Exploring the Impact of the Regolith on the Martian Water Cycle with a Global Climate Model

Ninth International Conference on Mars 2019 (LPI Contrib. No. 2089) 6410.pdf

Exploring the Impact of the Regolith on the Martian Water Cycle with a Global Climate Model. J. Weinmann1, P.-Y. Meslin1, M. Vals2, F. Forget2, E. Millour2. 1IRAP, UPS-CNRS, Toulouse (pmeslin at irap.omp.eu), 2Laboratoire de Météorologie Dynamique, Paris.

Introduction: The importance of regolith-atmos- amounts of water vapor [ref.]? More generally, how im- phere exchange in the diurnal, seasonal and inter-annual portant is the role played by the regolith in the current variability of the martian water cycle remains an open and past water cycles? question, although there is a consensus on its fundamen- Previous studies done with GCM models: To get tal role over geological timescales. Migration of ice a full and integrated view of all the processes that are across the planet upon variations of Mars’ orbital pa- involved, it is necessary to use a Global Circulation rameters and obliquity is one of the most significant ge- Model that describes the exchange of water with the ological processes that are still active and continuously perenial and seasonal polar caps, and its inter-hemi- reshaping the surface of the planet. spheric transport. Other GCMs interacting with an ac- Theoretical/numerical studies have studied the sub- tive regolith have been developed in the past [25, 26], surface ice stability over geological timescales [1-5] and but their model of regolith was quite simple (2-layers over the seasonal cycle [4,6]; theories or simplified model and no formation of subsurface ice in [26]) and static models have tried to explain the current distribu- gave conflicting results on its importance on the current tion of water-equivalent hydrogen measured by Mars water cycle. Böttger et al. [6] have implemented a more Odyssey (confirmed in situ by the Phoenix lander) in complex regolith scheme in the LMD Global Climate terms of topographic control [7], transient ground ice Model [27] and their results, although dependent upon [8], or exchange with hydrous minerals [9]; 1D or sim- initial conditions, demonstrated the potential im- plified 3D models of the 'breathing' of the regolith portance of regolith-atmosphere exchange on the sea- caused by adsorption of water vapor on the soil matrix sonal and inter-annual water cycle. For instance, they have been developed to explain the diurnal variations of found that the regolith can act as a buffer for water vapor the water column density measured from the orbit [10] leaving the northern high mid-latitudes during northern and of the H2O volume mixing ratio measured at the sur- summer, preventing some of the water from reaching face of Gale crater by the REMS suite [11]. the southern hemisphere via the Hadley circulation and A growing number of studies has also been filling thus enabling to maintain the observed North-South at- the need for experimental data on the adsorption or hy- mospheric water vapor asymmetry. This water, trapped dration properties of martian regolith simulants [12-18], in the soil during winter, is then released into the north- on the diffusion of water vapor through such porous me- ern spring time atmosphere and returns to the residual dia [19, 20] and on the thermal properties of ice-rich water ice cap through eddie transport. An active regolith materials under martian conditions [21]. These studies had therefore a stabilizing effect on the simulated water have enabled us to gain much insight into the dynamics cycle. However, they were unable to explain the pur- of sublimation of a subsurface ice table and on the pos- ported diurnal variations of the water column abundance sible exchange of water vapor with the regolith. by exchange with the regolith. The same subsurface Direct analyses of the martian soil by Curiosity have model has been used to evaluate the evolution of the also provided new mineralogical and chemical con- subsurface ice reservoirs over recent geological times straints about the possible phase(s) that can be involved [5]. in the exchange of water vapor [22-24]. Here, we upgraded the subsurface model developed Open questions: Despite all these studies, many by [28], which is coupled to the LMDZ GCM [27]. The questions on this topic remain (at least partly) unan- latter incorporates improved cloud microphysics [29], swered: how did subsurface ice at high latitudes em- and its description of the water cycle has been revised place? How did this ice evolve and migrate over time by parametrizing sub-grid scale clouds [30] and by re- between polar, equatorial and mid-latitude regions? fining its vertical resolution [31]. What is the nature of the hydrogen measured at mid- and Description of the subsurface model: The model, low-latitudes by Mars Odyssey? Does it reflect long- which is coupled to the subsurface thermal model, sim- term alteration of minerals by adsorbed water or frost? ulates the diffusion, adsorption and condensation of wa- How can we reconcile the fact that simulations of the ter molecules within the soil. It is an implicit model. It atmospheric water cycle without an “active” regolith fit is purely diffusive (it does not take into account advec- the atmospheric observations pretty well, while the reg- tion, surface diffusion, barodiffusion and thermal tran- olith contains a significant proportion of hydrated amor- spiration). The model uses either the H2O atmospheric phous component, whose terrestrial analogs are known mixing ratio as boundary condition, or the ice saturation to be very hygroscopic and likely to exchange large Ninth International Conference on Mars 2019 (LPI Contrib. No. 2089) 6410.pdf

pressure if water ice is present on the surface. The dif- From these initial states, we will explore the sensi- fusion coefficient (in a transition regime between Knud- tivity of the H2O atmospheric cycle (diurnal and sea- sen and molecular diffusion) takes into account the ob- sonal) to different types of adsorption properties. Con- struction of the pores by ice, following the numerical versely, a comparison to the observed atmospheric wa- work by [32]. The equilibrium between the vapor and ter cycle should provide constraints on these properties. adsorbed phases is described by a Langmuir-type of iso- We will also explore the sensitivity of the diurnal and therm, but linearized versions of Type II and III iso- seasonal cycles on the kinetics of the adsorption/desorp- therms, which are simplified versions of the isotherms tion process. measured by [12] are being implemented. Unlike exist- ing models, the kinetics of the adsorption process can also be taken into account. Although it has not yet been well constrained experimentally, its influence on the diurnal cycle can be investigated. The soil specific surface area (SSA) and the adsorption enthalpy are free param- eters of the model. A feedback of the ice content on the thermal properties of the regolith can also be activated,

based on the experimental results of [21]. Fig. 2: Typical profile of the ice content vs. depth obtained at high latitudes Results: Since the regolith can potentially hold a (saturation in the top layer has not been fully reached yet). very large amount of water, a critical step needed before coupling the subsurface model to the atmospheric cycle is the initialization of the subsurface water reservoir. Since the current atmospheric model matches the observations pretty well, we decided for this initialization to force the regolith to equilibrate with the H2O cycle obtained without an active regolith, by imposing time-var- iable near-surface H2O mixing ratios as boundary condition to the subsurface model. Small deviations from this initial state can then be explored. This procedure al- lows us to make predictions of the distribution of stable subsurface ice, both in terms of geographical extent and Fig. 3: Example of predicted map of regolith H2O content (adsorbed water + ice), for a Langmuir-type adsorption isotherm (saturation at 1 monolayer 2 -1 depth to the ice table (Fig. 1, 2). It can also be used to of H2O molecules, SSA = 17 m .g ). extract the time evolution of the ice content as a function of depth, to estimate timescales of ice growth and to ex- References: [1] Clifford S.M. and Hillel D. (1983), JGR, 88(B3). [2] Mellon M.T. and Jakosky B.M. (1993), JGR, 98(E2). [3] trapolate the ice filling process to geological times. It Mellon M.T. and Jakosky B.M. (1995), JGR, 100(E6). [4] Schor- can also be used to make predictions of global H2O ghofer N. and Aharonson O. (2005), JGR, 110(E05003). [5] maps for different adsorption isotherms (Fig. 3), to be Steele L.J. et al. (2017), Icarus, 284. [6] Böttger H.M. et al. (2005), Icarus, 177. [7] Feldman W.C. et al. (2005), JGR, compared to the actual H2O map derived by Mars Od- 110(E11009). [8] Jakosky, B.M. (2005), Icarus, 175. [9] Fialips yssey GRS. C.I. et al. (2005), Icarus, 178. [10] Zent A.P. et al. (1993), JGR, 98(E2). [11] Savijärvi et al. (2016), Icarus, 265. [12] Pommerol et al. (2009), Icarus, 204. [13] Zent A.P. and Quinn R.C. (1997), JGR, 102(E4). [14] Zent A.P. et al. (2001), JGR, 106(7). [15] Bish D.L. et al. (2003), Icarus, 164, 96-103. [16] Jänchen J. et al. (2006), Icarus, 180. [17] Bryson, K.L. et al. (2008), Icarus, 196. [18] Chevrier V. et al. (2008), Icarus, 196. [19] Hudson, T.L. et al. (2007), JGR, 112(E05016). [20] Chevrier V. et al. (2007), GRL, 34. [21] Siegler M. et al. (2012), JGR, 117, E03001. [22] Leshin L. et al. (2013), Science, 341. [23] Blake D.F. et al. (2013), Science, 341. [24] Meslin P.-Y. et al. (2013), Science, 341. [25] Houben H. et al. (1997), JGR, 102(E4), 9069-9083. [26] Richard- son M.I. and Wilson R.J. (2002), JGR, 107(E5), 5031. [27] Forget F. et al. (1999), JGR, 104(E10). [28] Meslin P.-Y. (2008), Mars Fig. 1: Map of stable subsurface ice (depth to the ice table) in equilibrium Atmosphere: Modeling and Observations, 9111. [29] Navarro T. with the LMDZ GCM atmospheric H2O cycle (resolution 32x24). et al. (2014), JGR. [30] Pottier A. et al. (2017), Icarus. [31] Vals M. et al. (2018), EPSC abstr., 12. [32] Meslin P.-Y., (2010), Soil Sci. Soc. Am., 74, 6.