FUTURE-FLOOD: New estimates of evolving UK flood risk for improved climate resilience

最近席卷英国和欧洲的毁灭性洪水凸显了让社会更能抵御洪水的必要性。考虑到预计洪水将随着气候变化变得更加频繁和具有破坏性，这一需求就更加迫切了。然而，目前对英国和其他地方未来洪水风险的估计是不可靠的，因为它们通常基于粗分辨率气候模型，无法捕捉导致洪水的短时间极端降雨。它们也没有捕捉到降雨事件的空间范围和时间聚类的关键变化，并且忽略了改变降雨和河流流量模式的关键物理过程。future - flood是一个雄心勃勃的项目，旨在提高我们对未来内陆洪水风险的理解，并在英国各地提供适合目的的新洪水估计。它将把国际上最先进的高分辨率气候预测与先进的洪水模拟能力结合起来。我们将利用一套新的连续的100年气候预测，提供1981-2080年12个集合成员在英国每2.2公里格框上逐小时的降雨量数据。这就像开始12个天气预报，每个预报100年。这些预估（“UKCP Local”的延伸，仅适用于三个20年周期）在空间和时间覆盖方面是独一无二的。它们将被用来获得对未来几年和几十年降雨变化的新认识，包括时间聚类的变化、事件的先决条件和空间范围。这种变化还没有被很好地理解，但可能对洪水风险至关重要。100年UKCP本地预测将用于驱动水文和洪水模型，首次提供全英国范围内对复合雨洪和河流洪水频率和严重程度变化的完整评估。UKCP将直接使用当地降雨数据，以便在模拟河流流量和洪水水位中捕获降雨模式和强度分布的复杂变化。最近的一项试点研究（由项目团队进行）表明，使用完整的UKCP局部时空变化降水场来驱动洪水模型，可能导致对未来洪水风险的估计与基于简化抬升方法的当前指导中所包含的估计完全不同。这项初步研究没有捕捉到复合效应，而且仅限于一个洪点（布里斯托尔）和两个洪点（泰特和迪菲），但证明了这里提出的全国规模研究的必要性。我们将把结果与使用标准方法和粗分辨率气候模型数据模拟的洪水进行比较，评估现有洪水预测的可靠性。额外的洪水模拟实验将使我们能够确定洪水的物理控制及其随时间的变化，包括降雨的时空变化及其与景观的相互作用。这种理解将是确定从业者在未来洪水风险评估中常用的改进的抬升方法的关键。提供超过100年的连续洪水预测是向前迈出的重要一步，它将使我们能够在气候变化的背景下解释个别观测到的洪水事件，并将结果转化为与具体政策相关的全球变暖水平的变化。我们将把新的洪水灾害信息与暴露和脆弱性的估计相结合，以估计洪水风险（例如，财产被淹没、关键基础设施受损、金钱损失）。这项估计将包括对社会经济变化的预测。我们将展示在国家层面的决策中使用这些新信息。我们还将与城市决策者共同开发一个地方规模的示范项目（最初为布里斯托尔），将新的洪水信息用于提高城市的复原力，评估适应性行动的范围和效益。

Recent devastating floods across the UK and Europe have highlighted the need to make society more resilient to flooding. This need is even more urgent given that flooding is predicted to become more frequent and destructive with climate change. However, current estimates of future flood risk in the UK and elsewhere are unreliable as they are typically based on coarse resolution climate models, which are unable to capture the short-duration rainfall extremes responsible for flooding. They also do not capture critical changes in the spatial extent and temporal clustering of rainfall events and neglect key physical processes in changing rainfall and river flow patterns. FUTURE-FLOOD is an ambitious project to advance our understanding of future inland flood risk and provide new flood estimates across the UK that are fit for purpose. It will bring together internationally state-of-the-art high resolution climate projections with advanced flood modelling capability. We will exploit a new set of continuous 100yr climate projections that provide rainfall data hour by hour, for every 2.2km grid box across the UK, for 1981-2080 for twelve ensemble members. This is like starting twelve weather forecasts and running each for 100yrs. These projections (an extension of "UKCP Local" only available for three 20yr periods) are unique in their spatial and temporal coverage. They will be exploited to gain new understanding of changes in rainfall over the coming years and decades, including changes in temporal clustering, antecedent conditions and spatial extent of events. Such changes are not well understood but are likely to be critically important for flood risk.The 100yr UKCP Local projections will be used to drive hydrological and flood models, providing a complete UK-wide assessment of changes in the frequency and severity of compound pluvial and fluvial flooding for the first time. UKCP Local rainfall data will be used directly such that complex changes to rainfall patterns and intensity distributions are captured in the simulated river flows and flood levels. A recent pilot study (carried out by the project team) showed that using the full UKCP Local space-time varying precipitation fields to drive flood models can lead to radically different estimates of future flood risks to those contained in current guidance based on simplified uplift methods. This pilot study did not capture compound effects and was limited to only one pluvial site (Bristol) and two fluvial catchments (Thet and Dyfi) but demonstrated the need for the national-scale study proposed here. We will compare results with flooding simulated using standard approaches and coarser resolution climate model data, assessing the reliability of existing flood predictions. Additional flood modelling experiments will allow us to identify the physical controls on flooding and its change through time, including the role of changes in the space-time variability of rainfall and its interactions with the landscape. This understanding will be key to identifying improved uplift approaches commonly used by practitioners for future flood risk assessment.Providing flood projections continuously over 100yrs is a major step forward and will enable us to interpret individual observed flood events in the context of climate change and translate results to changes for specific policy-relevant global warming levels. We will combine the new flood hazard information with estimates of exposure and vulnerability to estimate flood risk (e.g. properties flooded, damage to critical infrastructure, monetary loss). This estimation will include projections of socio-economic change. We will demonstrate the use of this new information in decision-making at national scale. We will also co-develop a local-scale demonstrator (initially for Bristol) with city decision-makers to take the new flood information through to improving city resilience, assessing the scope for and benefits of adaptive action.