IndicatoRs to Impacts for drought Surveillance and management (IRIS)

干旱是一种复杂、发展缓慢且代价高昂的自然灾害。检测它们的发生和追踪它们的发展可能很困难，因为它们通过水循环传播。尽管英国是一个典型的潮湿国家，但2012年和2018年的干旱对供水、农业和环境产生了重大影响。对未来气候变化的预估表明，它们将在未来几十年变得更加频繁和严重。因此，无论是现在还是在一个迅速变暖的世界，我们都迫切需要提高对干旱的抵御能力和更好的准备。干旱监测与预警（MEW）是有效干旱管理的重要组成部分，但由于定义干旱的挑战、在干旱影响非常严重之前识别干旱影响的困难以及使用干旱预警信息的广泛决策者的不同需求，使得干旱监测与预警工作变得复杂。在英国，由于英国生态与水文中心（UKCEH）领导的耗资1250万英镑的NERC干旱和水资源短缺计划（DWSP），以及UKCEH积极参与的其他联合国际研究项目，干旱研究在过去十年中取得了实质性进展。这导致了对干旱的理解和干旱新工具的开发取得了重大进展。这包括英国水资源门户网站，它提供实时水文气象数据和干旱指标。然而，像这样的干旱监测工具通常缺乏关于干旱影响的信息——尽管这是决策者采取行动所需的最重要的证据，正如我们与供水、农业、卫生、能源和环境等部门的主要利益攸关方所做的工作所强调的那样。了解MEW系统中常用的干旱指标（即描述物理干旱危害的指标，例如降雨或河流流量）与地面上看到的干旱影响之间的联系，一直是干旱科学家关注的焦点。然而，由于收集和记录干旱影响的挑战，分析通常是在大的空间尺度上进行的（例如，对整个威尔士来说），这与决策者不相关，他们倾向于在更局部的尺度上管理水，从田野到集水区尺度。在IRIS项目中，我们建议利用新的高分辨率干旱指标和干旱影响数据集在高空间分辨率上预测干旱影响，从而解决这一空间尺度问题。我们将通过三个综合工作包（WPs）来完成这项工作：WP1将侧重于从多个来源收集干旱指标和干旱影响的数据，包括英国前所未有的高分辨率的新作物产量数据。我们还将使用近年来获得的高分辨率遥感数据，如Sentinel-2，来获得干旱影响的代理（例如，植被指数和野火）。然后，WP2将使用在WP1中收集的数据来确定干旱指标与干旱影响之间的关系，并使用统计和机器学习方法构建定量描述这些关系的“影响函数”。这些关系可以用来预测潜在的干旱影响，利用现成的接近实时的指标。在第三阶段工作中，我们将与主要利益攸关方合作开展案例研究，并评估影响预测是否具备足够的技能，可用于管理干旱和减轻影响。这个项目的发现和结果有可能通过未来的资助机会扩大到一个全国性的干旱影响预测系统。基于影响的干旱预测能力将彻底改变英国管理和缓解干旱的方式，并具有向其他国家和环境转移的巨大潜力。

Droughts are complex, slow-evolving and costly natural hazards. Detecting their onset and tracking their development can be hard, as they spread through the water cycle. Although the UK is stereotypically a wet country, recent droughts in 2012 and 2018 had significant impacts on water supplies, agriculture and the environment. Projected changes to future climate suggests that they will become more frequent and severe in the coming decades. We therefore need to urgently be more resilient to and better prepared for droughts, both now and in a rapidly warming world. Drought Monitoring and Early Warning (MEW) is an important part of effective drought management, but this is complicated by the challenges in defining drought, the difficulties in identifying drought impacts before they are very severe, and the diverse needs of the wide range of decision makers that use drought MEW information. In the UK, drought research has advanced substantially over the past decade thanks to the £12.5m NERC Drought and Water Scarcity Programme (DWSP) led by the UK Centre for Ecology & Hydrology (UKCEH), as well as other allied international research projects in which UKCEH was heavily involved. This has led to significant progress in the understanding of droughts and the development of drought MEW tools. This includes the UK Water Resources Portal, which provides real-time hydro-meteorological data and drought indicators. However, drought monitoring tools like this typically lack information on drought impacts - despite the fact that this is the single most important piece of evidence required by decision makers to take actions, as has been highlighted in work we have done with key stakeholders from sectors including water supply, agriculture, health, energy and the environment. Understanding the link between drought indicators commonly used in MEW systems (i.e. that describe the physical drought hazard, e.g. in terms of rainfall or river flows) with the drought impacts seen on the ground has been the focus of drought scientists for some time. However, due to the challenges of collecting and recording drought impacts, the analysis has generally been carried out at large spatial scales (e.g. for Wales as a whole), which are not relevant to decision makers who tend to manage water at more local scales, from field to catchment scale. In the IRIS project, we propose to address this issue of spatial scale by using new high resolution drought indicators and drought impact datasets to predict drought impacts at a high spatial resolution. We will do this across three integrated Work Packages (WPs): WP1 will focus on gathering data from multiple sources of both drought indicators and drought impacts, including new crop yield data at an unprecedentedly high resolution for the UK. We will also use high resolution remote sensing data which have become available in recent years such as Sentinel-2 to derive proxies for drought impacts (e.g., vegetation indices and wildfires). WP2 will then use the data gathered in WP1 to identify the relationships between drought indicators and drought impacts, and build 'impact functions' that describe these relationships quantitatively, using statistical and machine learning approaches. These relationships can then be used to forecast potential drought impacts using indicators that are readily available in near-real-time. In WP3 we will work with key stakeholders to develop case studies and assess whether the impact forecasts have sufficient skill to be used to manage droughts and mitigate impacts. The findings and outcomes of this project have the potential to be scaled up into a nation-wide drought impact forecasting system through future funding opportunities. An impact-based drought forecasting capability would revolutionise the way droughts are managed and mitigated in the UK, and would have huge potential for transferring to other countries and environments.