A Parochial View of Groundwater Flow on Mars

Introduction: Groundwater on Mars has been considered to have participated in global flow with the presence of a northern ocean an " inevitable consequence " of early hydraulic and thermal conditions [1,2]. In contrast, we present evidence that an ocean is not hydrogeologically required and that groundwater flow was regional in scale at most. Groundwater Modeling: We seek to test several hypotheses regarding the length, depth, and time scales of groundwater flow on Mars as functions of epoch and the associated porosity and permeability structure of the crust and boundary conditions (sites of recharge and discharge). Here two simple preliminary models are presented. The planet's topography influences key boundary conditions. We assume that the general topography of Mars has been largely unchanged since the late Noa-chian. This assumption is supported principally by the observation that valley networks follow slopes determined by Tharsis loading [3], confirming earlier suppositions that Tharsis is very old [4]. MOLA topography [5] is represented in a Behrmann equal-area cylindrical projection. The current model has the large shield volcanoes removed but has not yet masked out the Valles Marineris for pre-Hesperian simulations. We use the USGS MODFLOW-2000 code [6]. Model cells are 46-km wide, leading to a 400x170-cell horizontal domain. The vertical direction is dis-cretized into 6 cells whose thickness increases with depth; the bottom of the model is 10 km. Because MODFLOW allows cell elevations to be specified, the entire finite-difference mesh can be effectively draped on the terrain. This also causes the bottom of the model to follow the terrain as would be expected for a compaction-limited aquifer. Unconfined model. If the cryosphere was absent during the Noachian, groundwater recharge to an un-confined aquifer can occur by precipitation and discharge occurs if the water table intersects the surface. Recharge is specified as a constant rate across the planet and each cell at the surface is treated as a drain of effectively infinite conductance. The equilibrium heads are determined by incrementally filling the aq-uifer and allowing a transient model to evolve to steady state. In this way the water table can naturally assume its form as a " subdued replica " of topgraphy without being artificially fixed. Results depend on