A tighter constraint on Earth-system sensitivity from long-term temperature and carbon-cycle observations

Abstract:

The long-term Earth-system response to changes in climate forcing hinges on a number of geophysical properties, including the Earth-system sensitivity (ESS). The ESS represents Earth’s long-term climate system response to a given CO₂ forcing, including both fast and slow feedback processes. Current ESS estimates are subject to substantial uncertainties. Long-term carbon cycle models can provide a useful avenue to investigate these uncertainties and constrain ESS, but previous efforts either lack a formal framework to integrate the models employed with paleo data, or focus on discrete paleoevents. Here, we improve on ESS estimates by using a Bayesian approach to fuse deep-time paleoclimate CO₂ and temperature data over the last 420 Myrs with a long-term carbon cycle model. We evaluate the ability of CO₂ and temperature data to constrain ESS individually and as a joint constraint, to assess the ability of these data sources to improve our knowledge of this key Earth-system property. Our best ESS estimates show a narrower range than previous assessments, implying increased learning, and highlight the critical role of temperature in constraining ESS. In light of apparent model biases relative to temperature proxy data during the Cretaceous period, we conduct a set of sensitivity analyses to determine what model parameterizations are responsible for these biases. Our sensitivity analyses indicate that during the Cretaceous, the efficiency of chemical weathering mediated by gymnosperms and a shift in the timing of gymnosperm- to angiosperm-dominated vegetation are key factors in the agreement between model output and temperature records. We find that research into improving the understanding about these plant-assisted weathering mechanisms and the tandem constraints of CO₂ and temperature proxy records provide potentially powerful avenues to further constrain ESS.