Abstract ID: 060
Using sub-seasonal forecasting to predict temporally compounding extreme events
Lead Author: James Carruthers
Newcastle University, United Kingdom
Keywords: Application, Compound events
Abstract: Temporally compounding extreme events refer to situations where a sequence of hazards contributes to impacts which may have been less severe had the hazards occurred independently in time. Recent examples of these events include the 2021 Haiti earthquake followed by Tropical Storm Grace, Japan’s summer of heatwaves, floods, landslides and earthquake’s in 2018, and the UK’s exceptionally stormy 2013/2014 winter. In these situations, the impacts of the primary hazard trigger a recovery period during which a secondary hazard may cause exacerbated impacts. This can be due to the afflicted community’s impaired capacity to prepare and respond and/or hazards interacting in unique and unexpected ways to produce novel multi-hazards. This work introduces a framework for using sub-seasonal forecasting in predicting temporally compounding events. Risk managers in various sectors use risk matrices to understand and communicate possible future risk based upon the likelihood and severity of different hazards. At any given time and place, the likelihood of an extreme weather event within the sub-seasonal forecasting range is limited in its certainty due to limited skill. However, in the context of an extreme event on a certain date, the value of information in the sub-seasonal forecast from that date increases. Although the likelihood of a second event may not change, the possible compounding impacts will. In these cases, the score on an impact risk matrix would move up the impact severity axis while remain constant on the likelihood axis. This work shows the utility in the framework in historical case studies, including providing prolonged warning of the Hurricane Laura and heatwave compound event Louisiana in 2021. Hurricanes have a severe impact on the capacity of individuals and emergency services to respond to heatwaves due to disrupted transport and energy infrastructure. In this example, over 90% of ECMWF model members predicted dangerous heat in the days following Hurricane Laura’s landfall with approximately 10 days lead time. This work represents a first step in highlighting the valuable information which sub-seasonal forecasts can provide on temporally compounding extreme events to risk managers.