New Models for Thermal Emission

In the past two years several conceptual breakthroughs have been made in the modeling Io's thermal emission. These improvements made it possible to obtain more accurate characterizations of the thermal anomalies.

It was well known that older thermal emission models satisfied neither the detailed flux history as Io cooled in eclipse and warmed upon returning to sunlight [e.g. see discussion in Sinton and Kaminski, 1988] nor the absolute level of Io's 20m emittance [McEwen et al., 1988]. While there were several possible reasons for these modeling problems, the most important involve some basic assumptions about what we thought we knew about Io. On the basis of 10m and 20m data indicating very rapid cooling of the surface upon entering eclipse, Io was believed to have a very porous or fluffy, low thermal inertia surface. Model hot spots had temperatures of 300 to 600 K which were believed to be determined solely by volcanic processes. In this ``world view'' of Io, little radiation from the hotspots was expected at 20m where the vast, cool, passive surfaces should dominate the thermal emission flux. Likewise, at shorter wavelengths, such as 5m, hot spot emission should dominate and there should be negligible flux from background emission. With more extensive data available, these assumptions turned out to be misleadingly simplistic approximations.

One compounding difficulty for the analysis of infrared data is the Earth's atmosphere. The key wavelengths for observing Io's thermal emission spectrum are from 2 to 30m. Unfortunately, the atmosphere is not transparent over much of this range. Observations are confined to ``windows'' where the atmospheric transmission is high. This limitation on the accessible spectrum made it more difficult to discover faults in the emission models and probably prevented an early recognition of the significant role played by the thermal pedestal effect. The thermal pedestal effect is the spectral blue shift which occurs in the thermal emission spectrum when sunlight is absorbed on an anomaly whose temperature is elevated by heat flow [Veeder et al., 1994b]. Recognition of this shifting effect leads to the concepts of active and passive components of the background spectrum. The power in the background spectrum is entirely due to the re-radiation of absorbed sunlight. The spectra for passive components can be calculated a priori, given the necessary properties of the surface. By contrast, spectra for active components cannot be computed until after the temperatures of the anomalies have been specified. Depending upon the temperature of the anomaly, the peak of the active background spectrum can be shifted in wavelength by a substantial amount. In the case of Io, the thermal anomalies occupy only several percent of the surface. Accordingly, one would normally assume that sunlight absorbed on them would contribute negligibly to the observed thermal emission. In terms of total power, this is true. However, the spectral emittance at a specific wavelength can be greatly affected. At 8.7m about 30% of the total observed radiation from Io is coming from these areas as a result of the thermal pedestal effect. At 4.8m the corresponding amount due to the heating of the thermal anomalies by sunlight is 13% of Io's thermal emission.

A second conceptual breakthrough was the realization that a significant amount of heat must be carried over to Io's nighttime hemisphere. Based on Io's rapid cooling in eclipse, earlier models assumed that Io's surface is very porous, with an extremely low thermal inertia---similar to other airless solar system bodies but even lower. This class of model results in temperatures droping to very low levels at night. The use of this relatively ``standard'' airless body model effectively blocked the consideration of models which could retain significant amounts of heat to be radiated later, long after local sunset. New data and the recognition of the thermal pedestal effect forced a reconsideration of these ideas. Significantly, it was discovered that the thermal pedestal effect can mimic some aspects of the temporal signature of the eclipse of a very low thermal inertia surface. As a consequence of this, it is now realized that the presently available eclipse-cooling measurements for Io place no useful constraint on thermal inertia.

The first of the new generation of models involves three types of surface units: 1) a relatively low albedo unit which is in instantaneous equilibrium with sunlight and contributes the fall off in emission when Io enters eclipse; 2) a high albedo thermal reservoir unit which contributes significant levels of thermal emission during the nighttime, and 3) the volcanic thermal anomalies. The reservoir, of course, has a very high thermal inertia. The other units are assumed to have relatively low inertias, allowing them to be in approximate equilibrium with sunlight. The three thermal units, in turn, actually produce five distinct spectral components: the emission due to radiation of absorbed sunlight for each of the three units, the emission due to heat flow, and reflected sunlight which is significant at shorter wavelengths, such as 4.8m. The equilibrium and reservoir thermal units are assumed to be interspaced uniformly over Io's disk (except at the sites of thermal anomalies). The anomalies, which are at specific locations, occupy several percent of the surface [Veeder et al., 1994b].

How well is the new modeling approach working? It successfully predicts the observed fluxes. At some wavelengths (e.g. 20m) the discrepancy previously had been more than a factor of two. Now the agreement between model and observation is about ten percent over the range of 5 to 20m. Consequently, the model parameters, including temperature and size, for the thermal anomalies are now more accurately known. The temperature at which an anomaly can be recognized has been lowered from about 300 K to a little less than 150 K. This, in turn, allows greater accuracy in calculating the heat flow from the global ensemble of thermal anomalies. Significantly, the areas and temperatures required by the new model based on telescope radiometry are in good agreement with a new independent analysis of spatially resolved spectra from the Voyager IRIS [McEwen et al., 1992].