Assessing the formation of valley networks on a cold early Mars: Predictions for erosion rates and channel morphology

Abstract

The ancient Noachian highlands of Mars host an extensive population of valley networks which formed predominantly during the early geologic history of the planet. Morphologic characteristics of the valley networks have been interpreted to indicate the formation of these features through precipitation-derived fluvial activity and therefore as evidence for a relatively warm and wet Noachian Mars climate. However, these interpretations conflict with the results of sophisticated global climate modeling studies, which suggest that early Mars was dominated by a cold and icy climate with conditions characterized by adiabatic cooling and regional ice sheets within the highlands. Difficulties in replicating the warm and wet early climate conditions interpreted from the valley networks has served as the basis for alternative suggestions for the formation of valley networks by transient heating and snowmelt in an otherwise cold and icy climate. Here we test a conceptual model for valley network formation and incision under cold and icy conditions with a substrate characterized by the presence of an ice-free, desiccated surface regolith and subjacent ice-cemented regolith, similar to that found in the Antarctic McMurdo Dry Valleys on Earth. We implement numerical thermal models, quantitative erosion and transport estimates, and morphometric analyses in order to outline and test predictions for: (1) the nature and structure of the cold and icy Noachian substrate, (2) valley network fluvial erosion and incision rates, and (3) resulting channel/valley morphology. Morphologic predictions are compared against observational data to determine if the valley networks characteristics are consistent with formation in a cold and icy climate. Through this approach we present and develop the underlying conceptual model, identify and generate the fundamental data inputs required to evaluate it, and perform a preliminary first-order assessment to serve as a basis for further investigation. While these analyses have been performed over a broad parameter space to address relative uncertainties, we report findings with a specific emphasis on the nominal results.We find that under cold conditions, the substrate is characterized by a kilometers-thick, globally-continuous cryosphere with a ∼ 50–100 m thick ice-free surface layer in diffusive equilibrium with the atmosphere and underlying ice-cemented regolith, the top of which forms the ice-table. Estimates of the potential infiltration capacity of the ice-free surface regolith layer indicate that surface runoff generation does not require excessively large precipitation rates, and thus does not preclude fluvial activity sourced by rainfall. Additionally, due to the predicted thicknesses of the ice-free surface regolith, any transient period of warming would need to be sustained in duration (∼ 2 kyr) to induce melting of ground ice to initiate solifluction/gelifluction activity. The predicted range of ice-table depths is exceeded by the incised depths of a majority of the valley network population, therefore suggesting interactions were possible. Evaluation of the rates of mechanical substrate erosion and thermal ice-cement erosion by valley network fluvial activity are subject to large uncertainties due to several poorly constrained parameters (i.e. the substrate erodibility factor and the valley network fluvial water temperature) but generally indicate that at low water temperatures, consistent with a cold and icy surface environment, thermal erosion of the ice-cement can be outpaced by mechanical erosion of the substrate. In this scenario, the relative efficiency of lateral erosion is predicted to be enhanced during incision below the ice-table, causing preferential channel/valley widening and increased width-to-depth ratios. Assessment of this prediction through a morphometric analysis of valley network width-to-depth ratios and occurrence of U-shaped cross sections indicates no significant correlation to ice-table depths. This result suggests that either: (1) permafrost was not present in the Noachian Mars substrate, indicating a warmer climate, (2) the permafrost ice-table did not significantly influence valley network width-to-depth ratios, possibly owing to more rapid thermal erosion rates than mechanical, or (3) additional factors not accounted for in this analysis are involved. Potential additional factors which could influence the results of this analysis include variable surface geothermal heat flux, substrate thermal conductivity, substrate composition and physical state as well as valley network preservation state, and morphometric sampling biases (due to data availability constraints). These factors represent areas for future investigation to refine assessments of valley network formation under cold early Mars conditions.