/F data were downloaded from the PlanetaryData System (PDS), corrected for the cosine of the incidence angle,and atmospherically corrected using a standard volcano scan cor-rection method (e.g., Mustard et al., 2008; Ehlmann et al., 2009;McGuire et al., 2009). Spectral parameter maps were calculatedfrom the atmospherically corrected data to highlight regions ofspectral diversity within each CRISM scene (Pelkey et al., 2007;Ehlmann et al., 2009; Viviano-Beck et al., 2014), and were usedas guides for manual analysis of CRISM data. Photometric andatmospheric corrections, and spectral parameter calculations werecompleted using the CRISM Analysis Toolkit (CAT), which is avail-able for download at http://pds-geosciences.wustl.edu/missions/mro/crism.htm.

During CRISM analysis, spectra from large, multi-pixel regionsof interest (ROIs) were averaged together in an attempt to reducespectral noise. These ROI spectra were then divided by the averageROI spectrum of a spectrally bland region within the same CRISMimage. This technique reduces residual instrument artifacts andnoise to emphasize unique spectral characteristics of the unit ofinterest (e.g., Mustard et al., 2008; Ehlmann et al., 2009). All

denominator spectra were taken from ROIs that cover the samecolumns within the CRISM image as the ROI of the unit of interestto reduce the effects of column-dependent instrument noise(Murchie et al., 2007, 2009a).

A particular focus for the CRISM data analysis was the identifi-cation of aqueous alteration minerals (e.g., phyllosilicates, sulfates,carbonates), including possible evidence of alteration mineralauthigenesis (e.g., Forsythe and Zimbelman, 1995; Bristow andMilliken, 2011). To complement the CRISM analyses, we alsosearched our catalog of closed-basin lakes for associated occur-rences of chloride-bearing material using the global distributionof chloride deposit identifications from Osterloo et al. (2010).

4. Results

4.1. Distribution of closed-basin lakes

Our final catalog includes 205 closed-basin lakes distributedacross the martian surface (Fig. 4; Table 1). This catalog is available

http://pds-geosciences.wustl.edu/missions/mro/crism.htm

http://pds-geosciences.wustl.edu/missions/mro/crism.htm

Fig. 3. Example sedimentary fan deposits at the mouths of closed-basin lake inlet valleys. (A) Sedimentary fan deposit within a closed-basin lake at �8.2�N, �159.5�E (basinnumber 102; Table 1). Portion of CTX image P02_001644_1713. North is down. (B) Sedimentary fan deposit within a closed-basin lake at 1.5�N, 116.3�E (basin number 92;Table 1). Portion of CTX image B16_015967_1814. North is to the right. (C) Sedimentary fan deposit within a closed-basin lake at 22.1�N, �8.2�E (basin number 2; Table 1).Portion of CTX image G19_025782_2021. North is down. (D) Sedimentary fan deposit within a closed-basin lake at 8.2�N, �49.3�E (basin number 73; Table 1). Portion of CTXimage P18_008167_1883. North is up.


for download from the Brown University Planetary GeosciencesGroup website (http://www.planetary.brown.edu/html_pages/data.htm) along with the open-basin lake catalog of Fassett andHead (2008b). The identified basins are almost exclusively con-fined to the southern highlands, although there are a few basinsalong the outer margins of the northern plains (Fig. 4). Of the205 identified basins, 175 (�85%) have short inlet valleys (e.g.,Fig. 1) and 30 (�15%) have long inlet valleys (e.g., Fig. 2). Of the30 closed-basin lakes with long inlet valleys, 12 also have shortinlet valleys that incise the host crater rim (Table 1). The shortand long inlet valley closed-basin lakes in our catalog have a sim-ilar global distribution (Fig. 4A), which also approximately corre-sponds to the distribution of open-basin lakes (Fassett and Head,2008b; Goudge et al., 2012). One exception to this is a clusteringof closed-basin lakes with short inlet valleys in the Arabia Terraand Xanthe Terra regions (Fig. 4A).

Our catalog includes 55 closed-basin lakes with sedimentaryfan deposits (e.g., Fig. 3). Nine of these fan deposits are containedwithin long inlet valley closed-basin lakes, and 46 are containedwithin short inlet valley closed-basin lakes (Fig. 4B). This differenceis readily explained by the overall numbers of closed-basin lakes,as both populations have approximately the same proportion ofbasins with sedimentary fan deposits – 30% (9 of 30 basins) for

long inlet valley closed-basin lakes, and �26% (46 of 175 basins)for short inlet valley closed-basin lakes. The distribution of bothshort and long inlet valley closed-basin lakes with fan depositsshows a clear association with the dichotomy boundary (e.g.,Hauber et al., 2009, 2013; Di Achille and Hynek, 2010; Fig. 4B).

We find no convincing evidence for lacustrine shorelines or ter-races in our analysis; however, from a study of terrestrial LatePleistocene shore landforms, Irwin and Zimbelman (2012) havesuggested that shoreline features on Mars might not be readilyobservable from orbit, or be expected to survive for long periodsof time due to erosion from impact gardening and aeolian activity.Many of the basins we identify also contain large deposits of dark,aeolian dunes in their interior (e.g., Figs. 1 and 2A, C). These depos-its are very similar to dark dune deposits studied extensively byTirsch et al. (2011), who suggest they are young, mafic aeolianmaterial sourced from distinct dark layers exposed in crater wallsand floor.

4.2. Closed-basin lake morphometry

The morphometric parameters collected for our catalog ofclosed-basin lakes are shown in Table 1 and summarized inFig. 5 and Table 2. There does not appear to be any systematic

http://www.planetary.brown.edu/html_pages/data.htm

http://www.planetary.brown.edu/html_pages/data.htm

Fig. 4. Global distribution of closed-basin lakes. Background for both images is MOLA gridded topography overlain on MOLA-derived hillshade. (A) Distribution of closed-basin lakes by inlet valley length. Closed-basin lakes with short inlet valleys (e.g., Fig. 1) are shown by red dots and those with long inlet valleys (e.g., Fig. 2) are shown by bluedots. Open-basin lakes mapped by Fassett and Head (2008b) and Goudge et al. (2012) are shown by small, yellow dots. Labeled, dashed white outlines indicate highconcentrations of short inlet valley closed-basin lakes in Xanthe Terra (XT) and Arabia Terra (AT). (B) Distribution of closed-basin lakes with (colored dots; basins with shortinlet valleys are red, and basins with long inlet valleys are blue) and without (white dots) sedimentary fan deposits in their interior (e.g., Fig. 3). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)


difference in the distribution of these parameters for closed-basinlakes with short as opposed to long inlet valleys. Maximum lakeareas span the range �27–23,000 km2 (Fig. 5A; Table 2), and max-imum lake volumes span the range �1–24,000 km3 (Fig. 5B;Table 2). While this range in maximum lake areas and volumes isquite large, it is within the range of morphometric parametersfor open-basin lakes on Mars; however, there are a small handfulof open-basin lakes that are substantially larger (up to�500,000 km2 in area and �200,000 km3 in volume) than anybasins in our catalog (Fassett and Head, 2008b). From maximumlake area and volume, the mean depth (or in this case the maxi-mum mean depth) can be calculated as volume/area. Mean depthvalues for our catalog of basins span the range �30–1075 m(Fig. 5C; Table 2).

4.3. Closed-basin lake sedimentary deposit mineralogy

Of the 55 basins in our catalog with sedimentary fan deposits(Fig. 4B), 22 of them have CRISM coverage of the fan (Table 3).Only four of these 22 fans display evidence for hosting alterationminerals (e.g., Fig. 6), while the other 18 examples have a spectralsignature that is similar to the surrounding terrain. This observa-tion is likely to indicate either the presence of unaltered basalticcrustal material (e.g., Mustard et al., 2005) within the fan deposits,or that many of these fan deposits are partially covered by a spec-trally obscuring dust cover. Indeed, all four fan deposits containingalteration minerals are located in relatively dust free regions asmapped by the global dust cover index of Ruff and Christensen(2002) calculated using data from the Thermal Emission

Table 1Closed-basin lake catalog. Table lists: basin number (assigned here); location of each closed-basin lake; the inlet type, where S = short (e.g., Fig. 1) and L = long (e.g., Fig. 2); the degree of confidence in identification as a closed-basin lake;whether or not a fan is present at the inlet valley mouth (Y = yes, N = no); the degradation class of the fan deposits (Class 1 = least degraded, Class 3 = most degraded); the maximum elevation (above the Mars datum) to which the basincould have filled before breaching; the maximum lake area; the maximum lake volume; the host crater diameter (not collected for basins defined by multiple or elliptical craters); the mean depth of the maximum lake extent, which isthe maximum volume divided by the maximum area; and references for original basin identifications.

Basin#

Lat.(�N)

Lon.(�E)

Inlet Basinconfidence

Fan Fan degradationclass

Max. elev. of flooding(m)

Max. lake area(km2)

Max. lake vol.(km3)

Host crater diam.(km)

Mean depth [V/A](m)

Reference

1 36.0 �8.1 S Certain N �3860 388.1 112.4 23.3 289.6 Cabrol and Grin (1999)2 22.1 �8.2 S Certain Y 1 �3695 1883.4 1093.4 51.7 580.53 14.8 �51.7 S Certain N �2270 1353.0 441.1 – 326.0 Cabrol and Grin (1999)4 12.0 �52.7 S Certain Y 1 �1095 1093.5 309.1 40.0 282.7 Cabrol and Grin (1999)5 10.2 �16.7 S Certain N �2280 1076.9 287.5 40.3 266.9 Cabrol and Grin (1999)6 7.1 38.5 La Certain Y 2 �160 6332.4 1947.1 95.3 307.5 Forsythe and Blackwelder

(1998)7 5.3 �58.6 L Certain Y 1 �100 1003.5 659.3 37.2 657.0 Forsythe and Blackwelder

(1998)8 �3.3 88.3 L Certain N 755 942.8 93.2 40.6 98.8 Forsythe and Blackwelder


(1998)10 �5.9 �149.5 S Certain Y 2 �1730 1649.6 57.8 48.2 35.1 Cabrol and Grin (1999)11 �7.0 31.8 S Certain N 1705 877.1 378.0 32.9 431.0 Cabrol and Grin (1999)12 �7.8 25.3 S Probable N 1680 5934.6 4303.8 91.7 725.2 Cabrol and Grin (1999)13 �8.3 128.7 La Certain N 1350 5606.5 967.4 87.3 172.5 Forsythe and Blackwelder


(1998)15 �13.3 176.6 S Certain N �1395 581.5 147.9 32.6 254.4 Cabrol and Grin (1999)16 �13.8 142.5 S Certain N 1455 4866.8 2838.8 – 583.3 Forsythe and Blackwelder

(1998)17 �15.2 61.3 La Certain N 1330 404.8 16.7 32.1 41.2 Forsythe and Blackwelder

(1998)18 �18.8 59.2 La Certain N 705 4890.0 1259.6 83.3 257.6 Cabrol and Grin (1999)19 �19.4 52.1 L Probable N 1320 1579.7 267.1 46.1 169.1 Forsythe and Blackwelder

(1998)20 �20.2 172.0 S Certain N 235 895.1 108.5 40.1 121.2 Forsythe and Blackwelder

(1998)21 55.0 �84.4 S Probable N �3135 2252.8 2240.3 56.9 994.422 47.5 �68.7 L Certain N �1440 6606.9 1804.5 118.1 273.1 Cabrol and Grin (1999)23 44.2 �57.1 S Certain N �1985 91.4 20.1 12.5 219.524 41.1 �2.9 S Probable N �3935 27.4 3.4 6.3 125.125 38.0 38.0 S Probable N �1025 187.5 54.3 15.8 289.826 38.4 47.4 L Certain N �60 1107.9 213.0 – 192.227 37.9 54.7 S Certain N �585 209.2 33.5 15.7 160.028 37.1 12.5 S Probable N �2565 1336.1 1240.0 41.6 928.1 Ori et al. (2000)29 35.8 �12.1 S Possible N �3775 143.0 22.6 15.0 158.1 Cabrol and Grin (1999)30 35.7 �55.2 S Possible N �2125 479.0 79.7 26.8 166.431 36.0 �141.8 S Probable N �2415 725.2 592.2 32.2 816.532 34.7 �137.5 S Possible N �2330 982.4 375.7 38.0 382.533 34.7 �55.3 S Possible N �1665 199.7 13.4 17.8 67.234 33.3 �54.7 S Certain Y 1 �2095 790.7 139.2 35.1 176.135 33.1 �9.2 S Possible N �2985 532.9 194.8 26.0 365.636 34.4 3.2 S Certain Y 3 �2480 685.6 199.9 29.9 291.537 33.6 37.9 La Certain N �920 3588.9 2083.5 64.6 580.538 32.2 52.0 S Certain N �1045 2815.1 1168.1 66.5 414.939 32.3 40.1 S Certain N �430 236.0 61.2 18.8 259.440 31.6 �5.7 S Possible N �2760 1061.8 842.4 38.8 793.341 29.1 17.1 S Certain N �1975 227.2 62.3 17.6 274.3

352T.A

.Goudge

etal./Icarus

260(2015)

346–367

Table 1 (continued)

Basin#

Lat.(�N)

Lon.(�E)




Max. lake area(km2)

Max. lake vol.(km3)


Mean depth [V/A](m)

Reference

42 29.6 56.5 S Certain Y 3 �205 976.0 311.6 – 319.343 26.5 43.9 S Certain N �1195 1143.1 751.8 41.7 657.644 26.2 37.5 S Certain N �1525 1124.6 671.7 40.9 597.345 26.2 24.1 S Certain N �1595 375.4 155.3 23.0 413.846 27.1 21.4 S Certain Y 2 �1945 3806.0 3360.6 78.3 883.047 26.5 19.1 S Certain N �1855 518.6 140.2 26.5 270.448 26.1 10.0 S Certain N �2400 524.4 191.8 28.4 365.849 27.7 �124.5 S Certain Y 1 260 224.7 121.1 17.2 538.950 25.1 �97.5 L Certain Y 1 2090 460.9 234.1 – 507.9 Fassett and Head (2007)51 31.4 �13.0 S Certain Y 1 �4085 944.3 629.5 38.6 666.6 Cabrol and Grin (1999)52 26.3 �7.8 S Probable N �2480 2207.0 1279.8 55.5 579.953 23.8 42.0 S Certain Y 3 �725 493.3 83.5 30.3 169.254 21.4 58.1 S Probable Y 2 225 2610.5 2570.8 64.7 984.855 21.4 37.8 S Probable Y 2 �620 2060.9 1317.4 52.6 639.3 Forsythe and Blackwelder

(1998)56 20.4 �20.9 S Certain N �2840 489.4 210.7 26.7 430.557 18.1 �23.0 S Certain Y 1 �2895 381.3 108.1 22.8 283.558 19.2 �18.6 S Possible N �2450 76.2 8.6 12.1 113.359 17.7 3.8 S Probable N �1990 807.1 129.8 33.5 160.860 19.0 4.9 S Certain N �2140 1630.3 1023.5 49.6 627.861 19.9 12.0 S Certain Y 1 �2200 1496.1 1208.6 47.4 807.962 16.7 26.9 S Certain N �920 671.9 110.4 30.1 164.363 16.2 �53.2 S Certain Y 1 �1775 243.6 122.7 19.3 503.5 Hauber et al. (2013)64 14.3 �24.4 S Certain Y 1 �2795 812.3 326.7 32.7 402.2 Hauber et al. (2013)65 13.0 �14.2 S Possible N �2095 985.9 534.3 35.4 541.966 13.2 9.2 S Certain Y 1 �1815 1562.7 1178.4 49.0 754.167 12.8 56.4 S Certain N 600 342.9 43.9 22.5 128.268 10.6 39.5 La Certain N 450 454.9 15.0 30.3 33.069 10.9 �14.0 S Certain N �2370 5440.0 3361.0 92.1 617.870 12.5 �49.0 S Certain N �1935 166.0 43.2 14.8 260.071 11.6 �51.3 S Certain Y 1 �1600 509.8 226.5 28.3 444.3 Di Achille and Hynek (2010)72 8.4 �56.9 S Possible N �425 772.5 245.6 32.3 317.9 Forsythe and Blackwelder

(1998)73 8.2 �49.3 La Certain Y 1 �1805 2136.2 1408.6 68.4 659.4 Di Achille et al. (2006b)74 9.9 �46.6 S Certain Y 3 �2150 264.0 126.3 19.4 478.575 7.8 �39.1 S Certain Y 1 �2635 618.5 313.3 31.3 506.6 Hauber et al. (2013)76 8.0 �26.2 S Certain Y 1 �2405 272.9 41.3 20.1 151.377 8.5 �15.8 S Certain Y 1 �2160 3616.4 3288.6 67.2 909.4 Di Achille and Hynek (2010)78 5.8 107.4 S Certain N �1400 2608.5 776.9 67.7 297.879 7.1 106.8 S Certain N �645 3253.8 1673.2 72.6 514.280 5.0 28.2 S Probable N 45 619.7 86.0 30.5 138.881 5.1 �50.8 S Possible N 165 185.8 41.7 16.5 224.382 7.0 �53.6 S Certain N 115 1384.6 1313.8 43.7 948.983 2.4 �51.6 S Certain Y 1 �900 2293.1 1742.5 62.0 759.9 Hauber et al. (2009)84 4.1 �40.5 S Certain N �1510 255.7 72.8 18.6 284.785 3.5 �40.2 S Certain N �1390 403.9 72.7 23.8 180.086 4.0 �38.6 S Certain Y 1 �2140 508.9 144.6 25.3 284.187 3.1 35.2 S Probable N 300 1413.1 435.4 42.3 308.188 3.9 33.3 S Possible N 185 559.5 61.0 30.7 109.089 3.0 45.5 L Certain N 1315 885.9 113.2 36.2 127.890 3.2 101.4 S Possible N �525 40.2 1.3 9.1 32.491 3.7 113.5 S Certain N �95 781.2 113.5 – 145.392 1.5 116.3 S Certain Y 1 340 136.9 65.0 13.9 475.0

(continued on next page)

T.A.G

oudgeet

al./Icarus260

(2015)346–

367353

Table 1 (continued)

Basin#

Lat.(�N)

Lon.(�E)




Max. lake area(km2)

Max. lake vol.(km3)


Mean depth [V/A](m)

Reference

93 �1.7 �49.2 S Certain N 535 689.6 106.4 31.2 154.394 �1.4 �39.7 S Certain N �1140 2612.2 1063.5 63.4 407.195 �1.4 �36.9 S Certain Y 3 �1160 193.4 77.7 16.8 401.796 �2.9 67.7 S Probable N 1610 5940.4 3769.2 97.2 634.5 Forsythe and Blackwelder

(1998)97 �5.2 137.8 La Certain Y 1 �3035 7760.4 5455.2 153.5 702.9 Forsythe and Blackwelder

(1998)98 �6.3 40.6 S Possible N 1645 1113.1 384.2 36.9 345.299 �4.8 2.0 S Possible N �1200 2192.8 2257.9 54.7 1029.7

100 �6.3 �4.4 S Probable N �1080 135.8 24.7 14.0 181.6101 �9.3 �159.5 S Certain N �1460 801.3 155.9 36.0 194.6 Ori et al. (2000)102 �8.2 �159.5 S Certain Y 1 �1825 2799.1 1357.1 73.8 484.8 Ori et al. (2000)103 �10.0 �158.1 S Probable Y 3 �1030 3831.7 2013.3 75.7 525.4104 �9.5 �148.0 S Certain Y 3 �200 364.7 108.5 22.7 297.4105 �7.7 �146.6 S Certain Y 1 �590 756.2 296.8 32.9 392.5 Forsythe and Blackwelder

(1998)106 �10.0 �53.7 S Certain Y 1 2105 911.4 588.0 39.1 645.2 Di Achille and Hynek (2010)107 �9.6 �11.1 S Certain N �1820 429.2 85.3 – 198.8108 �8.2 3.0 S Possible N �1035 1097.9 209.3 39.5 190.6 Forsythe and Blackwelder

(1998)109 �7.8 30.4 S Possible N 2095 809.0 45.4 36.9 56.1 Cabrol and Grin (1999)110 �9.0 38.1 S Probable N 1580 23137.7 23781.3 180.4 1027.8111 �9.6 144.1 S Certain Y 2 �660 390.0 265.0 24.1 679.4112 �9.4 148.8 S Certain Y 3 �385 858.0 182.7 39.1 212.9 Forsythe and Blackwelder

(1998)113 �10.6 139.7 S Certain Y 3 805 7058.0 2023.1 109.0 286.6114 �10.8 136.4 S Probable N 1635 2154.7 1174.5 54.4 545.1115 �11.3 131.4 La Certain N 1990 5644.4 1060.8 – 187.9 Forsythe and Blackwelder

(1998)116 �11.5 124.6 S Certain N 2345 382.5 22.0 23.0 57.5117 �12.0 123.7 S Certain N 2075 3529.0 2607.9 67.3 739.0 Forsythe and Blackwelder

(1998)118 �10.8 91.2 S Probable N 2495 2466.1 2155.3 58.5 874.0119 �11.1 18.8 S Possible N 1665 1426.4 357.6 45.9 250.7 Forsythe and Blackwelder

(1998)120 �11.6 16.5 S Certain N 1245 1812.9 447.6 – 246.9 Forsythe and Blackwelder

(1998)121 �11.6 12.0 S Certain N 125 1557.1 402.5 45.6 258.5 Forsythe and Blackwelder

(1998)122 �12.2 �17.0 S Possible N �1475 304.8 69.7 20.1 228.7123 �12.1 �163.4 S Certain N �1090 540.4 175.1 29.4 324.0124 �13.7 6.9 S Possible N 245 774.7 383.7 32.3 495.3125 �13.9 40.4 S Possible N 2700 182.1 27.7 15.8 152.0126 �13.3 50.5 L Certain N 1930 848.5 184.8 40.9 217.8 Forsythe and Blackwelder

(1998)127 �13.9 78.5 S Probable N 1230 299.1 135.5 20.7 453.1128 �13.8 99.1 S Certain N 1740 339.9 32.8 21.8 96.5129 �13.4 124.9 S Certain N 1860 2822.6 1686.7 64.0 597.6 Forsythe and Blackwelder

(1998)130 �16.7 169.4 S Certain N �780 6190.0 2963.7 98.1 478.8131 �15.6 79.3 S Certain N 1150 131.0 7.1 15.1 54.2 Forsythe and Blackwelder

(1998)132 �16.2 45.9 L Certain N 2585 348.4 49.9 22.5 143.1

354T.A

.Goudge

etal./Icarus

260(2015)

346–367

Table 1 (continued)

Basin#

Lat.(�N)

Lon.(�E)




Max. lake area(km2)

Max. lake vol.(km3)


Mean depth [V/A](m)

Reference

133 �16.9 45.9 S Possible N 1840 644.3 339.2 32.8 526.5134 �17.5 44.9 S Certain N 2270 218.5 25.6 17.8 117.2135 �16.7 28.9 S Possible N 2275 1054.0 477.2 38.4 452.7136 �16.5 25.7 S Possible N 2105 395.2 60.3 24.2 152.5137 �16.9 �37.8 S Certain N 645 1753.4 1521.4 48.6 867.7138 �15.5 �155.5 S Certain Y 1 565 875.0 367.1 35.3 419.5 Ori et al. (2000)139 �19.5 �13.3 S Certain N �1175 1085.9 243.7 38.6 224.4140 �19.9 �2.9 S Possible N �215 3508.8 1283.1 68.3 365.7 Forsythe and Blackwelder

(1998)141 �19.6 46.2 S Certain N 1940 604.2 323.3 29.9 535.1142 �19.4 53.8 L Certain N 1275 593.8 215.8 – 363.4 Forsythe and Blackwelder

(1998)143 �19.7 75.7 La Certain N 445 859.7 166.9 33.9 194.1 Forsythe and Blackwelder

(1998)144 �18.2 79.8 S Certain N 705 542.3 168.6 27.7 310.9145 �19.6 84.3 S Possible N 420 1127.5 226.3 49.4 200.7146 �20.4 176.9 S Possible N 605 323.4 45.1 23.3 139.4 Forsythe and Blackwelder

(1998)147 �21.8 162.7 L Certain N 660 1492.6 120.3 – 80.6 Forsythe and Blackwelder

(1998)148 �20.5 47.4 S Possible N 1125 343.3 78.4 24.8 228.4 Forsythe and Blackwelder

(1998)149 �22.6 �56.6 S Certain N 1500 710.4 409.1 36.1 575.9150 �23.5 �169.1 S Possible N 1390 1350.3 826.8 40.5 612.3 Forsythe and Blackwelder

(1998)151 �24.2 �44.9 S Probable N 1440 419.6 142.7 24.7 340.1152 �25.1 �8.4 S Possible N 425 233.7 54.4 18.6 233.0153 �23.8 52.6 S Certain Y 3 1005 301.5 214.7 22.3 712.3154 �25.3 71.9 S Certain N �2535 5980.0 3174.6 91.3 530.9 Forsythe and Blackwelder

(1998)155 �24.4 86.6 S Certain N �595 43.5 10.6 8.8 244.0 Forsythe and Blackwelder

(1998)156 �27.7 141.5 S Possible N 1490 928.9 252.7 37.6 272.0 Cabrol and Grin (1999)157 �27.3 127.7 L Certain N 2135 266.2 67.6 19.9 254.1158 �25.7 97.5 L Certain Y 2 440 163.0 81.8 17.1 501.7159 �27.4 67.5 S Certain N �2320 202.2 63.2 18.1 312.5160 �25.8 �173.8 S Certain N 1270 274.9 21.2 21.6 77.1161 �28.6 �167.2 S Certain N 1945 1010.4 88.5 40.0 87.5162 �29.3 �32.1 S Probable N 140 1676.6 1438.4 49.8 857.9163 �28.8 �6.6 S Certain N 1000 1278.7 330.4 44.2 258.4 Forsythe and Blackwelder

(1998)164 �29.0 139.1 S Certain N 2175 717.1 109.3 34.0 152.3165 �31.5 128.3 L Probable N 1700 296.7 50.1 22.3 168.8166 �30.9 35.8 S Certain Y 2 1300 395.0 126.1 23.7 319.3167 �30.2 32.5 S Certain Y 3 1200 2475.0 1781.2 57.4 719.7168 �31.3 20.1 S Probable N 1905 3839.6 519.9 76.7 135.4169 �32.5 10.6 S Certain N 1195 874.4 208.9 38.0 238.9170 �31.4 �10.9 S Certain N 940 1402.2 1088.2 – 776.1171 �31.5 �20.3 S Certain N 200 241.5 54.2 18.5 224.4172 �32.0 �22.2 S Probable N 45 311.0 85.2 21.0 274.0173 �30.7 �102.9 S Certain Y 2 5835 1778.5 1244.7 49.6 699.8174 �33.8 �48.6 S Certain Y 3 705 3416.9 3425.2 67.0 1002.4175 �34.9 �48.1 S Certain N 1425 1787.8 1737.5 49.8 971.8

(continued on next page)

T.A.G

oudgeet

al./Icarus260

(2015)346–

367355

Table 1 (continued)

Basin#

Lat.(�N)

Lon.(�E)




Max. lake area(km2)

Max. lake vol.(km3)


Mean depth [V/A](m)

Reference

176 �33.2 �32.4 S Probable N 730 381.7 21.1 23.4 55.2177 �34.6 �31.9 S Certain N 205 1408.4 336.2 – 238.7178 �34.4 �13.0 S Certain N 805 795.9 305.8 34.3 384.2179 �34.3 15.2 S Certain N 1075 5167.5 5380.2 87.2 1041.2180 �34.2 18.2 S Certain N 1375 1942.5 818.2 54.8 421.2181 �33.9 45.3 S Certain N 850 226.0 54.8 19.4 242.6182 �33.6 48.1 S Certain Y 2 �145 185.9 23.5 19.9 126.1183 �37.7 37.5 S Certain N 800 506.2 164.5 26.7 325.0184 �37.4 10.5 S Certain N 1305 1207.5 301.5 41.2 249.7185 �36.4 �7.5 S Probable N 1085 1149.0 641.8 41.0 558.6186 �37.9 �69.1 S Certain N 2615 773.6 239.1 – 309.1187 �36.6 �72.8 S Certain N 2610 587.9 282.4 33.0 480.3 Forsythe and Blackwelder

(1998)188 �39.5 �12.6 S Probable N 970 1282.6 98.5 47.6 76.8189 �39.1 23.9 S Certain N 1100 3015.0 875.2 73.6 290.3190 �40.7 43.6 L Certain N 115 127.2 13.6 15.8 107.2191 �41.6 �0.5 S Probable N 985 1040.8 192.7 37.8 185.2192 �41.3 �62.5 S Certain Y 3 1305 1797.0 1484.1 52.3 825.9193 �41.4 �76.3 S Certain N 2510 2589.1 1268.1 65.8 489.8194 �42.5 �150.7 S Certain N 1945 487.1 88.9 – 182.6195 �45.3 122.8 S Probable N 1140 438.7 108.4 28.1 247.1 Forsythe and Blackwelder

(1998)196 �46.3 �163.8 S Certain N 1145 927.3 581.3 37.6 626.9197 �48.3 �13.2 S Possible N 1220 1370.3 405.0 40.1 295.6198 �54.4 �100.4 S Certain Y 3 2120 1143.2 1019.5 – 891.8199 22.1 66.8 S Certain Y 3 �675 115.3 34.5 17.8 299.1200 �9.9 144.5 L Certain Y 1 �655 2532.2 446.2 65.1 176.2 Forsythe and Blackwelder

(1998)201 �12.0 125.1 S Certain N 2190 1764.3 678.0 47.6 384.3202 �37.5 �158.8 S Certain Y 1 2825 59.1 15.0 10.9 254.6203 �24.0 �33.3 La Certain Y 1 �1010 1738.9 566.1 – 325.5 Malin and Edgett (2003)204 �26.0 �34.0 La Certain Y 1 �1230 13985.5 12557.7 148.0 897.9 Grant and Parker (2002)205 20.7 75.8 S Certain Y 1 �470 2095.8 2252.4 58.0 1074.7 Mangold et al. (2007)

a Basin inlet is classified as long, however, host crater rim is also incised by short inlet valley(s).

356T.A

.Goudge

etal./Icarus

260(2015)

346–367

Fig. 5. Stacked histograms of closed-basin lake parameters separated by inlet valley length (long inlet valley basins shown in green and short inlet valley basins shown inred). Maximum lake area (A), maximum lake volume (B), mean lake depth (C), host crater diameter (D), latitude (E), and longitude (F). (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

Table 2Summary of closed-basin lake morphometric parameters. Table lists mean, median,standard deviation, minimum and maximum values for the major morphometricparameters of the closed-basin lake catalog (Table 1; Fig. 5).

Maximum lakearea (km2)

Maximum lakevolume (km3)

Host craterdiameter (km)

Mean depth[V/A] (m)

Mean 1550 840 41 389Median 874 239 36 307Std. dev. 2325 2053 26 257Min. 27 1 6 32Max. 23,138 23,781 180 1075


Spectrometer (TES) instrument (Fig. 7). However, fan depositswithout identified alteration minerals are located in dusty and rel-atively dust-free regions in similar numbers (Fig. 7). Therefore,while spectrally obscuring dust cover is likely to partially explain

our observations, we conclude that many of the examined fandeposits are composed of unaltered detrital material transportedinto the basin from the surrounding terrain. This conclusion is alsoconsistent with observations of sedimentary deposits withinopen-basin lakes (Goudge et al., 2012).

For the four fan deposits that contain alteration minerals(Table 3), three contain Fe/Mg-smectite (e.g., Fig. 6A) and one con-tains hydrated silica (Fig. 6B). Fe/Mg-smectite is identified by thepresence of narrow, vibrational absorptions at �1.4, 1.9 and2.3 lm (Clark et al., 1990; Bishop et al., 1999, 2002; Frost et al.,2002; Fig. 6C), although the CRISM spectra are also consistent withthe spectral signature of vermiculite (Clark et al., 1990). The�1.4 lm absorption is due to the first overtone of OH stretch, the�1.9 lm absorption is due to a combination tone of H–O–H bendand OH stretch, and the �2.3 lm absorption is due to a combina-tion tone of metal–OH bend and OH stretch (Clark et al., 1990;Bishop et al., 1999, 2002; Frost et al., 2002). The precise locationof the �2.3 lm absorption is related to the dominant cation in

Table 3Sedimentary fan deposits within closed-basin lakes (e.g., Fig. 3) that have overlapping CRISM coverage. Table lists: basin number; location of closed-basin lake; the inlet type,short (e.g., Fig. 1) or long (e.g., Fig. 2); the analyzed CRISM observation(s) covering the sedimentary fan deposit; what, if any, alteration mineral is detected within the fan deposit;and references for the alteration mineral detections.

Basin#

Lat.(�N)

Lon.(�E)

Inlettype

CRISM observation ID Alteration mineraldetected

Reference(s)

4 12.0 �52.7 Short HRL0000AA4B None6 7.1 38.5 Long FRT000183B1 None7 5.3 �58.6 Long HRL0001B67D None

49 27.7 �124.5 Short FRS0002C8CF None50 25.1 �97.5 Long FRT0000C360, FRT00008D6D, FRT00003621 None53 23.8 42.0 Short FRS00029E7A None55 21.4 37.8 Short FRS0002BB46, FRS0002CA26, FRS0002EB5B None63 16.2 �53.2 Short HRL0001900C None66 13.2 9.2 Short HRL0000BF52 None71 11.6 �51.3 Short FRT0000B0EC None76 8.0 �26.2 Short FRS00029DEE None83 2.4 �51.6 Short HRL0000985E, HRL0000927F Hydrated silica Popa et al. (2010); Carter et al. (2012); Hauber

et al. (2013); Wray et al. (2013)92 1.5 116.3 Short FRS00029E0C None

102 �8.2 �159.5 Short HRL0000C01C, FRT0000BCBD None103 �10.0 �158.1 Short HRL0000C577, FRS00029F40, FRS0002BD0F None105 �7.7 �146.6 Short FRT00018F1D, FRT0001BB1F None106 �10.0 �53.7 Short FRT00018FCF, HRL000080CB Fe/Mg-smectite Le Deit et al. (2012)138 �15.5 �155.5 Short FRT0000C165 None173 �30.7 �102.9 Short FRS00029DFB, FRS0002BBDF None203 �24.0 �33.3 Long FRT000060DD, FRT00009C06, HRS00002EF9,

HRS00003207Fe/Mg-smectite Milliken and Bish (2010)

204 �26.0 �34.0 Long FRT0000C1D1, HRS000030AF Fe/Mg-smectite Grant et al. (2008); Milliken and Bish (2010)205 20.7 75.8 Short FRS0002A9B2, FRS0002ADC4 None


the octahedral sheet of the smectite present. More Mg-rich smec-tites, such as saponite, have absorptions closer to �2.3 lm, whilemore Fe-rich smectites, such as nontronite, have absorptions closerto �2.29 lm (Clark et al., 1990; Bishop et al., 1999, 2002; Frostet al., 2002). Distinguishing between these two band center posi-tions can be difficult with CRISM data, and so we take a conserva-tive approach and refer to the broader classification ofFe/Mg-smectite, as is common for analysis of CRISM data (e.g.,Mustard et al., 2008; Ehlmann et al., 2009).

Hydrated silica is identified by the presence of vibrationalabsorption features at �1.4, 1.9 and �2.2 lm (Fig. 6D). The�1.4 lm absorption is due to the first overtone of OH stretch, the�1.9 lm absorption is due to a combination tone of H–O–H bendand OH stretch, and the �2.2 lm absorption is due to a combina-tion tone of Si–OH bend and OH stretch (Wu, 1980; Stolper,1982; Rice et al., 2013). Hydrated silica is distinguished fromAl-rich phyllosilicates, which also have an absorption at �2.2 lm,by the breadth of the �2.2 lm absorption, which is much largerfor hydrated silica than Al-rich phyllosilicates (Wu, 1980; Stolper,1982; Clark et al., 1990; Rice et al., 2013). We do, however, identifythe presence of an Al-rich phyllosilicate, such as kaolinite or mont-morillonite, in a small exposure of the wall of the inlet valley feed-ing this fan (Fig. 6B, black arrow). Al-rich phyllosilicate is identifiedby the narrower �2.2 lm absorption (Fig. 6D) caused by a combi-nation tone of Al–OH bend and OH stretch, in addition to absorp-tions at �1.4 and 1.9 lm (Clark et al., 1990).

Of our 205 closed-basin lakes, only two basins, numbers 16and 168 (Table 1), contain chloride deposits mapped byOsterloo et al. (2010). In these locations, the chloride depositsare located in isolated, patchy regions on the floors of the basins,and show no clear evidence for association with the fluvialand/or lacustrine activity associated with the formation of theinlet valleys (e.g., Fig. 8). Additionally, the large number of basinslocated in relatively dust-free regions of Mars (Fig. 7) suggeststhat dust cover is not likely to wholly explain the lack of chlo-rides mapped by Osterloo et al. (2010) contained within basinsin our catalog; however, dust cover is likely to be contributingto this result.

5. Discussion

5.1. Distribution of closed-basin lakes

Our updated catalog of closed-basin lakes shows that these fea-tures are far more widespread than previously mapped by Cabroland Grin (1999). The majority (�85%) of closed-basin lakes in ourcatalog are fed by short (<20 km) inlet valleys (Fig. 4A), an observa-tion also noted by Forsythe and Blackwelder (1998). Closed-basinlakes on Mars also have a broadly similar distribution toopen-basin lakes (Fassett and Head, 2008b; Goudge et al., 2012;Fig. 4A).

5.1.1. Closed-basin lakes in Arabia Terra and Xanthe Terra – links togroundwater activity?

One noteworthy exception to the similarity in distribution ofclosed-basin lakes and open-basin lakes (Fassett and Head,2008b; Goudge et al., 2012) is the clustering of closed-basin lakeswith short inlet valleys in the Arabia Terra and Xanthe Terraregions (Fig. 4A). Both of these regions have a relative paucity ofopen-basin lakes, while short inlet valley closed-basin lakes areabundant. The high concentration of fluvially incised craters inXanthe Terra has previously been noted by Hauber et al. (2009,2013).

Given the morphology of the short inlet valleys feeding theseclosed-basin lakes (Fig. 1), and in particular the steep,amphitheater-shaped headwalls, it is possible that these valleysformed from groundwater sapping as proposed for other martianfluvial valleys (Laity and Malin, 1985; Malin and Carr, 1999;Harrison and Grimm, 2005). The concentration of short inletclosed-basin lakes in Xanthe Terra and Arabia Terra may in factbe explained by modeling of global groundwater flow.Andrews-Hanna et al. (2007) has shown that Xanthe Terra andArabia Terra could potentially concentrate shallow groundwaterflow and upwelling due to the ‘bench’ topography along the dichot-omy boundary at these locations. Additionally, Warner et al. (2010)identified a suite of paleolake basins with outlet valleys but no

Fig. 6. Example alteration mineral detections in sedimentary fan deposits at the mouths of closed-basin lake inlet valleys (white arrows). North is up in both images. (A)Sedimentary fan deposit within a closed-basin lake at �10.0�N, �53.7�E (basin number 106; Table 1). Note the detections of Fe/Mg-smectite in the fan (yellow arrows) andshort inlet valley walls (black arrows). CRISM-derived parameter map (R = BD1900R, G = BD2200, B = D2300; Pelkey et al., 2007; Ehlmann et al., 2009) from observationHRL000080CB overlain on a portion of CTX image P12_005583_1700. Interpretive legend based on parameters shown in upper right. (B) Sedimentary fan deposit within aclosed-basin lake at 2.4�N, �51.6�E (basin number 83; Table 1). Note the detections of hydrated silica in the fan (yellow arrow), and Fe/Mg-smectite (orange arrow) and Al-rich phyllosilicate (black arrow) in the short inlet valley walls. CRISM-derived parameter map (R = BD1900R, G = BD2200, B = D2300; Pelkey et al., 2007; Ehlmann et al., 2009)from observation HRL0000985E overlain on a portion of CTX image P06_003539_1825. Interpretive legend based on parameters shown in upper right. (C) Top plot showsCRISM ratioed spectra of Fe/Mg-smectite detections from the fan deposit (blue spectra #3 and #4) and inlet valley wall (red spectra #1 and #2) shown in part (A). Numeratorspectra are extracted from CRISM image HRL000080CB at the locations shown by the numbered arrows in part (A). Bottom plot shows library spectra of saponite andnontronite from the USGS spectral library (Clark et al., 2007). Saponite spectrum is sample SapCa-1 and nontronite spectrum is sample NG-1.a. Dashed lines in both plots arelocated at �1.4, 1.92, and 2.3 lm. (D) Top plot shows CRISM ratioed spectra of aqueous alteration mineral detections from the fan deposit (green spectrum #3; hydratedsilica) and inlet valley walls (red spectrum #2 is Fe/Mg-smectite, and purple spectrum #1 is Al-rich phyllosilicate) shown in part (B). Numerator spectra are extracted fromCRISM image HRL0000985E at the locations shown by the numbered arrows in part (B). Bottom plot shows library spectra of kaolinite, montmorillonite, saponite, nontronite,and hydrated silica (opal) from the USGS spectral library (Clark et al., 2007). Kaolinite spectrum is sample KGa-1, montmorillonite spectrum is sample CM27, saponitespectrum is sample SapCa-1, nontronite spectrum is sample NG-1.a, and hydrated silica spectrum is sample TM8896. Dashed lines in both plots are located at �1.4, 1.92, 2.2,and 2.3 lm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


Fig. 7. Distribution of closed-basin lakes compared to the TES dust cover index of Ruff and Christensen (2002). Basins containing sedimentary fan deposits with CRISMcoverage (see Table 3) are shown in color. Red dots indicate fans with no identified alteration minerals, and green dots indicate fans that contain identified alterationminerals. Basins that do not contain sedimentary fan deposits or that contain sedimentary fan deposits without CRISM coverage are shown by white dots. Background is theTES dust cover index of Ruff and Christensen (2002) overlain on MOLA-derived hillshade. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)


inlet valleys in Xanthe Terra and Arabia Terra that they hypothe-sized were fed primarily by groundwater flow.

Forsythe and Blackwelder (1998) previously suggested thatclosed-basin lakes on Mars are predominantly sourced by ground-water flow. These authors modeled a scenario where impact cra-ters intersected a regional groundwater table, which thenexperienced drawdown from evaporation off the lake surface.This drawdown in the groundwater table, or base level, resultedin the incision of the inlet valleys via groundwater sapping. Thiswork relies on the major assumption that there was a regional,near-surface groundwater table in existence when each of theimpact craters that hosts a closed-basin lake formed; however,the validity of this assumption remains quite uncertain.Furthermore, the Forsythe and Blackwelder (1998) model resultsin conical (i.e., axisymmetric) drawdown of the regional ground-water table centered on the basin, and so it is not clear why thiswould result in the incision of a single inlet valley, as is oftenobserved in our catalog of short inlet valley closed-basin lakes(e.g., Fig. 1).

When considering whether short inlet valley closed-basin lakeson Mars are sourced by groundwater activity, it is informative toturn to terrestrial analogs. The Lonar Crater lake presents a partic-ularly interesting analog, as it is a hydrologically closed basin con-tained within the Lonar impact crater (Komatsu et al., 2014), a�2 km diameter impact crater formed within the Deccan basalttraps (Maloof et al., 2010). A detailed study of the hydrology ofthe Lonar Crater lake by Komatsu et al. (2014) showed that it issourced by both fluvial surface runoff and groundwater flow.Lonar Crater has one major inlet valley that breaches the craterrim, the Dhar valley, which extends <1 km into the surroundingterrain (Komatsu et al., 2014) and is similar in morphology to theshort inlet valleys observed in our catalog of closed-basin lakes(Fig. 1). While the exact scenario by which the Dhar valley was ini-tiated is unclear, Komatsu et al. (2014) hypothesize that incision ofthis valley is likely to be related to concentrated gully activity onthe interior crater rim sourced from surface runoff. Presently thereare groundwater springs in the Dhar valley walls, however majorsediment transport within the valley occurs only during the rainy

season associated with intense surface runoff (Komatsu et al.,2014). Therefore, it seems likely that incision of the Dhar inlet val-ley is primarily linked to surface runoff as opposed to groundwatersapping, although groundwater springs are an important source ofinflow for maintaining the lake in the basin interior (Komatsu et al.,2014).

Additional evidence arguing against the formation of short inletvalleys by groundwater comes from Lamb et al. (2006) who havehypothesized that groundwater sapping processes do not havethe erosive capacity to form bedrock valleys. Recent work byLamb et al. (2008, 2014) has also shown that rapid overland floodsmay be the primary mechanism for creating amphitheater-shapedheadwalls in bedrock valleys.

Therefore, we suggest that the source for the fluvial activityassociated with the short inlet valleys in our catalog ofclosed-basin lakes is not groundwater sapping, but is instead likelyto be overland flow, perhaps related to regional, large-scale flood-ing (e.g., Lamb et al., 2008, 2014); however, as at the Lonar Craterlake (Komatsu et al., 2014), groundwater inflow may have beenimportant in maintaining the water balance for any lakes thatdeveloped within these basins (e.g., Forsythe and Blackwelder,1998). This conclusion is also generally consistent with the workof Hauber et al. (2013), who showed that groundwater flow is unli-kely to be the major source of water for the formation of a numberof fan deposits in Xanthe Terra.

Although we conclude that the primary source of influx to themajority of closed-basin lakes in our catalog is likely to be overlandflow and not groundwater flow, this does not meangroundwater-fed closed-basin lakes did not form on Mars. In fact,recent work has suggested the presence of both hydrologicallyclosed (e.g., Wray et al., 2011; Michalski et al., 2013) and hydrolog-ically open (e.g., Warner et al., 2010) groundwater-fed paleolakes,which may have played an important role in the hydrologic evolu-tion of Mars.

5.1.2. Distribution of sedimentary fan depositsApproximately equal proportions of long (30%) and short

(�26%) inlet valley closed-basin lakes have sedimentary fan

Fig. 8. Example chloride deposit within a closed-basin lake at �13.8�N, 142.5�E(basin number 16; Table 1) fed by a short inlet valley (white arrow). North is up inboth images. (A) Overview of chloride detections mapped by Osterloo et al. (2010),indicated by purple boxes. Black box indicates the location of part (B). Mosaic ofHigh Resolution Imaging Science Experiment (HiRISE; McEwen et al., 2007) imageESP_034084_1655 and CTX image B18_016665_1641. (B) Detailed view of Osterlooet al. (2010) chloride detection. Note the isolated light-toned patches (purplearrows) that likely correspond to the chloride-bearing material mapped by Osterlooet al. (2010). Portion of HiRISE image ESP_034084_1655. (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version ofthis article.)


deposits in their interior, and these basins tend to cluster towardthe dichotomy boundary (e.g., Hauber et al., 2009, 2013; DiAchille and Hynek, 2010; Fig. 4B). One potential explanation forthis observation is preservation bias. Under this scenario, a muchlarger proportion of closed-basin lakes in our catalog may oncehave contained fan deposits; however, primarily fan depositsfound along the dichotomy boundary have been sufficiently resis-tant to erosion and modification to be preserved in the observablemorphologic record.

An alternative hypothesis is that sediment availability is par-tially controlled by the long wavelength topography of Mars, which

is generally directed toward the dichotomy boundary (e.g., De Hon,2010). This hypothesis would suggest that, on average, sedimentavailability for transport and deposition in terminal basins ishigher toward regional topographic lows (Fig. 4B). Sedimentshould be produced across the martian surface from a variety ofmechanisms, including eruption of volcaniclastic sediment (e.g.,tephra) and emplacement of ejecta from the formation of impactcraters (e.g., Grotzinger et al., 2011, 2013). Therefore, the hypoth-esis that sediment availability is higher near the dichotomy bound-ary requires a mechanism to preferentially transport sedimenttoward these margins. Fluvial activity, such as during the majorperiod of valley network formation (Howard et al., 2005; Irwinet al., 2005; Fassett and Head, 2008a), is a prime candidate for sucha mechanism, as it would tend to transport sediment downslope.Phillips et al. (2001) and Howard et al. (2005) have also shown thatthe long wavelength topography of Mars exerts a strong control onthe orientation of valley networks, which may result in higher sed-iment availability toward the northern lowlands basin.

5.2. Closed-basin lake morphometry

When analyzing the morphometric characteristic of our catalogof closed-basin lakes (Fig. 5), it is important to keep in mind thatthese parameters are calculated from the maximum possibleextent of the lake without overtopping, and therefore, a lake aslarge as implied by these morphometric parameters may not haveexisted. These morphometric parameters should instead be viewedas a maximum upper limit and only an approximation; however,maximum lake volume estimates provide an upper limit on thetotal volume of water for any single episode of continuous fluvialinflux to these basins. Unless water losses from groundwaterand/or evaporation were of a comparable magnitude to fluvialinputs, larger volumes of water than the measured upper limit vol-ume would have caused the basin to overtop and form an outletbreach.

Calculations of maximum lake volumes for our catalog allowsfor the calculation of a total, maximum reservoir of water thatthese basins may once have represented. Summing all of our calcu-lated maximum lake volumes yields a total volume of�1.7 � 105 km3, equivalent to a �1.2 m deep layer of water spreadevenly across the surface of Mars, or a global equivalent layer(GEL). For comparison, open-basin lakes on Mars represent a totalvolume of �3 m GEL (Fassett and Head, 2008b), the polar layereddeposits represent a volume of �22–32 m GEL (Smith et al.,1999), non-polar glacial landforms are estimated to contain a vol-ume of �0.9–2.6 m GEL (Levy et al., 2014), and the total currentsurface/near-surface water ice reservoir is estimated to be �34 mGEL (Carr and Head, 2015).

Therefore, if all of the closed-basin lakes in our catalog werefilled to their maximum extent simultaneously, they would repre-sent an appreciable, but minor, component of the total modernmartian water ice reservoir. However, our catalog of closed-basinlakes excludes any hydrologically closed paleolakes that may haveformed in tectonic basins associated with Valles Marineris (e.g.,Lucchitta et al., 1994; Di Achille et al., 2006a, 2007; Harrison andChapman, 2008, 2010; Lucchitta, 2010) or primarily from ground-water influx (e.g., Wray et al., 2011; Michalski et al., 2013). As such,our calculated total maximum reservoir of water for closed-basinlakes may be an underestimate. In particular, given the size ofthe tectonic basins associated with Valles Marineris (e.g.,Lucchitta et al., 1994; Di Achille et al., 2006a, 2007; Harrison andChapman, 2008, 2010; Lucchitta, 2010; Warner et al., 2013), suchpaleolakes may be volumetrically significant if they were filledcompletely, and may have constituted a large component of themartian water reservoir when they formed.


When looking at the frequency distribution of our maximumlake areas, it is particularly interesting how close it appears to alog-normal distribution (Fig. 5A). On Earth, both predicted(Wetzel, 2001) and observed (Downing et al., 2006; McDonaldet al., 2012) lake area cumulative frequency distributions follownegative power-law trends quite closely, where lakes are increas-ingly more abundant at smaller sizes. This is because processesthat create lake basins tend to make more small basins, such asthrough glacial damming or kettle lake formation, compared tolarge basins, such as through tectonic rifting (Wetzel, 2001).Large open-basin lakes on Mars (i.e., >�500 km2 in area) also havea cumulative frequency distribution described by a negativepower-law (Cabrol and Grin, 2010). The cumulative frequency dis-tribution of maximum lake area for our catalog of closed-basinlakes, however, does not conform to a negative power law trend,falling off at both small and large lake areas (Fig. 9).

Fig. 9. Cumulative frequency distribution of maximum lake areas, along with abest-fit power law trend.

This deviation is unexpected from a basin-forming-processstandpoint, as impact crater formation also follows a negativepower-law trend (e.g., Neukum et al., 1975; Ivanov, 2001).Instead, we suggest that for small basins a combination of searchmethods and erosion are the cause for this deviation. Our globalsearch was performed on grid cells with dimensions on the orderof 100s of kilometers, and so we think it is entirely possible thatsome of the very smallest closed-basin lakes were not detectedin our analysis at the map scale used for the initial survey.Additionally, small crater removal is widely cited as a significantprocess on Mars (e.g., Hartmann, 1971; Hartmann and Neukum,2001; Smith et al., 2008), and so this may also be responsible forthe removal of some of the smallest crater-containedclosed-basin lakes. Erosion and backwasting of crater rims andinterior walls (e.g., Craddock et al., 1997) may also have been sig-nificant enough to remove perceptible evidence of incised inlet val-leys, particularly for basins fed by short inlets. Cabrol and Grin(2010) also hypothesized that open-basin lakes on Mars deviatefrom a negative power-law cumulative frequency distribution atsmall basin sizes due to erosional processes.

It is more difficult to reconcile the lower-than-expected numberof large closed-basin lakes on Mars, as this is not observed for lakeson Earth (Downing et al., 2006; McDonald et al., 2012) oropen-basin lakes on Mars (Cabrol and Grin, 2010). One explanationwe propose is that this observation may be related to the age of thehost craters. Larger craters are preferentially older, and thus shouldbe more degraded (e.g., Craddock et al., 1997; Mangold et al.,

2012), which may obscure the identification of inlet fluvial valleys,in particular the short inlet valleys that make up the majority ofour catalog (Figs. 1 and 4; Table 1).

5.3. Closed-basin lake sedimentary deposit mineralogy

Our CRISM analyses indicate that many of the sedimentary fandeposits in closed-basin lake interiors are composed of unaltered,detrital martian crustal material, similar to observations foropen-basin lake sedimentary deposits (Goudge et al., 2012). Inaddition to these deposits, four fans contain confidently identifiedalteration minerals, including three detections of Fe/Mg-smectiteand one of hydrated silica (Table 3; Fig. 6). All the detections ofFe/Mg-smectite have been previously reported (Table 3), and theyinclude detections within the well-exposed fan deposits of theHolden and Eberswalde crater paleolakes (Grant et al., 2008;Milliken and Bish, 2010). The sedimentary Fe/Mg-smectite withinboth the Holden and Eberswalde basins has been hypothesizedby Milliken and Bish (2010) to be detrital in origin, with CRISM sig-natures of Fe/Mg-smectite in the catchment area of these paleo-lakes matching those observed within the fan deposits. It is alsoclear that the third detection of Fe/Mg-smectite is detrital in origin– there are large exposures of Fe/Mg-smectite in the walls of theinlet valley to this basin that match the spectral signature of thesmall exposures of Fe/Mg-smectite in the fan (Le Deit et al.,2012; Fig. 6A and C). Le Deit et al. (2012) also report the identifica-tion of small exposures of Al-rich phyllosilicate along the marginsof the inlet valley wall of this basin; however, there is no evidencefor exposures of similar Al-rich phyllosilicate within the fandeposit itself.

The hydrated silica detection shown in Fig. 6B has also beenreported previously (Popa et al., 2010; Carter et al., 2012; Hauberet al., 2013; Wray et al., 2013), although the origin of this mineralphase is not completely clear. The inlet valley walls of this basinhave small exposures of Fe/Mg-smectite (Fig. 6B, orange arrow)and an Al-rich phyllosilicate (Fig. 6B, black arrow), interpreted byWray et al. (2013) to be kaolinite. Wray et al. (2013) also report apotential exposure of plagioclase feldspar in the inlet valley wall.However, there are no clear exposures of hydrated silica in theCRISM image, aside from within the fan deposit. It is therefore pos-sible that the hydrated silica observed within this fan is authigenicin nature. Silica is an expected weathering product from the leach-ing of basalt and is predicted to be abundant on Mars, particularly insedimentary deposits (McLennan, 2003), and previous identifica-tions of hydrated silica in alluvial fans on Mars have been hypoth-esized as authigenic in origin (Carter et al., 2013). Alternatively, itis possible that the source hydrated silica, if detrital, lies outsideof the region of CRISM coverage. While definitively identifying thehydrated silica as authigenic may prove difficult from orbit, thisfan deposit certainly warrants further detailed study.

Regardless, on the basis of the CRISM analysis of these sedimen-tary fan deposits, it appears that major authigenic alteration min-eral formation in our catalog of closed-basin lakes on Mars wasrare, if it occurred at all. This is also supported by the observationthat chloride deposits, which have been widely identified acrossthe martian surface (Osterloo et al., 2008, 2010; Glotch et al.,2010), are found in only two basin interiors. In both of theseinstances, the chloride deposits are not located in the topographi-cally lowest portion of the basin, as would be expected for saltdeposits formed from lacustrine evaporative concentration (e.g.,Jones and Van Denburgh, 1966; Eugster and Hardie, 1978;Eugster, 1980). Instead, these deposits are located in isolatedpatches on the basin floors, and there is no clear evidence that theyare related to the fluvial activity that formed the inlet valley, norany lacustrine activity that occurred within the basin (e.g., Fig. 8).


Recent work by Hynek et al. (2015) and Osterloo and Hynek(2015) has hypothesized that a small number of the largest chlo-ride deposits identified by Osterloo et al. (2010) may be relatedto fluvial activity; however, the setting for these deposits is oftenin broad, inter-crater depressions, and not crater interiors(Osterloo and Hynek, 2015). Additionally, the deposit studied indetail by Hynek et al. (2015) is actually in an open-basin lake, asopposed to a hydrologically closed basin. While it is possible thatthese deposits are related to fluvial activity (Hynek et al., 2015;Osterloo and Hynek, 2015), the two chloride deposits containedin our catalog of closed-basin lakes have a very different geologiccontext (e.g., Fig. 8). These latter deposits may instead be expo-sures of more widespread, older chloride deposits underlying thecrater floor, or small deposits formed through atmospheric interac-tions, such as efflorescence (Osterloo et al., 2008, 2010).

Authigenic alteration mineral formation is common on Earthwithin highly saline and alkaline lakes (e.g., Jones and Weir,1983; Jones, 1986; Deocampo et al., 2009; Bristow and Milliken,2011), conditions typical of hydrologically closed basins (e.g.,Jones and Van Denburgh, 1966; Eugster and Hardie, 1978;Eugster and Jones, 1979; Eugster, 1980). When evaporation offthe lake surface is larger than the inflow, evaporative concentra-tion within hydrologically closed lakes causes the buildup of dis-solved solutes within the standing water of these closed basins(e.g., Jones and Van Denburgh, 1966; Eugster and Hardie, 1978;Eugster and Jones, 1979; Eugster, 1980). This can lead to cationsubstitution within phyllosilicates (e.g., Jones and Weir, 1983;Jones, 1986; Deocampo et al., 2009; Bristow and Milliken, 2011),authigenic production of phyllosilicates (e.g., Tettenhorst andMoore, 1978; Pozo and Casas, 1999; Bristow and Milliken, 2011),and, perhaps most commonly, the precipitation of evaporite salts(e.g., chlorides, sulfates, carbonates; Eugster and Hardie, 1978;Eugster and Jones, 1979; Eugster, 1980).

Therefore, when considering closed-basin lakes on Mars, onemight expect to observe similar authigenic alteration mineralsand evaporite salt deposits in their interiors (e.g., Forsythe andZimbelman, 1995); however, we find no evidence for widespreadformation of such deposits within our catalog of closed-basin lakes.We hypothesize that the lack of widespread evaporite deposits inour catalog of closed-basins indicates a general transience of anystanding body of water that formed within these basins. Any lakethat formed was not in existence for long enough to undergo suf-ficient evaporative concentration to form authigenic alterationminerals or salt deposits in its interior. This conclusion is also con-sistent with the results of Hauber et al. (2013) from a regionalstudy of fan deposits surrounding Chryse Planitia.

Although our catalog of closed-basin lakes on Mars does notshow definitive evidence for widespread authigenic alterationmineral formation, it is likely that our catalog does not containall hydrologically closed paleolake basins, as we only includedbasins defined by impact craters with incised inlet valleys.Recent work by Wray et al. (2011) and Michalski et al. (2013)has suggested authigenic alteration mineral formation in potentialgroundwater-fed closed-basin lakes. Wray et al. (2011) hypothe-sized that authigenic sulfates and kaolinite were formed in a pale-olake within Columbus crater, and possibly in nearbyhydrologically closed basins. Michalski et al. (2013) proposed agroundwater-fed paleolake within McLaughlin crater, whichshows possible evidence for authigenic carbonate formation.While these craters may have contained hydrologically closedpaleolakes, aside from the mineralogic evidence, it is difficult toconclude that fluvial activity occurred at these sites.

Many of the tectonic basins associated with Valles Marinerismay also have hosted hydrologically closed paleolakes (e.g.,Lucchitta et al., 1994; Di Achille et al., 2006a, 2007; Harrison andChapman, 2008, 2010; Lucchitta, 2010). Several of these basins

contain massive, interior-layered deposits of sulfates (e.g.,Gendrin et al., 2005; Bishop et al., 2009; Murchie et al., 2009b),which may be related to lacustrine evaporite deposition (Gendrinet al., 2005; Murchie et al., 2009b; Andrews-Hanna et al., 2010;Lucchitta, 2010); however, alternative mechanisms for the forma-tion of these sulfate deposits exist (e.g., Lucchitta, 2010; Michalskiand Niles, 2012), and their association with lacustrine activity isnot definitive.

Finally, our catalog does not include lacustrine deposits identi-fied in situ, such as the sulfate-rich Burns formation of MeridianiPlanum, which has been investigated by the Opportunity MarsExploration Rover (MER) (e.g., Squyres et al., 2004; Grotzingeret al., 2005; McLennan et al., 2005). Analysis of the sedimentologyof the Burns formation indicates deposition in an interdune playaenvironment (Grotzinger et al., 2005), and geochemical analysesfurther suggest authigenic sulfate formation (McLennan et al.,2005). Although there is in situ evidence for a playa lacustrine envi-ronment and sulfate authigenesis in the Burns formation (Squyreset al., 2004; Grotzinger et al., 2005; McLennan et al., 2005), there isminimal orbital evidence for the original interdune basin(s) withinwhich this deposit formed.

It is clear that our catalog of closed-basin lakes excludes a num-ber of basins that may exhibit evidence for authigenic alterationmineral formation (e.g., Squyres et al., 2004; Gendrin et al., 2005;McLennan et al., 2005; Murchie et al., 2009b; Andrews-Hannaet al., 2010; Wray et al., 2011; Michalski et al., 2013). However,when considering solely our catalog of closed-basin lakes onMars, there is a lack of major evidence for widespread authigenicalteration mineral formation. This result is particularly interestingwhen considering the geochemical environment of hydrologicallyclosed lakes on Earth (e.g., Jones and Van Denburgh, 1966;Eugster and Hardie, 1978; Eugster and Jones, 1979; Eugster,1980). We conclude that this observation indicates that any paleo-lake that formed within our catalog of basins was generally shortlived, and did not experience significant evaporative concentration.

5.4. Placing MSL results from Gale crater in global context

Our results provide the first catalog of closed-basin lakes com-piled using the wealth of high-resolution image and topographicdata made available in the last decade and a half (Cabrol andGrin, 1999). This catalog offers global-scale context for studyingGale crater (basin 97 in our catalog; Table 1; Fig. 10), long proposedas a closed-basin lake on the basis of its morphology as observedfrom orbit (Cabrol et al., 1999; Irwin et al., 2005), and the site ofexploration of the MSL Curiosity rover. Gale crater is a particularlyinteresting case from our catalog of closed-basin lakes, as it is fedby both a long inlet valley that breaches the southwestern craterrim (Fig. 10, red arrows) and two short inlet valleys that breachthe northern and northwestern crater rim (Fig. 10, white arrows),one of 12 basins fed by both long and short inlet valleys(Table 1). Additionally, Gale crater contains multiple fan depositsin its interior (Irwin et al., 2005; Anderson and Bell, 2010; LeDeit et al., 2013; Palucis et al., 2014; Fig. 10).

In its �3 years of exploration, MSL has found convincing in situevidence of both fluvial (Williams et al., 2013) and lacustrine activ-ity (Grotzinger et al., 2014) within Gale crater. These in situ resultslend confidence to the hypothesis that the features we have stud-ied here are associated with fluvial activity, and that they may alsobe associated with lacustrine activity. Our results also indicate thatGale crater, while a very fascinating site for exploration, is notlikely to be unique in its preservation of sedimentological evidencefor fluvial and lacustrine activity on the surface of Mars.

Examining Gale crater within the context of our results, it isimportant to consider whether the MSL results are related to flu-vial and lacustrine activity fed by the short inlet valleys or the long

Fig. 10. Closed-basin lake within Gale crater at �5.2�N, 137.8�E (basin number 97; Table 1). Basin is fed by both short inlet valleys (white arrows) and a long inlet valley (redarrows). The approximate MSL landing site is indicated by the yellow star in the northern portion of the crater floor. MOLA gridded topography overlain on a mosaic of CTXimages P04_002530_1745 and P22_009716_1773, and the THEMIS � 100 m/pixel global daytime infrared mosaic. North is up. (For interpretation of the references to color inthis figure legend, the reader is referred to the web version of this article.)


inlet valley. Given the location of the MSL exploration site in Galecrater (Fig. 10, yellow star), a testable hypothesis from in situ anal-yses is that both the fluvial activity indicated by the conglomeratedeposits on the crater floor (Williams et al., 2013) and the lacus-trine activity identified by MSL operations at Yellowknife Bay(Grotzinger et al., 2014) are related to the short inlet valleys inthe north and northwest of the crater. If shown to be correct, thishypothesis would imply that the majority of our catalog representsstrong possible candidates for both fluvial and lacustrine activityon the martian surface. As MSL traverses up the stratigraphic sec-tion toward Mt. Sharp, future analyses may be able to test thishypotheses by putting chrono-stratigraphic constraints on whenthe periods of surface aqueous activity (fluvial and lacustrine)occurred, and whether they are more likely to be linked to theshort or long inlet valleys that feed Gale crater. Such new resultscan then readily be placed into global context with our catalog ofclosed-basin lakes.

Additionally, any constraints from MSL analyses on the depth ofthe lake contained within Gale crater will have important implica-tions for how the catalog of closed-basin lakes presented here isinterpreted in the future. We have calculated a mean depth of�700 m for the closed-basin lake within Gale crater (Table 1).This mean depth is calculated for the maximum possible extentof a hydrologically closed lake within Gale crater and so it is

unclear how representative of a true lake depth this is; however,this depth is roughly consistent with previous estimates of paleo-lake depth in Gale crater using orbital remote sensing data (e.g.,Dietrich et al., 2013; Le Deit et al., 2013). Lake depth estimatesfrom MSL would provide in situ evidence for how close the lakewithin Gale crater got to the maximum topographic extent allow-able while remaining a closed basin (i.e., an elevation of �3035 mabove the Mars datum; Table 1). Such new results could then inturn be used to reassess our catalog of closed-basin lakes, and inparticular the volume estimates for these basins. As MSL continuesits exploration of Gale crater new in situ results can both be putinto a more global context using our catalog of closed-basin lakesand provide further understanding of our catalog.

6. Conclusions

We have presented a new catalog of 205 candidate closed-basinlakes contained within impact craters on Mars. While the lack of anoutlet valley associated with these basins means it is not possibleto confidently state whether or not they ever contained a standingbody of water, the presence of an inlet valley breaching the craterrim implies at least some period of fluvial activity at these sites.From analyses of the distribution, morphology, morphometry,


and mineralogy of these potential paleolakes and associated sedi-mentary fan deposits, we conclude the following:

1. High-resolution global mapping shows candidateclosed-basin lakes are far more widespread than previously rec-ognized using lower-resolution Viking data (Cabrol and Grin,1999).

2. Candidate closed-basin lakes can be divided into two majorclasses: those with short (<20 km) inlet valleys (e.g., Fig. 1)and those with long (>20 km) inlet valleys (e.g., Fig. 2).Candidate closed-basin lakes with short inlet valleys make upthe majority of the catalog (175; �85%) compared to basinswith long inlet valleys (30; �15%).

3. Candidate closed-basin lakes have approximately the same geo-graphic distribution as open-basin lakes (Fassett and Head,2008b; Goudge et al., 2012); however, candidate closed-basinlakes with short inlet valleys tend to cluster in the ArabiaTerra and Xanthe Terra regions (Fig. 4A).

a. This may be explained by predicted global groundwaterflow paths that would concentrate upwelling and shallowflow in these regions (Andrews-Hanna et al., 2007), whichmay result in groundwater-fed paleolakes (e.g., Warneret al., 2010). This hypothesis may also be consistent withthe amphitheater-shaped headwalls of the short inlet val-leys (Laity and Malin, 1985; Malin and Carr, 1999;Harrison and Grimm, 2005).

b. However, groundwater sapping may not be able to formbedrock valleys (Lamb et al., 2006), and we suggest theinlet valleys are instead more likely to have formed fromoverland flow, perhaps related to regional flooding (Lambet al., 2008, 2014). This is also consistent with work fromthe terrestrial analog Lonar Crater closed-basin lake, wheresurface runoff is hypothesized to be the dominant driver ofinlet valley incision (Komatsu et al., 2014).

4. Our catalog of candidate closed-basin lakes represent amaximum volume of water of �1.7 � 105 km3, or �1.2 m GEL,if all of the basins were filled to their maximum lake levelsimultaneously. This is an appreciable, although relativelysmall, component of the modern surface/near-surface reservoirof water ice on Mars, estimated to be �34 m GEL (Carr andHead, 2015).

5. Our catalog includes 55 candidate closed-basin lakes with sed-imentary fan deposits in their interiors (e.g., Fig. 3). Nine ofthese deposits are within basins fed by long inlet valleys, and46 are within basins fed by short inlet valleys, which representapproximately equal proportions of the total number of thesetwo classes of candidate closed-basin lakes.

6. Candidate closed-basin lakes with sedimentary fan deposits intheir interior cluster along the dichotomy boundary (e.g.,Hauber et al., 2009, 2013; Di Achille and Hynek, 2010;Fig. 4B). This may suggest either a preservation bias or that sed-iment availability is higher in these regions.

7. The majority of sedimentary fan deposits in candidateclosed-basin lake interiors are composed of detrital material,either altered or unaltered, consistent with previous work onboth closed- (Milliken and Bish, 2010) and open-basin lakes(Goudge et al., 2012).

8. The lack of widespread evaporite deposits within our catalog ofcandidate closed-basin lakes suggests that any lakes formedwithin these basins were too transient to undergo major evap-orative concentration.

9. Our catalog of candidate closed-basin lakes offers a broadercontext for the interpretation of results from MSL explorationof Gale crater, which is a closed-basin lake (Cabrol et al.,1999; Irwin et al., 2005) within our catalog. Gale crater is also

fed by both short and long inlet valleys, and future MSL resultswill be important in assessing the relationship between thesetwo types of valleys and paleolake activity within the basin.

Acknowledgments

The authors thank Nicholas Warner and Laetitia Le Deit for con-structive reviews that helped to improve the quality of this manu-script, and Jeffrey Johnson for editorial handling. We also thank J.L.Dickson for help with image and topographic data processing. Weexpress our appreciation for the superb work of the NASA MROproject team and the CRISM Science Operations Center (SOC).TAG gratefully acknowledges support from the Natural Sciencesand Engineering Research Council of Canada (NSERC)Postgraduate Scholarships Program (PGSD3-421594-2012). JFMand TAG acknowledge support of CRISM analysis through a sub-contract from the Johns Hopkins University Applied Physics Lab.We also acknowledge support from the Mars Data AnalysisProgram through grant NNX11AI81G to JWH, and for membershipon the ESA Mars Express High Resolution Stereo Camera (HRSC)Team through grant JPL-1488322 to JWH.

References

Anderson, R.B., Bell, J.F., 2010. Geologic mapping and characterization of Gale craterand implications for its potential as a Mars Science Laboratory landing site.Mars 5, 76–128.

Andrews-Hanna, J.C., Phillips, R.J., Zuber, M.T., 2007. Meridiani Planum and theglobal hydrology of Mars. Nature 446, 163–166.

Andrews-Hanna, J.C. et al., 2010. Early Mars hydrology: Meridiani playa depositsand the sedimentary record of Arabia Terra. J. Geophys. Res. 115, E06002.

Banfield, D., Donelan, M., Cavaleri, L., 2015. Winds, waves and shorelines fromancient martian seas. Icarus 250, 368–383.

Barlow, N.G., 1988. Crater size–frequency distributions and a revised martianrelative chronology. Icarus 75, 285–305.

Bishop, J.L. et al., 1999. Visible, Mossbauer and infrared spectroscopy ofdioctahedral smectites Sampor, SWy-1 and SWa-1. Proc. Int. Clay Conf. 11,413–419.

Bishop, J. et al., 2002. The influence of structural Fe, Al and Mg on the infrared OHbands in spectra of dioctahedral smectites. Clay Miner. 37, 607–616.

Bishop, J.L. et al., 2009. Mineralogy of Juventae Chasma: Sulfates in the light-tonedmounds, mafic minerals in the bedrock, and hydrated silica and hydroxylatedferric sulfate on the plateau. J. Geophys. Res. 114, E00D09.

Bristow, T.F., Milliken, R.E., 2011. Terrestrial perspective on authigenic clay mineralproduction in ancient martian lakes. Clays Clay Miner. 59, 339–358.

Cabrol, N.A., Grin, E.A., 1999. Distribution, classification, and ages of martian impactcrater lakes. Icarus 142, 160–172.

Cabrol, N.A., Grin, E.A., 2001. The evolution of lacustrine environments on Mars: IsMars only hydrologically dormant? Icarus 149, 291–328.

Cabrol, N.A., Grin, E.A., 2010. Searching for lakes on Mars: Four decades ofexploration. In: Cabrol, N.A., Grin, E.A. (Eds.), Lakes on Mars. Elsevier, Oxford, pp.1–29.

Cabrol, N.A. et al., 1999. Hydrogeologic evolution of Gale crater and its relevance tothe exobiological exploration of Mars. Icarus 139, 235–245.

Carr, M.H., 1987. Water on Mars. Nature 326, 30–35.Carr, M.H., Head, J.W., 2015. Martian surface/near-surface water inventory: Sources,

sinks, and changes with time. Geophys. Res. Lett. 42, 726–732.Carter, J. et al., 2012. Composition of alluvial fans and deltas on Mars. Lunar Planet.

Sci. XLIII. Abstract 1978.Carter, J. et al., 2013. Hydrous minerals on Mars as seen by the CRISM and OMEGA

imaging spectrometers: Updated global view. J. Geophys. Res. 118, 831–858.Christensen, P.R. et al., 2004. The Thermal Emission Imaging System (THEMIS) for

the Mars 2001 Odyssey mission. Space Sci. Rev. 110, 85–130.Clark, R.N. et al., 1990. High spectral resolution reflectance spectroscopy of

minerals. J. Geophys. Res. 95, 12653–12680.Clark, R.N. et al., 2007. USGS digital spectral library splib06a. U.S. Geol. Surv., Digital

Data Series, 231.Cohen, A.S., 2003. Paleolimnology: The History and Evolution of Lake Systems.

Oxford University Press, Oxford, 500pp.Craddock, R.A., Maxwell, T.A., Howard, A.D., 1997. Crater morphometry and

modification in the Sinus Sabaeus and Margaritifer Sinus regions of Mars. J.Geophys. Res. 102, 13321–13340.

De Hon, R.A., 2010. Hydrologic provinces of Mars: Physiographic controls ondrainage and ponding. In: Cabrol, N.A., Grin, E.A. (Eds.), Lakes on Mars. Elsevier,Oxford, pp. 69–89.

http://refhub.elsevier.com/S0019-1035(15)00331-0/h0005


















































Deocampo, D.M. et al., 2009. Saline lake diagenesis as revealed by coupledmineralogy and geochemistry of multiple ultrafine clay phases: PlioceneOlduvai Gorge, Tanzania. Am. J. Sci. 309, 834–868.

Di Achille, G., Hynek, B.M., 2010. Deltas and valley networks on Mars: Implicationsfor a global hydrosphere. In: Cabrol, N.A., Grin, E.A. (Eds.), Lakes on Mars.Elsevier, Oxford, pp. 223–248.

Di Achille, G. et al., 2006a. A steep fan at Coprates Catena, Valles Marineris, Mars, asseen by HRSC data. Geophys. Res. Lett. 33, L07204.

Di Achille, G. et al., 2006b. Geological evolution of the Tyras Vallis paleolacustrinesystem, Mars. J. Geophys. Res. 111, E04003.

Di Achille, G., Ori, G.G., Reiss, D., 2007. Evidence for late Hesperian lacustrineactivity in Shalbatana Vallis, Mars. J. Geophys. Res. 112, E07007.

Dietrich, W.E. et al., 2013. Topographic evidence for lakes in Gale crater. LunarPlanet. Sci. XLIV. Abstract 1844.

Downing, J.A. et al., 2006. The global abundance and size distribution of lakes,ponds, and impoundments. Limnol. Oceanogr. 51, 2388–2397.

Edwards, C.S. et al., 2011. Mosaicking of global planetary image datasets: 1.Techniques and data processing for Thermal Emission Imaging System(THEMIS) multi-spectral data. J. Geophys. Res. 116, E10008.

Ehlmann, B.L. et al., 2009. Identification of hydrated silicate minerals on Mars usingMRO-CRISM: Geologic context near Nili Fossae and implications for aqueousalteration. J. Geophys. Res. 114, E00D08.

Eugster, H.P., 1980. Geochemistry of evaporitic lacustrine deposits. Annu. Rev. EarthPlanet. Sci. 8, 35–63.

Eugster, H.P., Hardie, L.A., 1978. Saline lakes. In: Lerman, A. (Ed.), Lakes: Chemistry,Geology, Physics. Springer-Verlag, New York, pp. 237–293.

Eugster, H.P., Jones, B.F., 1979. Behavior of major solutes during closed-basin brineevolution. Am. J. Sci. 279, 609–631.

Fassett, C.I., Head, J.W., 2007. Valley formation of martian volcanoes in theHesperian: Evidence for melting of summit snowpack, caldera lake formation,drainage and erosion on Ceraunius Tholus. Icarus 189, 118–135.

Fassett, C.I., Head, J.W., 2008a. The timing of martian valley network activity:Constraints from buffered crater counting. Icarus 195, 61–89.

Fassett, C.I., Head, J.W., 2008b. Valley network-fed, open-basin lakes on Mars;Distribution and implications for Noachian surface and subsurface hydrology.Icarus 198, 37–56.

Forsythe, R.D., Blackwelder, C.R., 1998. Closed drainage crater basins of the martianhighlands: Constraints on the early martian hydrologic cycle. J. Geophys. Res.103, 31421–31431.

Forsythe, R.D., Zimbelman, J.R., 1995. A case for ancient evaporite basins on Mars. J.Geophys. Res. 100, 5553–5563.

Frost, R.L., Kloprogge, J.T., Ding, Z., 2002. Near-infrared spectroscopic study ofnontronites and ferruginous smectite. Spectrochim. Acta A 58, 1657–1668.

Gendrin, A. et al., 2005. Sulfates in martian layered terrains: The OMEGA/MarsExpress view. Science 307, 1587–1591.

Glotch, T.D. et al., 2010. Distribution and formation of chlorides and phyllosilicatesin Terra Sirenum, Mars. Geophys. Res. Lett. 37, L16202.

Goldspiel, J.M., Squyres, S.W., 1991. Ancient aqueous sedimentation on Mars. Icarus89, 392–410.

Goudge, T.A. et al., 2012. An analysis of open-basin lake deposits on Mars: Evidencefor the nature of associated lacustrine deposits and post-lacustrine modificationprocesses. Icarus 219, 211–229.

Grant, J.A., Parker, T.J., 2002. Drainage evolution in the Margaritifer Sinus region,Mars. J. Geophys. Res. 107 (E9), 5066.

Grant, J.A. et al., 2008. HiRISE imaging of impact megabreccia and sub-meteraqueous strata in Holden crater, Mars. Geology 36, 195–198.

Grotzinger, J.P. et al., 2005. Stratigraphy and sedimentology of a dry to wet eoliandepositional system, Burns formation, Meridiani Planum, Mars. Earth Planet.Sci. Lett. 240, 11–72.

Grotzinger, J. et al., 2011. Mars sedimentary geology: Key concepts and outstandingquestions. Astrobiology 11, 77–87.

Grotzinger, J.P. et al., 2013. Sedimentary processes on Earth, Mars, Titan, and Venus.In: Mackwell, S.J., Simon-Miller, A.A., Harder, J.W., Bullock, M.A. (Eds.),Comparative Climatology of Terrestrial Planets. University of Arizona Press,Tucson, pp. 439–472.

Grotzinger, J.P. et al., 2014. A habitable fluvio-lacustrine environment atYellowknife Bay, Gale crater, Mars. Science 343, 1242777-1–1242777-14.

Gwinner, K. et al., 2010. Topography of Mars from global mapping by HRSC high-resolution digital terrain models and orthoimages: Characteristics andperformance. Earth Planet. Sci. Lett. 294, 506–519.

Harrison, K.P., Chapman, M.G., 2008. Evidence for ponding and catastrophic floodsin central Valles Marineris, Mars. Icarus 198, 351–364.

Harrison, K.P., Chapman, M.G., 2010. Episodic ponding and outburst floodingassociated with chaotic terrains in Valles Marineris. In: Cabrol, N.A., Grin, E.A.(Eds.), Lakes on Mars. Elsevier, Oxford, pp. 163–194.

Harrison, K.P., Grimm, R.E., 2005. Groundwater-controlled valley networks and thedecline of surface runoff on early Mars. J. Geophys. Res. 110, E12S16.

Hartmann, W.K., 1966. Martian cratering. Icarus 5, 565–576.Hartmann, W.K., 1971. Martian cratering III: Theory of crater obliteration. Icarus 15,

410–428.Hartmann, W.K., Neukum, G., 2001. Cratering chronology and the evolution of Mars.

Space Sci. Rev. 96, 165–194.Hauber, R. et al., 2009. Sedimentary deposits in Xanthe Terra: Implications for the

ancient climate on Mars. Planet. Space Sci. 57, 944–957.Hauber, E. et al., 2013. Asynchronous formation of Hesperian and Amazonian-aged

deltas on Mars and implications for climate. J. Geophys. Res. 118, 1529–1544.

Head, J.W. et al., 2003. Recent ice ages on Mars. Nature 426, 797–802.Howard, A.D., Moore, J.M., Irwin, R.P., 2005. An intense terminal epoch of

widespread fluvial activity on early Mars: 1. Valley network incision andassociated deposits. J. Geophys. Res. 110, E12S14.

Hynek, B.M., Beach, M., Hoke, M.R.T., 2010. Updated global map of martian valleynetworks and implications for climate and hydrologic processes. J. Geophys.Res. 115, E09008.

Hynek, B.M., Osterloo, M.K., Kierein-Young, K.S., 2015. Late stage formation ofmartian chloride salts through ponding and evaporation. Lunar Planet. Sci. XLVI.Abstract 1045.

Irwin, R.P., Zimbelman, J.R., 2012. Morphometry of Great Basin pluvial shorelandforms: Implications for paleolake basins on Mars. J. Geophys. Res. 117,E07004.

Irwin, R.P. et al., 2005. An intense terminal epoch of widespread fluvial activity onearly Mars: 2. Increased runoff and paleolake development. J. Geophys. Res. 110,E12S15.

Ivanov, B.A., 2001. Mars/Moon cratering rate ratio estimates. Space Sci. Rev. 96, 87–104.

Jones, B.F., 1986. Clay mineral diagenesis in lacustrine sediments. In: Mumpton, F.(Ed.), Studies in Diagenesis, US Geol. Surv. Bulletin 1578. US Geological Survey,Reston, pp. 291–300.

Jones, B.F., Van Denburgh, A.S., 1966. Geochemical influences on the chemicalcharacter of closed lakes. Proc. Int. Assoc. Sci. Hydrol. Symp. Garda 70, 435–446.

Jones, B.F., Weir, A.H., 1983. Clay minerals of Lake Abert, an alkaline, saline lake.Clays Clay Miner. 31, 161–172.

Komatsu, G. et al., 2014. Drainage systems of Lonar Crater, India: Contributions toLonar Lake hydrology and crater degradation. Planet. Space Sci. 95, 45–55.

Laity, J.E., Malin, M.C., 1985. Sapping processes and the development of theater-headedvalley networks on the Colorado Plateau. Geol. Soc. Am. Bull. 96, 203–217.

Lamb, M.P. et al., 2006. Can springs cut canyons into rock? J. Geophys. Res. 111,E07002.

Lamb, M.P. et al., 2008. Formation of Box Canyon, Idaho, by megaflood: Implicationsfor seepage erosion on Earth and Mars. Science 320, 1067–1070.

Lamb, M.P., Mackey, B.H., Farley, K.A., 2014. Amphitheater-headed canyons formedby megaflooding at Malad Gorge, Idaho. Proc. Natl. Acad. Sci. 111, 57–62.

Le Deit, L. et al., 2012. Extensive surface pedogenic alteration of the martianNoachian crust suggested by plateau phyllosilicates around Valles Marineris. J.Geophys. Res. 117, E00J05.

Le Deit, L. et al., 2013. Sequence of infilling events in Gale crater, Mars: Results frommorphology, stratigraphy, and mineralogy. J. Geophys. Res. 118, 2439–2473.

Levy, J.S. et al., 2014. Sequestered glacial ice contribution to the global martianwater budget: Geometric constraints on the volume of remnant, midlatitudedebris-covered glaciers. J. Geophys. Res. 199, 2188–2196.

Lucchitta, B.K., 2010. Lakes in Valles Marineris. In: Cabrol, N.A., Grin, E.A. (Eds.),Lakes on Mars. Elsevier, Oxford, pp. 111–161.

Lucchitta, B.K., Isbell, N.K., Howington-Kraus, A., 1994. Topography of VallesMarineris: Implications for erosional and structural history. J. Geophys. Res.99, 3783–3798.

Malin, M.C., Carr, M.H., 1999. Groundwater formation of martian valleys. Nature397, 589–591.

Malin, M.C., Edgett, K.S., 2003. Evidence for persistent flow and aqueoussedimentation on early Mars. Science 302, 1931–1934.

Malin, M.C. et al., 2007. Context Camera investigation on board the MarsReconnaissance Orbiter. J. Geophys. Res. 112, E05S04.

Maloof, A.C. et al., 2010. Geology of Lonar Crater, India. Geol. Soc. Am. Bull. 122,109–126.

Mangold, N. et al., 2007. Mineralogy of the Nili Fossae region with OMEGA/MarsExpress data: 2. Aqueous alteration of the crust. J. Geophys. Res. 112, E08S04.

Mangold, N. et al., 2012. A chronology of early Mars climatic evolution from impactcrater degradation. J. Geophys. Res. 117, E04003.

McDonald, C.P. et al., 2012. The regional abundance and size distribution of lakesand reservoirs in the United States and implications for estimates of global lakeextent. Limnol. Oceanogr. 57, 597–606.

McEwen, A.S. et al., 2007. Mars Reconnaissance Orbiter’s High Resolution ImagingScience Experiment (HiRISE). J. Geophys. Res. 112, E05S02.

McGuire, P.C. et al., 2009. An improvement to the volcano-scan algorithm foratmospheric correction of CRISM and OMEGA spectral data. Planet. Space Sci.57, 809–815.

McLennan, S.M., 2003. Sedimentary silica on Mars. Geology 31, 315–318.McLennan, S.M. et al., 2005. Provenance and diagenesis of the evaporite-bearing

Burns formation, Meridiani Planum, Mars. Earth Planet. Sci. Lett. 240, 95–121.Michalski, J., Niles, P.B., 2012. Atmospheric origin of martian interior layered

deposits: Links to climate change and the global sulfur cycle. Geology 40, 419–422.

Michalski, J.R. et al., 2013. Groundwater activity on Mars and implications for adeep biosphere. Nat. Geosci. 6, 133–138.

Milliken, R.E., Bish, D.L., 2010. Sources and sinks of clay minerals on Mars. Philos.Mag. 90, 2293–2308.

Murchie, S. et al., 2007. Compact Reconnaissance Imaging Spectrometer for Mars(CRISM) on Mars Reconnaissance Orbiter (MRO). J. Geophys. Res. 112, E05S03.

Murchie, S.L. et al., 2009a. Compact Reconnaissance Imaging Spectrometer for Marsinvestigation and data set from the Mars Reconnaissance Orbiter’s primaryscience phase. J. Geophys. Res. 114, E00D07.

Murchie, S. et al., 2009b. Evidence for the origin of layered deposits in CandorChasma, Mars, from mineral composition and hydrologic modeling. J. Geophys.Res. 114, E00D05.






































































































































































Mustard, J.F. et al., 2005. Olivine and pyroxene diversity in the crust of Mars. Science307, 1594–1597.

Mustard, J.F. et al., 2008. Hydrated silicate minerals on Mars observed by the MarsReconnaissance Orbiter CRISM instrument. Nature 454, 305–309.

Neukum, G. et al., 1975. Cratering in the Earth–Moon system: Consequences for agedetermination by crater counting. Proc. Lunar Sci. Conf. 6, 2597–2620.

Neukum, G. et al., 2004. HRSC: The High Resolution Stereo Camera of Mars Express.Eur. Space Agency Spec. Publ. ESA SP-1240, pp. 17–35.

Ori, G.G., Marinangeli, L., Baliva, A., 2000. Terraces and Gilbert-type deltas in craterlakes in Ismenius Lacus and Memnonia (Mars). J. Geophys. Res. 105, 17629–17641.

Osterloo, M.M., Hynek, B.M., 2015. Martian chloride deposits: The last gasps ofwidespread surface water. Lunar Planet. Sci. XLVI. Abstract 1054.

Osterloo, M.M. et al., 2008. Chloride-bearing materials in the southern highlands ofMars. Science 319, 1651–1654.

Osterloo, M.M. et al., 2010. Geologic context of proposed chloride-bearing materialson Mars. J. Geophys. Res. 115, E10012.

Palucis, M.C. et al., 2014. The origin and evolution of the Peace Vallis fan system thatdrains to the Curiosity landing area, Gale crater, Mars. J. Geophys. Res. 119,705–728.

Pelkey, S.M. et al., 2007. CRISM multispectral summary products: Parameterizingmineral diversity on Mars from reflectance. J. Geophys. Res. 112, E08S14.

Phillips, R.J. et al., 2001. Ancient geodynamics and global-scale hydrology on Mars.Science 291, 2587–2591.

Pieri, D.C., 1980. Martian valleys: Morphology, distribution, age, and origin. Science210, 895–897.

Pike, R.J., 1971. Genetic implications of the shapes of martian and lunar craters.Icarus 15, 384–395.

Popa, C., Esposito, F., Colangeli, L., 2010. New landing site proposal for Mars ScienceLaboratory (MSL) in Xanthe Terra. Lunar Planet. Sci. XLI. Abstract 1807.

Pozo, M., Casas, J., 1999. Origin of kerolite and associated Mg clays in palustrine–lacustrine environments. The Esquivias deposit (Neogene Madrid Basin, Spain).Clay Miner. 34, 395–418.

Rice, M.S. et al., 2013. Reflectance spectra diversity of silica-rich materials:Sensitivity to environment and implications for detections on Mars. Icarus223, 499–533.

Robbins, S.J., Hynek, B.M., 2012a. A new global database of Mars impact craters P1km: 1. Database creation, properties, and parameters. J. Geophys. Res. 117,E05004.

Robbins, S.J., Hynek, B.M., 2012b. A new global database of Mars impact craters P1km: 2. Global crater properties and regional variations of the simple-to-complex transition diameter. J. Geophys. Res. 117, E06001.

Ruff, S.W., Christensen, P.R., 2002. Bright and dark regions on Mars: Particle size andmineralogical characteristics based on Thermal Emission Spectrometer data. J.Geophys. Res. 107 (E12), 5127.

Smith, D.E. et al., 1999. The global topography of Mars and implications for surfaceevolution. Science 284, 1495–1503.

Smith, D.E. et al., 2001. Mars Orbiter Laser Altimeter: Experiment summary afterthe first year of global mapping of Mars. J. Geophys. Res. 106, 23689–23722.

Smith, M.R., Gillespie, A.R., Montgomery, D.R., 2008. Effects of crater obliteration oncrater-count chronologies for martian surfaces. Geophys. Res. Lett. 35, L10202.

Squyres, S.W. et al., 2004. In situ evidence for an ancient aqueous environment atMeridiani Planum, Mars. Science 306, 1709–1714.

Stolper, E., 1982. Water in silicate glasses: An infrared spectroscopic study. Contrib.Miner. Petrol. 81, 1–17.

Strom, R.G., Croft, S.K., Barlow, N.G., 1992. The martian impact cratering record. In:Kieffer, H.H. et al. (Eds.), Mars. The University of Arizona Press, Tucson, pp. 383–423.

Tettenhorst, R., Moore, G.E., 1978. Stevensite oolites from the Green River formationof central Utah. J. Sediment. Petrol. 48, 587–594.

Tirsch, D. et al., 2011. Dark aeolian sediment in martian craters: Composition andsources. J. Geophys. Res. 116, E03002.

Viviano-Beck, C.E. et al., 2014. Revised CRISM spectral parameters and summaryproducts based on the currently detected mineral diversity on Mars. J. Geophys.Res. 119, 1403–1431.

Warner, N. et al., 2010. Late Noachian to Hesperian climate change on Mars:Evidence of episodic warming from transient crater lakes near Ares Valles. J.Geophys. Res. 115, E06013.

Warner, N.H. et al., 2013. Fill and spill of giant lakes in the eastern Valles Marinerisregion of Mars. Geology 41, 675–678.

Wetzel, R.G., 2001. Limnology: Lake and River Ecosystems, third ed. ElsevierAcademic Press, Sand Diego, 1006pp.

Williams, R.M.E. et al., 2013. Martian fluvial conglomerates at Gale crater. Science340, 1068–1072.

Wray, J.J. et al., 2011. Columbus crater and other possible groundwater-fedpaleolakes of Terra Sirenum, Mars. J. Geophys. Res. 116, E01001.

Wray, J.J. et al., 2013. Prolonged magmatic activity on Mars inferred from thedetection of felsic rocks. Nat. Geosci. 6, 1013–1017.

Wu, C.-K., 1980. Nature of incorporated water in hydrated silicate glasses. J. Am.Ceram. Soc. 63, 453–457.













































































Date post:	21-Jun-2020
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Classification and analysis of candidate impact crater ... · Classiﬁcation and analysis of...

Documents