ITS4.4/NH0.4 | Science as a service: delivering actionable research on climate impacts
Science as a service: delivering actionable research on climate impacts
Convener: Claire Burke | Co-conveners: James BrennanECSECS, Laura RamsamyECSECS, Nicholas Leach
Orals
| Thu, 27 Apr, 08:30–10:10 (CEST)
 
Room N1
Posters on site
| Attendance Thu, 27 Apr, 16:15–18:00 (CEST)
 
Hall X4
Orals |
Thu, 08:30
Thu, 16:15
Both society and industry are becoming increasingly aware of the physical risks and potential knock-on impacts posed by climate change. This awareness is leading to an increased demand for specific and actionable information about such risks across a wide range of sectors and domains. Awareness of and demand for data on climate-related impacts across multiple different hazards and across large geographical areas is also rising - in many nations this is driven by new and upcoming legislation requiring businesses or governments to understand and be prepared for any and all climate risk. Although researchers in climate and geosciences are best-informed to provide expertise, traditional research outputs such as publications and corresponding data are not necessarily useable by non-experts within these domains. In this session, we explore how focusing on the needs of non-expert users and decision makers changes the way in which research is carried out, disseminated, and scaled.

We welcome abstracts across all natural hazard types and climate impacts on a broad range of themes, including: how to work effectively with stakeholders and other users; how user-engagement and data-delivery requirements change the way
science goals are set and met; and how to scale climate risk research to provide information beyond localised case-studies.

Orals: Thu, 27 Apr | Room N1

Chairpersons: Claire Burke, James Brennan, Laura Ramsamy
08:30–08:40
|
EGU23-17453
|
ECS
|
On-site presentation
Sruti Modekurty, Arika Virapongse, Rupanwita Gupta, Zachary J. Robbins, Jonathan Blythe, and Ruth E. Duerr

Community resilience increases a place-based community’s capacity to respond and adapt to life-changing environmental dynamics like climate change and natural disasters. Timely access to environmental data is an important factor for community resilience. Most Earth science information is created for a particular science community for a specific scientific purpose, without much thought to who else could benefit from it and how they might use it. New approaches are needed to facilitate better data production and integration for community use.

In this session, we present the findings of a paper published by ESIP’s (Earth Science Information Partners) Community Resilience Cluster. As a convening space for over 150 member organizations across different sectors, ESIP’s biannual meetings, conference calls, and topic-driven clusters provided the infrastructure and expertise to support the Community Resilience cluster’s examination of the role of Earth science data for community resilience. This presentation highlights the challenges communities face when applying Earth science data to their efforts:

• Inequity in the scientific process,

• Gaps in data ethics and governance,

• A mismatch of scale and focus, and

• Lack of actionable information for communities.

Recommendations are made as starting points to address the challenges, along with examples of good practices from across the Earth science community. Given ESIP’s data stewardship efforts with large organizations and across domains, the recommendations are applicable at scale. We offer actionable steps for the Earth science community to help them produce data to better support community resilience.

How to cite: Modekurty, S., Virapongse, A., Gupta, R., Robbins, Z. J., Blythe, J., and Duerr, R. E.: Next Steps for Earth Science Contributions to Community Resilience, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-17453, https://doi.org/10.5194/egusphere-egu23-17453, 2023.

08:40–08:50
|
EGU23-11023
|
On-site presentation
Tina-Simone Neset, Katerina Vrotsou, Carlo Navarra, Fredrik Schück, Clara Greve Villaro, Magnus Mateo Edström, and Caroline Rydholm

In October 2021, the Swedish Meteorological and Hydrological Institute launched a novel national system for impact-based weather warnings, moving from the traditional format for meteorological, hydrological, and oceanographic warnings towards an assessment process that includes collaboration and consultation with regional stakeholders on the impacts that certain weather events would have for a specific geographic area and time frame. As part of this new system, local and regional administrative efforts are made to create assessment-support documentation drawing on local knowledge and providing support ahead of and during extreme weather events.

We present initial results from the ongoing research project ‘AI4ClimateAdaptation’ (https://liu.se/en/research/ai4climateadaptation), which explores the potential of employing AI-based image and text analysis to support the process and evaluate the precision of impact-based weather warnings. The project collects image and text data appropriate for subsequent use in AI-based analysis from citizen science campaigns and social media. The presentation focuses on the concept of integrating AI-based text and image analysis with the processes of the warning system, as well as the barriers and enablers that are identified by local, regional, and national stakeholders related to the role of AI in weather warning systems. We further discuss to what extent data and knowledge on historical extreme weather events can be integrated with local and regional climate adaptation efforts, and whether these efforts could bridge the divide between long-term adaptation strategies and short-term response measures related to extreme weather events. The results of this study are expected to contribute to the national system for impact-based weather warnings and to increase resilience to extreme climate-related weather events.

How to cite: Neset, T.-S., Vrotsou, K., Navarra, C., Schück, F., Greve Villaro, C., Mateo Edström, M., and Rydholm, C.: AI for Climate Adaptation?, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-11023, https://doi.org/10.5194/egusphere-egu23-11023, 2023.

08:50–09:00
|
EGU23-8097
|
ECS
|
On-site presentation
Sally Woodhouse, Claire Burke, Nick Leach, James Brennan, Graham Reveley, Laura Ramsamy, and Hamish Mitchell

Increasingly the financial sector is interested in understanding their risk to the impacts of climate change. This is driven both by governmental regulation that requires financial services to declare their risks due to climate change, as well as a desire to mitigate risks to profits that climate change poses.

To generate useful and accurate risks assessments users need access to high quality data of the projected changes to hazard due to climate change. However, there is typically a gap between scientific research and what our clients need to understand their risk. Many of the most damaging hazards, such as flooding and subsidence, are not directly modelled by climate models and require specialist hazard knowledge and well as climate data to assess. Scientific studies often focus on large scale changes or small regional studies, whereas clients need consistent high-resolution data across multiple regions. Additionally, a risk portfolio covers a wide range of climate related hazards, which all must be considered when understanding and attempting to mitigate risk. Users will often not have the inhouse knowledge to use data generated by the scientific community directly or the expertise to assess how this relates to the risks posed by different hazards. Therefore, the financial sector is turning to external data providers for this information, such as Climate X.

This talk will cover how at Climate X we make reliable and robust risk assessments of climate hazards that are presented in a way that is usable and useful for the financial sector as well as various other decision makers. The focus will be on how we use open-source climate model data to generate our heat risk metric. This will cover the definition of the metric, how it is calculated and how we how we present the data to users including accuracy and uncertainty. I will also present overview of the other hazards that we provide and the need for an interdisciplinary team to cover the broad range of physical hazards related to climate change.

How to cite: Woodhouse, S., Burke, C., Leach, N., Brennan, J., Reveley, G., Ramsamy, L., and Mitchell, H.: Climate X: Making climate risk data useful and usable for the financial sector, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-8097, https://doi.org/10.5194/egusphere-egu23-8097, 2023.

09:00–09:10
|
EGU23-6880
|
On-site presentation
|
Kristina Blennow and Johannes Persson

Effective support for people´s responses to climate change requires knowledge on the gap between physical climate change science and practices where the responses are realized. Studies have shown that individuals´ strong belief in local impacts of climate change is an important driver of climate change response (e.g. Blennow et al. 2012). Arguably this belief can be fortified by the belief that one has experienced the local impacts of climate change. However, a recent study shows that while responses to climate change correlate positively with the strength of belief that one has experienced negative local impacts of climate change, experience of positive local climate change impacts can either promote or inhibit the response (Blennow and Persson 2021). If the intention is adaptation to the impacts of climate change, positive experiences of climate change promote the response but if the intention is climate change mitigation, experience of positive impacts of climate change inhibit the response.

While strong belief in the local impacts of climate change is a prerequisite of climate change response, for adaptation, the agent also needs detailed knowledge of the causal links between climate change and the negative and positive values of expected climate change related impacts (Blennow et al. 2020). Decision-making in favor of adaptation to climate change generally increases with the absolute value of the net of positive and negative expected impacts in the absence of ‘tipping point’ behavior (Persson et al. 2020; Blennow et al. 2020). Tipping point behaviour occurs when adaptation is not pursued in spite of the strongly negative or positive net value of expected climate change impacts. For mitigation, moreover, it is important that the net value of expected impacts is negative and not positive (Blennow and Persson 2021). We discuss the implications of the results for policies aiming at supporting responses to climate change, such as communications that help the receiver subjectively attribute the causes of an event to climate change.

 

References

Blennow, K. Persson, J., 2021. To Mitigate or Adapt? Explaining Why Citizens Responding to Climate Change Favour the Former. Land, 10, 240. https://doi.org/10.3390/land10030240

Blennow, K., Persson, J., Tomé, M., & Hanewinkel, M., 2012. Climate change: believing and seeing implies adapting. PLOS ONE, 7(11):e50181. http://dx.plos.org/10.1371/journal.pone.0050182

Blennow, K. Persson, J., Gonçalves, L.M.S., Borys, A., Dutcă, I., Hynynen, J., Janeczko, E., Lyubenova, M., Merganič, J., Merganiová, K., Peltoniemi, M., Petr, M., Reboredo, F., Vacchiano, G., Reyer, C.P.O., 2020. The role of beliefs, expectations and values in decision-making favoring climate change adaptation – implications for communications with European forest professionals. Environmental Research Letters,15: 114061.  /doi.org/10.1088/1748-9326/abc2fa

Persson, J., Blennow, K., Gonçalves, L.M.S., Borys, A., Dutca, I., Hynynen, J., Janeczko, E., Lyubenova, M., Martel, S., Merganic, J., Merganicova, K., Peltoniemi, M., Petr, M., Reboredo, F., Vacchiano, G., Reyer, C.P.O., 2020. No polarization – expected values of climate change impacts among European forest professionals and scientists. Sustainability, 12, 2659; doi:10.3390/su12072659

How to cite: Blennow, K. and Persson, J.: The role of beliefs, expectations and values for decision-making in response to climate change, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-6880, https://doi.org/10.5194/egusphere-egu23-6880, 2023.

09:10–09:20
|
EGU23-14166
|
ECS
|
Highlight
|
On-site presentation
Peter Pfleiderer, Jana Sillmann, Robin Lamboll, Joeri Rogelj, and Carl-Friedrich Schleussner

Climate impacts have been studied intensively and our understanding of changes in climate impacts due to anthropogenic activity is impressive (see IPCC AR6). There is, however, a gap between the physical understanding of changes in climate impacts and availability of information that could directly be used by adaptation planners. We argue that this gap is to a large extent a result of the usual modeling chain that is based on a handful of representative emission scenarios.

Most climate change studies take a small, predefined set of emission scenarios (SSP2-45, SSP1-26, SSP5-85 etc.) and calculate the global and regional climate impacts resulting from these. Focusing on a limited set of emission scenarios allows us to compare results from different modeling groups and lets us run detailed climate models on each scenario. However, this modeling approach does not align with relevant research questions such as: “How much can be emitted to avoid a certain impact?” Or “what are the emission constraints to limit the probability of experiencing a certain event until 2050 to 10%?”

The presented reversal of the impact chain would help to answer these questions. The idea is to start from a clearly defined impact and evolve uncertainties backwards into the emission space. Doing so, we take the perspective of practitioners who know very well what impacts are of relevance and would like to know how these impacts are related to greenhouse gas emissions.

How to cite: Pfleiderer, P., Sillmann, J., Lamboll, R., Rogelj, J., and Schleussner, C.-F.: Reversing the impact chain, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-14166, https://doi.org/10.5194/egusphere-egu23-14166, 2023.

09:20–09:30
|
EGU23-13145
|
ECS
|
On-site presentation
Lydia Cumiskey, Denise McCullagh, Pia-Johanna Schweizer, and Sukaina Bharwani

Managing flood risk and adapting to climate change is complex where multiple actors need to work together across sectoral and disciplinary boundaries to capture synergies and manage trade-offs. A selection of governance mechanisms were found to influence actors’ capacity to work in partnership, break down silos and unlock opportunities.

Results from research conducted within the SYSTEM-RISK project identifies boundary spanning roles as governance mechanisms facilitating integrated flood risk management in England and Serbia (Cumiskey, 2020). Among other characteristics, the ‘reticulist’ was found to utlise networks and diplomacy to access funding, ‘entrepreneurs’ acted creatively to capture funding and test the flexibility of rules, ‘interpreters’ built interpersonal relationships and interpreted different professional languages, ‘organisers’ managed actor partnerships and ‘specialists’ were willing to engage and try new approaches. The availability of rules and resources influenced capacities to hire, train and sustain such boundary spanning staff.  Results highlighting the dynamic interdependencies between such roles and the governance system will be shared.

Place-based adaptation partnerships were found as another governance mechanism, strengthening collaboration, knowledge exchange and joint action across boundaries. The Climate Adaptation Partnership Framework1 was developed through the TalX project (Transboundary Adaptation Learning Exchange) to collate learning from applications in Ireland, Northern Ireland, Scotland, England and Wales and provide guidance for stakeholders interested in implementing such partnerships.  

The RISK-TANDEM framework is being developed within the DIRECTED project (Horizon Europe, 2022 - 2026) to enhance risk governance, knowledge co-production and interoperability across data, models and tools to enable disaster resilience in four Real World Lab regions. An initial version of the framework, which builds upon the existing Tandem Framework2 (among others) will be shared along with plans for implementation.   

The role of such governance mechanisms in integrating research, innovation and science in a collaborative way will be introduced, while opening the discussion on how to improve the application of such mechanisms to facilitate future engaged research.

 

Cumiskey, L. (2020). Embracing boundary spanning roles in Flood Risk Management. PhD Research Briefing Note 2. Middlesex University. Available at: https://eprints.mdx.ac.uk/30418/

1 Climate Adaptation Partnership Framework. Available at: https://talx2020.github.io/

2 The Tandem framework: a holistic approach to co-designing climate services. Available at: https://www.weadapt.org/knowledge-base/climate-services/the-tandem-framework

How to cite: Cumiskey, L., McCullagh, D., Schweizer, P.-J., and Bharwani, S.: Implications of governance mechanisms for spanning boundaries and managing risk , EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-13145, https://doi.org/10.5194/egusphere-egu23-13145, 2023.

09:30–09:40
|
EGU23-14730
|
Virtual presentation
What does the public talk about climate change, climate-health stressors, and citizen science?
(withdrawn)
Milan Kalas, Joy Ommer, Tommaso Sabbatini, Denis Kolokol, and Sasa Vranic
09:40–09:50
|
EGU23-11292
|
On-site presentation
Nicu Constantin Tudose, Mirabela Marin, Sorin Cheval, and Cezar Ungurean

The pressure on natural resources including water, energy and land is continuously growing through changes in climate and land use. Representatives of academia, industry, governments and society need to join forces in order to develop new pathways towards sustainable natural resource use and management. Such pathways start from the basic idea that natural resources are finite and interlinked and that human activities can affect these resources and links, with partly irreversible effects. We combine the water−energy−land nexus and the climate services concept and present a cross-sectoral approach of knowledge co-creation to inform natural resource use and management. The approach is tested in three case studies across Europe that face different challenges resulting from climate and socio-economic change. We present the process, applied methods and major results of knowledge co-creation for sustainable natural resource use and management, and we reflect on the challenges and opportunities from engaging multiple stakeholders. Even if a comprehensive, cross-sectoral approach encourages embedding the water−energy−land nexus into climate services and allows the development of pathways towards sustainable natural resource use and management, maintaining these achievements and partnerships beyond the lifetime of a research project remains challenging.

How to cite: Tudose, N. C., Marin, M., Cheval, S., and Ungurean, C.: Challenges and opportunities of knowledge co-creation for the water-energy-land nexus, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-11292, https://doi.org/10.5194/egusphere-egu23-11292, 2023.

09:50–10:00
|
EGU23-14580
|
ECS
|
On-site presentation
Riccardo Luporini, Marcello Arosio, Emanuele Sommario, and Mario Martina

Strategic climate change litigation is rising on a global scale as a tool to bridge the accountability and enforcement gap that is currently affecting climate change law. The vast majority of strategic climate cases concern mitigation, while adaptation is rarely addressed, and when it is, this is done in a rather residual and vague manner (Setzer and Higham, 2022). However, if it is true that states and corporate actors are lagging behind their emission reduction commitments, at the same time ‘at current rates of adaptation planning and implementation, the adaptation gap will continue to grow’ (IPCC, 2022). Accordingly, once strategic litigation is found to be a suitable tool to advance climate action, opportunities to litigate adaptation strategically should be further explored.

 The role of science in substantiating climate change litigation is very much under the spotlight when it comes to the determination of emission reduction targets, carbon budget and ‘fair shares’ in mitigation cases (BIICL and Sant’Anna, 2021). On the other hand, science does not yet provide accurate indicators of adaptation progress or lack thereof and this contributes to narrowing down opportunities for strategic litigation on adaptation (Berrang-Ford, Biesbroek et al, 2019).

Against this background, this study aims to investigate the role of geosciences in fostering strategic litigation on climate change adaptation. This objective is pursued via a case study. The study builds hypothetical strategic cases concerning public authorities’ liability for flood risk reduction and investigates the potential role of geosciences in such cases. How can geosciences help determine the impacts of climate change on flood risk in a given area and the consequent exposure and vulnerability of specific communities? What does a science-based assessment of given adaptation and flood risk reduction policies and measures look like? To what extent can geosciences determine the activities that public authorities should take to reduce flood risk in a certain area? And, finally, how far can existing commitments in the area of disaster risk reduction and human rights be used in order to distill concrete obligations in terms of adaptation to climate change-induced hazards? The study aims to address these questions by means of an interdisciplinary approach based on combining legal and policy practice with sound geoscience methodology.

References

Joana Setzer and Catherine Higham, ‘Global trends in climate change litigation: 2022 snapshot’, (2022) Grantham Research Institute on Climate Change and the Environment and Centre for Climate Change Economics and Policy, London School of Economics and Political Science

IPCC [Hans-O Pörtner et al. (eds)], Climate Change 2022 Impacts, Adaptation and Vulnerability. Working Group II Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, Summary for Policy Makers

A Holzhausen, R Luporini (Eds), The Role of Science in Climate Change Litigation: International Workshop Report, (July 2021)

Lea Berrang-Ford, Robbert Biesbroek, et al, Tracking global climate change adaptation among governments, Nature Climate Change 9, 440–449 (2019)

How to cite: Luporini, R., Arosio, M., Sommario, E., and Martina, M.: Strategic litigation on climate change adaptation: The case of public authorities’ liability in flood risk reduction, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-14580, https://doi.org/10.5194/egusphere-egu23-14580, 2023.

10:00–10:10
|
EGU23-16311
|
ECS
|
On-site presentation
Alice Gallazzi, Francesco Ballio, Daniela Molinari, Marina Credali, and Immacolata Tolone

The purpose of the study is to define how the models available for flood damage assessment in the Italian context can support cost-benefit or multi-criteria analyses of risk mitigation measures, in accordance with current laws and regulations on the subject. On the basis of the present situation in which risk mitigation measures are evaluated mostly according to their capability of reducing the hazard and by considering few simple exposure factors, the study aims at identifying more robust indicators to assess measures effectiveness based on results from flood damage modelling. State of the art flood damage models developed within the context of the project MOVIDA (MOdello per la Valutazione Integrata del Danno Alluvionale – Model for integrated evaluation of flood damage, https://sites.google.com/view/movida-project) were applied to evaluate the expected damage in several flood prone areas within the Lombardia Region (northern Italy), where mitigation actions are planned by the Regional Authority. Then, obtained results for these areas were analysed to define effectiveness indicators as well as their range of values. Finally, specific indicators were developed to evaluate the environmental impact of each intervention according to present policies to promote sustainable investments in the field of soil protection as well as contribute to achieve Green Deal goals. Results show that developed indicators increase the ability of local authorities in the definition of priorities of intervention, leading to a reduction of institutional and legislative inefficiencies and increasing the efficiency of disaster risk reduction policies.

How to cite: Gallazzi, A., Ballio, F., Molinari, D., Credali, M., and Tolone, I.: Flood risk assessment in support of the evaluation and selection of risk mitigation measures, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-16311, https://doi.org/10.5194/egusphere-egu23-16311, 2023.

Posters on site: Thu, 27 Apr, 16:15–18:00 | Hall X4

Chairpersons: Claire Burke, James Brennan, Laura Ramsamy
X4.1
|
EGU23-10040
Gijs van den Oord, Maarten van Meersbergen, Peter Kok, Jesus Garcia Gonzales, Sander van Rijn, Alessio Ciullo, Elco Koks, Ertug Ercin, Henrique Moreno Dumont Goulart, Esther Boere, Christian Otto, Patryk Kubiczek, Robin Middelanis, Carla Mauricio, Keren Prize Bolter, Dana Stuparu, and Bart van den Hurk

Disseminating the effects of climate change and its potential future impacts to a wider audience is a demanding task, yet of great importance to society. Moreover, quantifying causal chains emerging from global warming is often impeded by the growth of unknown parameters related to modeling socio-economic responses. One method to obtain insights into the complex consequences of climate change is the use of physical climate storylines. Conceptually, storylines correspond to reasonable choices for the unknowns within the modeled impact transmission chain. They allow us to understand and describe the unfolding of climate-induced extreme events, making the impacts of global warming tangible to a wide range of potential stakeholders.

The RECEIPT project develops and applies the concept of climate storylines to provide risk information on climate change effects with a remote origin and an impact on European socio-economic sectors. Sectors that are being addressed within RECEIPT are the European critical infrastructure, manufacturing chains, the food system, financial markets and European international cooperation with (developing) regions. Experts within the consortium construct credible storylines for these sectors, often starting from extreme, disrupting historical events and translating these to counterfactual climate and socio-economic futures. These analyses are being published in scientific journals, but the RECEIPT consortium envisions an alternative dissemination channel to target a larger community.

The storyline visualizer (https://www.climateimpactstories.eu) is an interactive, web-based user interface, aimed at communicating physical climate storylines to an audience of informed stakeholders. The visualizer enables storyline developers in RECEIPT to structure their message into a logical progression of sections, and support each page with text, pictures, geospatial data and interactive charts. The visualizer also allows the user to explore data used within the storyline and browse through counterfactual futures. Currently, five storylines have been visualized with this platform, describing:

  • the future impacts of sea level rise and storm surges upon critical infrastructure around the French Atlantic coast, based upon storm Xynthia;

  • increased impacts of cyclones upon European overseas territories and the sustainability of the European Solidarity Fund within this context;

  • soy production disruptions in a warming climate and their impact on the European food system;

  • multi-breadbasket harvest failures, locust infestations and their impact upon food security in the Greater Horn of Africa;

  • the impact of extreme hurricanes in the Houston metropolitan area for global manufacturing chains and European industry.

Implementing these studies as captivating climate storylines in the visualizer has taught us valuable lessons; one particular challenge has been to handle the growing complexity of the analyses when multiple socio-economic aspects are taken into account. Using a minimalist approach, shifting the focus towards the modeled impacts rather than the full academic reasoning, have appeared to be a useful path forward, resulting in accessible yet credible storylines of climate impacts. In this session, we plan to showcase the capabilities of the storyline visualizer, review lessons learned during the implementation process and discuss possible applications beyond RECEIPT.

How to cite: van den Oord, G., van Meersbergen, M., Kok, P., Garcia Gonzales, J., van Rijn, S., Ciullo, A., Koks, E., Ercin, E., Moreno Dumont Goulart, H., Boere, E., Otto, C., Kubiczek, P., Middelanis, R., Mauricio, C., Prize Bolter, K., Stuparu, D., and van den Hurk, B.: Communicating impacts of climate change with the RECEIPT storyline visualizer, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-10040, https://doi.org/10.5194/egusphere-egu23-10040, 2023.

X4.2
|
EGU23-9512
I-CHANGE (Individual Change of HAbits Needed for Green European transition) project
(withdrawn)
Antonio Parodi, Silvana Di Sabatino, Pinhas Alpert, Paolo Mazzetti, Francesco Pilla, Maria Carmen Llasat, Gert-Jan Steeneveld, Muhammad Adnan, Fredrik Wetterhall, Katriina Soini, Stine Skot, Katja Firus, Milan Kalas, Carlo Trozzi, Ephraim Broschkowski, and Seyni Salack and the I-CHANGE
X4.3
|
EGU23-15960
Lan Hoang and Michail Smyrnakis

Climate change is posing challenges for operating and designing critical infrastructure. Increasingly, AI has been used to enhance these decision making process. Reinforcement Learning has shown its advantages in dealing with difficult sequential decision making in games. When scaling to real life applications, their complexity and heterogenous nature potentially will require Multi Agent Reinforcement Learning (MARL) to provide adaptive capacity in a distributed manner. However, the human system is also characterised by the diverse belief of each individuals and groups - a feature that was captured in agent based models. AI/agent systems are evolving to work with human and become ubiquitous in real life/applications critical to society (such as health and transport). We argue that allowing belief transfer and full interactions across MARL actors in a three-layer model capturing data uncertainty, logical model and belief will help create a heterogeneous MARL system for better human-AI interaction that better aligns with human thoughts/values for actionable climate decisions.

How to cite: Hoang, L. and Smyrnakis, M.: Towards teaching multi agent system the concept of risks and safety for actionable climate decisions, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-15960, https://doi.org/10.5194/egusphere-egu23-15960, 2023.