A "CERN" for Climate Change
– An interview with the IMO Prize Lecturer
By Sylvie Castonguay, WMO Secretariat
The IMO Prize lecture delivered by Tim Palmer, Royal Society Research Professor in Climate Physics and a Senior Fellow at the Oxford Martin Institute, in June focused on ensemble forecasting for weather and climate, an area he helped to pioneer. Looking forward, Prof Palmer also advocated for “some kind of CERN [Conseil Européen pour la Recherche Nucléaire] for climate change where modelling centres around the world can pool human, computer and financial resources to develop a new generation of seamless kilometre-scale global weather/climate model.” It is a grand idea with a sense of urgency. The Bulletin interviewed Professor to learn why this idea has caught on with many climate scientists.
Bulletin: What more than existing climate models could a “Km-scale Climate-CERN” provide?
Professor Palmer: Scientists, and specifically climate scientists, have done a very good job warning society about the risk of climate change and global warming thanks to the global climate models that we have. But we urgently need to know much more than the existing models can provide: Countries require more information to guide infrastructure investments to adapt to climate change. Some countries are thinking about targeted or local geoengineering: injecting aerosols into clouds to make them brighter to reflect sunlight back into space, which could have a regional cooling effect. But what is the global implication of that? Would weakening a heat wave in one part of the world exacerbate a heat wave in a different region? Then there is the losses and damages discussion, which will be an important part of the upcoming 29th session of the Conference Of Parties (COP 29) of the United Nations Convention on Climate Change (UNFCCC) in Azerbaijan. A stronger scientific basis is required to attribute extreme events to climate change in a quantitative way. There is a lot of discussion in the media about tipping points – that we might do something irreversible to the ice sheets, ocean circulation or rainforests. Finally, tipping points have implications for mitigation policies: cuttings emissions after a tipping point has been reached will be completely ineffective. For these reasons and many, many more reasons a CERN for climate is needed now to gather scientists, experts and resources.
We need to move up a gear in the science from just being able to warn in general terms about climate change on a global basis to being very specific about what climate change will bring to individual regions and countries. The current generation of climate models are not capable of making very specific, very detailed predictions and projections, their resolution is too coarse. Why? Because universities and national institutions have had to tailor their ambitions to their computing capabilities – to their resources. So, unfortunately today, we just cannot answer key regional questions about climate change with confidence and clarity.
I and other climate scientists argue that getting climate models to the stage where they can really, reliably address questions about regional climate change will require a similar level of international collaboration and coordination as CERN. We need answers these questions in the next 5 to 10 years, not in the next 50 years.
Bulletin: Could you give us examples of answers that km-scale modelling could provide?
Professor Palmer: We are recording many extreme events like the incredible 50 °C temperatures in British Columbia, Canada, a couple of summers ago, extraordinary flooding in Pakistan, prolonged drought in southern Africa. We cannot answer precise questions about how climate change affected those extreme events simply because the models can’t simulate them. But it is not just extreme local events: over the last year or so, global mean temperatures have been higher month on month than predicted by the Coupled Model Comparison Project and other standard models, at over 1/10 to 2/10 °C, which is quite significant.
Bulletin: What are the roadblocks to a Climate-CERN solution?
Professor Palmer: Climate models started on the back of national weather forecast models. Today, many national centres have global models. The roadblock is getting the existing institutions to realize that the resolution of their models is too coarse to help with the important problems that I mentioned, and that national resources are unlikely to be able to move the problem forward, at least on a timescale that is relevant to the climate crisis. The only solution is to work together internationally now to do this. You could say it’s a political problem rather than a scientific problem.
Bulletin: What resources would we need for km-scale global Earth System Model?
Professor Palmer: Dedicated exascale computing – a billion billion arithmetic calculations per second – is needed. Available exascale computing power – for science at least – is currently shared between many different applications, a whole range of people are queuing up to use them. To demonstrate unambiguously that km-scale models will produce more reliable forecasts with lower systematic errors, and with more accurate representations of extreme events and so on, climate scientists need dedicated access to exascale computing.
So funding is the biggest issue with launching a Climate-CERN. Countries and philanthropists would have to contribute to the international effort. The finer the resolution of the climate models, the more accurate the representation of the laws of physics and the more specific and precise the information science can provide. But a sizeable initial investment is required – comparable but no bigger than CERN.
Bulletin: Time is of the essence. Could Climate-CERN be up and running in a short window of time?
Professor Palmer: I do strongly believe that if the politics can be solved and financing can be found, then scientists will do the job in a very short time. There are plenty of precedents. Just think of the Apollo program to put a man on the moon. Or much more recently, the production of COVID vaccines. Bulletin: Other than physics, what kind of expertise would be required?
Professor Palmer: Km-scale modelling of the climate system is one small part of Climate-CERN. Climate-CERN would also be about the impacts of climate on people and the environment, finding how to mitigate those impacts and identifying solutions. It would be critically important to have the whole spectrum of impact and application models, which would go from health through to agronomy, water resources, energy, food... To answer questions on what types of plants should be grown on a year-to-year basis, to planning urban environments...pretty much everything that is impacted by climate change. Even agent-based economic modelling should be fully integrated.
Artificial Intelligence would play a crucial role in identifying key variables in the wealth of climate data and how they should be fed into impact models to produce meaningful plain language outputs for policy and other decision-makers.
An international approach, funding and a wealth of expertise that exists across the world is needed to develop the next generation of climate models.
Bulletin: Professor, do you have any final words for our readers?
Professor Palmer: I started my research career in abstract theoretical physics. Then I thought, I don’t really want to spend that rest of my career doing stuff that doesn’t help many people. At the time, it was easier to do to pivot into a different field, such as weather and climate science, than it is today. We need to fund schemes to allow scientists (for example, at the postdoc level) to pivot from pure science to applied science. Fellowships to allow pure scientists to spend a year getting up to speed in their chosen applied field should be created.
Let me conclude with this. I cannot say how honoured I was to get the IMO Prize. When I read the list of prize winners over the years, it was of all my greatest heroes in meteorology and climatology. It is unbelievably humbling to be on such a list of esteemed people.