Geochemical Variability in Fossil Soils and Implications for Past Biogeochemical Cycling, Climates, and Atmospheres

Abstract

The co-evolution of the terrestrial biogeochemical cycle, the atmosphere, and the marine biosphere remain relatively poorly understood, with outstanding questions surrounding terrestrial-marine links, climate, and tectonics. In particular, the terrestrial sediment source (i.e., soils) remains understudied relative to the marine sediment sink, with the source essentially defined by the record in the sediment sink rather than being considered equally important. Both the sediment source and sink need to be well-constrained in order to understand global biogeochemical changes. Additionally, interpretations of trends in paleosol (fossil soil) geochemistry are only loosely constrained by large-scale modern soil chemical variability, limiting our ability to assess potential changes in biogeochemical cycling through time. This dissertation focused on two primary goals: improving quantitative constraints on terrestrial biogeochemical cycling and weathering over geologic time, and improving our ability to accurately interpret those records by understanding both modern context and what the paleosol record actually represents. To address these goals, I analyzed the geochemical composition of soils and paleosols (fossil soils) over the past three billion years. Because soils form in the ‘critical zone’—the intersection of the biosphere, geosphere, and atmosphere at Earth’s surface—they record surficial conditions more directly than other geologic records, providing valuable insight into past climates, atmospheres, and ecosystems. After providing generalized, quantitative constraints on geochemical and weathering variability in modern soils (Chapter II), I used the paleosol record to test for state changes in soil P (Chapter III) and weathering intensity (Chapter IV) on land during key biogeochemical transitions. I also explored a variety of processes that could bias the distribution of paleosols through space and time (e.g., preservation, sampling), which needs to be better constrained in order to interpret paleosols accurately. In modern soils, I found weaker than expected relationships between soil P and Fe geochemistry and key environmental factors (climate, vegetation, parent material), but weathering intensity, the presence of vegetation, and P concentrations were related. The weak relationships could be due to the continental rather than localized scale of analysis. While the latter might have provided predictive relationships between soil chemistry and soil-forming factors, a highly-localized scale is often not considered in deep-time biogeochemical modeling. In paleosols, I found that both the P composition and weathering intensity have been stable through time. Discrete, state changes in P composition or weathering intensity—as have been hypothesized based upon marine records—were not recorded. A discrete change was present in the concentration of Ca in paleosols, which increased in the Phanerozoic, perhaps reflecting a shift in pedogenic processes as vascular, rooting plants evolved. Roots and vascularity allowed plants to colonize more arid environments and facilitated the formation of pedogenic carbonate—an important C sink. Therefore, while the advent of land plants may not have led to a global state change in either terrestrial P retention or weathering intensity, plants facilitated the growth of the soil C sink. Because weathering intensity is consistent through time, other factors (e.g., land area, erosion rates) would have been dominant controls on marine nutrient supply through time, with shorter-term perturbations in weathering intensity occurring before returning to the stable baseline. Finally, the distribution of paleosols through time is uneven, with more paleosols being more common (a) towards the present and (b) during peaks in zircon ages, suggesting a formation and/or preservation bias related to the supercontinent cycle.