Impact attribution methods for complex extremes; learnings from the COMPASS Use Cases

Event attribution of single hazards has developed very rapidly over the last 20 years, with a wide range of methodologies now being commonly used. However, many devastating events occur as a result of more complex extreme hazards such as those of a compounding, cascading and sequential nature. In the COMPASS project the main goal is to produce a flexible and harmonised methodological framework for the impact attribution of compound extremes. This extends capability from both single-driver hazards to more complex extremes, and climate hazard attribution to the attribution of impacts. In this work we review the lessons learnt from a wide range of methods from the COMPASS Use Cases; including compound impacts from Storm Xynthia in France, Tropical Cyclones in East Africa, consecutive Hurricanes in Honduras and sequential storms and drought-heatwave impacts in the United Kingdom. The attribution approaches used in each Use Case vary significantly from storyline to probabilistic, covering a range of regions and event types. In this review, we summarise the key learnings from these Use Cases and make recommendations on the best methods for compound impact event attribution. The results emphasise the extra value of impact attribution, compared to attribution of the hazard alone, with significant non-linearities between changes in the hazard and societal impacts.