Determination of Tropical Oxidising Capacity through model calibration (DeTOX)

The oxidising capacity of the atmosphere determines the lifetime of major air pollutants and short-lived greenhouse gases such as methane, and is dominated by tropical regions that support active photochemistry. It is critical to understand the processes that govern tropical oxidation and how they vary over space and time to provide reliable estimates of the future evolution of key pollutants and their impacts on climate. Observations are unable to provide direct information on the processes governing oxidation at a global scale, or how they might change in future, and hence policy decisions on Net Zero targets and climate change mitigation depend on atmospheric models. However, current models show substantial biases in their simulation of oxidising capacity, under- or overestimating the lifetime of methane by as much as 40%, highlighting major gaps in current understanding. This model uncertainty has remained persistently high for more than two decades. The DeTOX project addresses these issues by taking an entirely new approach. We will provide the first formal constraints on oxidation in the tropics, bringing together the wealth of recent observations with global chemistry-transport models using innovative statistical and machine learning approaches. We will calibrate these models at a process level for the first time, providing new insight into the processes governing oxidation and reducing uncertainty in future projections. We will achieve this through rigorous uncertainty quantification and emulation of three independent models, using existing observations to constrain uncertainty in the processes governing the hydroxyl radical, OH, which dominates oxidation in the tropics. The lifetime of methane, the second most important greenhouse gas, is a key focus, and the project will enable improved estimates of the main sink of this gas and more robust projections of its future abundance. Our approach will provide improved quantification of the current oxidising capacity of the tropics and clear identification at a process level of how and why it is changing. Observational constraints will reduce uncertainty in estimates of the lifetime of key pollutants at both regional and global scales. New process-based understanding will allow us to identify the drivers of change and fully-coupled climate projections will permit improved assessment of future changes in OH and climate-relevant pollutants such as methane. Our approach will also highlight gaps in current understanding that require further investigation through new observations or targetted model development. We will use the emulators developed here with ongoing satellite measurements to provide a new dynamic observation-based estimate of oxidising capacity and its changes along with attribution of changes to the processes driving them. This project is based on a new collaboration between atmospheric scientists and experts in machine learning. In addition to improved insight into oxidation processes, a key outcome will be development of new statistical methodologies that will benefit the wider atmospheric chemistry and climate communities. This will be of particular value to data assimilation, reanalysis and inversion communities, allowing more reliable estimates of emissions and their variations. We will work with our international partners to encourage a paradigm shift in the way observations are used with models from evaluation towards formal constraint. Through our leadership roles in major international assessment projects such as IPCC, CMIP, TOAR and HTAP we will ensure that improved process understanding benefits future assessments of climate and air quality, providing more robust evidence in support of future environmental policy.