Climate and atmospheric composition during the Precambrian: from paleosols to numerical simulations simulations

To better understand the climatic impact of greenhouse gases and the evolution of atmospheric oxygen during the Archean, we developed a twofold approach comprising the study of paleosols from the Fortescue Group (Pilbara craton, Western Australia) and the modelling of climate and carbon cycle. We discovered at least six new paleosol occurrences formed between 2.76 and 2.69 Ga, which are distributed over large areas of several hundreds to thousands km2. Although largely reequilibrated by secondary hydrothermal processes, several of these paleosols preserve lithologic domains inherited from early weathering processes. Specifically, we report new mineralogical assemblages (ferric-montmorillonite, calcite, sulphate) preserved in the reference 2.76 Gyr old Mount Roe Basalt paleosols at Whim Creek. These mineral associations occur in lithofacies that were not documented in previous studies and raise questions about the interpretation that the Mount Roe paleosols were formed under a reducing atmosphere. Using a combination of mass balance calculations and sulphur isotope analysis of sulphate, we estimate the partial pressure of atmospheric oxygen (pO2) in equilibrium with the Mount Roe paleosols to range between 0.3 and 3% of the Present Atmospheric Level (PAL). We propose that the emergence of the Fortescue continental flood basalts by 2.76 Ga may be responsible for the accumulation of oxygen in the atmosphere. This process lasted long enough to trigger significant oxidative weathering as shown by numerous isotopic proxies in oceanic sediments between 2.7 and 2.5 Ga. As a consequence, thermodynamic models for pCO2 estimations that rely on the anoxic atmosphere hypothesis are irrelevant, and solely the mass balance model calculations (Sheldon, 2006) remain robust. Using climate modelling, we show that these estimates (pCO2 < 2.5 10-2 bar or 80 PAL, 1 PAL ≈ 300 ppmv) are consistent with an icefree surface at the end of the Archean without the need of invoking a high pCH4. We propose that paleoproterozoic ice ages were triggered by the collapse of atmospheric methane that follows the oxygenation of the atmosphere, but also by a decrease of pCO2 due to both an increase in weathering and a shutdown of volcanism. For older periods (3.5 – 2.5 Ga), we also show that the formation of clouds with large-sized dropplets, in an Archean atmosphere that is devoided of Cloud Condensation Nuclei (CCN), is able to maintain an ice-free surface even at very low pCO2 (6 PAL), despite the reduced solar luminosity. However, coupled climate-carbon simulations show that such a low pCO2 is unrealistic. In a general way, we argue that long-term pCO2 regulation is governed by the amount of continental surface exposed to weathering and their lithology. As a consequence, the evolution of climate during this period is strongly related to crustal growth models and associated lithologic changes.