Constraints on the concentration of heat sources in the Martian interior from viscous relaxation of topography

Abstract

Crustal thickness variations induce lateral pressure gradients which can drive flow in the lower crust if the temperatures there are sufficiently high. Gravity and topography data imply that the thickness of the Martian crust is not constant and that large lateral thickness variations are associated with ancient structures such as the Hellas impact basin. The abundance of radioactive isotopes in the Martian interior has a significant influence on the thermal structure of the crust and their concentration can therefore be constrained from the bounds on lower crustal temperatures placed by topographic relaxation considerations. We have used parametrized thermal evolution models which include contributions from secular cooling to determine the thermal structure of the crust underneath Hellas. We have calculated relaxation times for isostatically supported topography and the influence of radiogenic heating and hydrothermal crustal cooling on lower crustal temperatures have been investigated. Our results indicate that hydrothermal cooling of the crust is not necessary to explain the absence of lower crustal flow. Even a brecciated upper crust would be sufficiently conductive if the pore spaces were filled with ice or water. If global scale hydrothermal cooling did occur, radiogenic heating after core formation could have been larger than 60 pW kg^-1 without initiating lower crustal flow, but the question of the presence and/or extent of hydrothermal crustal cooling remains unresolved. If no additional crustal cooling occurred, the absence of lower crustal flow at the Hellas impact basin implies that radiogenic heating after core formation cannot have exceeded 45 pW kg^-1 and that a thin crust or low initial temperatures are required if heating exceeded 30 pW kg^-1, which corresponds to the chondritic heat production rate. In this case, compositional models for Mars with greater than chondritic abundances of K, U, and Th, e.g., the model by Lodders and Fegley (1997), are difficult to reconcile with the observations.