Machine learning (ML) is currently transforming data analysis and modelling of the Earth system. While statistical and data-driven models have been used for a long time, recent advances in machine learning now allow for encoding non-linear, spatio-temporal relationships robustly without sacrificing interpretability. This has the potential to accelerate climate science, by providing new physics-based modelling approaches; improving our understanding of the underlying processes; reducing and better quantifying climate signals, variability, and uncertainty; and even making predictions directly from observations across different spatio-temporal scales. The limitations of machine learning methods need to also be considered, such as requiring, in general, rather large training datasets, data leakage, and/or poor generalisation abilities, so that methods are applied where they are fit for purpose and add value.

This session aims to provide a venue to present the latest progress in the use of ML applied to all aspects of climate science and we welcome abstracts focussed on, but not limited to:
- Causal discovery and inference: causal impact assessment, interventions, counterfactual analysis
- Learning (causal) process, equations, and feature representations in observations or across models and observations
- Hybrid models (physically informed ML, emulation, data-model integration)
- Novel detection and attribution approaches, including for extreme events
- Probabilistic modelling and uncertainty quantification
- Super-resolution for climate downscaling
- Explainable AI applications to climate data science and climate modelling
- Distributional robustness, transfer learning and/or out-of-distribution generalisation tasks in climate science

Machine learning (ML) methods have emerged as powerful tools to tackle various challenges in ocean science, encompassing physical oceanography, biogeochemistry, and sea ice research.
This session aims to explore the application of ML methods in ocean science, with a focus on advancing our understanding and addressing key challenges in the field. Our objective is to foster discussions, share recent advancements, and explore future directions in the field of ML methods for ocean science.
A wide range of machine learning techniques can be considered including supervised learning, unsupervised learning, interpretable techniques, and physics-informed and generative models. The applications to be addressed span both observational and modeling approaches.

Observational approaches include for example:
- Identifying patterns and features in oceanic fields
- Filling observational gaps of in-situ or satellite observations
- Inferring unobserved variables or unobserved scales
- Automating quality control of data

Modeling approaches can address (but are not restricted to):
- Designing new parameterization schemes in ocean models
- Emulating partially or completely ocean models
- Parameter tuning and model uncertainty

The session welcomes also submissions at the interface between modeling and observations, such as data assimilation, data-model fusion, or bias correction.

Researchers and practitioners working in the domain of ocean science, as well as those interested in the application of ML methods, are encouraged to attend and participate in this session.

We welcome Julie Deshayes as a solicited speaker, presenting 'Ocean models for climate applications: progress expected from Machine Learning'

Cities are intricate multi-scale systems, composed of diverse sub-components such as population, energy, transport, and climate. These components interact on various time scales, from hourly to seasonal to annual and beyond. Effective urban models and digital twins, crucial for urban planning and policy-making, must account for these complex interactions as they govern the growth and functioning of cities, often giving rise to emergent large-scale phenomena. However, our ability to quantitatively describe city behaviour remains limited due to the myriad of processes, scales, and feedbacks involved.
This session invites studies focused on modelling and monitoring the dynamics of multiple sectors and city-biosphere interactions. Topics of interest include, but are not limited to:
• Demography
• Urban transport networks
• Energy consumption
• Anthropogenic emissions and Pollution
• Urban climate
• Urban hydrology
• Urban ecology

Our aim is to elucidate the complex dynamics within urban environments and explore how urban form and function can be optimised to enhance the liveability and well-being of their citizens.

Machine learning (ML) is being used throughout the geophysical sciences with a wide variety of applications. Advances in big data, deep learning, and other areas of artificial intelligence (AI) have opened up a number of new approaches to traditional problems.

Many fields (climate, ocean, NWP, space weather etc.) make use of large numerical models and are now seeking to enhance these by combining them with scientific ML/AI techniques. Examples include ML emulation of computationally intensive processes, data-driven parameterisations for sub-grid processes, ML assisted calibration and uncertainty quantification of parameters, amongst other applications.

Doing this brings a number of unique challenges, however, including but not limited to:
- enforcing physical compatibility and conservation laws, and incorporating physical intuition,
- ensuring numerical stability,
- coupling of numerical models to ML frameworks and language interoperation,
- handling computer architectures and data transfer,
- adaptation/generalisation to different models/resolutions/climatologies,
- explaining, understanding, and evaluating model performance and biases.
- quantifying uncertainties and their sources
- tuning of physical or ML parameters after coupling to numerical models (derivative-free optimisation, Bayesian optimisation, ensemble Kalman methods, etc.)

Addressing these requires knowledge of several areas and builds on advances already made in domain science, numerical simulation, machine learning, high performance computing, data assimilation etc.

We solicit talks that address any topics relating to the above. Anyone working to combine machine learning techniques with numerical modelling is encouraged to participate in this session.

Earth System Models (ESMs), climate forcing, and Earth system reconstructions are crucial for understanding climate dynamics. However, disparities in responses to forcing agents, system coupling - particularly across CMIP - as well as the integration of reconstructions, present significant challenges. This session combines insights from deep-time Earth system reconstructions with cutting-edge climate modeling to enhance our understanding of past, present, and future climate change. We highlight the role of anthropogenic and natural forcings, the importance of addressing model uncertainties in CMIP and beyond, opportunities to develop next-generation digital twins of our planet, and present CMIP7 forcings. This session features contributions that span the following themes:

1. Earth System Reconstructions and Digital Twins
- Integrating paleogeographic data and advanced modeling (e.g., machine learning) to reveal past environmental changes and major Earth system transitions.
- Building digital twins of the planet by fusing diverse datasets and numerical models, emphasizing open, community-driven approaches.

2. Anthropogenic and Natural Forcing for CMIP6, CMIP7, and beyond
- Developing and evaluating historical and future time series of climate drivers (e.g., greenhouse gases, aerosols, land-use changes).
- Investigating how changes in forcing propagate through the climate system, using both observational data and idealized or multi-model experiments (CMIP6, CMIP7, etc.).

3. Model Disparities and Uncertainty
- Identifying the causes of divergent outcomes within CMIP ensembles, including internal variability, parameterization, external forcings, and ESM architectures.
- Employing reduced-complexity models and emulators to capture underexplored regions of uncertainty and guide more robust climate projections.

4. Critical Model Development and Impact Research
- Refining ESMs to reduce uncertainties and improve model performance, with emphasis on interdisciplinary approaches.
- Addressing regional-scale challenges in using CMIP outputs for impact studies, ensuring that policymakers and non-experts can effectively interpret climate projections.

We encourage submissions that bridge these topics, highlight open research and interdisciplinary collaboration, and showcase the work of early career researchers.

Join us for an interdisciplinary session, where we will explore how cutting-edge omics technologies are transforming our understanding of ecosystems and their resilience in response to climatic change across all scales. Over billions of years, spatial and temporal shifts in environmental conditions have driven the evolution of diverse microbial, fungal, plant and animal species, shaping the ecosystems, atmosphere, and climate of Earth. Gaining insights into how these organisms and biomes function, adapt, and interact requires a deep understanding of their components and the complex feedback systems they form.

Technological innovations in measuring and interpreting “meta-omics” datasets are now providing unprecedented mechanistic insights across diverse organisms, scales, and environmental spheres. These advances also drive the development of next-generation models to predict ecosystem function. In this session, we bring together ecologists, geochemists, and evolutionary biologists to examine the available omics toolkits for studying organisms and communities and to discuss ongoing efforts to integrate this knowledge across biological and temporal scales to address pressing Earth system science questions.

By combining eco-evolutionary insights with ecosystem-level concepts like community traits and resilience, we aim to foster future ITS sessions that apply integrated omics approaches alongside geoscience techniques for a deeper, mechanistic understanding of ecosystems.

We welcome contributions studying all Earth’s spheres (Biosphere, Atmosphere, Hydrosphere, Cryosphere, Geosphere), using a wide range of omics datasets (metagenomics, metatranscriptomics, metabolomics, proteomics, lipidomics, spectranomics, ionomics, elementomics, and isotopomics) as well as other large datasets such as trait, phenotype, inventory, pollen, and fossil records. We are particularly interested in studies involving control experiments, long-term ecological surveys, or flux networks, as well as research that provides mechanistic insights and employs big data in Earth system models or machine learning to scale patterns across space and time.

Advances in forest system modelling and monitoring techniques are crucial for deepening our understanding of forest ecosystems and their dynamic responses to environmental stresses and disturbances. These advancements are instrumental in addressing global environmental challenges by improving predictions and adapting management strategies accordingly. This session aims to bring together scientists and researchers focused on the latest advancements in forest systems modelling, observational techniques, and analytical methodologies to enhance our understanding of forest structural dynamics, soil carbon (C) dynamics, and the impacts of natural disturbances such as wildfires, insect’s outbreaks, pathogens/disease, droughts, and windstorms. Specifically, this session covers the following topics:

• Advancements in Forest System Modelling: Presentations on new models or significant improvements in existing models, that help predict and analyse forest growth, structural dynamics, C sequestration in biomass and soils, and ecosystem resilience. This includes models that integrate hydrological, meteorological, and biological processes.

• Innovative Monitoring Techniques: Studies showcasing novel observational technologies or methodologies, including remote sensing, isotopic tracing, or ground-based monitoring systems that provide new insights into forest mortality, growth patterns, and C cycling.

• Impact of Natural Disturbances: Research on how wildfires, insect’s outbreaks, pathogens/disease, droughts, and severe wind events alter forest structure, soil C stocks, and overall ecosystem functions. Contributions may include forward-looking information, post-disturbance recovery processes, disturbance modelling, and strategies for disturbance mitigation and adaptation.

• Cross-Scale Integration: Contributions that demonstrate the integration of innovative integrations of data and models across different spatial and temporal scales to understand forest biomass and soil dynamics comprehensively.

• Implications for future Management Strategies: Insights into how advanced modelling and monitoring approaches can shape policy development, offer a range of adaptation strategies, and inform management practices to enhance forest resilience and C retention.

Data imperfection is a common feature in Geosciences. Scientists and managers alike are faced with uncertain, imprecise, heterogeneous, erroneous, missing or redundant multi-source data. Traditionally, statistical methods were used to address these shortcomings. With the advent of Big Data, Machine Learning methods, the development of new techniques in data mining, knowledge representation and extraction as well as artificial intelligence, new avenues are being offered to tackle the shortcomings of data imperfection.
This session aims to provide a venue to exchange on the latest progress in assessing, quantifying and representing data imperfection in all of its forms. We welcome abstracts focused on, but not limited to:
- Use cases and applications from all fields of Geosciences on missing value imputation, data fusion, imprecision management, model inversion. Examples may be built on any type of data: alpha-numerical time series, georeferenced field data, satellite, areal or ground imagery, geographical vector data, videos, etc...
- Theoretical developments for data fusion and completion; uncertainty assessment and quantification, knowledge extraction and representation from heterogeneous data, reasoning and decision making under uncertainty.
- Multi-disciplinary approaches including artificial intelligence and geosciences are encouraged. Contributions addressing data issues and solutions related to participatory sciences, crowd-sourced data and opportunistic measurements will be particularly appreciated.

Earth System Science is witnessing an ever-increasing availability of textual, digital trace, social sensing, mobile phone, opportunistic sensing, audiovisual, and crowdsourced data. These data open unprecedented new research avenues and opportunities but also pose important challenges, from technical hurdles to skewed coverage, difficulties in quality control, and reproducibility limits.
At the same time, large language models (LLMs) are revolutionising the field by enabling researchers to process and interpret complex geological, climatological, environmental, hydrological, and other earth systems data with unprecedented speed and accuracy, leading to new discoveries and insights.
The session scope spans data analysis methodologies, scientific advances from the analysis of emerging data, and broader perspectives on the opportunities and challenges that these data sources present. Specific topics include but are not limited to, for example: assessment of natural hazard impacts (e.g. floods, droughts, landslides, temperature extremes, windstorms), real-time monitoring of disasters, evidence synthesis, public sentiment analysis, policy and awareness tracking, discourse and narrative analyses, natural language processing, large language models, social media analysis, historical data rescue, image mining, deep learning, and machine learning.
This session will provide a platform for geoscientists to discuss the integration of LLMs and novel data types into their workflows, enhancing both efficiency and discovery while addressing challenges such as model accuracy and data bias. We invite presentations that explore the transformative potential of large language models and text data in the geosciences. Join us in contributing to this cutting-edge dialogue and helping shape the future of geosciences through AI.

Downscaling aims to process and refine global climate model output to provide information at spatial and temporal scales suitable for impact studies. In response to the current challenges posed by climate change and variability, downscaling techniques continue to play an important role in the development of user-driven climate information and new climate services and products. In fact, the "user's dilemma" is no longer that there is a lack of downscaled data, but rather how to select amongst the available datasets and to assess their credibility. In this context, model evaluation and verification is growing in relevance and advances in the field will likely require close collaboration between various disciplines.

Furthermore, epistemologists have started to revisit current practices of climate model validation. This new thread of discussion encourages to clarify the issue of added value of downscaling, i.e. the value gained through adding another level of complexity to the uncertainty cascade. For example, the ‘adequacy-for-purpose view’ may offer a more holistic approach to the evaluation of downscaling models (and atmospheric models, in general) as it considers, for example, user perspectives next to a model’s representational accuracy.

In our session, we aim to bring together scientists from the various geoscientific disciplines interrelated through downscaling: atmospheric modeling, climate change impact modeling, machine learning and verification research. We also invite philosophers of climate science to enrich our discussion about novel challenges faced by the evaluation of increasingly complex simulation models.

Contributions to this session may address, but are not limited to:

- newly available downscaling products,
- applications relying on downscaled data,
- downscaling method development, including the potential for machine learning,
- bias correction and statistical postprocessing,
- challenges in the data management of kilometer-scale simulations,
- verification, uncertainty quantification and the added value of downscaling,
- downscaling approaches in light of computational epistemology.

The advancement of Open Science and the affordability of computing services allow for the discovery and processing of large amounts of information, boosting data integration from diverse scientific domains and blurring traditional discipline boundaries. However, data are often heterogeneous in format and provenance, and the capacity to combine them and extract new knowledge to address scientific and societal problems relies on standardisation, integration and interoperability.
Key enablers of the OS paradigm are ESFRI Research infrastructures, of which ECCSEL (www.eccsel.org), EMSO (https://emso.eu/) and EPOS (www.epos-eu.org), are examples currently enhancing FAIRness and integration within the Geo-INQUIRE project. Thanks to decades of work in data standardisation, integration and interoperability, they enable scientists to combine data from different disciplines and data sources into innovative research to solve scientific and societal questions.
But while data-driven science is ripe with opportunity to groundbreaking inter- and transdisciplinary results, many challenges and barriers remain.

This session aims to foster scientific cross-fertilization exploring real-life scientific studies and research experiences from scientists and ECS in Environmental Sciences. We also welcome contributions about challenges in connection to data availability, collection, processing, interpretation, and the application of interdisciplinary methods.
A non-exhaustive list of of topics includes:
- multidisciplinary studies involving data from different disciplines, e.g. combining seismology, geodesy, oceanography and petrology to understand subduction zone dynamics;
- interdisciplinary works, integrating two or more disciplines to create fresh approaches, e.g. merging solid earth and ocean sciences data to study coastal/oceanic areas and earth dynamics;
- showcase activities enabling interdisciplinarity and open science, e.g. enhancing FAIRness of data and services, enriching data provision, enabling cross-domain AI applications, software and workflows, transnational access and capacity building for ECS;
- transdisciplinary experiences that surpass disciplinary boundaries, integrate paradigms and engage stakeholders from diverse backgrounds, e.g. bringing together geologists, social scientists, civil engineers and urban planners to define risk maps and prevention measures in urban planning, or studies combining volcanology, atmospheric, health and climate sciences.

The increasing frequency and severity of natural hazards, including floods, landslides, earthquakes, volcanic eruptions, droughts, wildfires, and ground subsidence, pose significant risks to the environment, infrastructure, and human societies. This trend is expected to continue, influenced by climate change and extreme weather events, underscoring the urgent need for improved disaster preparedness, environmental management, and resilient urban planning.

This session focuses on the use of advanced Geoinformatics technologies—such as Geographical Information Systems (GIS), Remote Sensing, and Artificial Intelligence (AI)—to understand and mitigate the impact of natural hazards. Emphasis will be placed on the application of explainable AI techniques (e.g., Shapley Additive Explanations, Local Interpretable Model-agnostic Explanations, and Explainable Boosting Machines) to enhance decision-making in disaster management. We invite contributions that explore the integration of new and historical data, remote sensing technologies, and innovative analytical methodologies aimed at understanding the manifestation and evolution of catastrophic events. Special attention will be given to successful case studies from diverse environments and climate scenarios, leveraging cutting-edge technologies to foster safer, more resilient societies.

In addition, the session will delve into the dynamic relationship between natural hazards and human activities—such as migration, construction, urban planning, and resource management—which influence and are influenced by environmental risks. Understanding these complex, spatiotemporal relationships is crucial for improving disaster risk reduction and building sustainable resilience. We encourage interdisciplinary contributions that combine Earth observation data (e.g., optical, hyperspectral, RADAR, GNSS, LiDAR) with historical, social, and demographic datasets to investigate these interconnections. The session seeks to bring together a broad range of experts, including geodesists, natural and social scientists, historians, anthropologists, engineers, urban planners, policymakers, and community workers, to promote transdisciplinary discussions on the integration of Earth observation data for disaster risk reduction and sustainable development.