A group of scientists from the University of Oxford have developed a new algorithm that reduces spin-up time by 90 %, allowing for faster and more reliable climate predictions. The research was published in the journal Science Advances.
Image Credit: Avigator Fortuner/Shutterstock.com
Earth System Models are intricate computer simulations that portray Earth's processes and their interactions, playing a pivotal role in predicting future climate changes. By replicating how our land, oceans, and atmosphere respond to human-induced greenhouse gas emissions, these models form the foundation for projecting potential extreme weather and climate events, including those outlined by the UN Intergovernmental Panel on Climate Change (IPCC).
However, climate modelers have long faced a major problem. As Earth System Models integrate many complicated processes, they cannot immediately run a simulation; they must first ensure that it has reached a stable equilibrium representative of real-world conditions before the Industrial Revolution. Without this initial settling period – referred to as the “spin-up” phase – the model can “drift,” simulating changes that may be erroneously attributed to manmade factors.
Unfortunately, this process is very slow because it involves running the model for thousands of model years, which can take up to two years on some of the most powerful supercomputers in the world for IPCC simulations.
However, a study published today in Science Advances by a University of Oxford scientist, supported by the Agile Initiative, unveils a novel computer algorithm. This algorithm, applicable to Earth System Models, remarkably cuts down spin-up time. During tests on models utilized in IPCC simulations, the algorithm proved to be an average of 10 times faster at initializing the model compared to current methods. This reduction in spin-up time decreases the duration needed to achieve equilibrium from several months to less than a week.
Minimizing model drift at a much lower cost in time and energy is obviously critical for climate change simulations, but perhaps the greatest value of this research may ultimately be to policymakers who need to know how reliable climate projections are.
Samar Khatiwala, Professor and Study author, Department of Earth Sciences, University of Oxford
Samar devised the algorithm.
Currently, climate scientists are unable to run their models at a higher resolution and define uncertainty through repeat simulations due to the long spin-up times of many IPCC models.
The new algorithm will allow researchers to investigate how small changes to the model parameters can change the output, which is important for characterizing the uncertainty of future emission scenarios by significantly reducing the spin-up time.
The novel algorithm developed by Professor Khatiwala makes use of a mathematical technique called sequence acceleration, which dates back to the renowned mathematician Euler. In the 1960s, D. G. Anderson used this concept to accelerate the solution of Schrödinger's equation, which describes the microscopic behavior of matter.
This problem is so important that over half of the world's supercomputing capacity is currently being used to solve it. One of the most widely used algorithms for this purpose is the “Anderson Acceleration” algorithm.
Professor Khatiwala discovered that Anderson Acceleration could potentially decrease model spin-up time as both issues involve iterative processes: generating an output and then reintegrating it into the model multiple times. The final solution is reached much faster by preserving past outputs and amalgamating them into a unified input using Anderson’s scheme.
This not only accelerates the spin-up process and reduces computing demands but also offers broad applicability to various models employed in policy formulation across diverse domains, spanning from biodiversity conservation to addressing ocean acidification.
In anticipation of the next IPCC report, which is expected to come out by 2029, research groups worldwide are starting to develop their own models. Professor Khatiwala is collaborating with several of these groups, including the UK Met Office, to test his methodology and software in their models.
Policymakers rely on climate projections to inform negotiations as the world tries to meet the Paris Agreement. This work is a step towards reducing the time it takes to produce those critical climate projections.
Helene Hewitt OBE, Professor and Co-Chair, Coupled Model Intercomparison Project Panel
Professor Colin Jones, Head of the NERC/Met Office-sponsored UK Earth system modeling, said, “Spin-up has always been prohibitively expensive in terms of computational cost and time. The new approaches developed by Professor Khatiwala have the promise to break this logjam and deliver a quantum leap in the efficiency of spinning up such complex models and, as a consequence, greatly increase our ability to deliver timely, robust estimates of global climate change.”
Journal Reference:
Khatiwala, S., et al. (2024) Efficient spin-up of Earth System Models using sequence acceleration. Science Advances. doi.org/10.1126/sciadv.adn2839