Paul Palmer (Professor at the University of Edinburgh)

Satellite observations reveal that Earth’s carbon cycle is shifting faster than expected, with abrupt declines in carbon uptake and surges in methane emissions across regions such as the Amazon and Eastern Africa. These changes, often linked to drought, heat extremes, and anomalous rainfall, highlight the vulnerability of ecosystems to climate stress. I will present recent work using atmospheric composition measurements and inverse modelling to diagnose these shifts, including how changes in atmospheric methane reflect variations in the hydroxyl radical, and how Amazon drought alters emissions of reduced‑carbon species such as isoprene, with consequences for regional atmospheric chemistry. I will also showcase new approaches that combine satellite observations of nitrogen dioxide and CO2 to infer fossil‑fuel CO2 emissions, enabled by a robust machine‑learning representation of nitrogen oxide chemistry, and describe our efforts to design and build the next generation of satellites to address current observational gaps. I will conclude by outlining a forward‑looking vision for physics‑informed AI methods. As satellite missions generate orders of magnitude more data than current models can assimilate or interpret, AI‑based approaches offer a way to learn governing relationships directly from observations, enabling more adaptive, data‑rich representations of the carbon cycle as the Earth system continues to change.

Paul Palmer is a Professor at the University of Edinburgh, where he leads a cross‑disciplinary group studying Earth’s atmosphere and the atmospheres of other planets using data, models, and theory. He is the Science Director of the NERC National Centre for Earth Observation, a science‑team member on NASA and Japanese greenhouse‑gas satellite missions, and the UK co‑lead of the French–UK MicroCarb mission. He also serves on the European Commission’s CO2 Task Force, providing scientific input to support Europe’s response to the Paris Agreement.