Answers are grounded in this paper and cite their sources.
As Earth system models (ESMs) become increasingly complex, independent validation of their predictions becomes increasingly important. Cloud and convective parameterizations are one of the most uncertain components of ESMs, and directly impact model sensitivity to CO2 changes. Simulations of past climate provide an opportunity to vet model performance outside of their original design conditions. This study uses paleoclimate data assimilation with a perturbed parameter ensemble (PPE) of ESM simulations with altered cloud and convective parameterizations to identify parameter choices that best match proxy data. Specifically, we use offline Kalman filter data assimilation to generate reconstructions of the Last Glacial Maximum (LGM) and pre-industrial (PI) periods using NASA GISS E2.1-G simulations and assimilating sea surface temperature and terrestrial water isotope proxy data for each time period. We find a LGM to PI global mean surface temperature (GMST) anomaly of-6.2±0.1°C (1𝜎), consistent with previous data assimilation studies. Subsequently, we apply an adaptation of the particle filter assimilation method to rank PPE members by their similarity to proxy data. We find the highest-ranked ensemble members have parameterizations that lead to reduced tropical precipitation and enhanced mid-latitude low cloud coverage relative to the ensemble mean. These members have equilibrium climate sensitivities (ECS) below 3.6°C and improve some of the precipitation biases seen in the GISS E2.1-G standard configuration. Our results demonstrate that paleoclimate data assimilation is a useful tool to select cloud and convective parameterizations in models that perform well across multiple climate states.
No takes yet. Share an insight, caveat, or question.
Tessler et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: