No abstract

No abstract

Fluvial erosion on early Mars was dominated by valley networks created through a combination of groundwater processes and surface runoff. A reduced greenhouse effect due to CO2 loss, together with a declining geothermal heat flux, promoted the growth of a cryosphere and a Hesperian hydrologic regime dominated by outflow channel formation. We test the hypothesis that the transition from valley network to outflow channel formation was preceded by a more subtle evolution characterized by a weakening of surface runoff, leaving groundwater processes as the dominant, final source of valley network erosion. Our hypothesis, supported by a terrestrial analog in the Atacama desert of Chile, is related to the groundwater sapping reactivation hypothesis for densely dissecting highland valley networks on Mars suggested by Baker and Partridge in 1986 and focuses on the age analysis of large, sparsely dissecting valley networks such as Nanedi Valles, Nirgal Vallis, valleys in fretted terrain, and tributaries of outflow channels and Valles Marineris chasmata. We find that these features are consistently late Noachian to Hesperian in age, younger than Noachian densely dissecting dendritic valley networks in the southern highlands. In the Tharsis region the observation of dense and sparse valley network morphologies on Hesperian terrain suggests that while surface runoff gave way to groundwater processes consistent with our hypothesis, the transition may have occurred later than elsewhere on the planet. The volcanic nature of Tharsis suggests that geothermal heat and volatile production led to episodically higher volumes of surface runoff in this region during the Hesperian.

[1] The large outflow channels of Chryse Planitia, Mars, are thought to have been carved by floodwaters discharged from an aquifer beneath a confining cryosphere. However, conventional models of groundwater flow require optimistically high permeabilities to produce, in a single flooding event, the discharge rates and volumes inferred from channel morphology. Additionally, discharge likely ceased upon refreezing of chaotic terrain fractures carrying it to the surface, further limiting the volume of water produced in a single flood. It is thus probable that multiple surface discharge events were required, and we quantify this hypothesis with regional groundwater simulations. Each discharge event is triggered by cryosphere disruption due to superlithostatic hydraulic head at the channel source region and is terminated by cryosphere fracture refreezing. Before the next event, head is allowed to recover with the aid of distal aquifer recharge constrained by previous global groundwater models. Our baseline model, which emulates the source region of Kasei Valles, yields a minimum of 2900 flooding events even for high depth-averaged aquifer permeabilities corresponding to near-surface values of 10 À9 m 2 . Although fewer events are required for other circum-Chryse channels, formation by a single flood remains improbable. We suggest that unrealistically high numbers of floods may be circumvented by an alternative model involving the local ponding of surface discharge to form standing bodies of water. Episodic failure of these bodies produces flood outbursts which erode the channels.

No abstract