Advancing climate predictions

At the heart of this are Earth System Models (ESMs), which are instrumental in understanding the complexities of our planet's interconnected systems. These models not only deepen our understanding of climate change impacts but also serve as crucial guides for shaping climate policies and actions worldwide.

Sophisticated as these tools are, though, there’s always room for improvement as our knowledge of Earth systems grows. Updating them will, therefore, allow scientists to improve climate simulations that can help to address key policy questions that remain to be answered.  

In response to this need, the EU-funded ESM2025 project is working to improve key processes in ESMs to target specific policy needs and make these models more relevant and useful for policy makers.  

We caught up with Roland Seferian, Research Scientist at the National Centre for Meteorological Research (CNRM) and ESM2025 Coordinator, to find out more. 

We're developing five new generation European Earth System Models, along with an improved Integrated Assessment Model and an open-source reduced complexity Earth system model. This new, more integrated and sophisticated generation of ESMs improves the consistency between different models and incorporates advanced techniques like machine learning. 

One of our key innovations is the development of emission-driven ESMs for CO2, CH4, and N2O. This will enable more comprehensive studies of greenhouse gas dynamics. Other innovations include using machine learning to reduce process uncertainty and optimise model calibration.

We're also improving the representation of crucial processes across the atmosphere, land, ocean and cryosphere and enhancing the coupling between ESM components. Finally, we’re creating enhanced representations of land-use and land-based mitigation strategies in both Earth System and Integrated Assessment Models. 

Among many other findings, these innovations will enable more reliable information on CO2-based and non-CO2-based mitigation strategies, including co-benefits for air quality and food security. Also, more reliable estimates of allowable carbon and methane emissions to stay below specific global warming thresholds. 

We’ve been able to shed light on some key questions, like our remaining carbon budget. This may be much smaller than previously thought, requiring more stringent and immediate reductions in greenhouse gas emissions to meet the target of limiting global warming to 1.5°C above pre-industrial levels. 

We’ve also detected a potential reduction in the efficiency of natural carbon sinks like forests and oceans as global temperatures rise. These natural systems currently absorb a significant portion of anthropogenic CO2 emissions, but their capacity may decrease as the climate warms, leading to increased atmospheric CO2 levels and exacerbating climate change.

Our research has also yielded important insights into land-based mitigation strategies and their complex effects on the climate system. We've found that while methods like afforestation, reforestation and bioenergy crops are valuable tools for reducing atmospheric CO₂, their overall impact on the climate system, including soil carbon, is still subject to debate. 

These strategies can have both positive and negative environmental effects, depending on how and where they're implemented, and may conflict with other sustainability goals. For instance, expanding bioenergy crop production often comes at the expense of forests or natural ecosystems, potentially reducing biodiversity and soil carbon storage.

We also have data on the need to reduce non-CO2 greenhouse gases like methane and nitrous oxide, which actually have higher potential to contribute to global warming. Understanding how methane sources and sinks (processes that add or remove methane) change over time is essential for predicting future climate conditions. And for assessing the effectiveness of mitigation strategies targeting anthropogenic methane emission reductions. 

The development of interactive methane cycle modeling in ESMs allows for dynamic simulation of methane's sources, sinks, and atmospheric concentrations, enabling real-time feedback between methane emissions and the climate system. 

This approach is crucial because methane is a powerful greenhouse gas with natural sources like wetlands, wildfires, and termites that are highly sensitive to climate conditions. Our models now account for how changes in climate variables influence natural methane emissions, and vice versa, leading to more accurate projections of future climate scenarios and better assessment of methane-targeted mitigation strategies.

As well as the critical areas I mentioned, we’ve provided policymakers with updated insights into issues like the impact of incremental warming, temperature overshoot above 1.5°C and the reliability of current overshoot predictions.  

These and other novel scientific data help policymakers better understand how Earth systems respond to policy actions, reduce uncertainties and provide a more robust basis for decision-making. We hope this knowledge will also help with climate negotiations. And, of course, our models provide valuable scientific insights to support the successful implementation of the Paris Agreement. 

We also engage with European, national, and local stakeholders to co-develop strategies and research priorities. Our efforts also highlight gaps between the EU and local decision-makers, which could help to bridge this divide. 

By nature, all models are imperfect, but some are still useful—our task as developers is to make them even more accurate and useful. 

Regarding usefulness, as demands for precision in Earth system modelling grow, we need to develop a new generation of more specialised models. The era of a single model serving all purposes is likely over. 

Challenges in model development are numerous, and fall into two major streams: Increasing resolution to better simulate climate-relevant processes, and representing missing processes, which is the focus of ESM2025.

As just one example, we’re enhancing emission-driven capabilities in ESMs by simulating the entire chain of processes, from greenhouse gas emissions to their impacts and feedback. This lets us make ESMs more suitable for mitigation studies and future projections. 

They provide climate researchers with much more powerful tools and realistic models to work with, so they can explore complex climate scenarios. They’ll be able to conduct more precise simulations, capture a broader range of nuanced, climate-relevant interactions and, potentially, provide more reliable projections for policymakers. 

Our work has already demonstrated the added value of emission-driven models and influenced multiple communities, including the Coupled Model Intercomparison Project (CMIP), an international initiative supporting IPCC assessments. Emission-driven models are now prioritised within CMIP, marking a significant success for our team. 

Crucially for the scientific community, we’re committed to open access. Our publications and data will support and accelerate open research in climate science based on more powerful tools. 

While science-policy dialogue is well-established, conveying up-to-date concepts to younger generations of EU citizens remains a challenge. To address this, we’ve organised outreach events targeting the teaching community. 

These efforts have been particularly successful and we’ve seen a faster response from teachers compared to international policymakers—the opposite of what you expect as a climate scientist. We’ve also co-developed novel open-source resources for teachers, supporting Article 12 of the Paris Agreement. 

We also organise regular stakeholder forums like our annual World Café workshops to identify key research priorities for climate policy. These are an opportunity for researchers and stakeholders to engage and dialogue, helping to align research objectives with policy needs.

Our regular policy briefings also summarise key insights from our research and stakeholder engagements, helping to translate complex scientific findings into accessible information for policymakers and other stakeholders.

And, last but not least, we have an active, ongoing dialogue with other research communities and international research bodies like the WCRP and IPCC, who are also our stakeholders.

One notable challenge is reconciling timelines—stakeholders often have many questions about climate change and projections but are overcommitted and have limited time. The same could be said for scientists, who also evolve on different timelines to policymakers or educators. So the challenge is how to bring them all together.  

To help address this, we initiated #ClimateResearchNet—a communication network bringing together a wide variety of EU climate change research projects. This platform aims to improve clustering between climate research projects to increase the impact of climate research communications. We’re also focusing on key stakeholders and teaming up with experts (like the OCE for education).  

Engage with platforms like #ClimateResearchNet that consolidate resources from various projects while offering access to scientific expertise when required. For scientists, understanding stakeholder constraints—such as time limitations—is crucial for effective communication. Collaboration networks like ours are one way to try to pool resources and bridge these gaps without reinventing the wheel.