Advances in attribution

Editorial https://doi.org/10.1038/s41558-024-02189-7 Advances in attribution Check for updates As attribution studies evolve in complexity and become more present in public discourse, care is needed to ensure that the associated uncertainties and relevant contexts remain clear. nature climate change The work of Chae Yeon Park and colleagues, also in this issue, considers the further impacts of climate change on human health, demonstrating that both absolute number and proportion of annual mortalities from fire fine particulate matter (PM2.5) attributable to climate change has risen. Attribution of human health outcomes to various climate change impacts is a relatively nascent field, which has, so far, largely focused on direct impacts of heat2. This contrasts with the vast range of direct and indirect climate-sensitive burdens on human health that remain largely unquantified (see for example the Comment article by Yin Long and colleagues in this issue discussing the health threats related to interruptions to fresh food supply). Park and colleagues highlight the uncertainty involved in such complex attribution, including substantial variances found between the three fire–vegetation models used, as well as the various roles that modelling assumptions and data availability play. Nonetheless, they show consistent trends in climate change increasing fire mortality in many key fire regions globally. Such studies are arguably key in underscoring the true and growing costs of climate inaction. Still, as attribution science naturally and necessarily evolves from attribution of meteorological events to more complex examples, further discussions are arising about the role of attribution within different fields (for example, legal and economic), and the increased need to clarify the limitations of the science and to consider it within broader social contexts . For example, recent discussions include whether extreme event attribution is4 or is not5 ready to inform loss and damage negotiations, particularly in the context of the higher data and modelling limitations in the countries that are most likely to be loss and damage fund recipients. As attribution studies get more complex, compounding uncertainties, in particular, need to be made apparent. For example, attribution of flooding involves uncertainties in rainfall modelling, in the conversion of rainfall to runoff, the role of protective infrastructure and so on, with inclusion of human outcomes adding further, unique uncertainties. These nuances are important as attribution studies not only cross scientific fields but also increasingly enter public discourse. Recent examples include reporting of the World Weather Attribution’s (a consortium of scientists who rapidly analyse extreme events) findings that climate change increased the likelihood and intensity of extreme rain in Europe in September and caused Hurricane Helene to be wetter and windier. While such broad recognition is a positive move for the field, it can also increase the risk of misinterpretation, particularly in the context of divided politics, and the real human impact of such events. While further development of the rapidly expanding field of attribution science will benefit climate action, care must be taken in communicating both the benefits and limitations of the field. Published online: 5 November 2024 References 1. Otto, F. E. Annu. Rev. Environ. Resour. 48, 813–828 (2023). 2. Carlson, C. J. et al. Detection and Attribution of Climate Change Impacts on Human Health: A Data Science Framework (The Wellcome Trust, 2024). 3. Raju, E., Boyd, E. & Otto, F. Commun. Earth Environ. 3, 1 (2022). 4. Noy, I. et al. Nat. Clim. Change 13, 1279–1281 (2023). 5. King, A. D., Grose, M. R., Kimutai, J., Pinto, I. & Harrington, L. J. Nat. Clim. Change 13, 415–417 (2023). Volume 14 | November 2024 | 1108 | 1108 Credit: US Army Photo / Alamy Stock Photo C limate change attribution involves the use of historical observations and modelling to understand how climate change has contributed to the likelihood and/or intensity of extreme weather events and their outcomes. In the past decade or so, the field has grown substantially, emerging as a broadly accepted tool in the scientific community, and simultaneously developing from attribution of single features of a meteorological event (for example, maximum temperature or rainfall intensity) to assessing more complex event types that depend on multiple variables (for example, wildfires and floods) and associated impacts (for example, human health and economic loss)1. As one example of complex event attribution, in this issue of Nature Climate Change, Chantelle Burton, Seppe Lampe and colleagues report the attribution of historical changes in global burned area (see also the associated Research Briefing). While climate change has been linked to changes in the frequency and intensity of fire weather, attribution of fire is further complicated by nonlinear interactions between fire weather and factors including vegetation growth, fuel changes, land-use changes, ignitions and human land management. Burton, Lampe and colleagues compare factual and counterfactual (without climate change) scenarios, to show that while climate change has increased global burned area by about 15.8% in the period 2003–2019, other direct human forcings have lowered burned area by 19.1%, masking the impact of climate change. 3  (...truncated)