Final Report Summary - ACRCC (Understanding the atmospheric circulation response to climate change)
My project sought to better understand the response of the atmospheric circulation to climate change, including how to characterize the uncertainties associated with that response. The latter are considerably greater than the uncertainties associated with the thermodynamic aspects of climate change such as surface warming and sea-level rise, because of the smaller signal-to-noise ratio of the climate-change response, and the impact of model uncertainty. This research is highly relevant to regional aspects of climate change and the associated climate-related risk.
A team of PhD students and post-doctoral research associates used both theoretical and computational methods to tackle this challenge from a number of different directions. A number of external collaborators were also involved. Later in the project, the team brought in social science expertise.
One line of attack was to improve climate models by diagnosing errors within them. We extensively examined the sensitivity of key features of atmospheric circulation simulated by the models, such as the strength and location of tropical rain belts and midlatitude storm tracks, to poorly constrained aspects of surface drag processes that need to be represented through parameterizations of unresolved processes. This is a comparatively unexplored aspect of climate model error. We overcame previous methodological limitations by developing new ways of constructing model hierarchies and of suppressing dynamical feedbacks in order to isolate key interactions between processes.
A second line of attack was to make better use of existing model projections of climate change, notably the CMIP5 multi-model archive of projections produced for the previous IPCC Assessment Report. The traditional approach to quantifying uncertainty in these projections is to take the ensemble mean as the central estimate, and the spread as a measure of uncertainty. This is widely admitted to be unfounded but is still the standard practice. We developed new ways to improve the signal-to-noise ratio of the atmospheric circulation response to climate change, and to identify its role in crucial climate-change impacts such as cold-season Mediterranean drying, including the different timescales of the response. We developed a novel ‘storyline’ approach to representing the uncertainty in the circulation response in a physically interpretable way.
A third line of attack was to better understand the physical processes and mechanisms behind atmospheric variability. This is challenging because climate noise is highly structured in space and time, with long-memory effects that can be difficult to separate from climate change. We critiqued the efficacy of standard statistical approaches using correlations of climate anomalies, which rely on an assumption of statistical stationarity. This assumption is generally invalid for the climate system. As an alternative we developed new physically based ways to diagnose causal inference, taking proper account of interannual variations in the seasonal cycle, and drawing on methods from artificial intelligence.
In addition to the achievements described above, the project involved substantial inter-disciplinary developments. These have raised the awareness of the importance of atmospheric circulation as a source of uncertainty in climate change risk, within the wider climate-change community. A notable application is the attribution of extreme weather and climate events in the context of climate change, which is a rapidly growing research area with considerable public interest. We have helped make the case for a paradigm shift towards a ‘storyline’ approach to unquantifiable aspects of climate risk, including its application to ecosystem and environmental catastrophes. We have further worked with psychologists to show that the approach to communication followed by the IPCC has significant limitations when it comes to conveying risk.
A team of PhD students and post-doctoral research associates used both theoretical and computational methods to tackle this challenge from a number of different directions. A number of external collaborators were also involved. Later in the project, the team brought in social science expertise.
One line of attack was to improve climate models by diagnosing errors within them. We extensively examined the sensitivity of key features of atmospheric circulation simulated by the models, such as the strength and location of tropical rain belts and midlatitude storm tracks, to poorly constrained aspects of surface drag processes that need to be represented through parameterizations of unresolved processes. This is a comparatively unexplored aspect of climate model error. We overcame previous methodological limitations by developing new ways of constructing model hierarchies and of suppressing dynamical feedbacks in order to isolate key interactions between processes.
A second line of attack was to make better use of existing model projections of climate change, notably the CMIP5 multi-model archive of projections produced for the previous IPCC Assessment Report. The traditional approach to quantifying uncertainty in these projections is to take the ensemble mean as the central estimate, and the spread as a measure of uncertainty. This is widely admitted to be unfounded but is still the standard practice. We developed new ways to improve the signal-to-noise ratio of the atmospheric circulation response to climate change, and to identify its role in crucial climate-change impacts such as cold-season Mediterranean drying, including the different timescales of the response. We developed a novel ‘storyline’ approach to representing the uncertainty in the circulation response in a physically interpretable way.
A third line of attack was to better understand the physical processes and mechanisms behind atmospheric variability. This is challenging because climate noise is highly structured in space and time, with long-memory effects that can be difficult to separate from climate change. We critiqued the efficacy of standard statistical approaches using correlations of climate anomalies, which rely on an assumption of statistical stationarity. This assumption is generally invalid for the climate system. As an alternative we developed new physically based ways to diagnose causal inference, taking proper account of interannual variations in the seasonal cycle, and drawing on methods from artificial intelligence.
In addition to the achievements described above, the project involved substantial inter-disciplinary developments. These have raised the awareness of the importance of atmospheric circulation as a source of uncertainty in climate change risk, within the wider climate-change community. A notable application is the attribution of extreme weather and climate events in the context of climate change, which is a rapidly growing research area with considerable public interest. We have helped make the case for a paradigm shift towards a ‘storyline’ approach to unquantifiable aspects of climate risk, including its application to ecosystem and environmental catastrophes. We have further worked with psychologists to show that the approach to communication followed by the IPCC has significant limitations when it comes to conveying risk.