 CRATER FLOOR POLYGONS (CFPs): SIGNS OF DESICCATED PALEOLAKES ON MARS?  M. R. El Maarry1, W. Markiewicz1, M. Mellon2, and W. Goetz1 .  1MPI für Sonnensystemforschung,  37191KatlenburgLindau, Germany, elmaarry@mps.mpg.de, markiewvicz@mps.mpg.de, goetz@mps.mpg.de.  2Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, CO 80309 USA, mellon@lasp.colorado.edu.  Introduction:  Crater-floor polygons have been first identified in high resolution images taken with Mars Orbital Camera (MOC) [1,2]. Crater-floor polygons (CFPs) show a range from 15 to 350 meters in width, and have variable shapes, with mostly an orthogonal pattern of intersecting troughs. This cracking pattern resembles, albeit on a larger scale, those of terrestrial desiccated lakes such as Coyote and Guano Lakes in the USA [3]. Despite this, these features have occasionally been interpreted as periglacial features on the grounds of their morphological similarity to thermal contraction and ice/sand-wedge polygons in addition to their "apparent" high latitude distribution. Using a larger updated database of high resolution images (Fig. 1), we carried out a thorough mapping of these features. An analytical model is used for the first time to show that under current climatic conditions, such polygonal systems can not form through seasonal thermal stresses alone. Complementing this analysis with geomophological investigation points to a desiccation processes as the most probable formation mechansim. Distribution & morphology:  Fig. 2 shows the locations of the craters that have been observed in this study (>250 craters) draped on a Mars MOLA terrain map. The northern lowlands show an apparent high latitudinal preference. In the case of the southern highlands, however, locations seem to be more scattered. Interesting to note is that the features extend down to the equator in both hemispheres. In addition, clustering of the features is evident in many localities. Notable examples are Nili Fossae, and Utopia in the north, and Hellas basin in the south. High resolution images show that the polygons form two distinct size groups: the main or primary troughs form polygons through their intersections that range in size from 15 to 350 meters in diameter, with a mean global size of 120 meters with a standard deviation of 40 meters. The secondary troughs; not always present, are narrower and embedded in the larger polygons dividing them into smaller ones. These fractures tend to intersect in a more random manner. The smaller polygons range in size from 4 to 20 meters, and as such; resemble more closely other thermal contraction (or ice/sand wedge) polygons in the Martian circumpolar regions. CFPs are found almost exclusively in the crater floors and less frequently on crater walls and associated with ejecta deposits. Depth to diameter ratios of variable sized impact craters containing CFPs are usually lower than the nominal ratio for fresh craters on Mars which indi- cates that the troughs are occurring in a secondary sediment fill that has accumulated inside the crater after its formation.  Fig. 1. Typical crater-floor polygons. Left panel: CTX image of a 14 km-sized crater. Upper right: high resolution HiRISE sub-image of same crater. Average polygon diameter is 120 meters. Lower right: HiRISE subimage of a single 100 meter-wide polygon. (Image IDs: CTX: P16_007372_2474, HiRISE: PSP_007372_2475).          Thermal contraction vs. desiccation:  Tensional cracks develop in different materials of geological interest, such as basalts, mud, and permafrost. They are a consequence of the build-up of stress to limits exceeding the material's tensile strength. The amount of relieved stress through fracture formation (and consequently, fracture spacing or polygon width) depends to a first order, on the depth of the fracture tip [4,5]. In order to estimate the thermal stress distribution in the ground, we adopted the stress profiles computed by [6] using a time-dependent viscoelastic model. [6] computed depth and latitude dependent stresses using three different models of thermal inertia.  In order to asses the role of thermal stresses we adopted the values in the high inertia model since they serve as un upper limit with regards to thermal stress values and thickness of the stressed zone. These values were fed to our analytical model using empirical equations relating the stress to final fracture tip depth [5]. Fig. 2. Locations of craters with CFPs draped on a MOLA topography map of Mars.   Results:  We focus in this work on the results at latitude 60° N. Our model shows that the maximum fracture depth should be 6.5 meters in the case of a frozen soil with a tensile strength 2 MPa, and 5 meters for a 3 MPa one. Considering the ratios in literature ranging from 1:4 to 1:10 for fracture depth versus fracture separation [3,5,7], an upper limit of 20 to 50 meters for a 3 MPa material, and 26-65 meters for a 2 MPa one can be given as a good estimate of the maximum size for thermal contraction polygons. The second set of values in particular, are in excellent agreement with a previous finite-element modeling of thermal contraction polygon sizes at the Phoenix landing site at 68° N [8] in addition to mean diameters of such polygons found on Mars [9]. Discussion & outlook:  CFPs are simply too large to have formed with thermal stresses alone under current climatic conditions, and as such, require a higher and deeper stress regime to form. One of the most plausible mechanisms could be desiccation. In that case, CFPs can act as excellent tracers of dried-up bodies of water inside crater basins. Indeed, an onging work documenting other geomorphological "lake" tracers shows that many of the craters in our database contain sedimentary structures that are indicative of ancient lakes such as deltas, terraces, shorelines and mounds [10]. As a result, we can divide such crater lakes into two types depending on the source of water: 1. Exogenic lakes, which are formed by breaching of the crater wall with inflow channels and later settling in the crater basin. In some cases, the inflow volume can exceed the crater's capacity leading to overspills or formation of outflow channels. A famous example of this category is the MER Spirit landing site; Gusev crater[11]. 2. Endogenic lakes on the other hand, lack a visible external source of water. These lakes can form shortly after the impact process, and can be a result of sapping if the impact was rather small, but has breached the ice/water table, or it can be the result of an impact generated hydrothermal system if the target material was volatile-rich, and the impact event was large enough to sustain such a system for a long time [12,13]. We plan on complementing our analysis by studying the mineralogy of the craters in our database that are currently mapped spectrally in high resolution, in an effort to look for minerals indicative of a waterrich environement such as clays, and evaporites.  References: [1] Kuzmin R. & Zabalueva E. (2003), LPSC Abstract #1912. [2] Seibert N. M. and Kargel J. S., (2001), GRL, 28, 899-902. [3] Neal J. et al., (1968), Geo. Soc. of America Bulletin v79, 69-90. [4] Lachenbruch A. H., (1961), JGR,  66, 4273-4292. [5] Lachenbruch A. H., (1962), Spec. Pap. Geological society of America, 70. [6] Mellon M., (1997), JGR, 102, 25,617-25,628. [7] Corte A., and Higashi A., (1964), CRREL Research report 66. U.S. Army Material Command, Hanover. [8] Mellon, M., et al., (2008), JGR, 113, doi:10.1029/2007JE003039. [9] Levy et al., (2009), JGR, 114, doi:10.1029/2008JE003273. [10] Cabrol N. A. and Grin E. A. (1999), Icarus, 142, 160172. [11] Grin E. A. and N. A. Cabrol, (1997),  Icarus, 130, 461-474. [12] Newsom et al., (1996), JGR, 101, 14951-14955.  [13] Abramov O.  & Kring D (2005), JGR, 110, E12S09, doi: 10.1029/2005JE002453.   
