Titan's diverse inventory of photochemically produced gases can be used as tracers to probe atmospheric circulation. Since the arrival of the Cassini–Huygens mission in July 2004 it has been possible to map the seasonal and spatial variations of these compounds in great detail. Here, we use 3.5 years of data measured by the Cassini Composite InfraRed Spectrometer instrument to determine spatial and seasonal composition trends, thus providing clues to underlying atmospheric motions. Titan's North Pole (currently in winter) displays enrichment of trace species, implying subsidence is occurring there. This is consistent with the descending branch of a single south-to-north stratospheric circulation cell and a polar vortex. Lack of enrichment in the south over most of the observed time period argues against the presence of any secondary circulation cell in the Southern Polar stratosphere. However, a residual cap of enriched gas was observed over the South Pole early in the mission, which has since completely dissipated. This cap was most probably due to residual build-up from southern winter. These observations provide new and important constraints for models of atmospheric photochemistry and circulation.
Saturn's largest moon, Titan, has the most substantial atmosphere of any planetary satellite in our Solar System. Titan's atmosphere has a surface pressure of approximately 1.5 bar and is composed mainly of molecular nitrogen (95–99%) and methane (1–5%). The atmospheric temperature structure is similar to the Earth's, although much colder, with a well-defined tropopause and stratopause (figure 1). Temperatures range from 70 K at the tropopause up to approximately 220 K in the winter stratosphere. Titan's low gravity and high surface pressure mean that its atmosphere is highly extended vertically.
Titan's orbit lies in Saturn's equatorial plane, so it has an axial tilt (obliquity) of 26.7° to the ecliptic and a year equal to 29.46 Earth years, the same as Saturn. Therefore, Titan experiences seasons and numerical models predict different atmospheric circulation regimes throughout the Titan year as the point of maximum solar heating shifts in latitude (Hourdin et al. 1995; Tokano et al. 1999; Luz et al. 2003; Hourdin et al. 2004; Rannou et al. 2005; Richardson et al. 2007). This results in a stratospheric circulation that switches from two symmetrical Hadley cells at the spring and autumn equinoxes, to a circulation dominated by one large hemisphere-to-hemisphere Hadley cell during inter-equinox periods. The stratosphere appears to have zonally symmetric temperature (Flasar et al. 1981, 2005) and composition (Teanby et al. 2006, 2008) structures—with very little longitudinal variation. This suggests that non-symmetric atmospheric features such as planetary waves, which are important for creating non-zonal structures on the Earth, are less significant on Titan.
Titan has a very photochemically active atmosphere. Solar UV photons and magnetospheric electrons break apart nitrogen and methane molecules to form radicals such as CH3 and N (Vuitton et al. 2007). These radicals combine to form more complicated photochemical daughter products, which in turn are broken up again to form even more complex species. The result is a rich atmospheric chemistry, with a vast array of trace gases. Titan's haze, which gives it an orange appearance at visible wavelengths, is also a product of this photochemistry. Most trace species are created in the upper atmosphere above 500 km (Wilson & Atreya 2004) and infiltrate lower altitudes by eddy diffusion. Eventually, most compounds condense near the tropopause cold trap and are removed from the stratosphere. This source–sink arrangement should result in stratospheric abundance profiles of trace species and haze that increase with increasing altitude, towards the production zone.
Most of Titan's trace gases have spectral features in the mid-infrared. Therefore, infrared remote sensing, either from spacecraft or ground-based telescopes, provides an excellent way to determine the global distribution of these gases. It is also possible to measure the composition directly using mass spectrometers, such as Cassini's INMS instrument (Waite et al. 2005) and the Huygens probe GCMS (Niemann et al. 2005). These direct samplings provide important constraints on atmospheric composition, but do not provide global coverage—this requires infrared remote sensing. Figure 2 shows an example of Titan's mid-IR spectrum and shows the vast array of compounds present in the atmosphere. Emission features in this region of the spectrum are produced in the middle and upper stratosphere in the 10–0.1 mbar pressure range and can thus be used to determine stratospheric temperature and composition.
The rich photochemistry of Titan is extremely interesting in its own right, as it has important implications for the Earth's early atmosphere, which is believed to have had a similar composition. However, in this paper we concentrate on exploiting variations in composition across Titan's disc to probe its equally fascinating atmospheric circulation. Because trace gases have a vertical gradient, they can be used as tracers of vertical motion in the atmosphere, as illustrated in figure 3. Upward motion of the atmosphere advects tracer-poor air from the lower atmosphere, and as a consequence of the vertical profile leads to a decrease in abundance at each atmospheric level. Conversely, subsidence advects tracer-rich air from the upper atmosphere to lower levels and leads to an increase in abundance. In this paper, we use this simple consequence of Titan's photochemistry to probe atmospheric circulation. Owing to the zonal symmetry of Titan's atmosphere, variations in composition with latitude are the most significant.
The first observations of latitudinal composition variations were from the Voyager I flyby of Titan in 1980 (Coustenis & Bézard 1995). These observations showed an enrichment in most species near the North Pole, which at that time was experiencing early spring. The next observations of composition variations were by Roe et al. (2004) using the Keck I telescope, which showed an enrichment of C2H4 over the South Pole from data taken in 1999–2002 (late northern autumn). A vast increase in our understanding of Titan's atmosphere was made possible by the arrival of the Cassini–Huygens spacecraft at the Saturnian system in July 2004.
The Composite InfraRed Spectrometer (CIRS) instrument on-board the Cassini orbiter has so far provided millions of far- and mid-IR spectra that can be used to determine the detailed composition of Titan's atmosphere. So far, latitudinal variations in composition using CIRS data have been determined by Flasar et al. (2005), Teanby et al. (2006), Coustenis et al. (2007), de Kok et al. (2007a) and Teanby et al. (2008). De Kok et al. (2007b) discussed variations in Titan's photochemical hazes. These studies all found enrichment of trace species or haze near the northern winter pole. Wind velocities derived from CIRS spectra by Flasar et al. (2005) and Achterberg et al. (2008) have indicated the presence of a North Polar vortex. This is corroborated by the composition mapping from Teanby et al. (2007). Therefore, current Cassini observations are consistent with subsidence occurring inside a North Polar vortex, leading to enrichment of trace species.
Vertical profiles of many of Titan's trace species and hazes were determined by de Kok et al. (2007a,b), Teanby et al. (2007) and Vinatier et al. (2007). These studies confirm that around the equator, the abundance of trace gases and haze increases with altitude. This corroborates disc-averaged ground-based studies in the sub-mm regime (e.g. Hidayat et al. 1997; Marten et al. 2002; Gurwell 2004; Livengood et al. 2006), which also inferred abundance profiles that increased with altitude. Near the North Pole the profiles become more complicated (Teanby et al. 2007; Vinatier et al. 2007).
Seasonal changes in composition were first noted at mid-northern latitudes by Roe et al. (2004) who compared ground-based C2H4 determinations with those from Voyager I. This was the first confirmation that Titan's atmospheric composition is indeed seasonally dependent. Comparison of Cassini and Voyager observations shows that some species were more enriched at the Voyager epoch (early northern spring) than at the Cassini epoch (northern winter). This also suggests seasonal variations. However, direct comparison of the results from different instruments requires us to assume that all instrument-related offsets have been accounted for. The first determination of seasonal variations using a single instrument (Cassini CIRS) was presented by Teanby et al. (2008), which showed a trend for decreasing trace species abundance in the south. This study has the advantage that it is based on continuous observations from one instrument, but has the disadvantage that it only covers 2.5 years or one-twelfth of a Titan year. In this paper, we extend the coverage to 3.5 years (almost the entire prime mission) and include variations of additional gases, CO2, C2H4 and C2H6. The results are used to investigate seasonal and latitude variations in composition and probe general features of Titan's atmospheric circulation.
2. Cassini CIRS observations and analysis
The CIRS instrument is described in detail by Kunde et al. (1996) and Flasar et al. (2004). Briefly, CIRS has three focal planes covering three different spectral regions: FP1 (10–600 cm−1); FP3 (600–1100 cm−1); and FP4 (1100–1400 cm−1). The spectral resolution is adjustable to have an apodized full-width half-maximum of 0.5–15 cm−1, but is generally used at 0.5, 2.5 and 15 cm−1 resolutions.
In this paper, we use 2.5 cm−1 resolution mapping and 0.5 cm−1 resolution single-swath nadir (downward looking) observations, which have sufficient spectral resolution to resolve individual gas peaks. Only data from FP3 and FP4 are used as these have better spatial resolution than far-IR FP1 data. Examples of latitude/longitude coverage for both observation types are shown in fig. 1 of Teanby et al. (2008). Typically, observations have a spatial resolution of 2–6° latitude.
Using two data resolutions provides complementary information and acts as a consistency check on the results. The 2.5 cm−1 mapping observations have better spatial and temporal coverage, whereas the 0.5 cm−1 swath observations contain more altitude information on the temperature structure and have better signal to noise because a sit-and-stare viewing geometry allows many more spectra to be taken at each location.
Composition mapping by Teanby et al. (2006, 2008) shows little variation with longitude. Therefore, to increase signal to noise, data were averaged into 10° wide latitude bins with restricted emission angle range following Teanby et al. (2008). The data cover 39 flybys for almost the full 4 years of Cassini's prime mission and are summarized in table 1.
The method used to determine atmospheric temperature and composition is described in detail in Teanby et al. (2008). In summary, we use a constrained iterative nonlinear inversion technique (Irwin et al. 2008) based on the correlated-k approximation (Lacis & Oinas 1991). The inversion is a two-stage process. First, inverting for a continuous temperature profile using the ν4-band of methane (1240–1360 cm−1). Second, fixing the temperature and inverting for gas abundances using the 600–1000 cm−1 spectral region and assuming constant gas volume mixing ratios above the condensation level.
In addition to the gases studied by Teanby et al. (2008) we also invert for C2H4 and C2H6 abundance using 790–1000 cm−1. For this region of the spectrum, we found that a constant infrared haze cross section with wavenumber gave the best fit to the data. For other spectral regions, a cross section generated using the refractive indices of Khare et al. (1984) was used. Spectral data for gases and haze are as described in Teanby et al. (2008) and de Kok et al. (2007b), except for C2H6, for which we use updated line parameters from Vander Auwera et al. (2007). Figure 4 shows example fits to the observed spectra.
3. Latitude variations
Figure 5 shows the variation in trace gas abundance as a function of latitude. In this plot, the results from all flybys have been averaged together using a cubic spline fit (Teanby 2007). All species (except CO2) show an increase in abundance towards the northern winter pole. This is in agreement with previous Cassini CIRS results and is consistent with subsiding air over the pole. Composition cross sections (Teanby et al. 2007) along with derived thermal winds (Flasar et al. 2005; Achterberg et al. 2008) indicate the presence of a polar vortex at 25–55° N encircling the North Pole at this time. The vortex acts to separate enriched from unenriched air. As would be expected, this separation is most effective for short lifetime gases such as HC3N and C4H2—long lifetime gases such as HCN have a more diffuse variation as they have time to escape the vortex. The fact that enrichments are strongest at 90° N implies that subsidence is strongest in the vortex core.
One striking difference between model predictions and the CIRS observations is that the massive enrichment of trace species predicted by numerical models in the south (e.g. Rannou et al. 2005) is not present in the data. In the models, enrichment is caused by a residual ‘cap’ left over from southern winter combined with a small secondary stratospheric circulation cell rising at approximately 60° S and descending at the South Pole. The discrepancy between models and measurements indicates the following. (i) There is no secondary stratospheric circulation cell near the South Pole at this time for levels above 10 mbar (where our data probe), and Titan's stratospheric circulation is best described by a single south–north Hadley cell. This is consistent with the lower-than-expected stratospheric temperatures at the South Pole observed by Flasar et al. (2005), which also suggests upwelling in this region. (ii) Any residual South Polar enrichment left over from southern winter has mostly dissipated. Gross offsets between the data and the model are not critical and can be reduced by adjusting the haze distribution and photochemical production rates (Crespin et al. 2008).
Trace gas enrichments allow us to test our assumption that vertical profiles are a result of photochemical production and vertical mixing. If this assumption is correct then species with shorter photochemical lifetimes should have steeper gradients, as they decay before mixing to lower altitudes occurs. This implies that the species with the shortest lifetime should be enriched by the largest factor in the north relative to equatorial regions. We test this simple idea by plotting the photochemical lifetime from Wilson & Atreya (2004) against the relative enrichment at 60, 70 and 80° N in figure 6. The hydrocarbons and carbon dioxide do indeed display the expected behaviour—with an almost linear trend in log–log space. However, the nitriles HCN and HC3N have relatively greater enrichment, which implies steeper profiles than expected from the photochemistry alone. Vinatier et al. (2007) also found that the HCN vertical profile was too steep to be explained by photochemical production and eddy mixing. This suggests that there is an additional sink for nitriles in the lower atmosphere—perhaps incorporation into photochemical hazes (Lebonnois et al. 2002; Wilson & Atreya 2003).
4. Seasonal variations
The 4 years of Cassini's prime mission provides us with a self-consistent dataset with which to probe seasonal variations in atmospheric composition covering just over half a Titan season (early–mid northern winter). However, due to Cassini's changing orbital geometry the temporal coverage is not identical for all latitudes. In particular, high northern latitudes were only visible during the 180° orbital transfer (mid-2006 to mid-2007). Looking for seasonal variations at northern latitudes is also complicated by the presence of large temperature and composition gradients. Therefore, the best place to look for seasonal variations with the current CIRS dataset is at southern latitudes. Figure 7 shows the variation of each species at 55° S along with those predicted from the numerical model of Rannou et al. (2005). This latitude is the furthest south we can probe while maintaining a well-sampled time series.
The data show a steady decrease in abundance of C2H2, C2H4, C3H4 and HCN until early 2006, when the southern stratospheric composition stabilizes. C4H2 and HC3N have very low abundance at southern latitudes and are hence very noisy. Over the same period, the temperature at 3 mbar increases by approximately 2 K. This proves that Titan's atmospheric temperature and composition are not in a steady state. The reduction in abundances could be due to dissipation of trace species built up at the South Pole by subsidence during southern winter and/or current upwelling at southern latitudes. Models also predict that abundances and temperatures should decrease and increase, respectively. However, for most species, the observed abundance decreases are less than predicted by the model, especially after early 2006, when the data show Titan's atmospheric composition to be stable. This implies that residual southern winter build-up had largely dissipated before the arrival of Cassini, with the last stage occurring around early 2006. Note that, as would be expected, long lifetime species such as CO2 and C2H6 show no significant seasonal variation.
Figure 5 shows that for most gases the variations with latitude are similar to those derived from the Voyager flyby, although greater enrichment was seen at the Voyager epoch for HCN and C2H4. However, CIRS is observing Titan earlier in the seasonal cycle and gas concentrations may continue to increase as the winter season progresses and subsidence continues. Another possibility is that the difference is due to the changes in the haze vertical distribution between Voyager and Cassini epochs (Porco et al. 2005)—if HCN is indeed important for haze formation then haze and HCN abundances will be linked.
During Cassini's orbital insertion in July 2004 it flew directly over the South Pole. This happened again in January 2007 during the 180° orbital transfer. These two datasets have the same viewing geometry and allow us to look for dissipation of any South Polar build-up (figure 8). C2H2, C2H4, C3H4 and C4H2 all show a slight residual cap of enriched gas over the South Pole in 2004, which is not present in 2007. This cap has not been recognized before in the CIRS data. In addition, South Polar C2H4 has been measured from ground-based measurements by Roe et al. (2004), who reported an enrichment of at least a factor of five over the South Pole during 1999–2002. In the CIRS data, we see a factor of two enrichment in 2004 and no enrichment by 2007. These observations would be consistent with a slowly dissipating South Polar build-up due to upwelling or photochemical decay, which are both consistent with the absence of a secondary circulation cell in the middle stratosphere. The time scale of these dissipations provides a powerful constraint for dynamical and photochemical models.
The first 3.5 years of data from Cassini CIRS has been used to determine the composition of Titan's atmosphere as a function of latitude and time for early–mid northern winter. If we assume that trace gas abundances generally increase with altitude due to high-altitude photochemical production combined with condensation at the tropopause, then subsidence leads to enrichment and upwelling leads to depletion in the stratosphere where gas emission lines are formed. This simple idea has been used to probe the atmospheric circulation, summarized as follows.
Trace species are more abundant in the north. This implies that subsidence is occurring there during the present epoch.
The relative enrichment of trace species increases with decreasing photochemical lifetime. This implies that shorter lifetime species have steeper vertical gradients, in keeping with ideas about processes that define the profile shapes.
HCN and HC3N have greater enrichment than hydrocarbons with comparable lifetimes, implying that these gases have a steeper vertical profile than expected from photochemistry alone. Perhaps this is due to an additional sink in the lower atmosphere such as incorporation into hazes.
Contrary to numerical model predictions, there is no enrichment of trace gases in the Southern Hemisphere in the time-averaged abundance measurements. This is a new constraint from Cassini (Voyager did not cover high southern latitudes) implying that there is no secondary stratospheric cell and that Titan's stratospheric circulation is best described as a single south–north Hadley cell. It also implies that any residual trace gas build-up from southern winter has now largely dissipated.
However, in the individual flyby data from July 2004 there is evidence for an enriched South Polar cap, most probably a residual from subsidence in southern winter, although the enrichment is much less than predicted by models. By early 2007, this polar cap had fully dissipated. Observations at 55° S indicate that a stable atmospheric composition at temperate southern latitudes is reached by early 2006.
The abundance of trace species at 55° S has decreased from 2004 to 2008, consistent with upwelling in the south at present and/or dissipation of southern winter build-up.
The rate of these seasonal variations provides important new constraints for dynamical models of Titan's atmosphere.
We gratefully acknowledge the financial support from the UK Science and Technology Facilities Council. The authors also thank the Cassini CIRS instrument team for making this research possible.
One contribution of 14 to a Discussion Meeting Issue ‘Progress in understanding Titan's atmosphere and space environment’.
- © 2008 The Royal Society