A very specific question addressed in this session, but one which is central to the problem of detecting extrasolar life, is the following: Would the detection of measurable amounts of ozone (O₃) in a planet's atmosphere be considered evidence for life? The argument is based on the fact that O₃ is produced photochemically from O₂, and the strongest source of the O₂ in Earth's atmosphere is photosynthesis. So, another way to word this question might be: What are the possible abiotic sources for O₂, and could such abiotic sources result in measurable amounts of O₂ and O₃ in a planetŐs atmosphere?

A consensus emerged from those present at the workshop that useful information about the existence of extrasolar life might be obtained from measurements of O₃. However, workshop participants stopped short of saying that the presence of O₃ was, by itself, a definitive indicator of life. Furthermore, it was pointed out that there are at least two clearly identifiable circumstances in which substantial amounts of O₂ and O₃ might be produced abiotically. To show what these circumstances are, and to show how the workshop participants arrived at this conclusion, let us briefly review the assumptions that were made.

First, we assumed that it was theoretically possible to observe spectroscopically the presence of at least three gases in the atmosphere of an extrasolar Earth-like planet: H₂O, CO₂, and O₃. All three gases have strong, clearly identifiable absorption bands in the thermal infrared between 6 mm and 17 mm. Other biogenic trace gases, such as CH₄ and N₂O, have weaker absorption features within this spectral region and, hence, would be much harder to observe. This is unfortunate because the participants agreed with the idea expressed originally by Lovelock that an optimal remote life detection method would be to look for the simultaneous presence of O₂ and a reduced gas, such as CH₄. But we assumed that it would not be possible to do this with a first- generation extrasolar planet telescope.

We also assumed that it would be possible with the same first generation telescope to obtain estimates for an extrasolar planet's radius (R), albedo (A), and effective radiating temperature (T_e). Te is obtained from the slope of the observed infrared emission curve. Then, R can be determined from the area under the curve, i.e. the total infrared flux (F), by using the Stefan-Boltzmann Law: F = sT_e⁴, where s is the Stefan-Boltzman constant. Finally, A is determined from the principle of planetary energy balance (absorbed stellar radiation = emitted infrared radiation), using the observed distance of the planet from its primary and the known luminosity of the star.

The additional physical information about an extrasolar planet is considered critical in understanding the possible significance of the detection of O₃ in its atmosphere. The presence of H₂O bands in a planet's atmosphere, combined with knowledge of the amount of stellar radiation it absorbs and its effective radiating temperature, should provide a good indication of whether liquid water exists at the planet's surface. Liquid water is considered essential for life as we know it. This direct information about the presence of water would augment purely theoretical estimates of whether the planet orbited within the star's liquid water habitable zone, or HZ. (See abstract by Kasting and references therein, this volume).

For a planet with a predominantly nitrogen atmosphere within the HZ around a star, it was generally agreed that there is no known abiotic mechanism for producing large amounts of O₂ and O₃. In an atmosphere with a terrestrial-like thermal structure, the production rate of O₂ from photodissociation of H₂O is relatively small because the stratosphere is relatively dry. O₂ would be consumed by reaction with reduced volcanic gases and by liquid water-mediated reactions with reduced materials (Fe²⁺, S^2-, organic C) at the planet's surface. Photochemical model calculations predict that atmospheric O₂ concentrations would be much too small to produce an observable amount of O₃ under these circumstances.

Several workshop participants (e.g., Leger, Allen, Kasting) pointed out that planets lying close to or within the HZ need not obey the same rule if the dominant atmospheric constituent is CO₂. Depending on a number of factors, photodissociation of CO₂ can produce significant O₂ and O₃ mixing ratios. The abundances of these free oxygen species will be modulated by the presence of H₂O and the consequential catalytic hydrogen photochemistry. An additional source of oxygen may result from the photodissociation of water and the escape of hydrogen to space. Planets like Venus that receive too much stellar radiation can develop wet stratospheres in which H₂O is rapidly photodissociated. Consequently, they may lose hydrogen to space at a rapid pace, building up O₂ in the residual atmosphere. Planets like Mars that are too cold at their surfaces to maintain liquid water and too cold inside to maintain active volcanism, could also accumulate O₂-rich atmospheres over time as a result of more gradual hydrogen loss. This would be especially true if the planet were slightly larger than Mars and had an intrinsic magnetic field, so that it did not lose oxygen by nonthermal escape mechanisms, such as dissociative recombination of O₂₊ and solar wind sputtering. Such Mars-like and Venus-like planets could presumably be identified from knowledge of their effective radiating temperatures and from measurements of H₂O absorption within their atmospheres.

Thus, the detection of O₂ and O₃ on extrasolar planets, in the absence of the detection of other key species, is likely to present ambiguous evidence for the presence of life. However, in principle, we could distinguish between abiotic and biotic sources of free oxygen with additional observational data, for example, temperature and abundances of H₂O and CO₂.

In addition to the interest in O₂-rich atmospheres, it was noted that some planets might resemble the Archean/Early Proterozoic Earth (3.8-2.2 Ga), which is thought to have been inhabited by microbial life, but on which the atmosphere remained essentially O₂-free. Earth's atmosphere during this time period may have contained several hundreds of ppm of CH₄, most of which was produced by methanogenic bacteria. These large amounts of CH₄ should be readily observed spectroscopically with the same first-generation space telescope that would be used to look for O₃. The presence of CH₄ in a planet's atmosphere would not be unambiguous evidence for life, as it is possible that this methane could have been supplied by volcanic outgassing at submarine hydrothermal vents. Nevertheless, biological methane sources are considerably larger than abiotic ones, so the abundance of CH₄ in a planet's atmosphere may provide some strong hints as to whether life is present.