Sensitivity of the Global Water Cycle on Mars to the physical timestep

Author: J. Naar

Date: 09/2021

Standard Global Water Cycle from state of the art Navarro et al. 2014 study

The implementation of Radiatively Active Clouds (RACs) in the GCM, along with a full cloud microphysics scheme, allows for a satisfying agreement between simulated Global Water Cycle and TES observations regarding atmospheric water vapor and cloud opacity. In this configuration, the Northern polar cap is set as an infinite ground ice reservoir avalaible for sublimation. The global water cycle is sensitive to four main parameters listed in the table below.

callphys.def from Navarro et al. 2014

PARAMETERS	COMMENTS	INTERACTION
albedo_h2o_ice=0.35	Frost & Northern water cap albedo	Ground ice
inert_h2o_ice= 800	Water ice thermal inertia	Ground ice
mteta = 0.95	Water contact parameter	Cloud microphysics
nuice_sed=0.1	Size distribution of nuclei	Cloud microphysics

This physical parametrization is tuned to work with a specific numerical scheme set in the run.def file :

day_step = 960
iphysiq = 10

Those lines mean that the temporal resolution for the integration of the dynamical equations is 1/960 martian sol (~24 martian hours), that is 1.5 martian minute. The physical computation is called every 10 dynamical step, that is with a 15 martian minutes resolution.

What is up with the Global Water Cycle and the RAC feedback ?

The PhD work of Margaux Vals unveiled the intrication between RAC feedback and turbulent mixing and cloud microphysics. As a result, there is a strong sensitivity to the physical resolution for the Global Water Cycle. Some specific features require an increased physical resolution, and in any case one should always use the more realistic parametrization. We find that dividing by 2 the iphysiq parameter substancially dries the Global Water Cycle.

Retuning and improving the Global Water Cycle

The reference study by Navarro et al. 2014 describes the sensitivity of the Global Water Cycle to each of the four main tunables parameters presented in the table above. We focus on the water contact parameter and the albedo of ground ice as they are poorly constrained and overly simplified in the GCM. Latent heat of ground ice sublimation is also now included by default in the GCM.

Temperature-dependant water contact parameter

This parameter describes the fraction of airborne dust nuclei avalaible for water condensation using simplified spherical geometry considerations. The water contact parameter can't be easily derived from theory as it is dependant of martian dust physical properties which vary over time, space and environmental conditions. It has been set constant for simplicity but laboratory experiments and theoretical work suggest a strong temperature dependance, as nucleation is disfavored at low temperatures. We have tested several parametrizations from Määttänen et al. (2014) to render the dichotomy between low and high temperatures for nucleation processes. We find that a simple linear parametrization, albeit oversimplifying, maintains a strong aphelion belt and reduces the polar hood thickness.

Albedo of frost and perennial ice

Observations suggest that the polar cap sublimation may be influenced by frost deposition during northern summer, because of its higher albedo. We implement a differentiation of the albedo depending on the nature of the ice (frost or perennial cap) which we use for a finer tuning of the Global Water Cycle. Frost deposition on perennial ice (grid cells where watercaptag=.true.) is limited by a threshold (frost_metam_threshold). If the amount of frost deposited in higher, the surplus is converted to perennial ice. Doing so, we ensure that the influence of frost deposition on albedo can be lifted if sublimation is strong enough.

Reference settings for acceptable global water cycle at increased resolution

For an acceptable Global Water Cycle using the increased iphysiq resolution (iphysiq = 5), one may use the following flags and parameters in callphys.def :

FLAG/PARAMETER	COMMENTS
latentheat_surfwater = .true.	Latent heat of ground ice sublimation (default is .true.)
cst_cap_albedo = .false.	Unchanged polar cap albedo even with frost deposits (default is .false.)
temp_dependant_m = .true.	Temperature-dependant water contact parameter
albedo_h2o_cap = 0.33	Albedo of watercaptag=.true. cells
frost_albedo_threshold = 0.005	kg/m2
albedo_h2o_frost = 0.36	Albedo of frost (if h2o_ice_s > frost_albedo_threshold)
frost_metam_threshold = 0.05	kg/m2
refill_watercap = .true.	(default true) if h2o_ice_s > frost_metam_threshold, h2o_ice_s = frost_metam_threshold and watercap = (watercap + h2o_ice_s - frost_metam_threshold)

10 year simulation using this set of parameters gives the following water cycle :

Reference papers

Navarro et al. (2014), ​https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2013JE004550
Määttänen et al. (2014), ​https://www.sciencedirect.com/science/article/pii/S2214242814000187