49th Lunar and Planetary Science Conference 2018 (LPI Contrib. No. 2083) 1132.pdf

CAN EARTH-LIKE PLATE TECTONICS OCCUR IN THE OUTER ICE SHELLS OF ICY SATELLITES? Samuel M. Howell1 and Robert T. Pappalardo1, 1Jet Propulsion Laboratory, California Institute of Technology ([email protected]).

Introduction: The outer H2O ice shells of some Figure 1. icy satellites show evidence for divergent, strike-slip, Geometry of and convergent tectonics. Researches propose that conceptual Earth-like plate tectonic processes may occur in the ice model consid- shells of Europa and Enceladus [1-3], raising the pos- ered in the sibility of a buoyantly driven cycle of tectonic resurfac- force balance. ing and lithosphere recycling on icy satellites [1]. See text for On Earth, a cycle of spreading and subduction is parameter supported by buoyancy forces transmitted elastically descriptions. through the lithosphere [4]. The majority of tectonic forcing comes from “slab pull,” driven by the negative buoyancy of cold lithosphere “slabs” intruding into the warm asthenosphere. “Ridge push” accounts for only a tenth of these forces, and is driven by the cooling of strength of the lithosphere, FY, the force required to lithosphere with increasing age for 10s Myr and over elastically deform the lithosphere, FE, and basal trac- 1,000s km, causing warmer, more buoyant material to tions that resist sliding between the lithosphere and slide downhill away from the ridge axis. asthenosphere, FA. Thus, subduction is permitted when On Europa, proposed "subsumption zones" at in- FFFFF t .(1) ferred convergent margins may allow old lithosphere to STYEA be reincorporated into the ice shell and recycled [2]. The buoyancy force exerted on the lithosphere by a Johnson et al. [1] recently demonstrated that a subduct- penetrating slab per along-trench distance is ing ice I slab on Europa may remain negatively buoy- H FgS,(2) ant for appropriate values of salt content, salt distribu- Sla UU sin M tion, and porosity of the subducting plate, providing a mechanism for slab pull. Like mid-ocean ridges, exten- where ȡl and ȡa are respectively the densities of the sional bands may locally bring warm isotherms to shal- cold ice lithosphere and warm ice asthenosphere, g is low depths [e.g. 5]. However, their limited width (~10s the gravitational acceleration, S is the penetration depth km) [6] requires either a short lifespan or slow opening of the downwelling slab beneath the base level of unde- rate, inhibiting the development of the large gradients formed lithosphere, H is the thickness of the litho- in plate thickness persistent through geologic time that sphere, and ij is the dip of the subducting slab. drive ridge push. Using the simplifying assumption that tidal stresses, To explore whether Earth-like, convectively driven ıT, are uniformly distributed through the ice shell, the plate tectonics can occur in the outer shells of icy satel- depth-integrated tidal forcing is lites, we employ a simple force balance to determine FHYY V ,(3) whether the buoyancy forces associated with slab pull The depth-integrated Mohr-Coulomb yield strength can overcome the strength of the lithosphere to initiate depends on the lithosphere's cohesion, C, its coefficient or sustain subduction. We consider a range in litho- of internal friction, ȝ, and the lithostatic stress at its sphere cohesion and coefficient of friction for pristine base, and weakened lithosphere, salt content, and tempera- ture to determine the depth a slab must reach before 1 FHCYl PU gH.(4) sustaining subduction. As a feasibility test, this force 2 balance incorporates the most optimistic first-order We maximize driving forces by assuming a vertical assumptions, promoting Earth-like plate tectonics to slab, and minimize resisting forces by negating elastici- the greatest possible extent. ty and assuming that the lithosphere slides freely over Force Balance Framework: Fig. 1 shows the ge- the asthenosphere. Thus, Eq.1 becomes ometry of the conceptual model used in the force bal- ance. Self-driving plate tectonics requires the sum of 1 UUlatgS VT C PU l gH .(5) driving forces to exceed the sum of resisting forces [7, 2 8]. In icy satellites, potential driving forces include the To investigate the density anomaly required for sus- forces associated with a negatively buoyant slab, FS, tained subduction, we solve Eqs. 1-6 for ȡl, and explore and tidal interactions, FT. In order to sustain subduc- the surface temperature, salt content, and external tion, tidal forces must persist through geologic time. stresses required to sustain plate tectonics. Resisting forces include the depth-integrated yield 49th Lunar and Planetary Science Conference 2018 (LPI Contrib. No. 2083) 1132.pdf

Optimistic Assumptions: Continuing optimistic >70 kg/m3 if the lithosphere has somehow already pen- assumptions that promote slab pull-driven subduction, etrated to the base. For comparison, Johnson et al. [1] the slab is assumed to be perfectly insulated from the find that on Europa, maximum sustained density con- warm interior, retaining its initial temperature contrast. trasts may be just ~30 kg/m3. The porosity of the slab and interior are ignored to If a slab penetrates far into the ice shell in response maximize the buoyancy contrast. The convecting as- to external forcing, subduction can be maintained only thenosphere is assumed free of salt, preferentially parti- if the subduction zone has greatly reduced strength. For tioned into the melt and extracted to the ocean. Fur- example, if cohesion is reduced to ~100 kPa and the thermore, we assume no diffusion of salt from the coefficient of internal friction is reduced to 0.1, sub- downwelling slab to the convecting asthenosphere. duction can be buoyantly maintained on Ganymede for The ice shells are taken to be sufficiently thick so a density contrast of 30 kg/m3 once the slab has pene- that the maximum slab penetration can be large (25 km trated ~10% of the ice shell thickness. While the gravi- for Europa, 50 km for Ganymede, 145 km for Encela- ty of Enceladus is low, the ice shell may be very thick. dus), and the lithosphere thickness is minimized at 3 If subduction zone cohesion is reduced to 100 kPa and 2 km (FB increases with H, while FY increases with H ). the coefficient of friction remains 0.6, subduction be- This is appreciably thinner than lithosphere thicknesses comes self-sustaining when the slab has penetrated predicted by thermal and tectonic models [e.g. 5]. ~50% of the ice shell thickness (~70 km). Results and Discussion: With this simple force The external forces required to initiate subduction balance, we investigate whether it is plausible for slab in pristine ice are 1-2 MPa. For comparison, diurnal pull to initiate and/or maintain subduction and drive tidal forces on Europa are on the order 100 kPa and Earth-like plate tectonics in outer H2O ice shells. We reverse sign cyclically [9]. Thus, sustained secular tidal perform this analysis for Europa, Enceladus, and Gan- stresses, e.g. non-synchronous rotation or true polar ymede, spanning from the smallest to the largest icy wander, would be required to initiate subduction for satellites with observed tectonics. realistic density contrasts [1] and to sustain subduction While the large gravity of Ganymede and potential- unless ice shell strength is greatly reduced. ly thick ice shell of Enceladus are more conducive to Conclusions: In the outer ice shells of icy satel- subduction than the ice shell of Europa, we find that lites, the density contrast between subducting litho- negative slab buoyancy alone is unlikely to allow the sphere and warm asthenosphere is small compared to initiation of subduction in any of the broad range of icy the mechanical strength of the lithosphere, even under satellites considered (Fig. 2). For example, initiating the most favorable assumptions. Thus, slab pull is un- subduction on Europa in pristine lithosphere (ȝ =0.6, likely to initiate subduction in the outer H2O ice shells C = 1 MPa) requires a density contrast of >400 kg/m3 of icy satellites, and will maintain ongoing subduction when the slab has penetrated 10% of the ice shell, and only when the subduction zone is weakened signifi- ȝ = 0 ȝ = 0.1 ȝ = 0.6 cantly compared to pristine lithosphere. An Earth-like 100 cycle of convectively driven plate tectonics is therefore C = 1 MPa Enc. ] 80 unlikely on any icy satellite.

-3 Eur. 60 C = 100 kPa Gan. If the strength of the ice shell is significantly lower 40 than considered here (i.e. if the cohesion and coeffi- [kg m

ȡ cient of friction are orders of magnitude smaller), the

ǻ 20 C = 0 negative buoyancy of the slab could plausibly drive 0 104 subduction. However, such a weak ice shell would have to be reconciled with the reliance of slab pull on 103 the elastic transmission of stresses, and inferences of 100 elastically supported topography from spacecraft ob- [kPa] T servations. ı 10 References: [1] B. C. Johnson et al. (2017, in 1 press) J. Geophys. Res. – Planets [2] S. A. Kattenhorn 0.01 0.1 1 0.01 0.1 1 0.01 0.1 1 Slab Penetration / Ice Shell Thickness and L. M. Prockter (2014) Nature Geosci., 10, 762-767 [3] M. T. Bland and W. B. McKinnon (2014), DPS Figure 2: Force balance predictions of the density Fall Meeting, Abstract #2817616. [4] D. Turcotte and contrast between the downwelling lithosphere slab and G. Schubert (2014), in Geodynamics, pp. 10-23. [5] S. warm asthenosphere (top) and regional driving stresses M. Howell and R. T. Pappalardo (2017) Europa Deep (bottom) required to satisfy Eq. 5. Colored lines show Dive I, Abstract #7002. [6] L. M. Prockter et al. failure curves for different cohesion values for Encela- (2009), in Europa, pp. 589. [7] C. E. Hall et al. (2003) dus (solid, bold), Europa (thin, dashed), and Ganymede EPSL., 212, 15-30 [8] D. P. McKenzie (1977) Island (thin, dotted). Columns show results for differing coef- Arcs, Deep Sea Trenches, and Back-Arc Basins, pp. ficients of internal friction. The black dashed line (top 57-61. [9] T. A. Hurford et al. (2007) Icarus, 186, 218- figures) shows the maximum sustained buoyancy con- 233. trast predicted by Johnson et al. [1].