Convection in Europa’s icy shell: the role of composite rheology and dynamic grain size evolution

Europa’s outermost layer is a shell of water ice with a probable thickness of a few to a few dozens of km. It is most likely underlain by a liquid water ocean in direct contact with mantle rock, which makes Europa a prime target for understanding habitability. Europa’s surface is heavily deformed and the mean surface age is low (< ~100 Myr), implying active resurfacing, perhaps even through subduction-like processes. While this requires future confirmation, convection in Europa’s icy shell is a viable mechanism to drive such processes. However, the pattern of convection and its link to resurfacing is poorly understood. Here, we use 2D numerical simulations to shed light on these aspects and implement a composite rheology featuring the different slip mechanisms potentially relevant for ice: diffusion creep, basal slip (BSL), grain-boundary sliding (GBS), and dislocation creep. We couple this to grain-size evolution (GSE) and test in basally and mixed basally-tidally heated cases in a 20 km-thick shell the parameters governing the deformation mechanism and GSE.

Without imposing a yield stress to modulate pseudo-plastic deformation, we typically observe an immobile layer at the top of the ice shell. This layer tends to deform via GBS/BSL and features very small grain-sizes (<40 µm), while grains are on the order of cm in the warmer deeper parts, due to stronger grain growth. The thickness of the immobile layer decreases with enhancing the rate of tidal heating and with the sensitivity of grain growth to temperature variations. The immobile layer is thinnest (10-20% of the total thickness), if grain growth in the interior is only moderately enhanced compared to the cold shallow parts, while a large contrast in grain growth increases the layer thickness until eventually convection in the ice shell ceases completely. The omnipresence of an immobile layer (no matter how thick) appears at odds with Europa’s strongly deformed surface and its low age, unless other processes can explain this aspect. Preliminarily, mobilization of the surface layer is possible in our models by imposing a small finite yield stress. Using a very low coefficient of friction, surface velocities can reach rates of up to tens of centimeters per year, under which the surface would be recycled efficiently.