Floods and droughts pose significant threats globally, causing widespread devastation and economic losses. While risk management strategies have demonstrably reduced vulnerability in recent decades, the increasing frequency and severity of these extreme events challenge existing approaches. This research addresses the critical gap in understanding how risk management performs under unprecedented conditions – events exceeding the scale and magnitude of previously experienced ones. The study’s core objective is to analyze the effectiveness of risk management in reducing the impacts of flood and drought events, particularly those of an unprecedented nature. Understanding this is crucial for developing effective adaptation strategies in the face of projected increases in extreme weather events due to climate change. The increasing impacts of floods and droughts underscore the urgent need for a thorough investigation of the factors influencing risk reduction efforts. Previous studies have shown promising reductions in vulnerability, yet the overall impact continues to rise. This disparity necessitates a deeper examination of the interplay between hazard, exposure, vulnerability, and the role of risk management itself. The significance of this study lies in its potential to inform future risk management strategies by providing empirical evidence on the effectiveness of different approaches and identifying critical areas for improvement.

Existing literature highlights the global benefits of adaptation in reducing vulnerability to river floods. Studies have shown a declining global vulnerability to climate-related hazards, with significant reductions in both human and economic vulnerability between 1980-1989 and 2007-2016. However, the increasing impacts of floods and droughts, particularly the economic damage projected to double for floods globally and triple for droughts in Europe with a 2°C temperature increase, necessitate further research. Previous analyses, including eight case studies, identified vulnerability as a key factor in reducing flood impacts. However, a lack of empirical data has hampered comprehensive assessments of the combined effects of various risk drivers and the overall effectiveness of risk management. Understanding the combined effects of hazard, exposure, and vulnerability is crucial for effective risk management. Past studies have pointed to increasing exposure as a primary driver of increasing impacts and vulnerability reduction as key for impact reduction. However, the combined effects of these drivers and the overall effectiveness of risk management remained poorly understood due to limited empirical data.

This study employs a novel global dataset comprising 45 pairs of flood or drought events that occurred in the same area, separated by an average of 16 years. This paired-event approach allows for a comparative analysis of changes in impact, hazard, exposure, and vulnerability between the two events within each pair. The dataset includes events across diverse socioeconomic and hydroclimatic contexts from all continents. Floods and droughts are analyzed together due to similarities in some management methods (e.g., warning systems, water reservoir infrastructure), potential trade-offs in risk reduction, and the value of cross-learning among management communities. Impact is quantified using a comprehensive approach that integrates direct (fatalities, monetary damage), indirect (e.g., disruption of traffic or tourism), and intangible impacts (e.g., impact on human health or cultural heritage). Three key drivers of impact are considered: hazard, exposure, and vulnerability. These drivers are quantified using a range of indices (detailed in Supplementary Table 1), including standardized precipitation indices, the number of houses in affected areas, and risk awareness. Management shortcomings are explicitly analyzed as a point of potential improvement through better strategies and policies. To account for data variability and uncertainty, the analysis focuses on indicators of change between the first and second events in each pair. These indicators are categorized as large decreases/increases (-2/+2), small decreases/increases (-1/+1), and no change (0) (Supplementary Table 2). A quality assurance protocol is implemented to minimize subjectivity and uncertainty in indicator assignment. Correlations between changes in impact and changes in the three drivers (hazard, exposure, vulnerability) and management shortcomings are analyzed using Spearman's rank correlation coefficient.

The analysis reveals that while most paired events show decreases in management shortcomings (71%) and vulnerability (80%), reflecting societal learning from extreme events and post-event vulnerability reduction, the impact of the second event frequently increased when the second event's hazard was significantly larger than the first. Changes in flood impacts are significantly and positively correlated with changes in hazard (r = 0.64, P ≤ 0.01), exposure (r = 0.55, P ≤ 0.01), and vulnerability (r = 0.60, P ≤ 0.01). For droughts, changes in impact are significantly correlated with hazard and exposure but not vulnerability, suggesting less success in vulnerability reduction for droughts compared to floods. The five paired events with a large decrease in impact are associated with decreases or no change in all three drivers. For paired events with decreasing impact, a decrease in hazard or no change in hazard is observed in most cases (90%). In paired events with increasing impact, a larger hazard is observed in the majority of cases (85%). Notably, when the second event is significantly more hazardous than the first (+1, +2), the impact of the second event is also larger in 61% of cases, despite small vulnerability decreases in some instances. Only two cases show a decrease in impact despite an increased hazard. These success stories are characterized by improved flood risk management governance, high investments in integrated management, and strongly improved early warning and emergency response. The correlation analysis also highlights how management affects the drivers of impact. For flood events, changes in management shortcomings are significantly positively correlated with changes in vulnerability and impact. This relationship is not significant for drought events, suggesting greater difficulties in reducing vulnerability and impact through improved risk management for droughts than for floods. 'Unprecedented' events (those where the second event was much more hazardous than the first), show increased impact in nearly all cases, highlighting the challenges of managing events of previously unseen magnitudes.

The findings underscore the limitations of current risk management approaches in dealing with unprecedented flood and drought events. While vulnerability reduction plays a critical role, it is often insufficient to offset the impact of dramatically increased hazard. The study demonstrates the importance of considering all three drivers—hazard, exposure, and vulnerability—for effective impact reduction. The significant correlations between changes in these drivers and changes in impact emphasize the need for holistic strategies. The difficulty in managing unprecedented events is likely due to two factors: (1) the capacity limits of infrastructure, such as levees and reservoirs, which can fail under extreme conditions, and (2) a reactive rather than proactive approach to risk management, influenced by cognitive biases and human risk perception. The two success stories highlight the importance of effective governance, substantial investments in integrated management, and robust early warning systems in mitigating the impact of even more hazardous second events. The findings suggest that proactive, integrated strategies that address hazard, exposure, and vulnerability simultaneously are essential for enhancing resilience to both expected and unprecedented extreme hydrological events.

This study provides compelling evidence that current risk management strategies are insufficient to cope with unprecedented flood and drought events. While reductions in vulnerability are important, they are not enough when hazard increases significantly. The success stories emphasize the importance of integrated management, substantial investment, and strong governance. Future research should focus on developing proactive strategies, particularly for unprecedented events, potentially incorporating advanced forecasting techniques and more nuanced understanding of human behaviour and risk perception.

The study relies on a specific dataset of paired events, which may not fully represent the diversity of flood and drought events worldwide. Subjectivity in assigning indicators of change and the limited number of cases in which only one driver changed might affect the interpretation of results. The analysis of unprecedented events is also limited by the subjective nature of defining such events and the relatively small number of cases in the dataset.