Effects of regolith-adsorption on the spatial distribution of the atmospheric water vapor simulated with a Mars GCM

Water is an important component of the atmosphere. Most of the water vapor of Mars exists in its lower atmosphere, including the planetary boundary layer, where the water vapor exchanges between the surface and the atmosphere. The regolith, which is a loose, unconsolidated material consisting of fine dust, sand, and fragmented rock at the planetary surface, has a water-adsorbing property (Fanale & Cannon, 1971; Zent & Quinn, 1995). 1-D model simulations suggest that the water vapor exchange between the regolith and the atmosphere through adsorption affects the diurnal variations of water vapor near the surface and obtain good agreement with lander observations (Zent et al., 1993; Savijärvi et al., 2016, 2019; Savijärvi & Harri, 2021). However, the observation of the daytime water vapor column map suggests that water is not well mixed in the atmosphere, with a significant part of the water vapor column apparently not influenced by the topography (Smith, 2002, Fouchet et al., 2007) as it should be if it were well mixed. This could be explained by an enrichment near the surface, possibly related to water exchanges with the regolith, or by an enrichment well above the surface, for instance, below the cloud level where ice sublimes (Fouchet et al., 2007). This study investigates the effects of the water vapor exchange between the regolith and the atmosphere to explain the observed spatial distribution of the atmospheric water vapor column.

This study uses a Mars Global Climate Model (MGCM) fully coupled with a regolith model. Our MGCM traces the Martian seasonal water cycle, including seasonal water ice caps and frost formation, turbulent flux in the atmospheric boundary layer (Kuroda et al, 2005, 2013), and simple cloud formation based on the large-scale condensation (Montmessin et al., 2004). The regolith model calculates water vapor diffusion, adsorption, and condensation in the regolith, using an adsorption coefficient as a free parameter (Kobayashi et al., 2025). The adsorption isotherm of Jakosky et al. (1997) is used to obtain the adsorbed water amount in the regolith, and the isotherm assumes palagonite on Mars. We examine several adsorption coefficients including zero (only considering pore ice) and the inhomogeneous adsorption coefficient using the regolith property model (Kobayashi et al., 2025). The annual mean water flux in high latitudes, where stable ice tables thermodynamically exist, is less than 10^-10 kg m^-2 s^-1 because pore ice fills pores and prevents water transport. We use the atmospheric water vapor column abundance normalized to a fixed pressure of 610 Pa to remove the effect of topography (Smith, 2002).

The subsurface layers are initialized with the subsurface water amount obtained from a spin-up run without the regolith-atmosphere interaction for approximately tens of thousands of Martian years. With the globally homogeneous adsorption coefficient, the subsurface adsorbed water amount increases with latitudes up to ±60° (corresponding to the annual mean surface temperature of 195 K) and rapidly decreases in higher latitudes. With the inhomogeneous adsorption coefficient, the subsurface adsorbed water amount is strongly controlled by the adsorption coefficient, leading to higher adsorbed water in areas of higher adsorption coefficient. We used those results as initial conditions of the subsurface water amount for the following simulations of the regolith-atmosphere interaction.

Our results show that a larger source of adsorbed water in the regolith supplies more water vapor into the atmosphere, with a small adsorption coefficient dependence. The daytime water vapor normalized to 610 Pa anti-correlates with the surface pressure and with thermal inertia in some seasons. The anti-correlations with the surface pressure and with thermal inertia are very slightly strengthened by the regolith-atmosphere interaction. There results indicate that the regolith adsorption contributes to the not-well mixed condition in the Martian lower atmosphere, and the effect is smaller than expected. The exchanged water flux at the surface in our simulation is approximately 10^-10-10^-9 kg m^-2 s^-1, corresponding to 10^-2-10^-1 pr-µm per 1 sol and increasing polewards, for any adsorption coefficient. Thus, we estimate that the regolith-atmosphere interaction integrated over hundreds of sols affects the change of up to several precipitable microns, and the amount becomes smaller due to transport. The magnitude is small or comparable to the sensitivity of the orbital observations nowadays (Fouchet et al., 2007). We conclude that the anti-correlation of water vapor with the surface pressure and with thermal inertia would not be fully explained by the regolith-atmosphere interaction alone, and it would be necessary to focus on transport near the ground surface.