Periodic Reporting for period 2 - CLINT (CLImate INTelligence: Extreme events detection, attribution and adaptation design using machine learning)
Berichtszeitraum: 2023-01-01 bis 2024-06-30
Climate Services are an essential component of adaptation and mitigation strategies as well as disaster risk management because extreme events, including tropical cyclones, heatwaves and warm nights, extreme droughts, along with compound events and concurrent extremes, are expected to increase in both frequency and intensity in many regions of the world over the next decades.
The implementation of effective local and regional adaptation strategies in line with the Paris Agreement and Sustainable Development Goals (e.g. climate action, clean water and sanitation, sustainable cities and communities, life on land, affordable and clean energy) is however challenged by the fact that extreme events are expected to be regionally more complex than that expected from thermodynamic changes alone. On the other hand, climate services can benefit from an unprecedented availability of data, in particular from the Copernicus Climate Change Service, and recent advances in Artificial Intelligence and Machine Learning offer a unique opportunity to exploit the full potential of these data with the aim of providing easily accessible, timely, and decision-relevant information to policy makers and end-users.
The main objective of CLINT is the development of an Artificial Intelligence framework composed of Machine Learning techniques and algorithms to process big climate datasets for improving Climate Science in the detection, causation, and attribution of Extreme Events. The CLINT AI framework also covers the quantification of the extreme impacts on a variety of socio-economic sectors under historical, forecasted, and projected climate conditions, and across different spatial scales (from European to local), ultimately developing innovative and sectorial AI-enhanced Climate Services. Finally, these services are being operationalized into Web Processing Services, according to the most advanced open data and software standards by Climate Services Information Systems, and into a Demonstrator to facilitate the uptake of project results by public and private entities for research and Climate Services development.
The implementation of effective local and regional adaptation strategies in line with the Paris Agreement and Sustainable Development Goals (e.g. climate action, clean water and sanitation, sustainable cities and communities, life on land, affordable and clean energy) is however challenged by the fact that extreme events are expected to be regionally more complex than that expected from thermodynamic changes alone. On the other hand, climate services can benefit from an unprecedented availability of data, in particular from the Copernicus Climate Change Service, and recent advances in Artificial Intelligence and Machine Learning offer a unique opportunity to exploit the full potential of these data with the aim of providing easily accessible, timely, and decision-relevant information to policy makers and end-users.
The main objective of CLINT is the development of an Artificial Intelligence framework composed of Machine Learning techniques and algorithms to process big climate datasets for improving Climate Science in the detection, causation, and attribution of Extreme Events. The CLINT AI framework also covers the quantification of the extreme impacts on a variety of socio-economic sectors under historical, forecasted, and projected climate conditions, and across different spatial scales (from European to local), ultimately developing innovative and sectorial AI-enhanced Climate Services. Finally, these services are being operationalized into Web Processing Services, according to the most advanced open data and software standards by Climate Services Information Systems, and into a Demonstrator to facilitate the uptake of project results by public and private entities for research and Climate Services development.
Building on the foundational activities of the initial phase mostly focused on the analysis of state-of-the-art literature and existing datasets along with interactions with climate services end-users, during the second 18 months of the project the CLINT partners have continued to harmonise their efforts, leveraging diverse professional backgrounds to maintain efficient coordination and integration. The development of AI and ML algorithms has progressed significantly, enabling the detection of relationships between extreme events and large-scale climatological fields. These advancements have furthered our ability to quantify causal interdependencies and isolate human fingerprints in relevant processes. As these analyses are consolidated, they will support the generation of sub-seasonal to seasonal forecasts for extreme events. Moreover, six CLINT Artificial Intelligence prototype(s) and two demonstrators of the enhanced Climate Services are under development.
As highlights, the following peer-reviewed articles on CLINT research advances have been already published in top-level journals:
Barriopedro, D., García‐Herrera, R., Ordóñez, C., Miralles, D. G., & Salcedo‐Sanz, S. (2023). Heat waves: Physical understanding and scientific challenges. Reviews of Geophysics, 61(2), e2022RG000780.
Du, Y., Clemenzi, I., & Pechlivanidis, I. G. (2023). Hydrological regimes explain the seasonal predictability of streamflow extremes. Environmental Research Letters, 18(9), 094060.
Torralba, V., Materia, S., Cavicchia, L., Álvarez-Castro, M. C., Prodhomme, C., McAdam, R., ... & Gualdi, S. (2024). Nighttime heat waves in the Euro-Mediterranean region: definition, characterisation, and seasonal prediction. Environmental Research Letters, 19(3), 034001.
Salcedo-Sanz, S., Pérez-Aracil, J., Ascenso, G., Del Ser, J., Casillas-Pérez, D., Kadow, C., ... & Castelletti, A. (2024). Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review. Theoretical and Applied Climatology, 155(1), 1-44.
Ascenso, G., Palcic, G., Scoccimarro, E., Giuliani, M., & Castelletti, A. (2024). A Systematic Framework for Data Augmentation for Tropical Cyclone Intensity Estimation Using Deep Learning (No. EGU24-8955). Copernicus Meetings.
Scoccimarro, E., Lanteri, P., & Cavicchia, L. (2024). Freddy: breaking record for tropical cyclone precipitation?. Environmental Research Letters, 19(6), 064013.
As highlights, the following peer-reviewed articles on CLINT research advances have been already published in top-level journals:
Barriopedro, D., García‐Herrera, R., Ordóñez, C., Miralles, D. G., & Salcedo‐Sanz, S. (2023). Heat waves: Physical understanding and scientific challenges. Reviews of Geophysics, 61(2), e2022RG000780.
Du, Y., Clemenzi, I., & Pechlivanidis, I. G. (2023). Hydrological regimes explain the seasonal predictability of streamflow extremes. Environmental Research Letters, 18(9), 094060.
Torralba, V., Materia, S., Cavicchia, L., Álvarez-Castro, M. C., Prodhomme, C., McAdam, R., ... & Gualdi, S. (2024). Nighttime heat waves in the Euro-Mediterranean region: definition, characterisation, and seasonal prediction. Environmental Research Letters, 19(3), 034001.
Salcedo-Sanz, S., Pérez-Aracil, J., Ascenso, G., Del Ser, J., Casillas-Pérez, D., Kadow, C., ... & Castelletti, A. (2024). Analysis, characterization, prediction, and attribution of extreme atmospheric events with machine learning and deep learning techniques: a review. Theoretical and Applied Climatology, 155(1), 1-44.
Ascenso, G., Palcic, G., Scoccimarro, E., Giuliani, M., & Castelletti, A. (2024). A Systematic Framework for Data Augmentation for Tropical Cyclone Intensity Estimation Using Deep Learning (No. EGU24-8955). Copernicus Meetings.
Scoccimarro, E., Lanteri, P., & Cavicchia, L. (2024). Freddy: breaking record for tropical cyclone precipitation?. Environmental Research Letters, 19(6), 064013.
CLINT is designed to develop an Artificial Intelligence framework for evolving Climate Science and Services by improving the understanding and predictability of extreme events and by quantifying their impacts on various targeted climate-related sectors under both historical and projected climate conditions and across different spatial scales, from the whole European to the local scale in different Climate Change Hotspots in the Netherlands deltas, Iberian Peninsula, Southern Africa, and the Italian Alps. Specifically, CLINT will (1) enhance the adaptive capacity in climate-sensitive sectors, from pan-European to local scale, by advancing the detection, attribution and quantification of future changes in extreme events based on a suite of machine learning techniques coupled with a physical understanding of the key processes; (2) reduce vulnerability to climate change, by developing prototypes of AI-enhanced CS to support EU policies, such as EU floods directive, the new Green Deal, the Climate Adaptation strategy, the Common Agricultural Policy; (3) enhance actions on climate change adaptation informed by continuous and consistent information on extreme events over different time horizons, ranging from sub-seasonal forecasts to climate projections, (4) advance the current scientific knowledge on detection, attribution, and causation of climate extreme events by strengthening the link with the Copernicus climate-related services and fostering effective knowledge transfer between researchers/scientists and service providers; and finally (5) inform CS and decision-making by engaging with local users in the Climate Change Hotspots to put forward AI-enhanced modelling chains for local impact-based predictions and projections of Extreme Events.