Mapping Valley Networks in the Noachian Terrain around Naktong Vallis, Mars: Topographic Control on Drainage Density

Introduction: Most of the cratered uplands of Mars are dissected by networks of branching valleys, whose evolutionary history is still controversial [e.g. 1, 2]. The valleys show branching patterns closely analogous to terrestrial fluvial valleys, which may indicate that the valleys were cut primarily by surface runoff after precipitation [2]. However, short, stubby tributaries and amphitheater heads of valleys may favor another hypothesis, where the valleys are cut by groundwater emerging from below a thick cryosphere [3, 4]. In addition, Carr suggested mass wasting aided by the presence of groundwater at shallow depths mainly formed valley networks [4]. If groundwater migrations had controlled the formation of the valley networks, the source of water should be regionally supplied. In this case, the presence of valley networks hardly suggests anything on a Noachian climate condition. The morphologic characteristics of valleys, such as U-shaped cross sections and relatively constant valley width downstream [5] were plausibly explained by all of these hypotheses, because the lithology and strati-graphy of early Mars were largely uncertain [2]. Thus, quantitative analyses of the distributions and spatial patterns of the valley networks, especially estimates of drainage densities, have been performed here to discuss the origin of the valleys [e.g., 6-12]. We focus on the variations of the drainage densities, because Craddock and Howard [2] qualitatively suggested the possibility that the mean slope is a dominant environmental factor for the drainage densities of valley networks. They suggested that correlation between drainage density and mean slope seems to exist, which might indicate that the particular spatial pattern of erosion involved precipitation, although their inspection was based on local area only and was not a quantitative analysis. Therefore, we test this hypothesis by mapping the valley networks of a larger region precisely in high-resolution images, estimating the drainage densities, and comparing them to local slopes. Method: We identify the valley networks in the Noachian cratered upland, mainly using THEMIS-IR daytime images (100 m/pix). We also use the Viking MDIM 2.1 (231 m/pix) to cover the image gaps existed in THEMIS-IR and use the MOLA topography (about 300 m/pix) to avoid false positives and to determine drainage divides and local slopes. We integrate these data into a GIS-based software, Arc GIS 9.2, where all of the mapping procedures were performed.