Accurate projections of climate recovery from a combination of historical data, simple models and AI

With the 10 warmest years on record all occurring in the last decade, we are on the brink of exceeding the Paris Agreement’s 1.5°C warming target. In response, the 2023 United Nations Climate Change Conference outlined targets to achieve net-zero emissions. Consequently, the need for projections of high-emission scenarios is waning, instead we need to know what climate recovery looks like. Yet, how climate will respond to a state of net-zero emissions and gradual CO2 removal from the atmosphere remains poorly understood. Hence, investigating the processes and implications of climate recovery becomes critical for informing long-term actions and policies.
To date, climate scientists have used complex earth system models that integrate a vast array of processes across the atmosphere, ocean, land, ice, and living organisms to answer this question, but their different representations of climate components and interactions, such as climate feedbacks, ocean circulation, and sea ice, have led to diverging projections at global and regional scales. This vital knowledge gap in our understanding limits our ability to guide accurate mitigation actions and policies.
With a decade of experience developing novel statistical tools and simple physical models to improve historical estimates of climate states, I am now positioned to lead a world-class team and develop new modelling methods that integrate simple models, historical data, and AI to better predict future climate recovery. My novel modelling approach will be more accurate and quicker to run, and hence responsive to the rapidly evolving landscape of climate actions as we strive to achieve net-zero emissions.
To achieve these aims, I will form an inter-disciplinary team including climate scientists, computer scientists, and stakeholder engagement experts. My team and I will first examine the behaviour of temperature, rainfall, humidity, wind, and ice in recovery scenarios simulated by complex earth system models, with a...