Quantifying carbon cycle feedbacks to past hyperthermal events

The magnitude of future climate change depends on how Earth's natural carbon reservoirs respond to the changing climate via carbon cycle feedbacks. Yet many of these feedbacks are poorly constrained and are widely acknowledged as a major source of uncertainty in climate projections, particularly into the long-term future. Whilst we can measure carbon cycle feedbacks over the historical period, the future pacing and strength of carbon cycle feedbacks remains uncertain. We do not yet know whether they will collectively amplify or dampen anthropogenic climate change in future or whether carbon cycle tipping point events will be triggered, releasing geologically sequestered carbon to the ocean-atmosphere. Hyperthermal events can serve as partial analogues to anthropogenic climate change and allow us to better constrain carbon cycle behaviour in response to global warming. However, most sedimentary proxy records of hyperthermals at the Earth’s surface record the net environmental change caused by both an initial ‘forcing’ and all subsequent ‘feedbacks’ to that forcing. Disentangling forcing and feedbacks signals across hyperthermals requires further independent constraints on some aspect of the system. The geological record is peppered with examples of past carbon emissions events from large igneous province (LIP) activity, many of which coincide with mass extinction and/or hyperthermal events. Here we show how carbon emissions from a large igneous province (the North Atlantic Igneous Province or NAIP) can be constrained at high resolution entirely independently from environmental proxy records. We further show how an Earth system modelling approach comparing NAIP carbon emissions predictions with proxy records of the Paleocene-Eocene Thermal Maximum (PETM, ~56 Ma) can be employed to constrain net global carbon cycle feedbacks to NAIP carbon emissions. Lastly, we show how the addition of carbon and trace metal isotope systems in this Earth system modelling framework has the potential to allow us to disentangle individual global carbon cycle feedbacks across events like the PETM, ‘fingerprinting’ the carbon reservoirs and quantifying their response to a known exogenic carbon input.