Geophysical Research Letters

Effects of vertically propagating thermal tides on the mean structure and dynamics of Mars' lower thermosphere

Authors


Abstract

[1] A general circulation model of Mars' atmosphere is used to elucidate the effects of vertically propagating thermal tides excited near the surface on the zonal mean density, temperature and wind structure of Mars' atmosphere between 90 and 160 km. The effects are substantial, and amount to order 10–40%, 10–50 K and 50–160 m s−1, respectively, at these altitudes under low dust conditions (dust visible optical depth = 0.3). These induced changes in the zonal mean structure are dominated by the non-migrating tides. Of particular note are the mean density perturbations, which represent an important consideration for modeling of the aerobraking region (100–170 km).

1. Introduction

[2] Vertically-propagating waves represent a well-known mechanism for transporting momentum in planetary atmospheres. In Earth's atmosphere tropospherically-generated gravity waves become sufficiently large above about 70 km that they become unstable and give up their momentum to the mean flow. In addition, vertically-propagating solar thermal tides undergo molecular dissipation at the base of the thermosphere (100–150 km), modifying the zonal mean wind field in this region [Miyahara, 1978; Miyahara and Wu, 1989; Angelats i Coll and Forbes, 2002]. The situation is somewhat different in Mars' atmosphere. The effects of gravity waves on the mean circulation of the middle and upper atmosphere are largely unknown, but existing evidence [Collins et al., 1997; Joshi et al., 1995; Theodore et al., 1993; Forget et al., 1999] suggests the consequences to be not nearly as profound as in the terrestrial atmosphere. However, in Mars' atmosphere a direct (thermally-driven) meridional circulation exists that extends from subsolar latitudes in the summer hemisphere to opposite latitudes in the winter one without the need for gravity wave forcing [Wilson, 1997a, 1997b; Bougher et al., 2006]. As noted by Wilson [1997a, 1997b] and Bougher et al. [2006] the main consequence of this Hadley cell is a polar warming at high winter latitudes, in the 30–80 km height region as well as at altitudes as high as 100–130 km. Further, polar warming effects are most notable at the solstice seasons and are intensified as the dust content in the summer hemisphere increases. Wilson [1997a, 1997b] also provides evidence that the momentum flux divergence due to dissipating tides enhances the inter-hemispheric solstitial Hadley cell, and may enable its extension beyond 70°N to the winter pole.

[3] Thermally-driven or solar tides appear to play a more dominant role in the dynamics of Mars' atmosphere than at Earth. The spectrum of tides is composed of in-situ generated tides, and tides propagating from the surface that are composed of migrating (westward-propagating, sun-synchronous) and non-migrating tides (non-Sun-synchronous) that result from the interaction of the Sun's diurnal heating cycle with the surface. Accelerometer data [Withers et al., 2003; Wang et al., 2006], radio occultation data [Hinson et al., 2008] and different modeling efforts [Wilson and Hamilton, 1996; Angelats i Coll et al., 2004; Lewis and Read, 2003; Moudden and Forbes, 2008] have contributed significantly to our understanding of the tidal spectrum. At aerobraking altitudes both migrating and eastward non-migrating tides coexist with comparable amplitudes. The non-migrating tides originate from the interaction of the migrating components with the surface zonal asymmetries, and the long vertical wavelengths of the eastward non-migrating tides enable them to propagate to higher altitudes than many of their westward counterparts. Thermal tides cause local variations with longitude and local time in upper atmospheric fields. We however focus in this study on the effect of thermal tides excited near the surface (not in-situ generated) on zonal mean properties. Lewis and Read [2003] utilize a GCM (general circulation model) to understand the effects of thermal tides on the zonal mean dynamics of Mars' atmosphere with emphasis on the formation of equatorial jets during the equinox season. They illustrate the formation of an eastward jet (superrotation of tens of m s−1) in the 0–30 km height regime that is driven by a momentum flux induced by the diurnal migrating tide [cf. Fels and Lindzen, 1974]. The effect is enhanced for increased dust loading of the atmosphere. At higher altitudes (>50 km) where the diurnal tide dissipates, a westward jet (also tens of m s−1) is formed [see Forget et al., 1999].

[4] The results from Lewis and Read [2003] are confined to altitudes below 80 km due to limitations of the model employed in their study. Relatively little is known about the effects of tides on the zonal mean thermal and dynamical structure of Mars' atmosphere above 80 km. Using a relatively simple model, Forbes and Miyahara [2006] demonstrate that the migrating semidiurnal tide may produce alterations in the mean temperatures and zonal winds of order 20–70 K and 20–200 m s−1 between 100 and 200 km. They also show that the diurnal eastward-propagating tide with zonal wavenumber 2 can affect the zonal mean wind at the 10–50 m s−1 level. These results suggest that the full spectrum of migrating and non-migrating tides might have aggregate effects. The question naturally arises as to whether dissipation of the full spectrum of upward-propagating tides in the lower thermosphere (ca. 100–150 km) of Mars can measurably modify the zonal mean winds, temperatures and densities at these altitudes. Such effects are estimated to be of some significance on Earth [e.g., Angelats i Coll and Forbes, 2002; Forbes et al., 2006] but are of unknown amplitude in Mars' upper atmosphere. It is the purpose here to provide this perspective. We demonstrate that important modifications of the zonal mean temperature and wind fields occur due mostly to dissipating eastward-propagating non-migrating tides. Moreover, the zonal mean density field is strongly altered; this latter result is important in that it provides improved understanding of the mechanisms that drive density variability of the atmospheric regime where aerobraking occurs (ca. 100–170 km).

2. Model

[5] The model used in this study is the Global Mars Multiscale Model (GMMM) [Moudden and McConnell, 2005]. It is a general circulation model based on the non-hydrostatic, semi-Lagrangian and semi-implicit dynamics of the Global Environmental multiscale Model (GEM) [Côté et al., 1998]. The physics and chemistry parameterizations are described in detail by Moudden and McConnell [2005, 2006] and Moudden [2007]. The experiments performed here use a uniform horizontal resolution of 9° by 9° and the atmosphere extends up to 160 km altitude. The time-step is 0.01 sol or 887.75 seconds (or 14.8 min). This model is also employed in a recent study [Moudden and Forbes, 2008] where numerical experiments are performed to better understand the origins of various waves in the aerobraking region (ca. 90–170 km), and in particular their connections with specific wavenumber components of the surface topography. As the focus here is on the tides excited near the surface, all experiments carried for this study use a diurnally averaged solar flux for the thermospheric ultraviolet heating thus eliminating in-situ generated tides. The control experiment includes a diurnal solar cycle and a surface topography, thermal inertia and albedo that include zonal asymmetries. In other simulations the diurnal cycle and/or the surface zonal asymmetries are removed to identify the effects of tides. The reader is referred to Moudden and Forbes [2008] for further information on the model and model inputs, validation of the model, and fidelity of the model in reproducing the zonal density variations and associated meridional trends observed by the Mars Global Surveyor accelerometer instruments [Wilson, 2002]. Our simulations assume a horizontally-uniform distribution of dust with visible optical depth τ = 0.3 at the surface and correspond to Ls = 60° (Ls refers to the solar aerocentric longitude, Ls = 0° corresponds to the start of northern spring).

3. Results

[6] We compare latitude-altitude cross sections of zonal mean density, temperature, zonal wind, and vertical wind for simulations without any tides and simulations with migrating and non-migrating tides. Figures 1a and 1d illustrate the distributions of zonal-mean temperature equation image and zonal wind equation image, calculated including full surface topography and diurnal cycle of heating (control simulation). Figures 1b and 1e illustrate equation image and equation image when zonally uniform topography, thermal inertia and albedo are employed (removing the source of non-migrating tides [see Moudden and Forbes, 2008]) and eliminating the diurnal cycle (removing the source of migrating tides). Comparing these two simulations reveals the influence of both migrating and non-migrating tides on the zonal mean temperature and wind structure of Mars' atmosphere. The difference in equation image and equation image attributable to thermal tides is shown in 1c and 1f. With the removal of solar thermal tides (Figures 1b and 1e), the Southern Hemisphere middle atmosphere (20–100 km) eastward jet is now totally confined to the southern hemisphere, while the northern hemisphere westward jet is intensified. Above 100 km a broad westward jet exists that extends nearly pole-to-pole with maxima over the equator of order 60–200 m s−1. The differences seen in Figure 1d as compared with 1 e are consistent with predominantly eastward momentum being dumped into the mean flow by the dissipation of eastward-propagating (i.e., non-migrating) tides. In the absence of solar thermal tides a warm region occurs in the temperature field (Figure 1b) above 100 km and poleward of −40° latitude. This feature is due to subsidence heating associated with a large circulation cell (see Figure 2a). With the addition of tides (Figure 1a) this warming is much diminished, indicating a counter effect of tides on the mean meridional circulation.

Figure 1.

(top) Zonal mean temperatures (in K) and (bottom) zonal winds (in m s−1) (a and d) with diurnal cycle and full surface topography and (b and e) with no diurnal cycle and zonally uniform topography, surface albedo and thermal inertia for low dust optical depth τ = 0.1. Differences between (c) Figures 1a and 1b and (f) Figures 1d and 1e indicate the influence of both migrating and non-migrating tides on the zonal mean temperature and wind structure of Mars atmosphere for low dust conditions. The X axis represents latitudes in degrees and Y axis the height in km.

Figure 2.

(a) Zonal average of mass stream function in 109 kg s−1 (control simulation) and difference fields illustrating the contributions of non-migrating and migrating thermal tides to (b) vertical velocity in cm s−1 and (c) density in %. Axes as in Figure 1.

[7] To better illustrate the above effects, and to bring total mass density into the discussion, Figures 1c and 1f illustrate difference fields obtained by differencing the zonal mean fields from the two simulations discussed above. Figure 1c indicates the changes in temperature that occur when neither migrating and non-migrating tides are excited in the simulation. The mean cooling poleward of −40° latitude and above 100 km is seen, as well as a cooling region above 130 km at equatorial and tropical latitudes and a less intense warming region just below it. These thermal effects generated by tides are all produced by adiabatic heating and cooling from the alteration of the planet-wide circulation. Figure 2a shows the meridional circulation represented by the mass stream function in units of 109 kg s−1. Momentum flux divergence by the dissipating tides alters the large scale horizontal winds that in turn modify the large scale vertical movements directly linked to the horizontal wind divergence through mass conservation. The difference in vertical velocity between the control simulation with thermal tides and one where they were not excited (Figure 2b) shows a structure consistent with the temperature differences in Figure 1c, which in turn explains the density changes in Figure 2c. Figure 2c illustrates the changes in zonal mean total mass density due to the effects of tides. Between 100–160 km there are reductions of 10–60% in background density at middle to high latitudes associated with the cooling indicated in Figure 1c. There is also an equatorially-centered region of 10–30% enhanced densities between 110 and 160 km. Figure 1f illustrates the corresponding changes in zonal mean zonal winds resulting from all vertically propagating tides. Note that these are predominantly positive, nearly symmetric about the equator, and range between 100–160 m s−1 above 100 km.

[8] The above changes in temperature, zonal wind and density are the total effects from both migrating and non-migrating tides. To help isolate these effects we performed another simulation where the surface topography, albedo and thermal inertia are set to their zonal average while keeping the diurnal cycle, thus eliminating the source of non-migrating tides (migrating tides only simulation). Figure 3 shows the effects of non-migrating (Figure 3 (top)) and migrating (Figure 3 (bottom)) tides in the zonal mean fields of temperature (Figures 3a and 3d), density (Figures 3b and 3e) and zonal wind (Figures 3c and 3f). The effect of non-migrating tides is obtained by substracting the zonal mean fields in the migrating tides only simulation from zonal mean fields in the control simulation. The effect of non-migrating tides is obtained by substracting the zonal mean fields in the no-tides simulation from zonal mean fields in the migrating tides only simulation. It is apparent that the non-migrating tides are dominating in the changes affecting all zonal mean fields. This is expected since it is known that by having longer vertical wavelengths, the non-migrating eastward propagating tides manage to reach higher altitudes before they dissipate. The density changes in Figures 3b and 3e are readily explained by the temperature difference in Figures 3a and 3d and the dependence of the density scale height on the temperature, thus being indirectly linked to the modification of the general circulation by tides. The effect of non-migrating tides on zonal winds is large: it reaches 200 m s−1 in our simulation at 140 km altitude, mostly in the winter hemisphere. The similarities between mean wind changes due to non-migrating tides above 60 km (Figure 3d) and the diurnal Kelvin waves [see amplitudes of these tides in] [Moudden and Forbes, 2008] lead us to suggest that these wind effects are reminiscent of Kelvin-like waves known to be strongly present in Mars' upper atmosphere, e.g., diurnal eastward propagating tides with zonal wavenumbers 1 and 2 and perhaps others [Wilson, 2002; Hinson et al., 2008]. Migrating tides produce a net momentum deposition in the westward direction that partially reverses the non-migrating effect.

Figure 3.

Difference fields illustrating the contributions of (top) non-migrating and (bottom) migrating tides to (a and d) temperature in K, (b and e) density in % and (c and f) zonal wind in m s−1. Axes as in Figure 1.

4. Conclusion

[9] Vertically propagating tides produce local variability in time and longitude of up to 30% in density and 20 K in temperature fields in Mars' thermosphere [see Moudden and Forbes, 2008] in low dust conditions. This local variability doesn't appear in the zonal and temporal averages. This study shows that tides excited near the surface do have a substantial effect on the mean wind, temperature and density through their irreversible momentum transport. This effect amounts to 50–160 m s−1, 10–50 K and 10–40% respectively in the thermosphere of Mars, thus being equal or larger than wave local amplitudes. By depositing their momentum, the tides with largest amplitudes produce a net effect that weakens westward winds and strengthens eastward ones. The change in zonal winds also alters the meridional circulation and the vertical velocities which in turn modify the thermal structure through adiabatic heating and cooling. The thermospheric polar regions in the winter hemisphere are significantly cooler which indicates a weakening of the winter polar warming by tides. The collapse of the atmosphere in colder regions and expansion in warmer ones due to the thermal effect of tides produce an average density deviation of nearly 50% near 160 km altitude. Most of the above effects are produced by non-migrating tides through eastward momentum deposition, the migrating tides have a small moderating effect from westward momentum deposition.

Acknowledgments

[10] This work was supported under grant ATM-0346218 from the National Science Foundation to the University of Colorado.