The preponderance of geologic evidence does not support carbon dioxide as the main carbon species degassed from early Earth, nor a carbon dioxide–rich early atmosphere. In fact, there are several problems that cannot be addressed by assuming either of these facets of what has become conventional wisdom about the early atmosphere. A careful examination of the conditions that most likely accompanied late accretion, incorporating the most probable average composition of accreting materials, suggests an early atmosphere produced by degassing of reduced carbon and nitrogen species, followed by photochemical processing to yield a surface environment rich in organic compounds. Recycling of these organics through hydrothermal and volcanic systems would have maintained a level of reduced gases (photochemically unstable as they may be) in the early atmosphere for an extended period, accompanied by a growing carbon dioxide component derived from mantle magmatism. Such a model for the early atmosphere is not only consistent with geological data, it also solves many problems of the early history of Earth.

Shaw postulates that Earth's early atmosphere was rich in reducing gases such as hydrogen, brought to Earth via impact events. This commentary seeks to place constraints on this idea through a very brief review of existing geological and geochemical upper limits on the reducing power of Earth's atmosphere prior to the rise of oxygen. While these constraints place tight limits on this idea for rocks younger than 3.8 Ga, few constraints exist prior to that time, due to a paucity of rocks of that age. The time prior to these constraints is also a time frame for which the proposal is most plausible, and for which it carries the greatest potential to explain other mysteries. Given this potential, several tests are suggested for the H 2 -rich early Earth hypothesis.

The mantle was probably oxidized early, during and shortly after accretion, and so the early atmosphere of Earth was likely dominated by CO 2 and N 2 , not by CH 4 and NH 3 . CO 2 declined from multibar levels during the early Hadean to perhaps a few tenths of a bar by the mid- to late Archean. Published geochemical constraints on Archean CO 2 concentrations from paleosols are highly uncertain, and those from banded iron formations are probably invalid. Thus, CO 2 could have been sufficiently abundant during the Archean to have provided most of the greenhouse warming needed to offset the faint young Sun. H 2 might have augmented this warming prior to the origin of methanogenic bacteria. Atmospheric CH 4 concentrations increased from at most tens of parts per million (ppm) on prebiotic Earth to hundreds of parts per million once methanogens evolved. CO was an important trace gas on prebiotic Earth because of its high free energy and its ability to catalyze key reactions involved in prebiotic synthesis. Large impacts could have made the atmosphere transiently CO rich, and this may have played a role in the origin of life and in fueling early biological metabolisms.

Oxygenic photosynthesis appears to be necessary for an oxygen-rich atmosphere like Earth's. However, available geological and geochemical evidence suggests that at least 200 m.y., and possibly as many as 700 m.y., elapsed between the advent of oxygenic photosynthesis and the establishment of an oxygen atmosphere. The interregnum implies that at least one other necessary condition for O 2 needed to be met. Here, we argue that the second condition was the oxidation of the surface and crust to the point where free O 2 became more stable than competing reduced gases such as CH 4 , and that the cause of Earth's surface oxidation was the same cause as it is for other planets with oxidized surfaces: hydrogen escape to space. The duration of the interregnum was determined by the rate of hydrogen escape and by the size of the reduced reservoir that needed to be oxidized before O 2 became favored. We speculate that hydrogen escape determined the history of continental growth, and we are confident that hydrogen escape provided a progressive bias to biological evolution.

The current understanding of the evolution of the atmosphere, hydrosphere, and biosphere on early Earth has been strongly influenced by the following six major paradigms for the geochemical cycles of oxygen, iron, and sulfur: (1) a dramatic change from a reducing to an oxidizing atmosphere at ca. 2.4–2.2 Ga, termed the “Great Oxidation Event” (GOE); (2) Fe-rich oceans until ca. 1.85 Ga; (3) a hydrothermal origin for the global oceanic Fe; (4) SO 4 2− -poor oceans before the GOE; (5) an atmospheric origin for the oceanic sulfur species; and (6) the existence of sulfidic Proterozoic oceans. Each of the six paradigms has been built on other paradigms, such as those concerning: (1) the behavior of Fe during soil formation, (2) the environments and processes required for the formation of Fe III oxides in banded iron formations (BIFs), and (3) the origins of siderite and pyrite, as well as (4) the origin of anomalous isotope fractionation of sulfur (AIF-S) in Archean sedimentary rocks. Here, we show that some of the paradigms contradict each other, and that each has serious flaws (contradictions, problems) when they are compared to a variety of observations (geological, mineralogical, or geochemical data from natural samples; laboratory experimental data; results of theoretical studies). In contrast, all of these observations are better explained by the Dimroth-Ohmoto model for Earth's evolution, which postulates that a fully oxygenated atmosphere-ocean system developed by ca. 3.5 Ga. Examination of the available data from natural and experimental systems has also led us to suggest the following: (1) The geochemical cycles of O, Fe, and S (and other redox-sensitive elements) through the atmosphere–ocean–oceanic crust–mantle–continental crust have been basically the same as today since at least ca. 3.5 Ga. (2) The anaerobic and aerobic microbial biospheres, both in the oceans and on land, developed by ca. 3.5 Ga, playing an important role in the geochemical cycles of nutrients and other elements. (3) The geochemistry of sedimentary rocks (shales, carbonates, cherts) has been basically the same since ca. 3.5 Ga. (4) Fe III oxides in BIFs were formed by reactions between locally discharged Fe 2+ - and silica-rich submarine hydrothermal fluids and O 2 -rich deep seawater. (5) Magnetite in BIFs was formed during high-temperature diagenetic stages of BIFs through reactions between primary goethite or hematite and Fe 2+ -rich hydrothermal fluids. (6) BIFs were formed throughout geologic history. (7) Sulfidic oceans (i.e., the “Canfield ocean”) did not exist during the Proterozoic Eon. However, regional sulfidic seas, like the Black Sea, have existed in globally oxygenated oceans throughout geologic history. (8) The primary carbonate in Archean oceans, as in younger oceans, was Fe-poor calcite. Furthermore, (9) the pre–1.8 Ga atmosphere was CO 2 rich with the p CO 2 level greater than ~100 present atmospheric level (PAL). CO 2 alone provided the green-house effect necessary to compensate for the young Sun's lower luminosity. (10) The Archean pH values were 4.0–4.5 for rainwater, between 4.5 and 6.0 for river water, and 7.0 ± 0.5 for ocean water. The oceans were saturated with calcite but under-saturated with siderite. (11) The δ 18 O of Archean oceans was ~0‰, as today. (12) Fe-rich carbonates (siderite, ankerite) have formed during the diagenesis of sediments throughout geologic history by reactions between the primary calcite and Fe 2+ -rich solutions, either hydrothermal solutions or those derived from biological or abiological dissolution of Fe III -(hydr)oxides within the sediments. Other suggestions include: (13) The ranges of δ 34 S values of pyrite and sulfates in Archean sedimentary rocks are much larger than those quoted in the literature and comparable to those in Proterozoic sedimentary rocks. (14) Pyrites in organic C–rich black shales associated with BIFs were formed by a reaction between Fe 2+ - and SO 4 2− -rich hydrothermal solutions and organic C–rich shales during early diagenetic stages of the host sediments. This reaction also created AIF-S in the pyrite and the residual SO 4 2− . (15) The AIF-S signatures in Archean and younger rocks were not created by the ultraviolet photolysis of volcanic SO 2 in a reducing atmosphere. AIF-S signatures are not evidence for a reducing atmosphere. (16) Contrary to a popular belief that AIF-S–forming events ceased at ca. 2.45 Ga, AIF-S was also formed at later geologic times. (17) The presence of AIF-S signatures in some pre–2.4 Ga rocks, but the lower abundance of AIF-S in post–2.4 Ga rocks, may reflect changes in the mantle-crust dynamics, including changes in the thickness and movements of oceanic lithosphere.

This is a commentary on the preceding chapter by Ohmoto et al., in which it is suggested that oxygen concentrations have been high throughout Earth history. This is a contentious suggestion at odds with the prevailing view in the field, which contends that atmospheric oxygen concentrations rose from trace levels to a few percent of modern-day levels around 2.5 b.y. ago. This comment notes that many of the data sets cited by Ohmoto et al. as evidence for a relatively oxidized environment come from deep-ocean settings. This presents a possibility to reconcile some of these data and suggestions with the overwhelming evidence for an atmosphere free of oxygen at that time. Specifically, it is possible that deep-ocean waters were relatively oxidized with respect to certain redox pairs. These deep-ocean waters would have been more oxidized than surface waters, thus representing an “upside-down biosphere,” as originally proposed 25 years ago by Jim Walker.

This review summarizes and assesses a series of papers presented at the Geological Society of America Annual Meeting in 2011 on the changing composition of Earth's early atmosphere. This is a developing field, with differing views, due largely to the facts of an incomplete rock record and negligible preservation of the gaseous components of the atmosphere. Nevertheless, there are constraints, available through geological proxies in the form of chemical sedimentary rocks that reflect the composition of the hydrosphere, and because the two are directly linked, the atmosphere. A review of the geological constraints on atmospheric conditions is presented for early Earth, from its formation at 4.56 Ga up to 1.8 Ga, followed by a developing model that links changing atmosphere/hydrosphere conditions and biosphere evolution to changes in planetary tectonics, including the evolving supercontinent cycle.