From single projection to many possibilities: A new tool for climate risk assessment

Whether planning flood protection, food production, or financial investments, decisions about climate change should be based on an ensemble of projections as they depend on understanding the full range of plausible futures. IIASA researchers and partners have developed an open-source tool that makes it much faster to generate regional climate and impact projections while capturing the uncertainties that shape future climate risks.

Assessing how climate change will affect different regions usually requires running a complex chain of models. While this approach provides valuable insights, it is expensive and can only cover a limited set of scenarios and model runs. Together with partners, IIASA researchers have developed a faster alternative that uses available projections from Earth system- or climate impact models to generate new regional climate and impact projections while capturing the uncertainty that is essential for understanding climate risk.

The new method, called RIME-X, combines a range of future warming projections from simple climate models with regional information from multi-model climate and impact model ensembles. Instead of producing a single projection, it generates scenario- and time-dependent probability distributions that account for uncertainty in the global climate response, differences between models, and natural climate variability.

“We needed a fast, flexible tool that captures all of that uncertainty at once,” says lead author Niklas Schwind, a researcher in the Integrated Climate Impacts Research Group of the IIASA Energy, Climate, and Environment Program. “Climate risk isn’t a single number, it’s a range of plausible outcomes, and RIME-X makes it practical to see and use that full range quickly for a wide range of regions, scenarios, or indicators.”

To test the new approach, the researchers validated RIME-X against climate model simulations that had been deliberately withheld during method calibration. Across a wide range of regions and indicators including temperature, precipitation, extreme heat, and crop yields, the emulator closely reproduced the distributions generated by the full climate model ensemble. Typical deviations were only 1-8% of the model spread, making them one to two orders of magnitude smaller than the spread between the models themselves.

“One of the strengths of RIME-X is that it combines global temperature projections with regional indicator distributions from weighted multi-model ensembles into a single framework,” says coauthor Edward Byers, a senior research scholar in the IIASA Energy, Climate, and Environment Program. “That means users can rapidly explore the full range of plausible regional climate impacts across many different scenarios without having to run the complete modeling chain.”

The new framework will benefit climate scientists and impact modelers who need rapid regional climate and impact projections, as well as practitioners in sectors such as finance and insurance that increasingly rely on climate risk assessments. It could also support policymakers and adaptation planners by providing quick, robust estimates without requiring access to supercomputing resources or specialist modeling expertise. The tool is openly available as a Python package and already powers the publicly available Climate Impact Explorer. To help researchers and practitioners get started, the team has also published a step-by-step tutorial demonstrating how to use RIME-X alongside the open-source software.

“Decision-makers need to understand not only what is most likely to happen, but also the range of outcomes they should prepare for. RIME-X helps make that information much easier to access,” notes coauthor Michaela Werning, who is also associated with the Energy, Climate, and Environment Program at IIASA.

The authors emphasize that RIME-X is not intended to replace comprehensive climate models. Rather, it complements them by making it easier to explore more scenarios, regions, and climate indicators than would otherwise be practical. The approach is applicable to regional indicators whose distributions are mainly determined by global warming levels and provides users with a computationally efficient way to assess future climate risks while capturing the uncertainty that shapes them.

“The method is open source and designed to be broadly reusable, rather than a one-off analysis,” says coauthor Mahé Perrette, from the Alfred Wegener Institute (AWI). “We hope it lowers the barrier to getting credible regional climate risk information quickly, whether for research, adaptation planning, or climate risk assessment.”

Reference
Schwind, N., Perrette, M., Byers, E., Högner, A., Lejeune, Q., Möller, T., Nicholls, Z., Pfleiderer, P., Schöngart, S., Werning, M., & Schleussner, C.-F. (2026). RIME-X v1.0: combining simple climate models, Earth system models, and climate impact models into a unified statistical emulator for regional climate indicators. Geoscientific Model Development DOI: 10.5194/egusphere-2025-5781

About IIASA:
The International Institute for Applied Systems Analysis (IIASA) is an international scientific institute that conducts research into the critical issues of global environmental, economic, technological, and social change that we face in the twenty-first century. Our findings provide valuable options to policymakers to shape the future of our changing world. IIASA is independent and funded by prestigious research funding agencies in Africa, the Americas, Asia, and Europe. www.iiasa.ac.at