Urban flood risk increases under rapid urbanization and climate change. Thus, it becomes crucial to assess current and future risk and potential adaptation strategies to minimize the consequences for society, ecology and economy, especially in the Global South where urbanization and vulnerabilities are particularly high. However, current assessment tools oftentimes struggle to perform integrated assessments of flood risk due to reasons like data scarcity, complexity of cities or the integration of different domains. Hence, current approaches usually apply a reduced perspective, e.g. in terms of the urban extent covered or the domains included. Here we propose an approach using urban structure types in combination with Bayesian networks to represent different environmental and socio-economic conditions throughout a city. The approach facilitates integrative flood risk assessments and allows to address questions of uncertainty, variability and explainability in complex and data-scare urban areas. The implementation of this new approach is presented and discussed. Results from our pilot in Mumbai, show that the approach is suitable for scenario evaluation in data-scarce contexts. The flexibility offered by the approach makes it relevant for policy and urban planning since different key drivers of urban flood risk can be integrated in assessments of adaptation strategies and decision-making.

中文翻译：

---

将城市结构类型和贝叶斯网络建模联系起来，对数据稀缺的特大城市进行综合洪水风险评估


快速城市化和气候变化导致城市洪水风险增加。因此，评估当前和未来的风险以及潜在的适应策略至关重要，以尽量减少对社会、生态和经济的影响，特别是在城市化和脆弱性特别高的南半球国家。然而，由于数据稀缺、城市复杂性或不同领域的整合等原因，当前的评估工具往往难以对洪水风险进行综合评估。因此，当前的方法通常采用简化的视角，例如就所覆盖的城市范围或所包括的领域而言。在这里，我们提出了一种使用城市结构类型与贝叶斯网络相结合的方法来代表整个城市的不同环境和社会经济条件。该方法有利于综合洪水风险评估，并可以解决复杂和数据匮乏的城市地区的不确定性、可变性和可解释性问题。介绍并讨论了这种新方法的实施。我们在孟买的试点结果表明，该方法适用于数据稀缺环境中的场景评估。该方法提供的灵活性使其与政策和城市规划相关，因为城市洪水风险的不同关键驱动因素可以纳入适应策略和决策的评估中。