New study simulates greatly reduced permafrost

Permafrost is an important component of the climate system – and not just for those people living in permafrost regions, who can see the negative effects of the dwindling ice on the existing infrastructure: Given the large amounts of carbon released when permafrost thaws, the loss of permafrost also has a warming effect on the global climate. Yet to date, many global climate models cannot yet adequately reflect the stability and distribution of permafrost, or the feedback effects on the global climate resulting from its loss. An international research team led by the Nansen-Zhu International Research Centre at the Chinese Academy of Sciences has now for the first time applied a more extensive ensemble of climate models (17 PlioMIP2 models) to quantify the distribution of permafrost in warm climates. In this regard, the experts focused on a warm period during the mid-Pliocene, roughly three million years ago – a time when the conditions were similar to those predicted by the end of the 21^st century in the SSP5-8.5 scenario, i.e., one characterised by unchecked use of fossil fuels.

“Our study puts the spotlight on permafrost and uses the mid-Pliocene as a laboratory to address the question of how well models can simulate warm climatic conditions,” says Dr Christian Stepanek, a paleoclimate expert at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI). “To do so, we take advantage of a unique aspect offered by paleoclimate research: We don’t have to rely on models that have only be validated using modern observational data. We can and do use model-independent proxy data from climate archives as reference values, which allows us to quantify model uncertainties, and to underscore the conclusions from those models that performed best,” Stepanek explains, before adding: “Warmer climatic conditions of the past and future could have very different qualities to what we know today. Accordingly, models that yield results on today’s climate that match the observational data won’t necessarily deliver reliable results for the future. That’s what makes it so important to also apply the models to warmer climates from the past, so as to gain a more robust understanding of the dynamics of a warming Earth system.”

In their work, the researchers combined the distribution of permafrost in two warm worlds – those of the Pliocene and the future scenario. This approach afforded them a proxy-data-supported outlook on permafrost behaviour in a future, warmer world. This is important, given the critical role of permafrost distribution and stability for the climate and society. The experts were amazed at how narrow the span of uncertainty regarding the distribution of permafrost in the mid-Pliocene was in those models successfully validated using proxy data. Based on this finding, simulations of the future climate are bleak: In the type of climate projected for the end of the 21^st century and assuming unchecked use of fossil fuels, the near-surface permafrost extent could decline by more than 70 percent.

Background:

PlioMIP2: The second phase of the Pliocene Model Intercomparison Project is an internationally coordinated research initiative in which the climate of the Pliocene, and the certainties and uncertainties in what we know about it, are assessed using a combination of model-based simulations and proxy data. PlioMIP2 is part of the Paleoclimate Model Intercomparison Project. The research conducted for the project also contributes to the IPCC Assessment Reports.

Original publication:

Guo, D., Wang, H., Romanovsky, V. E., Haywood, A. M., Pepin, N., Salzmann, U., Sun, J., Yan, Q., Zhang, Z., Li, X., Otto-Bliesner, B. L., Feng, R., Lohmann, G., Stepanek, C., Abe-Ouchi, A., Chan, W.-L., Peltier, W. R., Chandan, D., von der Heydt, A. S., Contoux, C., Chandler, M. A., Tan, N., Zhang, Q., Hunter, S. J., and Kamae, Y.: Highly restricted near-surface permafrost extent during the mid-Pliocene warm period; Proceedings of the National Academy of Sciences, PNAS (2023). DOI: 10.1073/pnas.2301954120.