4. Discussion and conclusion
The values of the CO mixing ratio that we have derived from the 1990 data are in good agreement with the values obtained previously with the 1988 and 1989 data (Billebaud et al., 1992), even if our error bars allow for some small variations (of the order of 50%). This would not necessarily be in contradiction with the photochemical models, as present models admit that, considering the high sensitivity of the CO abundance to the presence of water, this abundance could fluctuate over a timescale of a few years (Nair et al., 1994; see also Sect. 1).
The results we obtain with the 1991 data are more uncertain. It is indeed possible to find values of CO mixing ratios which would be compatible with all the points. Such values would be in the range from 5.5 10-4 to 11.5 10-4. And the dust opacity would then remain very small on all four locations: between 0.03 and 0.08. This range in CO mixing ratios is compatible with our previous determinations, therefore it has our preference, as in this case there is no need to invoke the presence of temporal or spatial variations of the CO abundance. However, from our results, it is not possible to firmly exclude the presence of possible variations, especially horizontal ones, over a scale of a few thousand kilometers, which corresponds to the resolution of our data. This is especially true as there could exist some variations of the dust optical depth. Indeed, Viking data showed that the dust optical depth is not constant all over the planet (Clancy and Lee, 1991). And as shown from our results (Fig. 7), a change in the dust optical depth results in a change in the CO mixing ratio. A comparison with the Rosenqvist et al. (1992) results would have been interesting, but it is unfortunately difficult, because we do not have the same spatial resolution (there is a difference of roughly a factor 3 between our resolution and the size of the high volcano regions). As mentioned previously, heterogeneous chemistry could account for some spatial variations of the CO mixing ratio, but the complexity and the uncertainties introduced by such a chemistry led several authors (Anbar et al., 1993b; Nair et al., 1994; Krasnopolsky, 1995) to exclude it from their models as long as the need to introduce it is not clearly justified. In all cases, in order to obtain a firm conclusion on this issue, new observations would be necessary, and in particular it would be interesting to resolve the high volcanoes (a spatial resolution of the order of 100 km should be enough).
We consider that we have reached the limits of 1D infrared spectroscopy concerning the monitoring of species such as CO. Choosing to make observations in the thermal and solar reflected components of the planetary emission field allowed us to experience both the advantages and drawbacks of the two. The thermal component near 5 had the advantage to be unaffected by dust scattering, but one has to carefully take the atmospheric thermal profile into account; in contrast, the solar reflected component is essentially independent from the thermal profile, but not from the presence of dust. Both component are very sensitive to the pressure, which is a key-parameter not easy to determine, especially for wide areas. In the case of observations performed in the solar reflected component of the planetary emission, a high spectral resolution and the availability of intrinsically strong and weak lines, not affected in the same way by the presence of dust, would be necessary to facilitate the interpretation of the data. The need for spatial resolution also implies to use 2D spectroscopy in order to investigate the possibility of spatial variations; and in addition, a higher spatial resolution would facilitate the determination of the average local pressure on a smaller area. Such observations, which depend on the availability of efficient spectro-imaging instruments will be performed as soon as possible.
© European Southern Observatory (ESO) 1998
Online publication: April 28, 1998