ANALYSIS OF STELLAR OCCULTATION DATA FOR PLANETARY-ATMOSPHERES .1. MODEL-FITTING, WITH APPLICATION TO PLUTO

An analytic model for a stellar-occultation light curve has been developed for a small, spherically symmetric planetary atmosphere that includes thermal and molecular weight gradients in a region that overlies an extinction layer. This work applies to the thermal structure of the upper part of Pluto's atmosphere probed by current stellar occultation data, so the issue of whether the lower part should be modeled as an extinction layer or sharp thermal gradient is not addressed. The model can be described by two equivalent sets of parameters. One set specifies the occultation light curve in terms of signal levels, times, and time intervals. Consequently, it is the more suitable set to use for fitting the light curve. The other set specifies physical parameters of the planetary atmosphere. Equations are given for the transforming between the sets of parameters, including their errors and correlation coefficients. Detailed numerical calculations are presented for a benchmark case. In order to establish the formal errors in the model parameters expected for datasets of different quality, least-squares fitting tests are carried out on synthetic datasets with different noise levels. This model has also been fit to the KAO data from the 1988 June 9 stellar occultation by Pluto. For the case with an isothermal constraint, the fitted parameters agree with our previous isothermal analysis [Elliot et al., Icarus 77, 148 (1989)]. Fits of these data that include a temperature gradient as a free parameter yield a temperature to molecular weight ratio T/mu = (3.72 +/- 0.75) K amu-1 and normalized gradient (dT/dr)/T = (- 4.9 +/- 7.0) x 10(-4) km-1 at r = 1250 km. Interpretation of these results depends on the mean molecular weight of the atmosphere. The values are 60 +/- 12 K and -0.029 +/- 0.040 K km-1 for the limiting case of pure CH4 (mu = 16.04) and 104 +/- 21 K and -0.051 +/- 0.070 K km-1 for the limiting case of pure N2 (mu = 28.01). Our result is consistent with the isothermal prediction of the "methane-thermostat" model of Pluto's atmosphere [Yelle & Lunine, Nature, 339, 288 (1989)]. However, Pluto's atmosphere could be isothermal in this region at a lower temperature than the 106 K predicted by the model, if the radiative cooling occurs at a wavelength longer than the 7.8-mu-m band of CH4. A summary of our current knowledge of Pluto's atmosphere and related parameters is tabulated. The ambiguity between the haze and thermal-gradient possibilities for Pluto's lower atmosphere limits the accuracy with which we now know Pluto's surface radius and bulk density. If the "haze model" is correct, then Pluto's surface radius is less than 1181 km and its bulk density is greater than 1.88 g cm-3. On the other hand, if the "thermal-gradient model" is correct, then Pluto's surface radius would be 1206 +/- 11 km and its density would be 1.77 +/- 0.33 g cm-3.