Paleosols and the evolution of atmospheric oxygen: a critical review

A number of investigators have used chemical profiles of paleosols to reconstruct the evolution of atmospheric oxygen levels during the course of Earth history (Holland, 1984, 1994; Kirkham and Roscoe, 1993; Ohmoto, 1996). Over the past decade Holland and his co-workers have examined reported paleosols from six localities that formed between 2.75 and 0.45 Ga. They have found that the chemical profiles of these paleosols are consistent with a dramatic change in atmospheric PO2 between 2.2 and 2.0 Ga from < or = 0.002 to > or = 0.03 atm (Holland, 1994). Ohmoto (1996) examined chemical data from twelve reported paleosols ranging in age from 2.9 to 1.8 Ga. He concluded that these chemical profiles indicate that atmospheric PO2 has not changed significantly during the past 3.0 Ga. We seek to resolve the conflict between these reconstructions through a broader examination of the paleosol literature, both to determine which reported paleosols can be definitively identified as such and to determine what these definite paleosols tell us about atmospheric evolution. We here review reports describing over 50 proposed paleosols, all but two are older than 1.7 Ga. Our review indicates that 15 of these reported paleosols can be definitively identified as ancient soils. The behavior of iron uring the formation of these 15 paleosols provides both qualitative and semiquantitative information about the evolution of the redox state of the atmosphere. Every definitely identified pre-2.44 Ga paleosol suffered significant Fe loss during weathering. This loss indicates that atmospheric PO2 was always less than about 5 x l0(-4) atm prior to 2.44 Ga. Analysis of the Hokkalampi paleosol (2.44-2.2 Ga) (Marmo, 1992) and the Ville Marie paleosol (2.38-2.215 Ga) (Rainbird, Nesbitt, and Donaldson, 1990) yield ambiguous results regarding atmospheric PO2. Loss of Fe during the weathering of the 2.245 to 2.203 Ga Hekpoort paleosol (Button, 1979) indicates that atmospheric PO2 was less than 8 x 10(-4) atm shortly before 2.2 Ga. The presence of red beds immediately overlying the Hokkalampi, Ville Marie, and Hekpoort paleosols suggests that by about 2.2 Ga there was an unquantified but substantial amount of oxygen in the atmosphere. Iron loss was negligible during formation of the 2.2 to 2.0 Ga Wolhaarkop (Holland and Beukes, 1990) and Drakenstein (Wiggering and Beukes, 1990) paleosols and during formation of all the later paleosols we previewed. Thus, atmospheric PO2 probably has been > or = 0.03 atm since sometime between 2.2 and 2.0 Ga.