It would appear from recent studies that the ozone screen was established prior to Silurian times and therefore was not directly linked with the spread of life onto land during that period.
One aspect of the early terrestrial environment that exerted a direct influence on the origin and evolution of life was the composition of the atmosphere. Earth's atmosphere is accepted to be of secondary origin, as evidenced by the observed large depletions of noble gases relative to their solar abundance. The bulk of our atmosphere was outgassed from the crust and mantle. The amount of time required to accumulate the current inventory of surface volatiles through the outgassing process is uncertain, but most authors agree that the initial degassing period lasted less than 1 billion years. Since that time, the total amount of surface volatiles has been maintained at a more or less constant level by dynamical exchange processes between the atmosphere, crust, and upper mantle. Figure 1 illustrates some of the more important processes (Walker, 1980). Volcanic gases, primarily H2O and CO2, are released into the atmosphere during eruptions. Water vapor precipitates into the oceans; carbon dioxide participates in weathering reactions that result in the deposition of carbonate sediments. When seafloor plates are subducted downward into the mantle, volatiles trapped in these sediments are released under the high pressure and are again recycled into the atmosphere through volcanic activity along the plate interface. Over the long term, these cycles are balanced so that the atmosphere and hydrosphere are maintained at approximately constant volume. This paper deals with the composition of the atmosphere after this steady-state situation was reached.
From a biological standpoint, the most important aspect of atmospheric composition was the amount of oxygen present. Free oxygen would probably have been a poison to primitive life forms, which presumably lacked sophisticated mechanisms for dealing with it. This, coupled with geological evidence indicating an absence of oxidized sediments before about 2 billion years ago, leads us to believe that Earth's atmosphere was basically reducing during the first half of its history. The first part of this paper is concerned with predicting the ambient oxygen level in such an atmosphere.
The tool used to make such a prediction is a one-dimensional coupled flow-photochemical computer model, similar to those used to study ozone in Earth's current atmosphere. Eight long-lived chemical families are included in the calculation: Ox O + O3, H, HxO2 O2 + HO2 + H2O2, NOx N + NO + NO2 + HNO + HNO2 + HNO3, H2, H2O, CO2, and CO. The individual species within each family are calculated by assuming photochemical equilibrium at each height step. Other short-lived species included in the model are OH, O(1D), and N(2D). The numerical calculation is carried out on a variable-spaced grid, with step sizes ranging from 0.5 km at the ground to 5 km at 180 km. Present-day N2 and CO2 concentrations are used in the results presented here, following Walker's (1977) suggestion that the  background levels of these constituents should not have varied greatly with time. Higher CO2 levels may have been required to keep Earth's surface temperature above freezing in the face of decreased solar luminosity; but this possibility will not be considered here. The temperature and eddy diffusion profiles used in the model are also consistent with present-day values, with the exception of the removal of the temperature bulge due to ozone heating in today's stratosphere. More details on the model, including a list of chemical reactions and rate constants used, are given in Kasting et al. (1979).
Given these assumptions, the amount of free oxygen present in the atmosphere is determined by the balance between the sources and sinks for oxygen atoms. Since molecular oxygen is not released by volcanoes, the only obvious net source of oxygen atoms is photolysis of water vapor:
followed by the escape of hydrogen into space. It has been shown by Hunten (1973a,b), Liu and Donahue (1974a,b,c), Liu et al. (1976), and Hunten and Strobel (1974) that, under a wide variety of conditions, the escape of hydrogen from the terrestrial atmosphere is governed by the principle of limiting flux. Mathematically,
where is the hydrogen escape flux, measured as the number of atoms escaping per square centimeter per second, and fT is the total hydrogen mixing ratio in the stratosphere. Since the total hydrogen mixing ratio remains constant above the cold trap at the tropopause, that is, above the height at which water vapor condenses out of the atmosphere, the production rate of oxygen atoms may be evaluated if one knows the water vapor mixing ratio at the tropopause. In our model, which assumes present-day tropospheric temperatures, the 3.8 ppm of H2O at 10 km yields a hydrogen escape flux of 1.47 x108 H atoms/cm2/sec, in agreement with Liu and Donahue. If no loss processes were operating, the oxygen left over as a result of this escape process would amount to 1 PAL (present atmospheric level) after about 3 billion years.
Two important loss processes for oxygen do operate: oxidation of reduced volcanic gases, primarily H2 and CO; and oxidation of crustal materials at Earth's surface, which may be written schematically as
The crustal loss rate is difficult to evaluate. It turns out not to be necessary to do so, however, since the H2 and CO outgassing from volcanoes appears  to be more than sufficient to overwhelm the production of O2 from H2O photolysis followed by hydrogen escape. By assuming that volcanic outgassing must in the long run balance the loss of surface volatiles via subduction of sea-floor plates down into the mantle, Walker (1977) estimated the H2O and CO2 volcanic outgassing rates to be, respectively, 3 x 10-4 and 1 x 10-4 gm/yr. Holland (1962) has predicted thermodynamic equilibrium ratios of
for present-day volcanic gases at 1225°C. These figures are in accord with averaged measured ratios of Hawaiian volcanic gases. Thus the estimated present-day outgassing rates for H2 and CO are both on the order of 2 x 108 atoms/cm2/sec, adding up to a combined reduced gas flux of about 4 x 108 atoms/cm2/sec. This potential loss rate for oxygen atoms should be compared to a production rate of 1.47 x 108/2 O atoms/cm2/sec from H2O photolysis followed by hydrogen escape. The reduced gas flux is clearly sufficient to dominate oxygen production, and may have done so to an even greater extent if the crust was being recycled at a faster rate during earlier periods of Earth's history. The oxygen and hydrogen budgets become balanced when H2 accumulates to a level such that the hydrogen escape rate balances the total (H2 + CO) outgassing rate. For an outgassing rate of 4 x 108 atoms/cm2/sec, the required H2 mixing ratio is 3.8 ppm X (8 x 108/1.47 x 108 -1) or about 17 ppm.
Once the hydrogen mixing ratio is known, the amount of oxygen present is determined from the photochemical model. The result for a solar zenith angle of 60° and the outgassing rate mentioned above is shown in figure 2. The O2 number density profile exhibits a peak around 40 km due to the photolysis of CO2 followed by
Below the peak, the O2 density decreases rapidly toward the ground as a result of the three-body reaction with atomic hydrogen:
The O2 mixing ratio at the ground amounts to less than 10-13 PAL, making oxygen a very rare constituent indeed. This result is in agreement with the predictions of low ground-level O2 densities made by Walker (1977).
Proceeding in this same fashion, oxygen profiles may be calculated for a variety of different hydrogen outgassing rates. Figure 3 shows how the ground-level O2 concentration varies as a function of total (H2 + CO) outgassing....
....rate. The vertical asymptote at an outgassing rate of 7.35 x 107 atoms/cm2/sec corresponds to the flux required to balance oxygen production exactly. If the outgassing rate should drop below this value, oxygen would begin to accumulate in the atmosphere, at least until crustal oxidation processes were able to offset the rate of oxygen production. For outgassing rates exceeding this critical flux, the O2 density decreases toward a limiting value (horizontal asymptote) of 5.4 x 104 atoms/cm3. The lower limit on oxygen is a result of the production of O2 in lightning discharges, which can be estimated by assuming thermodynamic equilibrium at high temperatures. (Chameides et al., 1977; Chameides and Walker, unpublished data). The net result is that, for reasonable hydrogen outgassing rates, the ground-level oxygen concentration is rather closely contained between the values 5 x 104 and about 5 x 105 atoms/cm2. At such a low density, free oxygen should have had little effect on any biological organisms that may have been evolving at the time.
The discussion thus far has neglected any influences early life may have exerted on its environment. During the earliest stages of biological evolution, these influences were probably small, so that such an omission is justifiable. At some point, however, perhaps as early as 3.3 billion years ago (Shidlowski...
...et al., 1975), the process of oxygenic photosynthesis was invented. By enabling organisms to split water molecules into their constituent atoms, photosynthesis exerted a drastic influence on the oxidation state of the atmosphere. Burial of reduced carbon in organic material produced by photosynthesis was accompanied by the release of a stoichiometrically equivalent amount of free oxygen into the environment. After exhausting any leftover surface reservoirs of reduced material, particularly dissolved ferrous iron which may have been present in the early Archean oceans (Cloud, 1972), this oxygen eventually began to accumulate in the atmosphere.
 We pick up the story again when the ground-level oxygen mixing ratio reached 10-5 PAL. Even at this level, oxygen was not yet well mixed in the lower atmosphere (see fig. 4). The oxygen bulge in the upper stratosphere due to CO2 photolysis persists until the mixing ratio exceeds 10-4 PAL.
Accompanying the buildup of O2 was a new atmospheric development that also had an important effect on the biosphere, namely, the emergence of an ozone layer. The presence of atmospheric ozone is essential to the existence of most land life since ozone is the only important absorber of solar near-ultraviolet radiation between 2000 and 3000 Å. This dependence led Berkner and Marshall (1964, 1965) and others to link the spread of life onto land in the late Silurian, about 420 million years ago, with the development of the ozone layer. An interesting and still unresolved question is whether this relationship was causal, with a rapid sequence of the two events, or whether the emergence of land life merely awaited the evolutionary advances necessary to make the transition well after the ozone screen had been established.
To examine the rise of atmospheric ozone, the computer model used to do the prebiological oxygen calculations is modified by the addition of nitrous oxide and methane, biogenic trace gases that are both important in....
 ....influencing ozone concentrations in the present-day atmosphere. Nitrous oxide, which is produced by bacterial dentrification processes in anaerobic soils, reacts with atomic oxygen in the metastable 1D state:
This reaction serves as an important source of odd nitrogen (NOX) compounds in the stratosphere. Odd nitrogen provides a catalytic destruction pathway for ozone via the sequence:
This destruction mechanism is the most important loss process for ozone in the present day middle stratosphere (Crutzen, 1970; Johnston, 1971).
Methane, produced primarily from fermentation and anaerobic decay processes in swamps and wetlands, is a source for ozone in today's troposphere (Fishman and Crutzen, 1977). Methane oxidation produces methyl peroxide and hydroperoxyl radicals which react with nitric oxide:
When these reactions are followed by photolysis of NO2, ozone is formed:
The NO in our model troposphere is produced in lightning discharges, according to the predictions of Chameides et al. (1977).
Including these species in our model necessitates extrapolation of their production rates back to times when the atmosphere contained much less oxygen than at present. The rate of NO production in lightning is scaled by assuming thermodynamic equilibrium at 2300 K in the immediate vicinity of the lightning discharge. This yields the following scaling factors relative to today's production rate:
PAL of O2
NO scaling factor
 The rates of biological generation of N2O and CH4 at lower oxygen levels are much more difficult to assess. The choice has been to hold the production rates constant by assuming a constant upward flux of each species at the ground. The decreasing validity of this assumption at the lowest O2 levels is compensated by the fact that both N2O and CH4 become less effective in influencing ozone densities when oxygen is less abundant in the atmosphere Increased tropospheric OH densities and rapid photolysis lead to decreased methane and nitrous oxide concentrations by the following reactions:
Making use of the above assumptions, model experiments have been carried out for oxygen levels ranging from 10-5 to 1 PAL of O2. The resulting ozone profiles are shown in figure S. A solar zenith angle of 45° was used, and the resulting photolysis rates were multiplied by a factor of 0.5 to account for the diurnal variation. The total ozone column depth for the 1 PAL case is 1.12 x 1019 molecules/cm2, which is somewhat higher than....
 ....the mean global value of 8.6 x 1018 molecules/cm2 reported by McClatchey et al. (1971). Experiments with higher solar-zenith angles yield smaller O3 column depths, so that these results represent a crude upper limit on the available ozone. The temperature profile used is "primordial" (i.e., we assume an isothermal stratosphere) for 10-1 PAL of O2 and below and present-day for 1 PAL. Thus the model includes a first-order approximation to the coupling between stratospheric temperature and ozone abundance.
The ozone density for the present atmosphere peaks at a height of 24 km, where the number density is 5.37 x 1012 molecules/cm3. At lower oxygen levels the ozone peak moves downward due to the increased depth of penetration of solar ultraviolet radiation. At 10-3 PAL of O2 and below, the O3 peak is found at 10 km, the height of the tropopause in our model. Below 10 km, photolysis of water vapor leads to large concentrations of odd hydrogen (HOX = H + OH + HO2) radicals, which destroy ozone via a number of catalytic reactions.
The characteristic of these ozone profiles which is important from a biological standpoint is the total column depth since that is what determines how much ultraviolet radiation may leak through. The original estimate by Berkner and Marshall (1964, 1965) was that an effective ultraviolet shield would be provided by a column depth of 0.2 atm cm, or 5.4 x 1018 molecules/cm2 (1 atm cm = 2.687 x 1019 molecules/cm2). This column depth is sufficient to reduce the ultraviolet flux at the surface of Earth to less than 1 erg/cm2/sec at 50 Å. Ratner and Walker (1972) argue that this flux is still unacceptably high; hence they adopt a somewhat more stringent lower limit on ozone column depth of 7 x 1018 molecules/cm2.
Figure 6 shows the ozone column depths calculated by our model for various oxygen levels. Also shown are the original results of Berkner and Marshall (1964, 1965) and those of Levine (1977), who performed a similar set of calculations with a one-dimensional model less sophisticated than ours. The differences between our results and those of Levine are mostly due to the fact that he modeled O3 separately rather than including it as part of odd oxygen (Ox = 0 + O3), as is done in current stratospheric ozone models. Also, Levine did not include H2O and CO2 in attenuating the solar ultraviolet flux, which caused his results to be defective at the lower oxygen levels.
The ozone column depth curve derived from our model crosses Ratner and Walker's "critical level" at an O2 mixing ratio of about 0.05 PAL. This again is for an assumed solar-zenith angle of 45°, so that the predicted ozone column depth represents an upper limit. For a solar-zenith angle of 57.3°, corresponding to that used by Ratner and Walker (1972) and Levine (1977), the critical level is not passed until the O2 mixing ratio reaches about 0.1 PAL - the value predicted originally by Berkner and Marshall, which is considerably higher than the 10-2 PAL found by Levine or the 10-3 PAL estimated by Ratner and Walker.
The significance of this result may be ascertained by comparing our critical O2 level with the estimate by Rhoads and Morse (1971) for the minimum atmospheric oxygen content during the Cambrian period (600 million years ago). By studying modern anaerobic marine basins, Rhoads and Morse determined that calciferous fauna require a dissolved oxygen concentration of at least 1 ml/liter, compared to the 4-9 ml/liter that would be m equilibrium with the present atmosphere. Since the Cambrian period was marked by the sudden appearance of abundant shelled organisms, they conclude that the atmospheric oxygen content must have been at least as high as 0.1 PAL  during that time. Their lower bound on Cambrian O2 is the same as our estimate for the amount of oxygen necessary to produce a biologically effective uItraviolet shield. Thus it would appear that the ozone screen was established before the Silurian and therefore was not directly linked with the spread of life onto land during that period. However, the uncertainty inherent in both caIculations leaves open the possibility that a causal relationship between the evolution of atmospheric ozone and the appearance of land life did indeed exist.
- Berkner, L. V.; and Marshall, L. C.: The History of Oxygenic Concentration in the Earth's Atmosphere. Disc. Faraday Soc., vol. 37,1964, pp. 122-141.
- Berkner, L. V.; and Marshall, L. C.: On the Origin and Rise of Oxygen Concentration in the Earth's Atmosphere. J. Atmos. Sci., vol. 22, May 1965, pp. 225-261.
- Chameides, W. L.; Stedman, D. H.; Dickerson, R. R.; Rusch, D. W.; and Cicerone, R. J.: NOX Production in Lightning. J. Atmos. Sci., vol. 34, Jan.1977, pp. 143-149.
- Cloud, P.: A Working Model of the Primitive Earth. Amer. J. Sci., vol. 272, June 1972, pp.537-548.
- Crutzen, P. J.: The Influence of Nitrogen Oxides on the Atmospheric Ozone Content. Quart. J. R. Meteor. Soc., vol. 96, Apr. 1970, pp. 320-325.
- Fishman, J.; and Crutzen, P.: A Numerical Study of Tropospheric Photochemistry Using a One-Dimensional Model. JGR, vol. 82, Dec. 1977, pp. 5897-5906.
- Holland, H. D.: Model of the Evolution of the Earth's Atmosphere. In: Petrologic Studies: A Volume to Honor A. F. Buddington, Geological Soc. of Amer., New York, 1962, pp. 447-477.
- Hunten, D. M.: The Escape of H2 from Titan. J. Atmos. Sci., vol. 30, May 1973a, pp. 726-732.
 - Hunten, D. M.: The Escape of Light Gases from Planetary Atmospheres. 1. Atmos. Sci., vol. 30, Nov.1973b, pp. 1481-1494.
- Hunten, D. M.; and Strobel, D. F.: Production and Escape of Terrestrial Hydrogen. J. Atmos. Sci., vol. 31, Mar.1974, pp. 305-317.
- Johnston, H. S.: Reduction of Stratospheric Ozone by Nitrogen Oxide Catalyst from SST Exhaust. Science, vol.1 73, Aug.1971, pp. 517-522.
- Kasting, J. F.; Liu, S. C.; and Donahue, T. M.: Oxygen Levels in the Prebiological Atmosphere. JGR, vol. 84, June 1979, pp. 3097-3107.
- Levine, J. S.: The Evolution of Stratospheric Ozone. Ph.D. dissertation, University of Michigan, 1977.
- Liu, S. C.; and Donahue, T. M.: The Aeronomy of Hydrogen in the Atmosphere of the Earth. J. Atmos. Sci., vol.31, May 1974a, pp.1118-1136.
- Liu, S. C.; and Donahue, T. M.: Mesospheric Hydrogen Related to Exospheric Escape Mechanisms. J. Atmos. Sci., vol. 31, July 1974b, pp.1466-1470.
- Liu, S. C.; and Donahue, T. M.: Realistic Model of Hydrogen Constituents in the Lower Atmosphere and Escape Flux from the Upper Atmosphere. J. Atmos. Sci., vol. 31, Nov.1974c, pp. 2238-2242.
- Liu, S. C.; Donahue, T. M.; Cicerone, R. J.; and Chameides, W. L.: Effect of Water Vapor on the Destruction of Ozone in the Stratosphere Perturbed by CLX or NOX Pollutants. J. Geophys. Res., vol. 81, June 1976, pp. 3111-3118.
- McClatchey, R. A.; Fenn, R. W.; Selby, J. E. A.; Volz, F. E.; and Garing, J. S.: Optical Properties of the Atmosphere. Tech. Rept., AFCRL-710279, Air Force Cambridge Res. Labs, Bedford, MA,1971.
- Ratner, M. I.; and Walker, J. C. G.: Atmospheric Ozone and the History of Life. J. Atmos. Sci., vol. 29, July 1972, pp. 803-808.
- Rhoads, D. L.; and Morse, J. W.: Evolutionary and Ecologic Significance of Oxygen-deficient Marine Basins. Lethaia, vol. 4, 1971, pp. 413-428.
 - Schidlowski, M.; Eichmann, R.; and Junge, C. E.: Precambrian Sedimentary Carbonates: Carbon and Oxygen Isotope Geochemistry and Implications for the Terrestrial Oxygen Budget. Precambrian Res., vol. 2, 1975, pp. 1-69.
- Walker, J. C. G.: Evolution of the Atmosphere. New York: Macmillan, 1977.
- Walker, J. C. G.: Atmospheric Constraints on the Evolution of Metabolism. In: Origins of Life, vol. 10, June 1980, pp. 93-104.