Modeling the global water cycle on Mars with improved physical parametrization

Introduction: For many years we have been working on the modeling of the martian water cycle in the Laboratoire de Météorologie Dynamique martian Global Climate Model [1,2,3,4]. This has been challenging because of destabilizing feedbacks such as the coupling between temperatures, water vapour, clouds, and temperature changes induced by radiatively active water clouds [2].

It is interesting to accurately model Mars’s water cycle not only to understand present day Mars but also to simulate past climate changes induced by obliquity and orbital variations. This study is partly motivated by the ERC Mars Through Time project which notably aims at modeling recent (<10 Myrs) martian paleoclimates. We have run control simulations using the set of water-cycle related parameters from [2] which show good agreement with satellite measurements. However, we have recently discovered that the coupling between the microphysics and other physical processes such as turbulent and convective mixing induce a non-negligible sensitivity to the GCM physical timestep. Of course it is better to use the shortest timestep but this causes an significant drying of the modeled climate system. This motivated us to implement more realistic physical processes which, in addition to being prominent for paleoclimatic studies, allow for an improved agreement with atmospheric water-related data measured by the Thermal Emission Spectrometer (TES) and Spectroscopy for the Investigation of the Characteristics of the Atmosphere of Mars (SPICAM) instruments [5,6,7] The improved physics include a temperature-dependant water contact parameter, latent heat of ground-ice sublimation and distinct albedo for water frost and perennial ground ice.

Sensitivity to the physical timestep: The standard physical timestep used in [1] was 15 martian minutes, subsequently reduced to 7.5 minutes. The water cycle shows a significant sensitivity to the physical timestep. This is because the dust and water physics and transport are strongly coupled with boundary layer parametrization radiative transfert, and cloud microphysics.  Cloud microphysics uses a dedicated subtimestep for computing efficiency, but the diversity of processes involved bind us to increase physical resolution altogether. We will discuss the influence of an even smaller physical timestep to the water cycle, and the distinct effect of the dynamical timestep.

Temperature-dependant water contact parameter:  The water “contact parameter” controls the effective fraction of nuclei activated for water condensation. The higher it is, the more water ice nucleation is favored and water-ice clouds form. Experimental data and literature [8] tend to show a strong dependence of the contact parameter with temperature. However, GCMs usually use constant water contact parameter. We find that using a simple linear fit from [8], albeit oversimplifying the bimodal behavior between low  and high temperatures, attenuates the thickness of the northern polar hood during the second part of the year while allowing for a thick aphelion cloud belt, which was a key modeling issue in the LMD GCM (Navarro et al. 2014).

Latent heat of ground ice sublimation:  The latent heat of ground water ice sublimation had been neglected in our climate model, because of the low amount of water and thus energy flux involved. The sublimation of water was hitherto simply computed using surface temperature and water vapor equilibrium. Preliminary studies of martian climates at higher obliquities have shown that the energy fluxes involved may become of prime importance, as accounting for the latent heat of ground ice strongly inhibits the summer sublimation of the polar caps. Its calculation in a GCM requires an implicit numerical scheme to properly account for the coupling between temperature and sublimation during the GCM timestep. We show that the latent heat of ground ice sublimation has a minor influence on the global present-day water-cycle but may be considered for specific phenomenons such as the perenniality of crater-induced water ice ejectas [8].

Differentiation of perennial ice albedo and water frost albedo:  Through sublimation, the northern polar cap around summer solstice is the dominant atmospheric source of water vapor in our GCM. It shows a strong dependence of summer surface temperature with albedo and thus sublimation [2]. While acceptable water ice albedos range from 0.3 to 0.5 on Mars [9], monitoring the northern polar cap show that the start and end of sublimation phases are correlated with albedo variations [10]. The outlier region of the polar cap is actually the main source of sublimation in the GCM, while the central cap is a cold trap where water vapor condenses into frost. The albedo of perennial water ice and frost, previously identical, are now separated in our GCM to take into account the fact that fresh frosts can exhibit higher albedos which tend to slow down their sublimation . Future enhancements may include a time-varying frost albedo, as it changes with the evolution of ice particules size (metamorphism) and dust accumulation, and thus represent more accurately the insolation on the northern polar cap during summer. These processes may play a key role in paleoclimate studies.

Conclusion:  Because of the intricate complexity of the feedbacks between dynamics, dust and water transport and cloud microphysics, increasing the temporal resolution in the GCM is mandatory. We have addressed the disruptions it causes to our reference water cycle simulation by taking into account previously neglected physical processes, which allow for a satisfying agreement with TES and SPICAM datasets (figure 1).

 

Figure 1: Reference simulation (MY26) compared to TES data for water vapor column and cloud opacity at 2 PM.

References:

[1] Montmessin et al. (2004),  Journal of Geophysical Research: Planets, 109(E10). [2] Navarro et al. (2014),  Journal of Geophysical Research: Planets, 119(7), 1479-1495. [3] Pottier et al. (2017), Icarus, 291, 82-106. [4] Vals (2019), Doctoral dissertation, UPMC. [5] Smith (2004), Icarus, 167(1), 148-165.  [6] Maltagliati et al. (2011), Science, 333(6051), 1868-1871. [7] Fedorova et al. (2020), Journal of Geophysical Research: Planets, e2020JE006616. [7] Määttänen et al. (2014), GeoResJ, 3, 46-55.  [8] Byrne et al. (2009), Science, 325(5948), 1674-1676. [9] Wilson et al. (2007), Geophysical Research Letters, 34(2). [10] Hale, A. S. et al. (2005),  Icarus, 174(2), 502-512.