EXTREME EVENTS: ARTIFICIAL INTELLIGENCE FOR DETECTION AND ATTRIBUTION

Projektinformationen

XAIDA

ID Finanzhilfevereinbarung: 101003469

DOI

10.3030/101003469

Projekt abgeschlossen

EK-Unterschriftsdatum 27 Mai 2021

Startdatum 1 September 2021

Enddatum 28 Februar 2026

Finanziert unter

• SOCIETAL CHALLENGES - Climate action, Environment, Resource Efficiency and Raw Materials

Gesamtkosten

€ 5 999 241,75

EU-Beitrag

€ 5 999 241,75

Koordiniert durch

CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRS
France

Periodic Reporting for period 2 - XAIDA (EXTREME EVENTS: ARTIFICIAL INTELLIGENCE FOR DETECTION AND ATTRIBUTION)

Berichtszeitraum: 2023-03-01 bis 2024-08-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Climate alters the frequency or intensity of several types of weather extremes around the globe. Extreme heat waves and heavy rainfall are increasing in intensity on a global scale, trends which will continue with future global warming. Often, extreme events thus provide insights on future climate conditions and their implications, but not all extremes are harbingers of the future. In order to be useful for adaptation, a causal link between events and human influence on climate must be scientifically established or refuted. XAIDA aims to fill a number of knowledge gaps in making links between different types of extreme events and climate change. The main goals are to:

- Characterize, detect, attribute and project extreme events with a novel impact-based approach,
- Assess their underlying causal pathways and physical drivers,
- Assess new event types, and develop storylines of yet unseen, but physically plausible events for present and future climates,
- Provide new tools for model assessment to investigate causes of disagreements between models and observations.

XAIDA is designed to provide new methodologies, tools, and datasets and to demonstrate their use in a number of case studies. This is expected to lead to improved attribution and projection services for extreme events. During the first 18 months of the project, the focus has been on a number of actions to prepare the analysis of new extreme events using innovative frameworks.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

We have achieved the expected goals for the first reporting period:

1. Identifying 8 stakeholder groups

We have identified the key stakeholder groups for which co-design of questions would be carried out (MS4):
- Humanitarian organizations and disaster risk reduction groups
- Insurance, or reinsurance
- Industry (among others water and energy)
- Youth climate groups
- Teachers & educators
- Journalists and knowledge brokers
- Legal professionals
- Weather and climate services

2. Shaping key extreme questions and defining 6 central case studies

Interactions with stakeholders and internal discussions in different groups helped to prioritize six central case studies of the highest relevance:
- Pacific-North American Heatwave in June 2021
- Cold air outbreak in Europe in spring 2021
- Extreme rainfall from convective cells in the Mediterranean
- European Drought/Heatwave 2022, including wildfire risks
- Compound winter cold spell and wind-drought, Tropical Cyclone Irma (Aug-Sept 2017)

3. Designing new tools and approaches for the analysis of extreme events

Several new tools have been developed, to strengthen the use of AI techniques. An important development is the extreme events detection toolbox developed by WP3, a generic tool including a number of supervised learning methods to classify events based on meteorological variables. WP4 has developed several new elements of the suite of tools (in the Tigramite python package) for assessing causality chains in the development of extreme events.

4. Preparing events databases and review of events’ impact data bases, vulnerability & exposure information sources

WP2, WP3 and WP8 have allowed various sources of information to be used in extreme events studies and, further, in applications: an overview of global, publicly available databases that indicate exposure and vulnerability to extreme weather events (WP2); a comprehensive database on climate and impacts by combining information and building upon existing databases (WP3); a strategy to select severe convective events and a list of high precipitation events and floods as well hurricanes, tropical storms and tropical-like cyclones(WP8). WP3 developed an AI-based methodology for predicting maps of ecosystems impacts from remote sensing.

5. Launching the communication on extremes

Communication on extremes was done through the project web site (https://xaida.eu) and a series of 6 briefs on extreme events . These briefs have been actively disseminated via social media (particularly Twitter) and have in total received 3470+ unique views. This communication will progressively be enriched with the new methods and their applications.

6. Developing several actions for internal communication and integration across the project

For young scientists and students, a summer school in Trieste took place in June 2022. In addition, we are organizing a continuing series of monthly internal webinars where scientific progress is actively shared. The Trieste summer school was a great success (148 participants, 20 lecturers).

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

The project has produced 21 scientific research papers so far. Several major results, clearly beyond the state of the art, were obtained by the XAIDA teams related to recent extreme events including the central case studies and framework development. For instance, frameworks and methodologies for complex events such as compound events were set up and applied to different cases (Bevacqua, 2021, 2022; Zscheichler et al., 2021, Li et al., 2022). The freak event of 2021 in British Columbia was analyzed and shown to result from a conjunction of factors including dynamics, diabatic heating and dry soils (Schumacher et al., 2022). Attribution analyses of two central case studies, the Summer 2022 heat and drought case and the April 2021 frost demonstrated the role of human activities in such events (Schumacher et al., 2022; Vautard et al., 2023).

An emerging topic is the role of atmospheric circulation changes in altering extreme events. The compounding role of atmospheric circulation changes was highlighted in recent extreme events changes (cyclones and summer heat) (Faranda et al., 2022, 2023), and boreal forest fires (Scholten et al., 2022).

The development of advanced statistical methods to detect or reconstruct cyclones is ongoing (Gardoll et al., 2022; Faranda et al., 2023, in review). Such original advanced statistical methods were also applied to precipitation detection and attribution questions (Egli et al., 2022; de Vries et al., 2023).

For the coming period integration and application of these developments will be a priority. We identify as priorities for the second phase of the project:

- Pursuing application of the different developments to central case studies using new methods. Causality and boosting methodologies, for building storylines, are now ready to be applied to long-lasting events (cold spells, droughts), following the concerns of stakeholders;

- Strengthening the links to stakeholder groups in order to collect more framing questions on groups where a rising interest exist (eg. insurance, lawyers).

- Identifying the potential to feed current climate services (eg. through C3S), and development of new climate services methodologies. Storyline methodologies bear a high potential for applications. Also, triggered by the establishment of a Loss & Damage Fund at COP27, we will explore the question of how event attribution can provide information for loss and damage. For this we will develop a framework combining several methods (from full storylines to fully probabilistic), to provide most useful information.

General illustration for XAIDA

Logo of the Project