5.1. Qualitative analysis
The most prominent feature visible on Saturn's disk, especially in the ethane band, is the north-polar emission (NPE). It is located above 50o N and it seems to extend all around the pole. Due to the geometry of Saturn, we cannot see if the NPE is centered on the north pole, but it is likely that it is present at all longitudes. The NPE has also been observed by Gezari et al. (1989) on images recorded in March 1989 at 7.8, 11.6 and 12.4 m. Unlike these two recent observations, Gillett and Orton (1975), Rieke (1975) and Tokunaga et al. (1978) (see Table 1 for the characteristics of the different observations considered here) pointed out from infrared observations that the south pole of Saturn was very bright. The geometry of Saturn during these latter observations was the opposite of ours, and their south-polar emission was related to the greater insolation time of the south pole at the time of their observations. Thus we conclude that a greater insolation time tends to increase the temperature of the insolated pole, but with a phase lag due to the inertia of the atmospheric response (Fig. 2 of Conrath & Pirraglia, 1983). Indeed, this is clear from Table 1, where we see that Voyager observations showed a warm south pole during early northern spring. It must be noticed that limb-brightening also contributes to the enhanced emission at the pole: for instance, at 12.47 m, the flux for a viewing angle of 80o is 41% greater than for 0o. Nevertheless, this is not enough to explain the observed features and in the next section, we will show that this feature could be due to an enhancement of the hydrocarbon abundances or a hotter stratosphere.
Table 1. The characteristics of the different observations of Saturn. Ls refers to the solar longitude. The seasons correspond to those of the northern hemisphere. The last column indicates the main feature seen on the corresponding observation.
As noticeable as the NPE is the north-equatorial belt (NEB). This structure is also visible at all wavelengths, except at 10.91 m where it is less evident, and is located all around the planet between 10o and 30o N, its area varying with the longitude and wavelength. Its maximum is located around 20o N. This feature can be associated with the thermal structure of the atmosphere. Indeed, from Voyager/IRIS spectra, Conrath and Pirraglia (1983) retrieved the latitudinal temperature variations at three pressure levels. In their Fig. 1, we see a strong increase of the temperature at the 150 mbar level between 10o and 20o in both hemispheres. We logically assume that these thermal structures are not confined to a pressure level around 150 mbar and extend to levels sounded on our images, which are located between 100 and 500 mbars (Table 2), and therefore are associated with the NEB.
Table 2. Pressure levels (in mbar) of the maxima of the contribution function. Except for 10.91 m, there is a maximum of the contribution function in the troposphere (Tropo.) and in the stratosphere (Strato.). These pressure levels have been determined for two different viewing angles, 0o and 70o. The sixth column indicates the corresponding molecular band.
The dark structure south of the equator is due to the rings. The observations of Voyager 2 indicated that the rings' opacity in the infrared is important, especially that of the B ring (Hanel et al. 1982). Moreover, the rings are colder than Saturn's upper troposphere and thus they would appear in absorption at 10 microns. Another effect, which we did not study, is the cooling of the part of the atmosphere which lies in the shadow of the rings (see Barnet et al. (1992) which have included the different effects of the rings in a radiative-dynamical model to study the thermal structure of the Saturnian atmosphere).
Many other features are visible on Saturn, especially on the NEB. These small-scale structures are not well defined and it is difficult to identify them from one image to another. For instance, on the images of December 8, at 11.69 m, a bright spot seems to be present at a longitude of 160o. This longitude range is visible on the second image of December 8, but unfortunately not on the December 9 images. Such a bright spot is also visible on the former, at a longitude of 150o. The imprecision on the geometry of Saturn could have led to a difference of 10o on the longitude, but we cannot be sure that it is the same structure we see on the two images. These possible features could be associated with temperature variations or changes in the cloud structure.
5.2. Quantitative analysis
Due to their lower quality, we do not study the December 8 images. Nevertheless, the observed stuctures are almost the same on the two sets of images. In this work, we only study the latitudinal variations of the infrared flux, and we limit ourselves to the northern hemisphere. Indeed, we have a well-defined latitudinal structure (10o N-30o N: the north-equatorial belt, 30o N-50o N: a darker region, and above 50o N: the north-polar emission) at nearly all longitudes. We do not study the southern hemisphere because of the less favourable geometry and the ring occultation. We also restrict our analysis to the central meridian: it corresponds to a longitude around 240o for the December 9 images at 10.91, 11.69 and 12.47 m, and around 275o for the December 9 images in the acetylene band (all the latitudinal structures are visible on the central meridian). The purpose of this analysis is to reproduce the observed variations of the infrared fluxes at all wavelengths. We choose to calculate the flux for five different viewing angles: 0o, 15o, 30o, 40o and 50o. Since the planetocentric latitude of the sub-Earth point is equal to 17o N (exactly 16.7o N)at the time of our observations, these viewing angles correspond to planetocentric latitudes equal to 17o N, 32o N, 42o N, 52o N and 62o N. The location of the peak flux at 10.91, 11.69 and 12.47 m occurs around 15o N in the NEB (Figs. 6 and 7). Therefore, on these images, a viewing geometry of 0o also approximately corresponds to the maximum of the NEB. At 13.09 and 13.29 m, the maximum of the NEB is located around 20o N, and around 30o N at 13.48 m. A possible interpretation is that the thermal structure of Saturn's atmosphere or hydrocarbon abundances shows an evolution with the latitude. But, we have seen that the images at 13.09, 13.29 and 13.48 m are affected by an additional noise which affects some sets of pixel raws, and, in particular, one of this line-shaped features affects the signal on the NEB. Moreover, the signal-to-noise ratio is quite low for these images which are thus more sensitive to this line-shaped features than the other images. The maximum of the contribution function (Table 2) at 13.29 and 13.48 m are very close, and it would be difficult to reproduce the observed relative variations of the flux at 13.29 and 13.48 m with a realistic thermal profile or realistic abundance variations. Therefore we interpret the differences in the location of the maximum of the NEB as due to noise rather than being a real structure of the atmosphere. Finally, the quantitative analysis will essentially focus on the images at 10.91, 11.69, 12.47 and 13.09 m, which show the best signal-to-noise ratio and are not affected by the additional noise. Figa. 6 and 7 present the comparison between the observed and calculated latitudinal relative variations, and we see that we obtain very good results at 10.91, 11.69, 12.47 and 13.09 m, while no fit at 13.29 and 13.48 m is obtained.
The contribution functions show a different behaviour for the different studied wavelengths. At 10.91 m, the emission is centered in the thermal continuun. Owing to the low spectral resolution, this wavelength range is also sensitive to phosphine, but its contribution is very small and does not influence the outgoing flux. The maximum of the contribution function at 10.91 m (Table 2) is located in the troposphere near 540 mbar. The image at this wavelength helps us to constrain the tropospheric temperature, since it is independant of ethane and acetylene abundances (we consider the phosphine mixing ratio as constant). At all the other wavelengths, the contribution functions have two maxima: one in the troposphere and another one in the stratosphere. These wavelengths are thus sensitive to the abundances of hydrocarbons as well as to the stratospheric temperature. Therefore a constraint is missing: with the images at 10.91 m, we have a constraint on the tropospheric temperature, but the stratospheric temperature is unknown. We then choose to consider two different hypotheses. The analysis of previous observations of Saturn's atmosphere in the infrared always considered that the hydrocarbon abundances were constant and attributed the observed structures to thermal effects. Thus, our first hypothesis states that acetylene and ethane have the same abundance at all latitudes and that the observed features are only due to variations of the temperature profile: the resulting temperature profile is then defined by , where varies with the altitude z. The second hypothesis states that the shape of the thermal profile is constant, i.e. the temperature profile is shifted by the same values at all altitude levels: , where does not vary with altitude. is determined to obtain a fit with the 10.91 m images, and the observed variations of the infrared flux are then due to variations of ethane and acetylene abundances. Table 3 displays the parameters used to calculate the flux at each latitude in the two cases.
Table 3. Parameters used to model the radiative transfer at the different latitudes. Variable temperature & Constant abundances refers to hypothesis 1 (we fix the ethane and acetylene mole fraction and the thermal profile varies), and Variable abundances & Constant shape to hypothesis two (the shape of the thermal profile is constant and ethane and acetylene abundances vary). The abundance of C2H6 and C2H2 correspond to the mole fraction profile of Fig. 4, multiplied by the indicated factor. The indicated temperature corresponds to the variation with respect to the standard profile constructed from the observations of Lindal et al. (1985) and Hubbard et al. (1997).
First of all, we have explored the possibility that the cloud structure is at the origin of the observed structures. We have made some tests with different cloud distributions: they clearly show that a cloud located in the lower stratosphere could produce a quite different thermal infrared flux. The point is that it has the same effect at all wavelengths, in particular at 10.91 m. Therefore, if the NPE was due to a different cloud distribution, it would also have been observed at 10.91 m (if we suppose that the infrared properties of the clouds are the same at all the wavelengths concerned here). Therefore, a different cloud structure or composition cannot be at the origin of the observed structure.
5.2.1. Thermal structure vs. latitudes
In the case of constant ethane and acetylene abundances, the agreement between the variations of the observed and the calculated flux is obtained with the temperature profiles shown in Fig. 5. Fig. 6 presents the relative variations of the observed flux with latitude and the diamonds show the calculated variations obtained with the thermal profile corresponding to each latitude. At 17o N, the thermal profile has the same shape as the standard profile but the temperatures are increased by at all altitudes z: . This profile leads to a good agreement with the absolute flux at 10.91 m, but not at other wavelengths. We did not try to find an exact fit with other wavelengths because in this case, the fluxes depend on two parameters (temperature and abundances), while at 10.91 m, the flux depends only on the thermal profile. Moreover, the calibration procedure leads to fluxes which are of the order of magnitude of the computed one, but there is some evidence for large absolute calibration uncertainties. Nevertheless, we only study the relative variations and we did not try to find an exact fit with the absolute flux. Consequently, we do not retrieve here any absolute parameter, such as the temperature at a given level, but we only give the relative variations of temperature (and abundances in Sect. 5.2.2.).
From the comparison between the relative variations of the observed and computed fluxes, we retrieve the latitudinal variation of the thermal profile. Between 17o N and 32o N, the temperature decreases at all levels by 1K (). Then, at 42o N, the profile is 0.5K cooler than at 32o N. At 52o N, the fit is obtained with temperatures decreased by 1.5K relative to 42o N at all levels below the tropopause (), while the stratospheric temperatures increase by 6.5K relative to the 42o N at all altitudes (). At 62o N, the tropospheric temperatures are the same as at 52o N, and the stratospheric ones increase again (+5K at all levels above the tropopause with respect to the 52o N profile: ). The amplitude of the latitudinal variation of the stratospheric temperature is thus about 12K at a given altitude in the stratosphere whereas that of the tropospheric temperature is about 3K.
5.2.2. Ethane and acetylene abundances
We now consider that the shape of the thermal profile is the same at all latitudes ( constant with the altitude). The 10.91 m flux is independant of the abundances of ethane and acetylene and allows us to retrieve the tropospheric temperature. Then the hypothesis of a constant temperature profile shape constrains the stratospheric temperature. The variations of the flux are then reproduced with variations of ethane and acetylene abundances. We have seen in the previous section that between 17o N and 42o N, the shift of the temperature is the same in the troposphere and in the stratosphere. The temperature profile has thus the same shape in between these latitudes, and the abundances of ethane and acetylene are constant. At 52o N, we obtain a good fit with a thermal profile decreased at all altitudes by 1K relative to 42o N (), and with abundances of ethane and acetylene respectively multiplied by 2.5 and 3 with respect to the calculated abundances. At 62o N, the thermal profile is the same as at 52o N, and the abundances of ethane and acetylene are multiplied by respectively 5.5 and 6 with respect to the calculated abundances. Fig. 7 shows the relative variations of the flux at all wavelengths and the computed variations.
© European Southern Observatory (ESO) 2000
Online publication: March 28, 2000