ROBUST - Enabling better management of UK multi-hazard risk

英国受到多种与天气有关的灾害的严重影响，特别是洪水和极端大风。2015-16年的冬天说明了英国对这种极端天气的脆弱性，估计成本为13 - 19亿英镑（EA，2018）。 政府机构的高调报告现在认识到，当洪水和极端风同时发生时，风险可能会加剧，但目前大多数英国主要部门（例如保险，关键基础设施）认为大多数高影响天气风险是独立的现象。从单一风暴和季节的研究进展，我是第一个证明英国冬季洪水和极端风之间存在系统性联系的人（Hillier et al，2015），即使洪水和破坏性风通常发生在不同的风暴中（De Luca & Hillier，2017）。这种关系尚不清楚，但它可能反映了这种风暴与急流相互作用的方式（Hillier &狄克逊，2020）。我最近的工作是基于估计损失，并使用保险公司熟悉的图表（例如概率与损失）来传达其发现，使其含义比标准科学论文中报告的输出更容易获得。因此，它被认为是重要的，但还需要采取另一个步骤来改变金融实践;引用伦敦劳合社的一位高级风险经理的话，一位对这些危险有深入了解的人--“这是很好的工作，但我如何使用它？".鲁棒性是对这个问题的回答。 具体来说，英国央行（BoE）监管英国金融服务的部门审慎监管局（Prudential Regulation Authority）也问了我同样的问题。他们关注的是确保英国的金融服务公司，如保险公司，能够抵御自然灾害的影响。在这个角色中，他们需要决定是否要求保险公司持有更多的资本，以支付比预期更大的自然灾害的账单。然而，持有更多的资本成本保险公司的钱，所以英国央行不想强加这一要求不必要的。为了应对这一困境，我与英格兰银行和Sayers LLP（一家领先的洪水风险建模公司）共同设计了ROBUST，以了解这种观察到的严重内陆洪水和极端风灾在潮湿多风的年份共同发生的趋势的财务影响。 ROBUST的主要目的是使银行能够满足其就可能需要什么监管发表声明的需要。我们的试点研究（Hillier，BoE，Aon，CatInsight）表明，在财务损失的统计模型中简单地应用科学观察到的相关性无法得出明确的答案-所需的假设会产生很大的不确定性。需要在气候模型（例如英国气象局的UKCP 18）的驱动下，对洪水和极端风力进行联合物理建模。然而，这不是一个简单的努力，然后英国央行需要将其从危险（例如洪水）扩展到损失，然后再扩展到损失的影响。也许这就是为什么，尽管2015年像德斯蒙德这样的风暴令人印象深刻，但英国洪水和极端风之间的关联在保险“待办事项”列表中徘徊了十年而没有完成。 为了解决英国央行的问题，ROBUST采用务实的知识交流方法，与包括英国央行，Sayers LLP和英国气象局在内的组织合作。它开发了几个现有的UKRI资助的洪水（AquaCAT - Sayers）和风灾（WISC - Met Office）项目的输出，以改变我们如何理解它们的共同发生，然后明确进展到适合于英国金融稳定决策的指标。我的技能和经验非常适合这项任务。 此外，为了产生更广泛的影响，我将通过与保险公司合作，并与公用事业公司和交通基础设施提供商合作，更好地管理英国多种灾害风险，促进跨部门学习。

The UK is severely impacted by multiple weather-related hazards, particularly flooding and extreme winds. The winter of 2015-16 illustrates the UK's vulnerability to such weather extremes, with estimated costs of £1.3-1.9 billion (EA, 2018). High-profile reports by government agencies now recognise that risk might be exacerbated when flooding and extreme wind co-occur, but at present most major UK sectors (e.g. insurance, critical infrastructure) consider most high-impact weather risks as independent phenomena. Progressing from studies of single storms and seasons, I was the first to demonstrate a systematic link in the UK between wintertime flooding and extreme winds (Hillier et al, 2015), even if floods and damaging winds typically occur in different storms (De Luca & Hillier, 2017). This relationship is not well understood, but it may reflect the way in which such storms interact with the jet stream (Hillier & Dixon, 2020). My recent work is based around estimating losses and communicates its findings using plots familiar to insurers (e.g. of probability vs loss), making implications more accessible than the outputs reported in standard scientific papers. Consequently, it is recognized as important, but another step is needed to alter financial practice; to quote a senior risk manager at Lloyd's of London, a person with in-depth knowledge of these hazards - "It's good work, but how do I use it?". ROBUST is a response to that question. Specifically, the Prudential Regulation Authority, a division of the Bank of England (BoE) that regulates UK financial services asked me exactly the same thing. Their concern is to ensure that the UK's financial service companies, such as insurers, are resilient to the impacts of natural disasters. In this role they need to decide whether or not to require insurers to hold more capital to pay the bill for larger than expected natural disasters. However, holding more capital costs insurers money, so the BoE do not want to impose this requirement unnecessarily. In response to this dilemma, I co-designed ROBUST with the Bank of England and Sayers LLP (a leading flood risk modeller) to understand the financial implications of this observed tendency for severe inland flooding and extreme wind damage to co-occur in wet and windy years. ROBUST's main aim is to enable the Bank to fulfil its need to deliver a statement on what regulation might be required. Our pilot study (Hillier, BoE, Aon, CatInsight) demonstrated that simply applying the scientifically observed correlation within statistical models of financial loss cannot reach a definitive answer - the assumptions needed create large uncertainties. Joint physical modelling of flooding and wind extremes is needed, driven by a climate model (e.g. the Met Office's UKCP18). However, this is not a simple endeavour, and then the BoE needs to extend it from hazard (e.g. flooding), to loss, and then to the implications of that loss. Perhaps this is why, despite being spectacularly illustrated by storms like Desmond in 2015, the association between flooding and extreme winds in the UK has lingered for a decade at the top of the insurance 'to do' list without being done. To solve the BoE's problem, ROBUST adopts a pragmatic, knowledge exchange approach, partnering with organisations including the BoE, Sayers LLP and the Met Office. It develops outputs from several existing UKRI funded projects on flooding (AquaCAT - Sayers) and wind damage (WISC - Met Office) to create a step-change in how we understand their co-occurrence, and then explicitly progresses to metrics suitable for decisions about the UK's financial stability. My skills and experience are ideal for this task. Moreover, to create wider impacts, I will promote cross-sector learning by taking the lessons about better management of UK multi-hazard risk gained with insurers and engaging with utility companies and transport infrastructure providers.