Drivers and Impacts of Extreme Weather Events in Antarctica (ExtAnt)

ExtAnt将首次全面评估南极洲当前和未来的高影响极端天气事件及其相关风险。主要风险包括极端天气对脆弱冰架的影响，冰架的破裂会加速地面冰的流动，影响全球海平面，以及对高度专业化的南极生物多样性的影响。这个雄心勃勃的项目汇集了英国和国际领先的科学家，利用新的建模资源和方法来阐明极端事件的驱动因素。将开发新的模拟能力，以量化极端事件对冰架表面融化的影响。这些进步将使目前对极端现象的认识发生重大变化。ExtAnt的遗产将包括一个数据集，用于推进更广泛影响的研究，例如对生态系统的影响。现存研究计划围绕三个核心研究目标（RAs）和实现这些目标的具体目标制定，概述如下。RA1。量化南极极端事件主要驱动因素的相对贡献，并确定在具体观测案例中人为强迫的作用。已经和正在创建的大量新的高分辨率区域气候模式（RCM）再分析驱动的预测将与观测一起使用，以建立迄今为止最全面的南极极端天气记录。这将通过量化特定大尺度和天气现象（包括气旋和大气河流）的作用，为评估极端事件的驱动因素提供基础。新的诊断模式能力将提供前所未有的低纬度湿气来源的定量信息。将采用快速（使用现有模拟）或延迟（生成新的整体模拟）模式对极端事件归因的多方法方法评估人为驱动因素。号。解决极端事件的当今趋势和变率及其影响：特别是评估臭氧空洞、温室气体浓度和内部气候变率模式的作用。南极洲具有地球上最大的内部气候变率，因此，为了回答与极端事件发生/严重程度的趋势和变率有关的问题，我们将使用气候模式模拟的大集合（LEs）。低碳是在相同的外部强迫（如温室气体或臭氧）下多次运行气候模式（通常是10次）而产生的。这将使我们能够解决最有利于极端事件发生的内部气候变率阶段，以及在南极洲的哪些地区发生极端事件，以及对温室气体和平流层臭氧的人为强迫的响应。RA3中。量化未来南极极端事件的严重程度和频率，及其对脆弱冰架影响的相关风险，并为评估对生态系统的影响提供相关信息。将对21世纪极端事件的行为进行预估，以探索与内部变率和外部强迫相关的一系列可能的未来。将开发一个统计rcm模拟器，用于帮助将LE气候模式模拟的输出转化为与当地影响相关的尺度。这将为一个新整合的融湖模式提供输入/强迫，允许评估不同内部变率实现和不同强迫情景对冰架稳定性的影响。先进的统计方法，包括机器学习(ML) /人工智能（AI），将用于选择LE预测的代表性样本进行缩减。此外，将提供与生态系统有关的极端事件的信息，以供正在进行和未来对生态/更广泛影响的研究。

ExtAnt will provide the first comprehensive assessment of present day and future high impact extreme weather events in Antarctica, and associated risks. Key risks include impacts of extreme weather on vulnerable ice shelves, the breakup of which can speed up flow of grounded ice and affect global sea level, and on the highly specialised Antarctic biodiversity. This ambitious programme brings together leading UK and international scientists to use new modelling resources and methods to elucidate drivers of extreme events. New modelling capability will be developed to quantify impacts of extreme events on surface melt of ice shelves. These advances will bring a step change over current knowledge of extremes. ExtAnt's legacy will include a dataset for advancing research into broader impacts, for example on ecosystems.The ExtAnt programme of research has been developed around three core research aims (RAs) and specific objectives to achieve these, outlined as follows. RA1. Quantify the relative contributions of key drivers of Antarctic extreme events and determine the role of anthropogenic forcing in specific observed cases.A wealth of new high resolution regional climate model (RCM) reanalysis-driven hindcasts that have been and are being created will be used, along with observations, to establish the most comprehensive record to date of Antarctic weather extremes. This will provide a foundation for assessing drivers of extreme events by quantifying the roles of specific large-scale and synoptic phenomena, including cyclones and atmospheric rivers. New diagnostic model capabilities will provide unprecedented quantitative information on lower-latitude sources of moisture. Anthropogenic drivers will be assessed using a multi-method approach to extreme event attribution in rapid (use existing simulations) or delayed (generating new ensemble simulations) mode. RA2. Resolve present day trends and variability of extreme events and their impacts: in particular, assess the roles of the ozone hole, GHG concentrations and modes of internal climate variability.Antarctica exhibits the largest internal climate variability on earth, therefore in order answer questions relating to trends and variability of the occurrence / severity of extreme events, we will use large ensembles (LEs) of climate model simulations. LEs are created by running a climate model many times (often 10s of times) with the same external forcing (e.g. greenhouse gases (GHGs) or ozone). This will allow us to resolve the phases of internal climate variability most conducive to the occurrence of extremes and over which parts of Antarctica, alongside responses to anthropogenic forcing from GHGs and stratospheric ozone. RA3. Quantify the severity and frequency of future Antarctic extreme events, and associated risks related to their impacts on vulnerable ice shelves, and provide information relevant for assessing impacts on ecosystems.Projections of 21st century behaviour of extreme events will be developed to explore a range of possible futures associated with both internal variability and external forcing. A statistical RCM-emulator will be developed and used to help translate output from LE climate model simulations to local impact-relevant scales. This will provide input/forcing for a newly integrated melt lake model allowing impacts of different realisations of internal variability and different forcing scenarios on ice shelf stability to be assessed. Advanced statistical methods, including machine learning (ML) / artificial intelligence (AI), will be used to select a representative sample of LE projections for downscaling. Additionally, information on extremes relevant to ecosystems will be provided for ongoing and future research into ecological/broader impacts.