To better predict the long-term climate impact of future anthropogenic CO2 emissions, it is essential that we quantify mechanisms of climate system and carbon cycle coupling over centennial to millennial timescales. The last 50,000 years offer an ideal natural laboratory to investigate such climate-carbon cycle feedbacks. This period was punctuated by multiple abrupt climate shifts and large-scale climatic reorganizations, each mediated by changes in the partitioning of carbon across various atmosphere, ocean, and terrestrial reservoirs. However, the spatial and temporal dimensions of carbon sinks versus sources, including their relationship to atmospheric CO2, remain poorly quantified.

Climate model experiments are widely used to evaluate climate-carbon cycle feedbacks over short time frames, but require additional validation from paleoenvironmental proxies across centennial and longer timescales. To more explicitly facilitate the testing and constraining of models beyond the limits of present-day climate scenarios, this project will apply paleo-data assimilation to generate the first proxy-constrained, spatiotemporally continuous global carbon-cycle reanalysis of the last 50,000 years. This will be done by combining transient simulations from climate system models of intermediate complexity with geochemical proxies to investigate drivers and modes of carbon-cycle variability over centennial and millennial timescales.

The successful candidate will i) perform transient climate model simulations using a wide range of forcing and boundary conditions to produce an ensemble of climate and carbon-cycle states from our models, and ii) apply data assimilation techniques to produce a complete reanalysis of climate and carbon-cycle change for the past 50,000 years.