SEASONAL DISAPPEARANCE OF FAR-INFRARED HAZE IN TITAN'S STRATOSPHERE

A distinguishing characteristic of Titan's atmosphere is its ubiquitous haze. It has been known since the Voyager encounters in the early 1980s that a thick photochemical haze obscures the surface and lower atmosphere from view in visible images (Smith et al. 1981). The haze, which is layered in structure, is not only a repository of chemicals created in the upper atmosphere, but is also a source of the precipitation that forms refractory organic deposits on the surface. Because it mediates the heating and cooling of the atmosphere and surface, the haze plays a major role in the generation of weather and atmospheric circulation. During the two northern winter–spring seasons in which Voyager and Cassini have observed Titan, a haze "hood" has covered latitudes beyond about 55°N (Sromovsky et al. 1981; Porco et al. 2005). Related to this polar hood, Griffith et al. (2006) identified a tropospheric cloud at 51°N to 68°N which they suggested was composed of condensed ethane. The polar haze hood, together with a global north–south asymmetry of the haze, indicates that the haze structure has a seasonal dependence. Indeed, West et al. (2011) reported a decrease in altitude, at equinox, of a detached haze layer seen at latitudes outside the northern polar region. The Composite Infrared Spectrometer (CIRS) on Cassini has identified thermal emission from stratospheric aerosols and condensation clouds at a variety of latitudes on Titan, with enhancement northward of about 50°N (de Kok et al. 2007; Vinatier et al. 2010a, 2012b; Samuelson et al. 2007; Anderson & Samuelson 2011; Cottini et al. 2012). The northern haze region also coincides with increases in abundances of many organic vapors (Coustenis & Bézard 1995; Coustenis et al. 2007; Coustenis et al. 2010; Teanby et al. 2010).

We report here measurements from Cassini of a marked seasonal decrease in the intensity of a prominent haze emission feature in far-infrared spectra of Titan's polar region. The origin and chemical composition of this particular haze is unknown, despite numerous attempts at its identification. Attributed to a volatile or refractory material, it is probably a condensate formed from gases downwelling at the winter pole. Seen originally in spectra recorded by the Infrared Interferometer Spectrometer on Voyager 1 (Kunde et al. 1981), this band at 220 cm⁻¹ was first characterized by Coustenis et al. (1999), who found it to be confined to latitudes northward of 50°N and to altitudes above 100 km. Until now it had not been recognized that this haze emission is time dependent.

We compiled spectra from CIRS (Flasar et al. 2004) during 45 Titan flybys between 2004 October and 2012 February. Examples of the data are shown in Figure 1. We used spectra recorded at 15 cm⁻¹ resolution and covering 10–600 cm⁻¹. To stay inside the northern haze hood we confined the observations to latitudes northward of 69°N and imposed no restriction on longitude. Observations during each flyby were averaged into single measurements, with nadir and limb data kept separated. Our measurements are shown in Figure 2. The number of spectra in the disk averages ranged from 6 to 713 and in the limb averages from 3 to 189 (the error bars in Figure 2 were determined from the number of spectra in each measurement). Nadir averages included all 0°–90° emission angles, while limb averages covered 40–160 km tangent height (the distance at the tangent point of the field-of-view center above the surface). The 220 cm⁻¹ emission forms at 80–150 km altitude and peaks at approximately 140 km (Anderson et al. 2012; de Kok et al. 2007).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Our data set includes wide ranges of spacecraft distance, emission angle, and tangent height. The spacecraft distances to Titan were as small as 2200 km and as large as 188,000 km. At most distances there was significant mixing of nadir, limb, and deep space within individual fields of view. These diverse observing geometries made modeling the spectra difficult or impossible. Figure 1 shows the effect on the spectra of different observing geometries. The two spectra from 2005 and 2011 were observed at similar emission angles, 85° and 78°, while the spectrum from 2008 was observed at 60° emission angle. The spacecraft distances for the three spectra varied considerably (see the figure caption). In the 2005 and 2011 spectra, the intensity of the feature at 220 cm⁻¹ grew smaller over the period, whereas the intensity of the C₃H₄ feature at 325 cm⁻¹ remained about the same. On the other hand, in the 2008 spectrum both features appear weaker than they did in 2005. By examining spectra throughout 2004–2011 we determined that when the viewing geometries were similar the C₃H₄ peak changed very little. We found that to a good approximation we could use the relative constancy of C₃H₄ to compensate for the dependence on observing geometry. We therefore took as our measurements the ratios of the peak radiances of the two features, haze/C₃H₄, and these are plotted in Figure 2 (the peaks were measured with respect to their local continua). Since the C₃H₄ intensity was steady, the figure shows the change in the 220 cm⁻¹ peak intensity over time. The validity of this method derives from the fact that (1) the C₃H₄ emission arises in an altitude range similar to that of the 220 cm⁻¹ emission, (2) the two emissions are optically thin, and (3) C₃H₄ has only a small seasonal variation. We checked the zero level in our measurements by taking similar ratios in spectra from the south where there has been no detectable 220 cm⁻¹ haze. The ratios in the south fell below zero by −0.25; we assumed this to be a bias and used it to correct all of the ratios in Figure 2.

The peak radiance of the 220 cm⁻¹ emission band dropped by about a factor of four over approximately three Titan months between 2004 December and 2012 February. Judging from the slope in Figure 2, the decrease was already underway when Cassini arrived at Saturn. The falloff is approximately linear over the period, with perhaps a slight lessening of the slope with time. Extrapolating backward, the ratio may have been considerably higher at northern winter solstice in 2002 October and the steep falloff rate must have been reached within about two years. A comparison of 2010 Cassini data with Voyager spectra from 1981 (Coustenis et al. 1999) shows that, at the same seasonal phase after one Titan year, the intensity of the feature was approximately the same.

Part of the radiance at 220 cm⁻¹ arises from gaseous C₄H₂ and C₂N₂ (Coustenis et al. 1999). However, judging from the narrow Q-branches visible in CIRS high-resolution spectra, the emissions from these compounds remained nearly constant throughout the period. We estimate that these gases contribute about 0.5 to each of the ratios in Figure 2.

The change in the northern distribution of the 220 cm⁻¹ intensity during the Cassini mission is shown in Figure 3. Again, the figure is a plot of the ratios of 220 cm⁻¹ emission to C₃H₄ emission (the assumption that C₃H₄ was constant with latitude is approximately valid). Going north, the 220 cm⁻¹ emission began to appear at about 50°N and reached a maximum near the pole. This shape remained about the same during the mission while the intensities at all latitudes decreased. This behavior is very different from that reported by West et al. (2011) for the high-altitude detached haze layer seen outside the polar region. That layer changed abruptly at equinox, dropping in height from 500 to 380 km. Our decrease also runs counter to the results of Vinatier et al. (2012a), who have found that after equinox there was an unexpected enrichment of gases and aerosols in the north above 200 km, probably linked to seasonal shifting of global atmospheric circulation cells.

Zoom In Zoom Out Reset image size

There have been a number of attempts to identify the composition of the 220 cm⁻¹ haze. Ices of H₂O, HC₃N, and C₂H₅CN have been proposed as candidates (Samuelson 1985; Coustenis et al. 1999; Dello Russo & Khanna 1996; Khanna 2005), but low abundances, lineshape mismatches, and the absence of associated spectral features have kept these in question (Samuelson et al. 2007; de Kok et al. 2008). Another possibility is that the haze that forms the 220 cm⁻¹ emission is composed of a variety of nitrile condensates and that the larger of these tend to have bands at this spectral position. Although the simplest nitriles do not have bands at 220 cm⁻¹, many complex nitriles do have –C–CN bending vibrations in this vicinity (Socrates 2001; Coustenis et al. 1999; Compton & Murphy 1981; Schrumpf & Martin 1983). Gaseous HC₃N might serve as a tracer for complex nitriles and, indeed, CIRS measurements of HC₃N lend support to the possibility that the haze is associated with nitriles. Vinatier et al. (2010b, 2012a) found a decrease by over an order of magnitude in HC₃N at 190 km between 2005 and 2009; at 150 km they found an order of magnitude decrease between 2007 and 2009. (They also found the change in C₃H₄ to be relatively small.) Among gases observed by CIRS, only HC₃N shows such a large change. This behavior is similar to that of the 220 cm⁻¹ feature and is also consistent with nitrile enhancement in the polar shadow (Yung 1987), making nitriles a strong candidate for the source of the 220 cm⁻¹ haze material.

Several mechanisms might have contributed to the seasonal decrease in the 220 cm⁻¹ emission. First, it is possible that the source of high-altitude gases that condense to form the haze gradually turned off as sunlight returned to the polar region. According to Yung (1987), the elevated abundances of nitrile compounds in the north on Titan result from a lack of photolytic destruction within the northern winter shadow. This supply of nitriles would have decreased as solar exposure increased and, if the 220 cm⁻¹ haze is of nitrile origin, the haze would have been depleted. Indeed, by the time of the 2009 equinox most of the intensity change at 220 cm⁻¹ had already taken place. Second, the haze may have dissipated as the stratospheric temperatures increased. At the altitudes of the 220 cm⁻¹ haze, winter polar temperatures were depressed by 20–40 K below the HASI-like profile typical of equatorial and southern latitudes (P. J. Schinder et al. 2012, in preparation; Coustenis et al. 2010; Achterberg et al. 2008; Fulchignoni et al. 2005). The reduced stratospheric temperatures promoted formation of the 220 cm⁻¹ haze because some gases then reached their condensation temperatures at higher altitudes than they would have with the HASI profile. For example, the lower polar temperatures caused HCN, HC₃N, C₂N₂, C₄N₂, and CH₃CN to condense at a height of 150–170 km, well above the ∼90 km height expected outside the polar region (Lara et al. 1996) but consistent with the ∼140 km height of the 220 cm⁻¹ haze (Anderson et al. 2010, 2012; Anderson & Samuelson 2011). As the temperature depression became shallower after solstice the high-altitude condensation would have ceased, eliminating the source of the 220 cm⁻¹ material. As summer returns to the north, a condensate layer may appear at 90 km. A third possible factor in the thinning of the 220 cm⁻¹ haze is that subsidence in the north was undergoing a seasonal decrease. Achterberg et al. (2011) reported a warming at 1 mbar in 2007 followed by an anomalous cooling in 2008–2009, which they attributed to a weakening of the polar vortex. A waning of the subsidence would have slowed the delivery of material to the haze formation altitude. The radiative time constant at 3 mbar is approximately one Titan season (Achterberg et al. 2011), consistent with the timescale of observed changes in haze and temperature.

In 2011 December, the Cassini Imaging Science Subsystem recorded an image that showed a change in the photochemical haze structure at the south pole (Cassini image PIA14913). This was the first sign that a polar hood was forming in the south. We searched the CIRS data at 70°S–90°S and found that, as of 2012 February, no sign of the 220 cm⁻¹ emission had appeared in the south. If the emission feature in the south behaves similarly to that in the north, it should reach a maximum near the winter solstice in 2017. By that time, the 220 cm⁻¹ emission in the north should have disappeared. At the current rate of decrease in Figure 2, zero emission could be reached in 2014–2015.

We acknowledge support from NASA's Cassini mission and Cassini Data Analysis Program. V.C. was supported by the NASA Postdoctoral Program.