James Schaeffer

Mars atmospheric dynamics as simulated by the NASA Ames General Circulation Model: 1. The zonal‐mean circulation

This is the first in a series of papers that will discuss Mars atmospheric dynamics as simulated by the NASA Ames General Circulation Model (GCM). This paper describes the GCMs zonal-mean circulation and how it responds to seasonal variations and dust loading. The results are compared to Mariner 9 and Viking observations, and the processes responsible for maintaining the simulated circulation are discussed. At the solstices the zonal-mean circulation consists of a single cross-equatorial Hadley circulation between 30°S and 30°N. For relatively modest dust loadings (τ=0.3), the associated peak mass flux is 100 × 108 kg s−1 at northern winter solstice and 55 × 108 kg s−1 at southern winter solstice. At both seasons, westerlies dominate the winter hemisphere, and easterlies dominate the summer hemisphere. Maximum zonal winds occur near the model top (∼47 km) and are about the same at both seasons: 120 m s−1 in the winter hemisphere and 60 m s−1 in the summer hemisphere. Mean surface westerlies of 10–20 m s−1 are predicted at the middle and high latitudes of the winter hemisphere, as well as in the summer hemisphere near the rising branch of the Hadley circulation. The latter has the structure of a “jet” and is particularly strong (>20 m s−1) at northern winter solstice. With increasing amounts of dust (up to τ=5), the zonal mean circulation at northern winter solstice intensifies and gives no indication of a negative feedback. Dust can easily double the mass flux of the Hadley circulation. In the solstice simulations, the mean meridional circulation is the main dynamical contributor to the heat and momentum balance; the eddies play a relatively minor role. There is no evidence in these simulations for a polar warming. At the equinoxes the zonal mean circulation is more Earth-like and consists of two roughly symmetric Hadley cells with westerly winds in the mid-latitudes of each hemisphere and easterlies in the tropics. The simulated zonal winds are about half as strong as they are at solstice. However, the strength of the mean meridional circulation is much less than at solstice and averages between 5 and 10 × 108 kg s−1. At these seasons, the eddies and mean circulation make comparable, but opposing, contributions to the heat and momentum balances.

General circulation model simulations of the Mars Pathfinder atmospheric structure investigation/meteorology data

The NASA Ames Mars General Circulation Model is used to interpret selected results from the Mars Pathfinder atmospheric structure instrument/meteorology (ASI/MET) experiment. The present version of the model has an improved soil thermal model, a new boundary layer scheme, and a correction for non-local thermodynamic equilibrium effects at solar wavelengths. We find good agreement with the ASI/MET entry data if the dust observed at the Pathfinder site is assumed to be distributed throughout the lowest five to six scale heights. This implies that the dust is globally distributed as well. In the lower atmosphere the inversion between 10 and 16 km in Pathfinders entry profile is likely due to thermal emission from a water ice cloud in that region. In the upper atmosphere (above 50 km), dynamical processes, tides in particular, appear to have a cooling effect and may play an important role in driving temperatures toward the CO2 condensation temperature near 80 km. Near-surface air temperatures and wind directions are well simulated by the model by assuming a low surface albedo (0.16) and moderately high soil thermal inertia (336 SI). However, modeled tidal surface pressure amplitudes are about a factor of 2 smaller than observed. This may indicate that the model is not properly simulating interference effects between eastward and westward modes.

On the possibility of liquid water on present‐day Mars

Using a validated general circulation model, we determine where and for how long the surface pressure and surface temperature on Mars meet the minimum requirements for the existence of liquid water in the present climate system: pressures and temperatures above the triple point of water but below the boiling point. We find that for pure liquid water, there are five “favorable” regions where these requirements are satisfied: between 0° and 30°N in the plains of Amazonis, Arabia, and Elysium; and in the Southern Hemisphere impact basins of Hellas and Argyre. The combined area of these regions represents 29% of the planets surface area. In the Amazonis region these requirements are satisfied for a total integrated time of 37 sols each Martian year. In the Hellas basin the number of degree days above zero is 70, which is well above those experienced in the dry valley lake region of Antarctica. These regions are remarkably well correlated with the location of Amazonian paleolakes mapped by Cabrol and Grin [2000] but are poorly correlated with the seepage gullies found by Malin and Edgett [2000]. In both instances, obliquity variations may play a role. For brine solutions the favorable regions expand and could potentially include most of the planet for highly concentrated solutions. Whether liquid water ever forms in these regions depends on the availability of ice and heat and on the evaporation rate. The latter is poorly understood for low-pressure CO2 environments but is likely to be so high that melting occurs rarely, if at all. However, even rare events of liquid water formation would be significant since they would dominate the chemistry of the soil and would have biological implications as well. It is therefore worth reassessing the potential for liquid water formation on present day Mars, particularly in light of recent Mars Global Surveyor observations.

Mars atmospheric dynamics as simulated by the NASA Ames General Circulation Model: 2. Transient baroclinic eddies

A large set of experiments performed with the NASA Ames Mars general circulation model (GCM) have been analyzed to determine the properties, structure, and dynamics of the simulated transient baroclinic eddies. The Mars GCM simulations span a wide range of seasonal dates and dust loadings and include a number of special sensitivity experiments (e.g., with flat topography). There is strong transient baroclinic eddy activity in the extratropics of the northern hemisphere during the northern autumn, winter, and spring seasons. The eddy activity remains strong for very large dust loadings, though it shifts northward. The eastward propagating eddies are characterized by zonal wavenumbers of 1–4 and periods of ∼2–10 days. In several simulations, the eddy variance is dominated by a single zonal wavenumber and a narrow range of periods. The longer (wavenumbers 1 and 2) transient eddies have a very deep vertical structure, exhibiting a maximum kinetic energy density at the model top (∼45–50 km) in many of the simulations. In a simulation for early northern spring the eddies are quite shallow, however. The transient eddies generate the bulk of their energy baroclincally via large meridional and vertical heat fluxes, at both lower and upper levels. This is despite the fact that their vertical structure is typically close to equivalent barotropic above the lowest 10 km. The eddies also appear to generate a substantial amount of energy barotropically, via large meridional momentum fluxes at both lower and upper levels. In the tropics and northern subtropics at upper levels (∼25–50 km) there are strong transient eddy motions with structures resembling those characteristic of inertially unstable modes. This eddy activity appears to be a response to the forcing of a region of marginal inertial stability by the extratropical transient baroclinic eddies, as the wavenumbers and periods are the same as those in the extratropics. A major surprise is the presence of very weak transient eddy activity in a number of the southern winter simulations. It appears that this is partly a consequence of the stabilizing effects of the zonally symmetric topography in the GCM, but it also must be associated with certain aspects of the zonal-mean circulation in southern winter. This is indicated by the presence of relatively large amplitude eddies in simulations for early southern autumn and spring and in a southern winter solstice simulation incorporating a different topography (derived from the Mars Digital Terrain Model). This topography differs from that used in most of the GCM simulations in not being characterized by steep symmetric slopes (which are stabilizing to baroclinic instability) in southern high latitudes. It is hypothesized that the very large extent of the southern seasonal polar cap and high elevations in the south both might contribute to weakening the transient eddy activity. Large zonally symmetric topography in the northern hemisphere of the Mars GCM also appears to have a strong impact on the transient eddies, acting to increase the dominant zonal wavenumbers and phase speeds. The properties of the GCM baroclinic eddies in the northern extratropics are compared in detail with analogous properties inferred from Viking Lander meteorology observations. The GCM eddies are found to be very similar to those observed in most respects. A notable exception is that the eddy amplitudes in the highly dusty GCM simulations are much larger than those observed by Viking during the 1977 winter solstice dust storm. This is almost certainly at least partly due to the relatively small latitudinal expansion of the Hadley circulation in the highly dusty GCM experiments.

Simulations of the general circulation of the Martian Atmosphere: 2. Seasonal pressure variations

We have simulated the CO2 seasonal cycle of the Martian atmosphere and surface with a hybrid energy balance model that incorporates dynamical and radiation information from a large number of general circulation model (GCM) runs. This information includes heating due to atmospheric heat advection, the seasonally varying ratio of the surface pressure at the two Viking landing sites to the globally averaged pressure (rk), the rate of CO2 condensation in the atmosphere, and solar heating of the atmosphere and surface. The GCM runs collectively covered a full set of seasonal dates and a large range of dust optical depths. We have compared the predictions of the energy balance model with the seasonal pressure variations measured at the two Viking landing (VL)sites and the springtime retreat of the seasonal polar cap boundaries. Numerical experiments with the energy balance model indicate that the following quantities have a strong influence on the VL seasonal pressures: albedo Ais of the seasonal CO2 ice deposits, emissivity eis of this deposit, atmospheric heat advection, and the pressure ratio rk. This last factor does not enter into the seasonal CO2 condensation/sublimation cycle in a significant way. The numerical experiments also indicate that the following factors have only a minor effect on the VL pressures: (1) the net radiative effects (solar plus thermal) of atmospheric dust at the latitudes of the polar caps, and (2) the subsurface heat conduction. The significant influence of the pressure ratio rk on the VL seasonal pressures is due to large seasonal variations in the global distribution of surface pressure. At low and mid-latitudes, these “weather” variations are engendered by seasonal changes in the Hadley circulation and by seasonal changes in the atmospheric scale height close to the surface. Comparison of the VL1 and VL2 pressures with one another provide direct evidence for the presence of such a “weather component” in the measured pressures. The differential weather component (VL2-VL1) derived from the data is reproduced approximately by the energy balance model. We find that the seasonal weather variations account for about 20% and 30% of the seasonal pressure variations measured at VL1 and VL2, respectively, that dynamical and scale height variations make comparable contributions to the weather component during years without global dust storms, and that the dynamical contribution is the larger one during years with global dust storms. Interannual variations in the weather component, rather than variations in CO2 condensation rates, are the dominant sources of the observed interannual variations of pressure during the season of global dust storms. Optimum fits to the Viking pressure measurements and the data on the polar cap boundaries are achieved with values of about 0.45 and 0.75 for Ais and eis, respectively. The former value is consistent with available photometric determinations of the albedo of the seasonal caps, while the latter value, especially in light of infrared thermal mapper brightness temperatures at high latitudes, may reflect, in part, the influence of the polar hood on the radiation balance of the winter polar regions.

Three‐dimensional numerical simulation of Martian global dust storms

We present results from the first numerical simulations of simultaneously evolving three-dimensional thermal, dynamical, and radiatively active suspended dust fields in the Martian atmosphere. Simulations of southern summer dust storms (arising from a prescribed southern subtropical surface dust source) conducted with a Mars general circulation model (GCM) illustrate the important role of dust transport by atmospheric eddies. Both traveling and stationary eddies contribute to dust transport to high latitudes in both hemispheres. These hemispheric differences arise from seasonal and topographic effects. Transport into the south polar regions is accomplished primarily by thermally and topographically forced standing eddies. Both traveling and stationary eddies transport dust to middle and high northern (winter) latitudes. Atmospheric wave motions are affected by the developing storms. Thermal tidal amplitudes increase at storm onset, with the calculated pressure response at a model grid point corresponding to the location of the Viking Lander 1 site in good agreement with observations. In qualitative agreement with observations, winter hemisphere baroclinic waves weaken during the early stages of the storm, but as the storm wanes, amplitudes of these waves increase. A slowly westward propagating (9 degrees of longitude per sol) zonal wavenumber one feature in the temperature and geopotential fields at middle northern latitudes amplifies rapidly during the initial sols (Martian solar days) of the simulated storms. This feature is suggestive of the observed north polar warming which occurred during the 1977B global dust storm, but the simulations produce a much weaker polar warming (∼10 K at 0.5 mbar) than was observed (40–50 K). The globally integrated CO2 condensation rate decreases by 15–20% during the simulated dust storm onset and would likely be decreased more if a stronger polar warming were produced. During the initial stages of the simulated storms, surface stress values in the southern subtropics intensify due primarily to the intensification of the Hadley circulation and thermally driven tides. This supports the hypothesis that these components of the general circulation contribute strong positive feedbacks to the developing storms.

Mars atmospheric dynamics as simulated by the NASA Ames general circulation model: 3. Winter quasi‐stationary eddies

A set of simulations with the NASA Ames Mars general circulation model (GCM) has been analyzed to define the basic properties and dynamics of quasi-stationary eddy circulations in the winter hemisphere. These circulations, differing substantially from those in low latitudes and the summer hemisphere, extend from the surface to the model top at ∼47 km; typically, the largest geopotential and wind perturbations are found at the highest model level. Near the surface the eddies are of largest amplitude in lower latitudes, but above ∼5–10 km the amplitudes are a maximum in middle and high latitudes. The vertical structure of the eddies is nearly equivalent barotropic; the phase variation in latitude is also relatively small. Zonal wavenumbers 1 and 2 dominate the stationary circulations, with wave 3 being of significance only at low levels (below ∼10 km). The quasi-stationary eddies differ substantially in northern and southern winters, and undergo considerable changes with increasing dustiness in northern winter. The southern winter eddy circulation is dominated by zonal wavenumber 1, while the northern eddies have a strong wavenumber 2 component at low levels of dust loading; at high dust levels wavenumber 1 becomes dominant. This change may be at least partly the result of the dusty zonal-mean state becoming more responsive to wave 1 forcing. The standing eddies in a GCM simulation for Ls ∼ 0° exhibit considerable similarity in basic structure and amplitude to those revealed by an recent analysis of Viking IRTM data [Banfield et al., 1996], though there are differences in the phases of the eddy patterns.

Meteorological predictions for the Mars Pathfinder lander

The NASA Ames Mars general circulation and boundary layer models are used as guides to forecast the meteorological environment of the Pathfinder lander site. Based on these models we predict that for a Viking-like atmospheric dust loading, significant vertical wave structure will be seen in the entry temperature profile above 50 km. Temperatures in this region will oscillate by up to 30 K about a mean value of ∼145 K. At the surface during the primary mission, winds are expected to rotate clockwise during the day at 2–10 m/s, while air temperatures will range from overnight lows in the mid 180s K to daytime highs near 250 K. If the atmosphere is dust free at arrival, as appears possible from recent Earth-based observations, then we expect a cooler upper atmosphere (130–140 K) with less wave structure during entry, weaker surface winds, and a slight increase in near surface air temperatures. Regardless of dust loading, we predict that the daily averaged surface pressure will reach its annual minimum about 15–20 sols after landing. During the extended mission, the basic character of Pathfinders meteorology will change from that of a summertime quasi-regular regime to that of a wintertime highly variable regime. This transition is predicted to occur by sol 60. The strongest winds (30 m/s) and lowest daily-averaged temperatures (∼200 K) will be recorded during early winter.