Abstract:We have witnessed and experienced increasing compound extreme events resulting from simultaneous or sequential occurrence of multiple events in a changing climate. In addition to a growing demand for a clearer explanation of compound risks from a hydrological perspective, there has been a lack of attention paid to socioeconomic factors driving and impacted by these risks. Through a critical review and co-production approaches, we identified four types of compound hydrological events based on autocorrelated, multivariate, and spatiotemporal patterns. A framework to quantify compound risks based on conditional probability is offered, including an argument on the potential use of generative Artificial Intelligence (AI) algorithms for identifying emerging trends and patterns for climate change. Insights for practices are discussed, highlighting the implications for disaster risk reduction and knowledge co-production. Our argument centres on the importance of meaningfully considering the socioeconomic contexts in which compound risks may have impacts, and the need for interdisciplinary collaboration to effectively translate climate science to climate actions.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to quantify, understand, and manage the compound risks of hydrological extreme events in the context of climate change. Specifically, the paper focuses on the following aspects: 1. **Definition and Classification**: The paper first explores the definition of compound hydrological events (CHE) and proposes four types of compound hydrological events based on autocorrelation, multivariable, and spatio - temporal patterns: - **Emerging Preconditioned Events**: When weather or climate conditions exacerbate the impact of disasters, for example, heavy precipitation causes floods, followed by landslides. - **Temporally Compounding Events**: A series of disasters occur sequentially in time, for example, floods and runoff droughts that occur within six months. - **Spatially Compounding Events**: Disasters that occur simultaneously at interconnected locations and jointly produce cumulative impacts, for example, floods that occur simultaneously on the east and west coasts of the United States. - **New Combination of Multivariate Events**: Multiple driving factors or disasters converge to produce an impact, for example, the impact of El Niño on precipitation and runoff leads to floods in Texas. 2. **Quantification Methods**: The paper proposes a framework based on conditional probability to quantify compound risks, including using generative artificial intelligence algorithms to identify emerging trends and patterns. Specific methods include: - **Joint Probability Distribution**: Directly model the joint probability distribution of multivariable disasters \( P(H_1 \leq h_1, \ldots, H_n \leq h_n) \), where \( \mathbf{H}=(H_1, \ldots, H_n) \) represents multivariable disasters. - **Conditional Forecast Joint Probability Distribution**: When the modulators \( \mathbf{m} \), driving factors \( \mathbf{d} \), and climate change variables \( \mathbf{c} \) are known, estimate the conditional forecast joint probability distribution \( P(H_1 \leq h_1, \ldots, H_n \leq h_n|\mathbf{d}, \mathbf{m}, \mathbf{c}) \). - **Distribution Regression Model**: Predict the conditional cumulative distribution function (CDF) \( F(\mathbf{h}|\mathbf{d}, \mathbf{m}, \mathbf{c}) = M_F(\mathbf{d}, \mathbf{m}, \mathbf{c}, \theta) \). - **Semi - Parametric Model**: Predict the statistical function of the conditional probability distribution \( T(F(\mathbf{h}|\mathbf{d}, \mathbf{m}, \mathbf{c})) = M_T(\mathbf{d}, \mathbf{m}, \mathbf{c}, \theta) \). 3. **Socio - economic Factors**: The paper emphasizes the importance of considering socio - economic factors in evaluating and managing compound risks, including aspects such as population structure, housing conditions, resource access, social networks, institutional capacity, and infrastructure readiness. These factors can amplify the impact of compound events and weaken the social foundation for disaster response. 4. **Future Research Directions**: The paper discusses future research directions, including developing new statistical theories to explain the emergence of compound risks, and using artificial intelligence and statistical downscaling techniques to improve the transparency and accuracy of models. In summary, this paper aims to comprehensively understand and manage compound hydrological risks in the context of climate change through an integrated approach from multiple perspectives.

Unravelling compound risks of hydrological extremes in a changing climate: Typology, methods and futures