1

The fate of the photosynthetic free oxygen produced each year, however, presents a serious problem for this proposal. Accepting the minimum biogenic-methane input of 3 × 10¹³ mol year¹ assumed by Catling et al. [notes 9 and 10 and equation 3 in (1)] requires that twice that amount of net primary productivity, or 6 × 10¹³ mol C, be recycled annually to CH₄. Organic carbon escaping recycling and buried in sediments represents net production of another 10¹³ mol C year¹ [table 1 of (1)]. Thus, at least 7 × 10¹³ mol net cyanobacterial C are produced annually under the model, with 7 × 10¹³ mol free O₂ being simultaneously released into the global anoxic environment. The oxidation of "graphite" settling into the troposphere following stratospheric photolysis of CH₄ [equation 4b and table 1 of (1)] would leave 6.3 to 6.9 × 10¹³ mol free O₂ to be consumed each year (2). Given any geologically and biologically plausible constraints for the Archean, no resources are available to scavenge this annual output of free oxygen. Oxidizing methane [note 9 of (1)] is obviously counterproductive; this would recycle what was produced and eliminate the "greenhouse," the hydrocarbon smog itself, or both (3). It would also simultaneously add isotopically light CO₂ to the atmosphere and diminish hydrogen escape. Highly reactive reduced gases (H₂S, H₂) were either insufficient to accomplish the necessary scavenging or are unsupported altogether by the rock and isotope records (4-8). Scavenging the oxygen by oxidizing the necessarily large amounts of dissolved Fe²⁺ also yields implausible rocks (6-9). The "back reaction of O₂ and CH₂O via respiration" [note 10 of (1)], essential to the model, would also use the same carbon twice by oxidizing the photosynthetic carbon used to produce the methane in the first place--and, in any event, is an aerobic process requiring pO₂ greater than ~0.002 atm (6-8, 10).

There is abundant evidence to support a low-O₂ oxic atmosphere (~0.003 atm) for the Archean (6-8, 11, 12)--and it remains debatable whether other evidence or models to the contrary (1) can indeed distinguish with any certainty between pO₂ values of either <0.0008 atm or ~0.003 atm and a "transitional" value of ~0.03 atm [reference 1 of (1)]. Both the amount of oxygen in early Earth's atmosphere and the rise of oxygen on the early Earth remain problematic.

*Senior Scientist, Emeritus, Smithsonian Institution

Kenneth M. Towe
230 West Adams Street
Tennille, GA 31089, USA
E-mail: towe{at}accucomm.net

REFERENCES AND NOTES

1.	D. C. Catling, K. J. Zahnle, C. P. McKay, Science 293, 839 (2001) [Abstract/Free Full Text] .
2.	A loss of CH4 hydrogen to space irreversibly affects the net oxidation of the Earth as a whole (1), but does not produce free oxygen directly and thus does not significantly affect the annual carbon cycle. The values used by Catling et al. [table 1 of (1)] assume an established 100 to 1000 ppm CH4 atmosphere.
3.	J. Kasting, Science 261, 1060 (1993) .
4.	D. J. Des Marais, Science 289, 1703 (2000) [Free Full Text] .
5.	J. C. G. Walker and P. Brimblecombe, Precambrian Res. 28, 205 (1985) [CrossRef] [Web of Science] [Medline].
6.	K. M. Towe, Nature 348, 54 (1990) [CrossRef] [Medline] .
7.	___, Global Planet. Change 97, 113 (1991) .
8.	___, in Early Life on Earth, S. Bengtson, Ed. (Columbia Univ. Press, New York, 1994), pp. 36-47.
9.	The upwelling of 3 ppm ferrous iron from the deep ocean would remove dissolved oxygen at the rate of 2.4 × 1013 mol O2 year1 [note 10 of (1)]. With 4 mol Fe required per mol O2 [reference 7 of (1)], an enormous 5.4 × 1015 g Fe year1 would be added to global sediments as "hematite." The required concurrent downwelling of O2 could double this annual deposition. Organic matter would be buried at the rate of 1.2 × 1014 g year1 in sediments that contain an average ~1% kerogen carbon. Thus, these same rocks would average an unacceptably high 45% total Fe, all of which must be added to marine sediments each year.
10.	J. Kasting (13, 14), presenting a box model of an Archean ocean, avoids the aerobic respiration dilemma by assuming that ~20% of the surface waters represent local, high-productivity "oxygen oases," but does not specifically identify the rocks that would characterize these special oxic places or present distinguishing features that would separate them from the normal rocks formed in the other, anoxic 80% of the globe.
11.	K. M. Towe, Adv. Space Res. 18, (12)7 (1996).
12.	H. D. Holland, The Chemical Evolution of the Atmosphere and Oceans (Princeton Univ. Press, NJ, 1984), pp. 353, 385, 392, 395, 403.
13.	J. Kasting, Global Planet. Change 97, 125 (1991) .
14.	___, Science 276, 923 (1997) [Abstract/Free Full Text] .

Response: Towe suggests that much more O₂ could accumulate in the early atmosphere than is consistent with the geologic record. We suggest, however, that kinetic losses of O₂ operated at a greater rate than he allows, resolving the discrepancy--and thus, as is explained below, that early atmospheric CH₄ was important (1).

Towe deduces a net annual oxygen flux of 6.3 to 6.9×10¹³ mol O₂ (2-4). The dilemma that he poses, however, has its roots in the dismissal in his analysis of important O₂ sinks and in his neglect of kinetics, or different rates of reactions between different species. The assumed CH₄ flux of ~3 × 10¹³ mol year¹ could be entirely consumed by 6 × 10¹³ mol O₂ year¹ via CH₄ + 2O₂ = 2H₂O + CO₂. But not all CH₄ would be oxidized in this way, because substantial O₂ (of order 10¹³ mol year¹) would be scavenged by more reactive reductants supplied to the early environment by hydrothermal, volcanic, metamorphic and weathering fluxes. Such reductants include Fe²⁺, H₂, CO, H₂S, and SO₂. Thus, on the more reducing early Earth, the annihilation of CH₄ by O₂ would have been incomplete. Excess CH₄ would accumulate to a level at which its photolytic destruction promoted rapid escape of hydrogen to space. That the bulk of the CH₄ reacts with O₂ in this scenario is not "counterproductive" but is expected (5, 6). Also, oxidation of CH₄ to CO₂ would not significantly affect carbonate isotopes, as Towe implies, because, as with the present day, such CO₂ would be mixed with larger cycling of CO₂ from gross photosynthetic productivity.

In (1), we argued that the sink on O₂ from reductants emanating from the crust was greater in the Archean to account for geological evidence of low atmospheric O₂. Carbon isotopes show that roughly 20% of the CO₂ flux into the biosphere has been fixed biologically and buried as organic carbon since at least ~3.2 billion years ago (Ga) [note 8 of (1)]. The constant burial flux of photosynthetic carbon implies a constant O₂ supply rate [via CO₂ + H₂O = CH₂O (buried) + O₂]. Given that unchanging O₂ source, for O₂ to rise at 2.4 to 2.2 Ga, the O₂ sink must have decreased (1, 7). Today, reduced volcanic and metamorphic volatiles consume about a third of the O₂ flux associated with organic carbon burial (8). The recycling of more reduced Archean crust via weathering and metamorphism would have provided a greater sink on O₂ than the recycling of today's more oxidized crust. The flux of reductants need only have been around three to four times greater, relative to the CO₂ flux, to consume the ancient O₂ flux associated with organic burial plus some O₂ associated with CH₄ production (9). CH₄ would then accumulate to 100 to 1000 times today's abundance (1, 6, 7). Abundant CH₄ explains why the O₂ sink eventually diminished: The crust slowly lost reducing power to space via CH₄-induced hydrogen escape (1). Once the O₂ sink dropped below the O₂ source, pO₂ rose to a new equilibrium.

Thus, the early rock record is most plausibly reconciled by loss of O₂ to excess reductants (7). Such conditions stabilize abundant CH₄ (1). Abundant CH₄ may help explain several major issues in Earth history--why the Earth was not frozen over when the Sun was 20 to 30% less luminous, why Earth irreversibly oxidized, and why low-latitude glaciation occurred in the Paleoproterozoic (given that greenhouse CH₄ would be lost upon the Paleoproterozoic increase of O₂).

David Catling
Department of Atmospheric Sciences/
Astrobiology Program
Box 351640
University of Washington
Seattle, WA 98195, USA
E-mail: davidc{at}atmos.washington.edu
Kevin Zahnle
Christopher McKay
NASA Ames Research Center
M/S 245-3
Moffett Field, CA 94035, USA

REFERENCES AND NOTES

1.	D. C. Catling, K. J. Zahnle, C. P. McKay, Science 293, 839 (2001) .
2.	In (1), we hypothesized that a global Archean CH4 flux was produced from photosynthetically produced organic matter in sediments, in which (as they do today) anaerobes converted organic carbon to CH4 (3). Thus, the overall reaction is CO2 + 2H2O = CH4 +2O2 (1). Several assumptions made by Towe differ from ours. We considered a CH4 flux 0.1 to 1 times the modern flux, noting that it could have been greater. Today, most CH4 is microbially oxidized with O2 or sulfate in the water column; for example, only 1 in 70 CH4 molecules emerging from Black Sea sediments reaches the atmosphere. In the low O2, low sulfate Archean ocean, it is probable that proportionately more CH4 reached the atmosphere [note 9 of (1)]. Also in (1), we assumed that much of the photosynthesized organic carbon was probably respired with O2 by aerobes in symbiotic proximity to photosynthesizers to produce CO2, contributing no net O2. The production of CH4 and O2 is stoichiometrically balanced in 1:2 ratio only after the loss of O2 to respiration is subtracted from the full organic carbon cycle. However, the presence of aerobic microorganisms in the Archean implies little about early atmospheric O2 concentrations. Globally low levels of free atmospheric O2 do not imply that O2 levels were everywhere and always low. Oases or strata in which pO2 is locally high would be common, in the same way that today one easily finds places where H2S or CH4 is locally high despite an O2-rich atmosphere.
3.	M. A. K. Khalil, Atmospheric Methane: Its Role in the Global Environment (Springer-Verlag, New York, 2000).
4.	The oxidation of carbon left over from H escape would occur in the gas phase [with net reaction CH4 + O2 = CO2 + 4H(space)] without the "graphite" intermediate suggested by Towe. Equation 4b in (1) is a schematic representation of what, in reality, is the sum of many photochemical redox reactions (1).
5.	The presence of photochemical oxidation of CH4 does not totally eradicate it. Thus, CH4 is not exactly at zero abundance in today's atmosphere but at ~1.75 ppm by volume (ppmv). CH4 reaches a level at which its input rate equals its kinetic loss rate. Photochemical models that include biogenic O2 and CH4 production suggest that CH4 would be stable at 102 to 103 ppmv in the Archean atmosphere because of the effect of low O2 on the hydrogen budget of the atmosphere's chemistry (6).
6.	A. A. Pavlov, J. F. Kasting, L. L. Brown, J. Geophys. Res. 106, 23267 (2001) [CrossRef].
7.	J. F. Kasting, Science 293, 819 (2001) [Free Full Text] .
8.	H. D. Holland, The Chemistry of the Atmosphere and Oceans (Wiley, New York, 1978), pp. 291-292.
9.	Oxidation rates of oceanic Fe2+ do not imply unfeasible amounts of sedimentary Fe3+ [note (9) in Towe's comment). In calculating oxidation rates of Fe2+ [note 10 of (1)], we stated a simplifying assumption to ignore all other kinetic losses of O2 (through reaction with H2, CO, H2S, SO2, the crust, etc.). Thus, our purpose was to demonstrate that to maintain ~3 ppm Fe2+ in the deep Archean ocean required a low pO2 value (whatever the exact sinks on O2 may have been), given oceanic upwelling rates and Fe2+ oxidation rates. A large Fe2+ oxidation rate implies that Fe2+ oxidation was limited by pO2, not by oceanic upwelling, and supports the idea that other sinks on O2 must have been greater than today.