Modern challenges of climate change, disaster management, public health and safety, resources management, and logistics can only be effectively addressed through big data analytics. Advances in technology are generating vast amounts of geospatial data on local and global scales. The integration of artificial intelligence (AI) and machine learning (ML) has become crucial in analysing these datasets, leading to the creation of various maps and models within the various fields of geosciences. Recent studies, however, highlight significant challenges when applying ML and AI to spatial and spatio-temporal data along the entire modelling pipeline, including reliable accuracy assessment, model interpretation, transferability, and uncertainty assessment. This gap has been recognised and led to the development of new spatio-temporally aware strategies and methods in response to the promise of improving spatio-temporal predictions, the treatment of the cascade of uncertainties, decision making and facilitating communication.
This session discusses challenges and advances in spatial and spatio-temporal machine learning methods and the software and infrastructures to support them.

Recent breakthroughs in machine learning, notably deep learning, that facilitate massive amounts of data with data-driven AI models have led to an unprecedented potential for large-scale environmental monitoring through remote sensing. Despite the success of existing deep learning-based approaches in remote sensing for many applications, their shortcomings in jointly leveraging various aspects of Earth observation data prevent fully exploiting the potential of remote sensing for the environment. Namely, integrating multiple data modalities and remote sensing sensors, leveraging deep learning methods over multi-spatial/spectral resolution Earth observation data, and modeling space and temporality together offer remarkable opportunities for a comprehensive and accurate understanding of the environment. Throughout this session, we aim to gather the community to delve into the latest scientific advances that leverage these multi-dimensional approaches to tackle pressing environmental challenges.

Earth, its weather and climate constitute a complex system whose monitoring and forecasting have witnessed remarkable progress in recent years. In particular, enhanced spaceborne observations with the integration of Machine/Deep Learning (ML/DL) techniques are key drivers of innovation in Earth System Observation and Prediction (ESOP) for Weather and Climate. In parallel, the concept of Digital Twins (DTs) of the Earth has emerged as a revolutionary approach to address climate resilience, disaster risk management, and sustainable development through highly detailed, high fidelity digital replicas of the Earth system.

Recently, ML/DL techniques have attracted significant attention and increased adoption within the ESOP community due to their ability to enhance our simulation and prediction capabilities of the Earth's complex dynamics. At the same time, DTs serve as comprehensive monitoring, simulation, and prediction systems that enable us to analyse and better comprehend the intricate interactions between natural phenomena and human activities.

The focus of the session is on exploring new data sources and benchmarks for weather and climate modelling, the adaptation of large-scale physics- or data-driven Earth system models, the integration of real-time multi-disciplinary data, and demonstrations of practical applications of these systems in addressing climate impacts, resilience and sustainability. This session invites experts from diverse fields to discuss how recent advances innovate on established ESOP approaches, to better address current challenges and to identify opportunities for future work as well as synergies across domains.

A key emphasis will be placed on the societal implications of these technologies, showcasing how ML-enhanced ESOP and Earth Digital Twins can empower policymakers with tailored insights for optimizing resource management, designing effective adaptation strategies, and building resilience against severe weather and climate challenges.

Geospatial Foundation Models (GeoFMs) have shown great promise in a wide range of applications for Earth Observation (EO) and Earth System Modelling (ESM), as well as for Weather and Climate. With the increasing number of models being published, model inter-comparison is key to identify the best GeoFM for deployment. This session aims to highlight efforts on model development, benchmarking, fine-tuning, and their best practices for utilizing GeoFMs in real-world applications. We invite submissions focused on creating GeoFMs to leverage multi-modal, multi-temporal, and multi-resolution datasets towards sensor-independence. Diverse FMs for EO, ESM, Weather, and Climate can revolutionize data analysis by handling text, imagery, and time-series, enabling insights into natural hazards and climate resilience. Our session will cover advances in data curation, model architecture, scaling, benchmarking, pretraining, fine-tuning, and MLOps for GeoFMs, including use cases and deployment strategies.

The topics of our session revolving around GeoFMs are:
1. Benchmarks & Evaluation: Establish standardized fair evaluation metrics and benchmarks to assess the performance and capabilities of GeoFMs in multi-modal data analysis, ensuring reliability and efficiency.
2. Pre-Training Strategies & Best Practices: Discuss efficient data sampling strategies, proxy tasks, and scalable model training for efficient pre-training of GeoFMs. Guidelines for using existing pre-trained GeoFMs for a diverse set of applications, with focus on how to decide which models are best for certain use cases.
3. Sensor Independence: GeoFMs can process data from various sensors, enabling comprehensive analysis of the Earth's dynamics holistically.
4. Multi-Modal/Temporal: GeoFMs offer novel approaches to multi-modal data analysis and spatio-temporal change detection.
5. Scientific Insights: Highlighting the scientific insights enabled through the creation of GeoFMs, particularly in relation to geo-physical principles and causal relations.
6. Community Involvement & Impact: How to build an open-science community around GeoFMs that is easily accessible to all while keeping an eye on potential societal, environmental, and economic impacts when deploying GeoFMs.

We aim to foster discussions on current applications, challenges, and opportunities of GeoFMs seeking contributions from AI and domain researchers, climate modelers, industry experts, and stakeholders in AI, HPC, and Big Data.

The recent growing number of probes in the heliosphere and future missions in preparation led to the current decade being labelled as "the golden age of heliophysics research". With more viewpoints and data downstreamed to Earth, machine learning (ML) has become a precious tool for planetary and heliospheric research to process the increasing amount of data and help the discovery and modelisation of physical systems. Recent years have also seen the development of novel approaches leveraging complex data representations with highly parameterised machine learning models and combining them with well-defined and understood physical models. These advancements in ML with physical insights or physically informed neural networks inspire new questions about how each field can respectively help develop the other. To better understand this intersection between data-driven learning approaches and physical models in planetary sciences and heliophysics, we seek to bring ML researchers and physical scientists together as part of this session and stimulate the interaction of both fields by presenting state-of-the-art approaches and cross-disciplinary visions of the field.

The "ML for Planetary Sciences and Heliophysics" session aims to provide an inclusive and cutting-edge space for discussions and exchanges at the intersection of machine learning, planetary and heliophysics topics. This space covers (1) the application of machine learning/deep learning to space research, (2) novel datasets and statistical data analysis methods over large data corpora, and (3) new approaches combining learning-based with physics-based to bring an understanding of the new AI-powered science and the resulting advancements in heliophysics research.
Topics of interest include all aspects of ML and heliophysics, including, but not limited to: space weather forecasting, computer vision systems applied to space data, time-series analysis of dynamical systems, new machine learning models and data-assimilation techniques, and physically informed models.

One of the big challenges in Earth system science consists in providing reliable climate predictions on sub-seasonal, seasonal, decadal and longer timescales. The resulting data have the potential to be translated into climate information leading to a better assessment of global and regional climate-related risks.
The main goals of the session is (i) to identify gaps in current climate prediction methods and (ii) to report and evaluate the latest progress in climate forecasting on subseasonal-to-decadal and longer timescales. This will include presentations and discussions of developments in the predictions for the different time horizons from dynamical ensemble and statistical/empirical forecast systems, as well as the aspects required for their application: forecast quality assessment, multi-model combination, bias adjustment, downscaling, exploration of artificial-intelligence methods, etc.
Following the new WCRP strategic plan for 2019-2029, prediction enhancements are solicited from contributions embracing climate forecasting from an Earth system science perspective. This includes the study of coupled processes between atmosphere, land, ocean, and sea-ice components, as well as the impacts of coupling and feedbacks in physical, hydrological, chemical, biological, and human dimensions. Contributions are also sought on initialization methods that optimally use observations from different Earth system components, on assessing and mitigating the impacts of model errors on skill, and on ensemble methods.
We also encourage contributions on the use of climate predictions for climate impact assessment, demonstrations of end-user value for climate risk applications and climate-change adaptation and the development of early warning systems.
A special focus will be put on the use of operational climate predictions (C3S, NMME, S2S), results from the CMIP5-CMIP6 decadal prediction experiments, and climate-prediction research and application projects.
An increasingly important aspect for climate forecast's applications is the use of most appropriate downscaling methods, based on dynamical, statistical, artificial-intelligence approaches or their combination, that are needed to generate time series and fields with an appropriate spatial or temporal resolution. This is extensively considered in the session, which therefore brings together scientists from all geoscientific disciplines working on the prediction and application problems.

In recent years, technologies based on Artificial Intelligence (AI), such as image processing, smart sensors, and intelligent inversion, have garnered significant attention from researchers in the geosciences community. These technologies offer the promise of transitioning geosciences from qualitative to quantitative analysis, unlocking new insights and capabilities previously thought unattainable.
One of the key reasons for the growing popularity of AI in geosciences is its unparalleled ability to efficiently analyze vast datasets within remarkably short timeframes. This capability empowers scientists and researchers to tackle some of the most intricate and challenging issues in fields like Geophysics, Seismology, Hydrology, Planetary Science, Remote Sensing, and Disaster Risk Reduction.
As we stand on the cusp of a new era in geosciences, the integration of artificial intelligence promises to deliver more accurate estimations, efficient predictions, and innovative solutions. By leveraging algorithms and machine learning, AI empowers geoscientists to uncover intricate patterns and relationships within complex data sources, ultimately advancing our understanding of the Earth's dynamic systems. In essence, artificial intelligence has become an indispensable tool in the pursuit of quantitative precision and deeper insights in the fascinating world of geosciences.
For this reason, aim of this session is to explore new advances and approaches of AI in Geosciences.

The complexity of hydrological and Earth systems poses significant challenges to their prediction and understanding capabilities. The advent of machine learning (ML) provides powerful tools for modeling these complex systems. However, realizing their full potential in this field is not just about algorithms and data, but requires a cooperative interaction between domain knowledge and data-driven power. This session aims to explore the frontier of this convergence and how it facilitates a deeper process understanding of various aspects of hydrological processes and their interactions with the atmosphere and biosphere across spatial and temporal scales.

We invite researchers working in the fields of explainable AI, physics-informed ML, hybrid Earth system modeling (ESM), and AI for causal and equation discovery in hydrology and Earth system sciences to share their methodologies, findings, and insights. Submissions are welcome on topics including, but not limited to:

- Explainability and transparency in ML/AI modeling of hydrological and Earth systems;
- Process and knowledge integration in ML/AI models;
- Data assimilation and hybrid ESM approaches;
- Causal learning and inference in ML models;
- Data-driven equation discovery in hydrological and Earth systems;
- Data-driven process understanding in hydrological and Earth systems;
- Challenges, limitations, and solutions related to hybrid models and XAI.

Deep Learning has seen accelerated adoption across Hydrology and the broader Earth Sciences. This session highlights the continued integration of deep learning and its many variants into traditional and emerging hydrology-related workflows. We welcome abstracts related to novel theory development, new methodologies, or practical applications of deep learning in hydrological modeling and process understanding. This might include, but is not limited to, the following:
(1) Development of novel deep learning models or modeling workflows.
(2) Probing, exploring and improving our understanding of the (internal) states/representations of deep learning models to improve models and/or gain system insights.
(3) Understanding the reliability of deep learning, e.g., under non-stationarity and climate change.
(4) Modeling human behavior and impacts on the hydrological cycle.
(5) Deep Learning approaches for extreme event analysis, detection, and mitigation.
(6) Natural Language Processing in support of models and/or modeling workflows.
(7) Applications of Large Language Models and Large Multimodal Models (e.g. ChatGPT, Gemini, etc.) in the context of hydrology.
(8) Uncertainty estimation for and with Deep Learning.
(9) Advances towards foundational models in the context of hydrology and Earth Sciences more generally.
(10) Exploration of different training strategies, such as self-supervised learning, unsupervised learning, and reinforcement learning.

In this short course we will address the increasing role of artificial intelligence (AI) in geoscientific research, guiding participants through the various stages of the research process where AI tools can be effectively implemented, however with responsibility. We will explore freely available AI tools that can be used for data analysis, model development, and research publication. Additionally, the course aims to provoke reflections on the ethical implications of AI use, addressing concerns such as data bias, transparency, and the potential for misuse. Participants will engage in interactive discussions to explore what constitutes responsible and acceptable use of AI in geoscientific research, aiming to establish a set of best practices for integrating AI into scientific workflows.

Model Land is a conceptual place within the boundaries of a model. When we confuse a Model Land for the real world, we can be ignorant to the assumptions, limitations, uncertainties, and biases inherent in our models. These things need to be carefully understood and considered before we use models to inform decisions about the real world and by doing so we can escape from our Model Lands (Thompson, 2019).
However, in order to escape, we need to explore our Model Lands, mapping them and developing a deeper understanding of their rules and boundaries. In this short course we will present a framework inspired by tabletop roleplay games (TTRPGs) that will bring Model Lands to life. Either using your own model or one of our examples you will learn how to build a world that follows its rules, how to investigate what it would be like to exist within that world, and how to share with others what you have learnt.
Please bring along a pen and paper and be prepared to share your Model Lands. We want to encourage creative expression, so if you have a flair for drawing, poetry, games design, or interpretive dance, feel free to bring along the means to share your creations through whatever medium you prefer.

Data assimilation (DA) is widely used in the study of the atmosphere, the ocean, the land surface, hydrological processes, etc. The powerful technique combines prior information from numerical model simulations with observations to provide a better estimate of the state of the system than either the data or the model alone. This short course will introduce participants to the basics of data assimilation, including the theory and its applications to various disciplines of geoscience. An interactive hands-on example of building a data assimilation system based on a simple numerical model will be given. This will prepare participants to build a data assimilation system for their own numerical models at a later stage after the course.
In summary, the short course introduces the following topics:

(1) DA theory, including basic concepts and selected methodologies.
(2) Examples of DA applications in various geoscience fields.
(3) Hands-on exercise in applying data assimilation to an example numerical model using open-source software.

This short course is aimed at people who are interested in data assimilation but do not necessarily have experience in data assimilation, in particular early career scientists (BSc, MSc, PhD students and postdocs) and people who are new to data assimilation.

Over the last decade, a flurry of machine learning methods has led to novel insights throughout geophysics. As wide as the applications are the data types processed, including environmental parameters, GNSS, InSAR, infrasound, and seismic data, but also downstream structured data products such as 3D data cubes, earthquake catalogs, seismic velocity changes. Countless methods have been proposed and successfully applied, ranging from traditional techniques to recent deep learning models. At the same time, we are increasingly seeing the adoption of machine learning techniques in the wider geophysics community, driven by continuously growing data archives, accessible codes, and software. Yet, the landscape of available methods and data types is difficult to navigate, even for experienced researchers.

In this session, we want to bring together machine learning researchers and practitioners throughout the domains of geophysics. We aim to identify common challenges connecting different tasks and data types and formats, and outline best practices for the development and use of machine learning. We also want to discuss how recent trends in machine learning, such as foundation models, the shift to multimodality, or physics informed models may impact geophysical research. We welcome contributions from all fields of geophysics, covering a wide range of data types and machine learning techniques. We also encourage contributions for machine learning adjacent tasks, such as big-data management, data visualization, or software development in the field of machine learning.

Assessing the spatial heterogeneity of environmental variables is a challenging problem in real applications when numerical approaches are needed. This is made more difficult by the complexity of Natural Phenomena, which are characterized by (Chiles and Delfiner, 2012):
- being unknown: their knowledge is often incomplete, derived from limited and sparse samples;
- dimensionality: they can be represented in two- or three-dimensional domains;
- complexity: deterministic interpolators (i.e., Inverse Distance Weighted) may fail in providing exhaustive spatial distribution models, as they do not consider uncertainty;
- uniqueness: invoking a probabilistic approach, they can be assumed as a realization of a random process and described by regionalized variables.
Geostatistics provides optimal solutions to this issue, offering tools to accurately predict values and uncertainty in unknown locations while accounting for the spatial correlation of samples.

The course will address theoretical and practical methods for evaluating data heterogeneity in computational domains, exploiting the interplay between geometry processing, geostatistics, and stochastic approaches. It will be mainly split into 4 parts, as follows:
- Theoretical Overview: Introduction to Random Function Theory and Measures of Spatial Variability
- Modeling Spatial Dependence: An automatic solution to detect both isotropic and anisotropic spatial correlation structures
- The role of Unstructured Meshes: Exploration of flexible, robust, and adaptive geometric modeling, coupled with stochastic simulation algorithms
- Filling the Mesh: Developing a compact and tangible spatial model, that incorporates all alternative realizations, statistics, and uncertainty

The course will offer a comprehensive understanding of key steps to create a spatial predictive model with geostatistics. We will also promote MUSE (Modeling Uncertainty as a Support for Environments) (Miola et al., STAG2022) as an innovative and user-friendly open-source software, that implements the entire methodology. Tips on how to use MUSE will be provided, along with explanations of its structure and executable commands. Impactful examples will be used to show the effectiveness of geostatistical modeling with MUSE and the flexibility to use it in different scenarios, varying from geology to geochemistry.

The course is designed for everyone interested in geostatistics and spatial distribution models, regardless of their prior experience.

Since the breakthrough of datacubes as a contributor to Analysis-Ready Data, a series of implementations have been announced, and likewise services. However, often these are described through publications only.

In this session, hands-on demos are mandatory. Speakers must spend maximum 50% of their time on presenting slides etc., and minimum 50% to live demos of their tools. To guarantee fair and equal conditions, only in-person presentation will be allowed. Presenters are invited to share their latest and greatest functionality and data, but must balance this with their confidence that the demo will work out of the box.

This enables the audience to see first-hand what datacube features are effectively implemented, how stable they are under strong timing conditions, and what their real-life performance. The expected outcome for attendees is to get a realistic overview on the datacube tools and service landscape, and to assess how much each tool supports their individual needs, such as analysis-readiness.

In a changing climate world, extreme weather and climate events have become more frequent and severe, and are expected to continue increasing in this century and beyond. Unprecedented extremes in temperature, heavy precipitation, droughts, storms, river floodings and related hot and dry compound events have increased over the last decades, impacting negatively broad socio-economic spheres (such as agriculture), producing several damages to infrastructure, but also putting in risk human well-being, to name but a few. The above have raised many concerns in our society and within the scientific community about our current climate but our projected future. Thus, a better understanding of the climate and the possible changes we will face, is strongly needed. . In order to give answers to those questions, and address a wide range of uncertainties, very large data volumes are needed across different spatial (from local-regional to global) and temporal scales (past, current, future), but sources are multiple (observations, satellite, models, reanalysis, etc), and their resolution may vary each other. To deal with huge amounts of information, and take advantage of their different resolution and properties, high-computational techniques within Artificial Intelligence models are explored in climate and weather research. In this short-course, a novel method using Deep Learning models to detect and characterize extreme weather and climate events will be presented. This method can be applied to several types of extreme events, but a first implementation on which we will focus in the short-course, is its ability to detect past heatwaves. Discussions will take place on the method, and also its applicability to different types of extreme events. The course will be developed in python, but we encourage the climate and weather community to join the short-course and the discussion!

Machine Learning (ML) is on the rise as a tool for cryospheric sciences. It has been used to label, cluster, and segment cryospheric components, as well as emulate, project, and downscale cryospheric processes. To date, the cryospheric community mainly adapts and develops ML approaches for highly specific domain tasks. However, different cryospheric tasks can face similar challenges, and when an ML method addresses one problem, it might be transferable to others. Thus, we invite the community to share their current work and identify potential shared challenges and tasks. We invite contributions across the cryospheric domain, including snow, permafrost, glaciers, ice sheets, and sea ice. We especially call for submissions that use novel machine learning techniques; however, we welcome all ML approaches, ranging from random forests to deep learning. Other contributions, such as datasets, theoretical research, and community-building efforts, are also welcome. By identifying shared challenges and transferring knowledge, we aim to channel resources and increase the impact of ML as a tool to observe, assess, and model the cryosphere.

In the era of big data, Environmental and Earth Sciences depend on advanced digital tools, seamless data integration, and interdisciplinary collaboration to tackle pressing global challenges. Sensors, surveys, and experiments produce vast quantities of data, researchers face increasing demands for effective data sharing, large computational resources, and interoperability. Research and e-Infrastructures, together with Virtual Research Environments (VREs) are transforming science by providing cohesive ecosystems that enable global collaboration, support the full research lifecycle, and facilitate shared access to data, computational resources, and communication networks.
We particularly encourage case studies demonstrating how researchers, data scientists, and engineers have successfully utilised infrastructures such as ENVRIs, the European Open Science Cloud (EOSC), and e-Infrastructures such as EGI and D4Science, as well as other relevant platforms and frameworks supporting interdisciplinary collaboration. Experiences from Virtual Access and Transnational Access programmes are also of interest, as are discussions on policies for infrastructure utilisation, software implementation, and lessons learned.
By highlighting collaborative frameworks, this session aims to stimulate discussion on how VREs and e-Infrastructures can enhance interdisciplinary research, streamline workflows, and deliver solutions to complex global environmental challenges. Join us as we explore how these digital ecosystems are helping innovate research in Environmental and Earth Sciences, enabling faster, more accurate models and fostering a global scientific community dedicated to addressing pressing ecological and climate-related issues.

Researchers in Earth System Science (ESS) address complex, interdisciplinary challenges that require analysis of diverse data across multiple scales. Robust and user-friendly Research Data Infrastructures (RDIs) are crucial for supporting data management and collaborative analysis, addressing societal issues. This session will explore how RDIs can bridge the gap between user needs and sustainable ESS data solutions by fostering interdisciplinary collaboration and addressing key challenges in data management and interoperability.

We welcome contributions on the following themes:
- User-Centric Infrastructure Development: This includes user stories, storylines and use cases that demonstrate the importance of cross-disciplinary and cross-scale data usage, as well as innovative infrastructure concepts designed to meet specific user needs. This includes methods for developing high-quality user interfaces and portals.
- Interdisciplinary data fusion and stakeholder engagement: Contributions are welcome that address how RDIs and data centers can facilitate the seamless integration of diverse ESS data to tackle complex societal challenges. This includes exploring interdisciplinary data fusion techniques, strategies for engaging different stakeholders and approaches for integrating stakeholder knowledge into RDI development and data management practices.
- Sustainable software solutions and interoperability: This theme focuses on approaches to building and reusing sustainable software solutions that meet the diverse needs of ESS researchers, including interoperability challenges between different data sources and platforms, and considering appropriate building blocks. It also includes discussion of operation and sustainability models for diverse ESS data centers and strategies for fostering cooperation and interoperability.
- Transdisciplinary research and public engagement: We encourage contributions that explore how RDIs can support transdisciplinary research on sustainability challenges (e.g., climate change and its impacts, etc.) and facilitate public engagement with ESS issues through initiatives such as citizen science.
- Fostering cultural change and collaboration: This theme focuses on strategies for promoting cultural change within research communities to encourage data sharing, collaboration, and the adoption of FAIR principles. This also includes approaches to international collaboration and the development of effective collaboration patterns.

Seismological and Geophysical research consistently uses sophisticated tools for data analysis, modelling, and interpretation. Evidently, the rapid development and diversification of research software pose challenges in maintaining code quality, ensuring comprehensive documentation, achieving reproducibility of results, and enabling uninterrupted workflows comprising various tools for seamless data analysis. As researchers increasingly rely on complex computational tools, it becomes essential to address these challenges in scientific software development, to avoid inefficiencies and errors and to ensure that scientific findings are reliable and can be built upon by future researchers.
We welcome contributions that introduce software tools/toolboxes and their real-world applications, showcasing how they have advanced the field, providing practical insights into the development/application process. Additionally, we seek presentations that discuss methodologies for software testing, continuous integration in software projects, upgrades and deployment. Moreover, we are looking for case studies demonstrating the successful implementation of these tools in various seismological/geophysical problems and how these can bring value to the community.
Sharing of resources, toolboxes, and knowledge is encouraged to improve the overall quality and (re)usability of research software. We encourage the inclusion of demonstrations to showcase usability and functionality examples, as well as videos to illustrate proposed workflows. Videos and other resources can be added as supplementary material and will be available after the conference. Depending on the technical setup and the time available, we will also support live demonstrations for the on-site participants.
We warmly invite seismologists, geophysicists, software developers, and researchers to participate in this session and share their insights, experiences, and solutions to elevate software development standards and practices in our field. Join us to contribute to and learn from discussions that will drive innovation and excellence in seismological and geophysical research.

Recent Earth System Sciences (ESS) datasets, such as those resulting from high-resolution numerical modelling, have increased both in terms of precision and size. These datasets are central to the advancement of ESS for the benefit of all stakeholders, and public policymaking on climate change. Extracting the full value from these datasets requires novel approaches to access, process, and share data. It is apparent that datasets produced by state-of-the-art applications are becoming so large that even current high-capacity data infrastructures are incapable of storing, let alone ensuring their usability. With future investment in hardware being limited, a viable way forward is to explore the possibilities of data compression and new data space implementation.

Data compression has gained interest for making data more manageable, speeding up transfer times, and reducing resource needs without affecting the quality of scientific analyses. Reproducing recent ML and forecasting results has become essential for developing new methods in operational settings. At the same time, replicability is a major concern for ESS and downstream applications and the necessary data accuracy needs further investigation. Research on data reduction and prediction interpretability helps improve understanding of data relationships and prediction stability.

In addition, new data spaces are being developed in Europe, such as the Copernicus Data Space Ecosystem and Green Deal Data Space, as well as multiple national data spaces. These provide access to data, through streamlined access, cloud processing and online visualization generating actionable knowledge enabling more effective decision-making. Analysis ready data can easily be accessed via API transforming data access and processing scalability. Developers and users will share opportunities and challenges of designing and using data spaces for research and industry.

This session connects developers and users of ESS big data, discussing how to facilitate the sharing, integration, and compression of these datasets, focusing on:
1) Approaches and techniques to enhance shareability of high-volume ESS datasets: data compression, novel data space implementation and evolution.
2) The effect of reduced data on the quality of scientific analyses.
3) Ongoing efforts to build data spaces and connect with existing initiatives on data sharing and processing, and examples of innovative services that can be built upon data spaces.

Cloud computing has emerged as a dominant paradigm, supporting industrial applications and academic research on an unprecedented scale. Despite its transformative potential, transitioning to the cloud continues to challenge organizations striving to leverage its capabilities for big data processing. Integrating cloud technologies with high-performance computing (HPC) unlocks powerful possibilities, particularly for computation-intensive AI/ML workloads. With innovations like GPUs, containerization, and microservice architectures, this convergence enables scalable solutions for Earth Observation (EO) and Earth System Modeling domains.
Pangeo (pangeo.io) represents a global, open-source community of researchers and developers collaborating to tackle big data challenges in geoscience. By leveraging a range of tools—from laptops to HPC and cloud infrastructure—the Pangeo ecosystem empowers researchers with an array of core packages, including Xarray, Dask, Jupyter, Zarr, Kerchunk, and Intake.
This session focuses on use cases involving both Cloud and HPC computing and showcasing applications of Pangeo’s core packages. The goal is to assess the current landscape and outline the steps needed to facilitate the broader adoption of cloud computing in Earth Observation and Earth Modeling data processing. We invite contributions that explore various cloud computing initiatives within these domains, including but not limited to:
This session aims to:
• Assess the current landscape and outline the steps needed to facilitate the broader adoption of cloud computing in Earth Observation and Earth Modeling data processing.
• Inspire researchers using or contributing to the Pangeo ecosystem to share their insights with the broader geoscience community and showcasenew applications of Pangeo tools addressing computational and data-intensive challenges.
We warmly welcome contributions that explore:
• Cloud Computing Initiatives: Federations, scalability, interoperability, multi-provenance data, security, privacy, and sustainable computing.
• Cloud Applications and Platforms: Development and deployment of IaaS, PaaS, SaaS, and XaaS solutions.
• Cloud-Native AI/ML Frameworks: Tools designed for AI/ML applications in EO and ESM.
• Operational Systems and Workflows: Cloud-based operational systems, data lakes, and storage solutions.
• HPC and Cloud Integration: Converging workloads to leverage the strengths of both computational paradigms.
In addition, we invite presentations showcasing applications of Pangeo’s core packages in:
• Atmosphere, Ocean, and Land Modeling
• Satellite Observations
• Machine Learning
• Cross-Domain Geoscience Challenges
This session emphasizes real-world use cases at the intersection of cloud and HPC computing. By sharing interactive workflows, reproducible research practices, and live executable notebooks, contributors can help map the current landscape and outline actionable pathways toward broader adoption of these transformative technologies in geoscience.

Mitigating earthquake disasters involves several key components and stages, from identifying and assessing risk to reducing their impact. These components include: a) Long-term and time-dependent analysis of hazards: anticipating the space-time characteristics of ground shaking and its cascading events. b) Vulnerability and exposure assessment c) Risk management: preparedness, rescue, recovery, and overall resilience. A variety of seismic hazard and risk models can be adopted, at different spatial and temporal scale, that incorporate diverse observations and require multi-disciplinary input. Testing and validating these methodologies, for all risk components, is essential for effective disaster mitigation.
From the real-time integration of multi-parametric observations is expected the major contribution to the development of operational time-Dependent Assessment of Seismic Hazard (t-DASH) systems, suitable for supporting decision makers with continuously updated seismic hazard scenarios. A very preliminary step in this direction is the identification of those parameters (seismological, chemical, physical, etc.) whose space-time dynamics and/or anomalous variability can be, to some extent, associated with the complex process of preparation of major earthquakes.
This session includes studies on various aspects of seismic risk research and assessment, observations and/or data analysis methods within the t-DASH and Short-term Earthquakes Forecast perspectives:
- Studies on time-dependent seismic hazard and risk assessments
- Development of physical/statistical models and studies based on long-term data analyses, including different conditions of seismic activity
- Application of AI to assess earthquake risk factors (hazard, exposure, and vulnerability). Exploring innovative data collection and processing techniques, such as statistical machine learning
- Estimating earthquake hazard and risk across different temporal and spatial scales and assessing the accuracy of these models against available observations
- Earthquake-induced cascading effects such as landslides and tsunamis, and multi-risk assessments
- Studies devoted to the description of genetic models of earthquake’s precursory phenomena
- Infrastructures devoted to maintain and further develop our present observational capabilities of earthquake related phenomena also contributing to build a global multi-parametric Earthquakes Observing System (EQuOS) to complement the existing GEOSS initiative

Database documentation and sharing is a crucial part of the scientific process, and more scientists are choosing to share their data on centralised data repositories. These repositories have the advantage of guaranteeing immutability (i.e., the data cannot change), which is not so amenable to developing living databases (e.g., in continuous citizen science initiatives). At the same time, citizen science initiatives are becoming more and more popular in various fields of science, from natural hazards to hydrology, ecology and agronomy.

In this context, distributed databases offer an innovative approach to both data sharing and evolution. These systems have the distinct advantage of becoming more resilient and available as more users access the same data, and as distributed systems, contrarily to decentralised ones, do not use blockchain technology, they are orders of magnitude more efficient in data storage as well as completely free to use. Distributed databases can also mirror exising data, so that scientists can keep working in their preferred Excel, OpenOffice, or other software while automatically syncing database changes to the distributed web in real time.

This workshop will present the general concepts behind distributed, peer-to-peer systems. Attendees will then be guided through an interactive activity on Constellation, a scientific software for distributed databases, learning how to both create their own databases as well as access and use others' data from the network. Potential applications include citizen science projects for hydrological data collection, invasive species monitoring, or community participation in managing natural hazards such as floods.

Interdisciplinarity is becoming a common approach to solve socio-ecological problems, but datasets from different disciplines often lack interoperability. In this SC we will explore interoperability levels in the context of integrated research infrastructure services for the climate change crisis.

WEkEO offers a single access point to all of the environmental data provided by the Copernicus programme, as well as additional data from its four partner organisations. While data access is the first step for research based on EO data, the challenges of handling data soon become overwhelming with the increasing volume of Earth Observation data available. To cope with this challenge and to tame the Big Earth Data, WEkEO offers a cloud-based processing service for Earth Observation data coming from the Copernicus programme and beyond.
This course will explain new trends and developments in accessing, analysing and visualizing earth observation data by introducing concepts around serverless processing, parallel processing of big data and data cube generation in the cloud.
The session will begin with a theoretical introduction to cloud-based big data processing and data cube generation, followed by a demonstration how the participants can utilize these concepts within the WEkEO environment using its tools. Participants will have the opportunity to apply the concepts and tools in multi-disciplinary environmental use cases bringing together different kinds of satellite data and earth observation products as data cubes in the cloud.
The course will start with a beginner-level introduction and demonstration before introducing more advanced functionalities of the WEkEO services. Prior knowledge of satellite data analysis/Python programming would be an advantage but is not a prerequisite. Comprehensive training material will be provided during the course to ensure that participants with varying degree of knowledge of data processing can follow and participate.

The analysis and visualisation of data is fundamental to research across the earth and space sciences. The Pangeo (https://pangeo.io) community has built an ecosystem of tools designed to simplify these workflows, centred around the Xarray library for n-dimensional data handling and Dask for parallel computing. In this short course, we will offer a gradual introduction to the Pangeo toolkit, through which participants will learn the skills required to scale their local scientific workflows through cloud computing or large HPC with minimal changes to existing codes.
The course is beginner-friendly but assumes a prior understanding of the Python language. We will guide you through hands-on jupyter notebooks that showcase scalable analysis of in-situ, satellite observation and earth system modelling datasets to apply your learning. By the end of this course, you will understand how to:
- Efficiently access large public data archives from Cloud storage using the Pangeo ecosystem of open source software and infrastructure.
- Leverage labelled arrays in Xarray to build accessible, reproducible workflows
- Use chunking to scale a scientific data analysis with Dask
All the Python packages and training materials used are open-source (e.g., MIT, Apache-2, CC-BY-4). Participants will need a laptop and internet access but will not need to install anything. We will be using the free and open Pangeo@EOSC (European Open Sicence Cloud) platform for this course. We encourage attendees from all career stages and fields of study (e.g., atmospheric sciences, cryosphere, climate, geodesy, ocean sciences) to join us for this short course. We look forward to an interactive session and will be hosting a Q&A and discussion forum at the end of the course, including opportunities to get more involved in Pangeo and open source software development. Join us to learn about open, reproducible, and scalable Earth science!
Preparation: We recommend learners with no prior knowledge of Python review resources such as the Software Carpentry training material and Project Pythia in advance of this short course. Participants should bring a laptop with an internet connection. No software installation is required as resources will be accessed online using the Pangeo@EOSC platform. Temporary user accounts will be provided for the course and we will also teach attendees how to request an account on Pangeo@EOSC to continue working on the platform after the training course.

Addressing global environmental and socio-technical challenges requires interdisciplinary, data-driven approaches. Today’s research produces unprecedented volumes and complexity of value-added research data and an increasing number of interactive data services, putting traditional information management systems to the test. Collaborative infrastructures are challenged by their dual role of advancing research and scientific assessments while facilitating transparent data and software sharing.

Since the breakthrough of datacubes as a contributor to Analysis-Ready Data, a series of implementations have been announced, and likewise services. However, often these are described through publications only and without publicly accessible deployments to evaluate.

We invite abstracts from all data stakeholders that highlight innovative platforms, frameworks, datacube tools, services, systems, and initiatives designed to enhance access and usability of data for research on topics such as climate change, natural hazards, sustainable development, etc. We welcome presentations describing collaborations across national and disciplinary boundaries as well as live demos of datacube tools and services that contribute to building trustworthy and interoperable data networks, guided by UNESCO’s Open Science recommendations, the FAIR and CARE data principles. The expected outcome for attendees is to get a realistic overview on the datacube tools, service landscape and ongoing collaborations that enable researchers worldwide to address pressing global problems through data.

Almost a decade ago, the FAIR data guiding principles were introduced to the broader research community. These principles proposed a framework to increase the reusability of data in and across domains during and after the completion of e.g. research projects. In subdomains of the Earth System Sciences (ESS), like atmospheric sciences or partly geosciences, data reuse across institutions and geographical borders was already well-established, supported by community-specific and cross-domain standards like netCDF-CF, geospatial standards (e.g.OGC). Further, authoritative data producers such as CMIPs were already using Persistent Identifiers and corresponding handle systems for data published in their repositories – so it was often thought and communicated this data is “FAIR by design”.

However, fully implementing FAIR principles, particularly machine-actionability—the core idea behind FAIR—has proven challenging. Despite progress in awareness, standard-compliant data sharing, and the automation of data provenance, the ESS community continues to struggle to reach a community-wide consensus on the design, adoption, interpretation and implementation of the FAIR principles.

In this session, we invite contributions from all fields in Earth System Sciences that provide insights, case studies, and innovative approaches to advancing the adoption of the FAIR data principles. We aim to foster a collaborative dialogue on the progress our community has made, the challenges that lie ahead, and the strategies needed to achieve widespread acceptance and implementation of these principles, ultimately enhancing the future of data management and reuse.

We invite contributions focusing on, but not necessarily limited to,
- Challenges and solutions in interpreting and implementing the FAIR principles in different sub-domains of the ESS
- FAIR onboarding strategies for research communities
- Case studies of successful FAIR data implementation (or partial implementation) in ESS at infrastructure and research project level
- Methods and approaches to gauge the impact of FAIR data implementation in ESS
- Considerations on how AI might help to implement FAIR
- Future direction for FAIR data in ESS

Performing research in Earth System Science is increasingly challenged by the escalating volumes and complexity of data, requiring sophisticated workflow methodologies for efficient processing and data reuse. The complexity of computational systems, such as distributed and high-performance heterogeneous computing environments, further increases the need for advanced orchestration capabilities to perform and reproduce simulations effectively. On the same line, the emergence and integration of data-driven models, next to the traditional compute-driven ones, introduces additional challenges in terms of workflow management. This session delves into the latest advances in workflow concepts and techniques essential to address these challenges taking into account the different aspects linked with High-Performance Computing (HPC), Data Processing and Analytics, and Artificial Intelligence (AI).

In the session, we will explore the importance of the FAIR (Findability, Accessibility, Interoperability, and Reusability) principles and provenance in ensuring data accessibility, transparency, and trustworthiness. We will also address the balance between reproducibility and security, addressing potential workflow vulnerabilities while preserving research integrity.

Attention will be given to workflows in federated infrastructures and their role in scalable data analysis. We will discuss cutting-edge techniques for modeling and data analysis, highlighting how these workflows can manage otherwise unmanageable data volumes and complexities, as well as best practices and progress from various initiatives and challenging use cases (e.g., Digital Twins of the Earth and the Ocean).

We will gain insights into FAIR Digital Objects, (meta)data standards, linked-data approaches, virtual research environments, and Open Science principles. The aim is to improve data management practices in a data-intensive world.
On these topics, we invite contributions from researchers illustrating their approach to scalable workflows as well as data and computational experts presenting current approaches offered and developed by IT infrastructure providers enabling cutting edge research in Earth System Science.

In recent decades, the advent in geoinformation technology has played an increasingly important role in determining various parameters that characterize the Earth's environment. These technologies often combined with conventional field surveying and spatial data analysis methods and/or simulation process models provide efficient means for monitoring and understanding Earth’s environment in a cost-effective and systematic manner. This session invites contributions focusing on modern open-source software tools developed to facilitate the analysis of mainly geospatial data in any branch of geosciences for the purpose of better understanding Earth’s natural environment. We encourage the contribution of any kind of open source tools, including those that are built on top of global used commercial GIS solutions. Potential topics for the session include the presentation of software tools developed for displaying, processing and analysing geospatial data and modern cloud webGIS platforms and services used for geographical data analysis and cartographic purposes. We also welcome contributions that focus on presenting tools that make use of parallel processing on high performance computers (HPC) and graphic processing units (GPUs) and also on simulation process models applied in any field of geosciences.

Code is read far more often than it's written, yet some still believe that complex, unreadable code equates to a better algorithm. In reality, the opposite is true. Writing code that not only works but is also clear, maintainable, and easy to modify can significantly reduce the cognitive load of coding, freeing up more time for scientific research. This short course introduces essential programming practices, from simple yet powerful techniques like effective naming, to more advanced topics such as unit testing, version control, and managing virtual environments. Through real-life examples, we will explore how to transform code from convoluted to comprehensible.

Software plays a pivotal role in various scientific disciplines. Research software may include source code files, algorithms, computational workflows, and executables. It refers mainly to code meant to produce data, less so, for example, plotting scripts one might create to analyze this data. An example of research software in our field are computational models of the environment. Models can aid pivotal decision-making by quantifying the outcomes of different scenarios, e.g., varying emission scenarios. How can we ensure the robustness and longevity of such research software? This short course teaches the concept of sustainable research software. Sustainable research software is easy to update and extend. It will be easier to maintain and extend that software with new ideas and stay in sync with the most recent scientific findings. This maintainability should also be possible for researchers who did not originally develop the code, which will ultimately lead to more reproducible science.

This short course will delve into sustainable research software development principles and practices. The topics include:
- Properties and metrics of sustainable research software
- Writing clear, modular, reusable code that adheres to coding standards and best practices of sustainable research software (e.g., documentation, unit testing, FAIR for research software).
- Using simple code quality metrics to develop high-quality code
- Documenting your code using platforms like Sphinx for Python

We will apply these principles to a case study of a reprogrammed version of the global WaterGAP Hydrological Model (https://github.com/HydrologyFrankfurt/ReWaterGAP). We will showcase its current state in a GitHub environment along with example source code. The model is written in Python but is also accessible to non-python users. The principles demonstrated apply to all coding languages and platforms.

This course is intended for early-career researchers who create and use research models and software. Basic programming or software development experience is required. The course has limited seats available on a first-come-first-served basis.

Python is one of the fastest growing programming languages and has moved to the forefront in the earth system sciences (ESS), due to its usability, the applicability to a range of different data sources and, last but not least, the development of a considerable number of ESS-friendly and ESS-specific packages.

This interactive Python course is aimed at ESS researchers who are interested in adding a new programming language to their repertoire. Except for some understanding of fundamental programming concepts (e.g. loops, conditions, functions etc.), this course presumes no previous knowledge of and experience in Python programming.

The goal of this course is to give the participants an introduction to the Python fundamentals and an overview of a selection of the most widely-used packages in ESS. The applicability of those packages ranges from (simple to advanced) number crunching (e.g. Numpy), to data analysis (e.g. Xarray, Pandas) to data visualization (e.g. Matplotlib).

The course wi