Modern challenges of climate change, disaster management, public health and safety, resources management, and logistics can only be addressed through big data analytics. A variety of modern technologies are generating massive volumes of conventional and non-conventional geospatial data at local and global scales. Most of this data includes geospatial data components and are analysed using spatial algorithms. Ignoring the geospatial component of big data can lead to an inappropriate interpretation of extracted information. This gap has been recognised and led to the development of new spatiotemporally aware strategies and methods.
This session discusses advances in spatiotemporal machine learning methods and the softwares and infrastructures to support them.

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 and feature representations in observations or across models and observations
- Hybrid models (physically informed ML, emulation, data-model integration)
- Novel detection and attribution approaches
- Probabilistic modelling and uncertainty quantification
- Explainable AI applications to climate data science and climate modelling
- Distributional robustness, transfer learning and/or out-of-distribution generalisation tasks in climate science

Please note that a companion session “ML for Earth System modelling” focuses specifically on ML for model improvement, particularly for near-term time-scales (including seasonal and decadal) forecasting, and related abstracts should be submitted there.

The increasing amount of data from an increasing number of spacecraft in our solar system shouts out for new data analysis strategies. There is a need for frameworks that can rapidly and intelligently extract information from these data sets in a manner useful for scientific analysis. The community is starting to respond to this need. Machine learning, with all of its different facets, provides a viable playground for tackling a wide range of research questions in planetary and heliospheric physics.

We encourage submissions dealing with machine learning approaches of all levels in planetary sciences and heliophysics. The aim of this session is to provide an overview of the current efforts to integrate machine learning technologies into data driven space research, to highlight state-of-the art developments and to generate a wider discussion on further possible applications of machine learning.

A Digital Twin of the Earth (DTE) is an interactive, dynamic digital replica of our planet that combines observations with simulations from physical models and advanced AI-based analysis. It aims to replicate the Earth's complex ecosystem, allowing us to estimate our planet’s response to changes under both the current climate state and future climate projections. A DTE is an emerging concept that is capable of simulating what-if scenarios before they occur, which is crucial for natural hazard mitigation and adaptation plans (e.g., floods, heatwaves, wildfires, droughts, etc.). A DTE calls for an advanced, federated, multi-computing architecture that works across organizations and agency boundaries. For a DTE to be successful, it needs to be open source and developed for and by the community. It needs to be an extendable framework that encourages continuous contributions, integration, and development of advanced AI solutions in order to have an accurate digital representation of the physical environment. This session welcomes presentations on current open-source frameworks and enabling technologies for DTE including, but not restricted to:
• Open-source DTE framework
• Computer infrastructure to move virtual data from DTE repositories to service platforms
• Surrogate models for missing observations and unresolved physical processes
• Hybrid AI / physics-based modelling
• Extreme value predictions
• Uncertainty quantification and representation
• Post-processing (event detection and downscaling/super-resolution)

Proper characterization of uncertainty remains a major research and operational challenge in Environmental Sciences, and is inherent to many aspects of modelling impacting model structure development; parameter estimation; an adequate representation of the data (inputs data and data used to evaluate the models); initial and boundary conditions; and hypothesis testing. To address this challenge, methods for a) uncertainty analysis (UA) that seek to identify, quantify and reduce the different sources of uncertainty, as well as propagating them through a system/model, and b) the closely-related methods for sensitivity analysis (SA) that evaluate the role and significance of uncertain factors (in the functioning of systems/models), have proved to be very helpful.

This session invites contributions that discuss advances, both in theory and/or application, in methods for SA/UA applicable to all Earth and Environmental Systems Models (EESMs), which embraces all areas of hydrology, such as classical hydrology, subsurface hydrology and soil science.

Topics of interest include (but are not limited to):
1) Novel methods for effective characterization of sensitivity and uncertainty
2) Analyses of over-parameterised models enabled by AI/ML techniques
3) Single- versus multi-criteria SA/UA
4) Novel approaches for parameter estimation, data inversion and data assimilation
5) Novel methods for spatial and temporal evaluation/analysis of models
6) The role of information and error on SA/UA (e.g., input/output data error, model structure error, parametric error, regionalization error in environments with no data etc.)
7) The role of SA in evaluating model consistency and reliability
8) Novel approaches and benchmarking efforts for parameter estimation
9) Improving the computational efficiency of SA/UA (efficient sampling, surrogate modelling, parallel computing, model pre-emption, model ensembles, etc.)

Understanding Earth’s system’s natural processes, especially in the context of global climate change, has been recognized globally as a very urgent and central research direction which needs further exploration. With the launch of new satellite platforms with a high revisit time, combined with the increasing capability for collecting repetitive ultra-high aerial images through unmade aerial vehicles, the scientific community have new opportunities for developing and applying new image processing algorithms to solve old and new environmental issues.

The purpose of the proposed session is to gather scientific researchers related to this topic aiming to highlight ongoing research and new applications in the field of satellite and aerial time-series imagery. The session focus is on presenting studies aimed at the development or exploitation of novel satellite times series processing algorithms and applications to different types of remote sensing data for investigating longtime processes in all branches of Earth (sea, ice, land, atmosphere).

The conveners encourage both applied and theoretical research contributions focusing on novel methods and applications of satellite and aerial time-series imagery in all geosciences disciplines, including both aerial and satellite platforms (optical and SAR) and data acquired in all regions of the electromagnetic spectrum.

Among various fields of exploration in artificial intelligence (AI), the availability of high-quality training datasets is an exciting area of research that holds great potential to make accurate predictions or perform a desired task. Training data is the initial dataset used to train machine learning algorithms and models. Training data is also known as training dataset (TDS), learning set, and training set. The goal of this session is:
1) to discuss the cutting-edge topics of machine learning training data for the geospatial community;
2) to describe the spatial, temporal and thematic representativeness of TDS and their uncertainties;
3) to focus on sharing and reusability of TDS to increase the adaptation of TDS for geospatial analysis.

This session will focus on the following topics around training datasets:
-How to describe a training dataset to enable efficient re-use in ML/AI applications?
-What are the main characteristics of the training dataset, and what additional information needs to be provided to sufficiently understand the privacy, nature and usability of the dataset?
-Exploring the effect of training data accuracy level, uncertainty of the measurement, labelling procedure used to generate the training data, original data used to create labels, external classification schemes for label semantics, e.g. ontologies or vocabularies;
-What metadata is required, recommended, or optionally provided?
-How to express the quality of a TDS? Is it possible to auto-generate quality indicators?
-Evaluating the effect of training data size, spatial resolution and structure, temporal resolution and currency, the spectral resolution of imagery used for annotation, and annotating accuracy.
-Methods for documenting, storing, evaluating, publishing, and sharing the training datasets;
-Transfer learning and impact of combining various training datasets;
-Open standards and open source training datasets;
-How to enable FAIR (findable, accessible, interoperable and reusable) data principles to be at the heart of future TDS standardization.

Machine learning (ML) applied to earth observation (EO) data provides an ample source to distill insights about our planet and societal activities. Typically, such investigations run as scientific research projects or as industrial proof-of-concept studies with significant manual interaction. In practice, corresponding solutions operate on a local or regional scale considering individual events or limited time periods. Advancing platform technologies and adherence to Open Science principles to enable scalable and reproducible workflows of high complexity are key to drive innovation in EO science and applications.

In our session presenters discuss the design of platforms and methods to scale-up and develop end- to-end repeatable, reusable and/or reproducible ML-model workflows based on multi-modal EO data to global and real-time services. These methods support the organization of input data, the efficient model training, continuous evaluation & testing, and deployment for federated operations on hybrid compute systems.

In particular, the following five topics will be addressed:
1. big geospatial data hubs for efficient preparation of analysis-ready data and features, 2. large-scale ML training on high-performance computing and cloud infrastructure,
3. frameworks for ML-operations at global scale considering complex workflows and hybrid systems,
4. reusability and reproducibility of complex EO-based workflows across platforms, as well as
5. big geospatial data and GeoML-model federation to reach maximal scale by efficient data sharing and model training & inference across institutions around the globe.

We target to not exclusively provide insights into frameworks and methods, but will also discuss the challenges faced en route from research experiments to a successfully integrated, real-time and global service.

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. Abstracts are solicited 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) Integrating deep learning with process-based models and/or physical understanding.
(3) Improving understanding of the (internal) states/representations of deep learning models.
(4) Understanding the reliability of deep learning, e.g., under non-stationarity.
(5) Deriving scaling relationships or process-related insights with deep learning.
(6) Modeling human behavior and impacts on the hydrological cycle.
(7) Extreme event analysis, detection, and mitigation.
(8) Natural Language Processing in support of models and/or modeling workflows.

Unsupervised, supervised, semi-supervised as well as reinforcement learning are now increasingly used to address Earth system-related challenges for the atmosphere, the ocean, the land surface, or the sea ice.
Machine learning could help extract information from numerous Earth System data, such as in-situ and satellite observations, as well as improve model prediction through novel parameterizations or speed-ups. This session invites submissions spanning modeling and observational approaches towards providing an overview of state-of-the-art applications of these novel methods for predicting and monitoring the Earth System from short to decadal time scales. This includes (but is not restricted to):
- The use of machine learning to reduce or estimate model uncertainty
- Generate significant speedups
- Design new parameterization schemes
- Emulate numerical models
- Fundamental process understanding

Please consider submitting abstracts focused on ML applied to observations and modeling of the climate and its constituent processes to the companion "ML for Climate Science" session.

Artificial intelligence (in particular, machine learning) can be used to predict and respond to natural disasters. The ITU/WMO/UNEP Focus Group AI for Natural Disaster Management (FG-AI4NDM) is building a community of experts and stakeholders to identify best practices in the use of AI for data processing, improved modeling across spatiotemporal scales, and providing effective communication. This multidisciplinary FG-AI4NDM-session invites contributions addressing challenges and opportunities related to the use of AI for the detection, forecasting, and communication of natural hazards and disasters. In particular, it welcomes presentations highlighting innovative approaches to data collection (e.g., via sensor networks), data handling (e.g., via automating annotation), data storage and transmission (e.g., via edge- and cloud computing), novel modeling or explainability methods (e.g., integrating quantum computing methods), and outcomes of operational implementation.

Landslides, debris flows and avalanches are common types of unsteady bulk mass movements. Globally, the risk from these mass movements is expected to increase, due to changes in precipitation patterns, rising average temperatures and continued urbanisation of mountainous regions. Climate change also reduces the power of site-specific empirically-based predictions, requiring updated approaches for effective and robust management of the associated risk.

Given sustained improvements in computational power, the techniques involving artificial intelligence and explicit hydromechanical modelling are becoming more and more widespread. Both techniques have the advantages of reducing our dependence on empirical approaches. This session thus covers two main domains:

1) New approaches and state-of-the-art artificial intelligence techniques on remote sensing data for creating and updating landslide inventories.
2) Advances in hydromechanical numerical models and digital tools for geophysical mass flows.

The ultimate goal of both is integration into the wider context of hazard and/or risk assessment and mitigation.

Contributions to this session may involve:
(a) Regional scale analysis for landslide detection and applications for establishing multi-temporal inventories.
(b) Data processing, fusion, and data manipulation, as well as novel AI model tuning practices.
(c) Evaluating the quality of landslide detection through AI techniques.
(d) Comparing the performance of different AI segmentation models.
(e) Novel constitutive and hydromechanical modelling of flows, both at the field- and laboratory-scales.
(f) Hydromechanical modelling of the interaction of mass movements with structural countermeasures.
(g) Advances in risk analysis through the integration of digital technologies and multidisciplinary viewpoints (potentially including combining AI and hydromechanical modelling techniques).

Landslide susceptibility, the spatial likelihood of occurrence of landslides, is the subject of countless scientific publications. They use heterogeneous data, and apply many different methods, mostly falling under the definition of statistical and/or machine learning with the common feature of considering many input variables and a single target output, denoting landslide presence. It is a classification problem: given N input variables assuming different values, each combination associated with a 0/1 possible outcome, a model should be trained on some dataset, tested, and eventually it applied to unseen data.
Relevant input data (“predictors”, “factors”, “independent variables”) is usually a mixed set of topographic, morphometric, environmental, climatic, and a landslide inventory. Choice of a specific method depends on software availability, personal background, and existence of relevant literature in the area of interest. New methods are proposed regularly and very often is difficult to judge their relative performance based with respect to existing methods.
A meaningful comparison of many different methods would require a common dataset – a benchmark - to train and test each of them in a systematic way. This is a standard procedure in machine learning science and practice, for virtually all the fields: benchmark datasets exist for medical sciences, image recognition, linguistics, and in general any classification algorithm. The “Iris dataset” is a famous example of a benchmark in classification of numerical data into three different variants of the flower Iris. This session aims at establishing one or more benchmark datasets that could be helpful in landslide susceptibility research, to compare the plethora of existing methods and new methods to come.
We propose an interactive session: the organizers will single out benchmark datasets, share them with participants at due time, prior to the conference venue. We expect abstract proposals to describe the method(s) they intend to apply, the type of data it requires, and an independent case study for which the method proved successful. Participants should be ready to disclose minimal computer code (in any major programming language) to run their method, to apply the code to the benchmark dataset prior to the conference, and present their results. We aim at collecting all of the results in a journal publication, including datasets, benchmark and computer codes in collaboration with the participants.
Download dataset at: http://dx.doi.org/10.31223/X52S9C

Geophysical methods have great potential for the characterization of subsurface properties and processes to inform geological, reservoir, hydrological, and biogeochemical studies. In these contexts, the classically used geophysical tools only provide indirect information about the characteristics and heterogeneities of subsurface rocks and their associated processes (e.g., flow, transport, biogeochemical reactions). Petrophysical relationships hence have to be developed to provide links between physical properties (e.g. electrical conductivity, seismic velocity, or attenuation) and the intrinsic parameters of interest (e.g. fluid content, hydraulic properties, pressure conditions). In addition, geophysical methods are increasingly deployed in increasingly complex environments for time-lapse, continuous, and distributed monitoring. Here again, there is a great need for accurate and efficient physical relationships such that geophysical data can be correctly interpreted (e.g., included in fully coupled inversions). Establishing such petrophysical models requires multidisciplinary approaches and diverse theoretical frameworks. While each physical property has its own intrinsic dependence on pore-scale interfacial, geometrical, and biogeochemical properties or on external conditions such as pressure or temperature, each associated geophysical method also has its own specific investigation depth and spatial resolution. Such complexity poses great challenges in combining theoretical developments with laboratory validations and scaling laboratory observations to field practices. This session consequently invites contributions from various communities to share their models, their experiments, or their field tests and data in order to discuss multidisciplinary ways to advance petrophysical relationship development and to improve our knowledge of complex processes in the subsurface.

Non-destructive testing (NDT) methods are employed in a variety of engineering and geosciences applications and their stand-alone use has been greatly investigated to date. New theoretical developments, technological advances and the progress achieved in surveying, data processing and interpretation have in fact led to a tremendous growth of the equipment reliability, allowing outstanding data quality and accuracy.

Nevertheless, the requirements of comprehensive site and material investigations may be complex and time-consuming, involving multiple expertise and equipment. The challenge is to step forward and provide an effective integration between data outputs with different physical quantities, scale domains and resolutions. In this regard, enormous development opportunities relating to data fusion, integration and correlation between different NDT methods and theories are to be further investigated.

This Session primarily aims at disseminating contributions from state-of-the-art NDT methods and new numerical developments, promoting the integration of existing equipment and the development of new algorithms, surveying techniques, methods and prototypes for effective monitoring and diagnostics. NDT techniques of interest are related–but not limited to–the application of acoustic emission (AE) testing, electromagnetic testing (ET), ground penetrating radar (GPR), geoelectric methods (GM), laser testing methods (LM), magnetic flux leakage (MFL), microwave testing, magnetic particle testing (MT), neutron radiographic testing (NR), radiographic testing (RT), thermal/infrared testing (IRT), ultrasonic testing (UT), seismic methods (SM), vibration analysis (VA), visual and optical testing (VT/OT).

The Session will focus on the application of different NDT methods and theories and will be related –but not limited to– the following investigation areas:
- advanced data fusion;
- advanced interpretation methods;
- design and development of new surveying equipment and prototypes;
- real-time & remote assessment and monitoring methods for material and site inspection (real-life and virtual reality);
- comprehensive and inclusive information data systems for the investigation of survey sites and materials;
- numerical simulation and modelling of data outputs with different physical quantities, scale domains and resolutions;
- advances in NDT methods, numerical developments and applications (stand-alone use of existing and state-of-the-art NDTs).

The session gathers multi-disciplinary geoscientific aspects such as dynamics, reactions, and environmental/health consequences of radioactive materials that are massively released accidentally (e.g., Chernobyl and Fukushima nuclear power plant accidents, wide fires, etc.), future potential risk of leakage (e.g., Zaporizhzhia nuclear power plant) and by other human activities (e.g., nuclear tests).

The radioactive materials are known as polluting materials that are hazardous for human society, but are also ideal markers in understanding dynamics and physical/chemical/biological reactions chains in the environment. Therefore, man-made radioactive contamination involves regional and global transport and local reactions of radioactive materials through atmosphere, soil and water system, ocean, and organic and ecosystem, and its relations with human and non-human biota. The topic also involves hazard prediction, risk assessment, nowcast, and countermeasures, , which is now urgent important for the nuclear power plants in Ukraine.

By combining long monitoring data (> halftime of Cesium 137 after the Chernobyl Accident in 1986, 12 years after the Fukushima Accident in 2011, and other events), we can improve our knowledgebase on the environmental behavior of radioactive materials and its environmental/biological impact. This should lead to improved monitoring systems in the future including emergency response systems, acute sampling/measurement methodology, and remediation schemes for any future nuclear accidents.

The following specific topics have traditionally been discussed:
(a) Atmospheric Science (emissions, transport, deposition, pollution);
(b) Hydrology (transport in surface and ground water system, soil-water interactions);
(c) Oceanology (transport, bio-system interaction);
(d) Soil System (transport, chemical interaction, transfer to organic system);
(e) Forestry;
(f) Natural Hazards (warning systems, health risk assessments, geophysical variability);
(g) Measurement Techniques (instrumentation, multipoint data measurements);
(h) Ecosystems (migration/decay of radionuclides).

The session consists of updated observations, new theoretical developments including simulations, and improved methods or tools which could improve observation and prediction capabilities during eventual future nuclear emergencies. New evaluations of existing tools, past nuclear contamination events and other data sets also welcome.

Europe’s green transition and response to environmental challenges will depend on a parallel digital transition, supporting decision-makers and actors with fit-for-purpose digital technologies and assets. The EU’s “European strategy for data” and Data Governance Act identify data spaces as the instruments to achieve a single market for data, global competitiveness and data sovereignty through “a purpose- or a sector-specific or cross-sectoral interoperable framework of common standards and practises to share or jointly process data”. Shortly, European data spaces will ensure that more data becomes available for use in research, economy and society while keeping data rights-holders in control.

Several projects and initiatives are building thematic data spaces, allowing researchers, industries, and governments to access high-quality, interoperable data and related services from multiple providers and giving data holders and providers tools to manage, control and provide access to their data. The benefits of data spaces for a FAIR data ecosystem and potential users are clear, but a deeper understanding of the design, set up and evolution of data spaces is needed.

This session seeks contributions from any group, project or initiative that has established or is establishing a data space in the context of environmental and Earth sciences. Talks in this session should provide a general overview of the designing, building, running, and governing data spaces and share best practices in this respect. Practical use cases on different levels (regional, national, European or global) demonstrating the value of data spaces for access, combining data from various sources and flexible environmental/Earth system data processing are also welcome including initiatives bridging into data spaces such as Destination Earth Data Lake. Finally, we welcome presentations from projects and initiatives working to consolidate the complex landscape of different data ecosystems, both within and beyond environmental and Earth sciences.

Several projects and initiatives are building thematic data spaces, allowing researchers, industries, and governments to access high-quality, interoperable data and related services from multiple providers and giving data holders and providers tools to manage, control and provide access to their data. The benefits of data spaces for a FAIR data ecosystem and potential users are clear, but a deeper understanding of the design, set up and evolution of data spaces is needed.

This session seeks contributions from any group, project or initiative that has established or is establishing a data space in the context of environmental and Earth sciences. Talks in this session should provide a general overview of the designing, building, running, and governing data spaces and share best practices in this respect. Practical use cases on different levels (regional, national, European or global) demonstrating the value of data spaces for access, combining data from various sources and flexible environmental/Earth system data processing are also welcome including initiatives bridging into data spaces such as Destination Earth Data Lake. Finally, we welcome presentations from projects and initiatives working to consolidate the complex landscape of different data ecosystems, both within and beyond environmental and Earth sciences.

Research data infrastructures (RDIs) aim to manage and share research products and metadata systematically to enable research across all scales and disciplinary boundaries. Their services support researchers throughout the entire research lifecycle, especially during data management and collaborative analysis, and they foster FAIRness and openness, e.g., by applying established standards for metadata, data, and/or scientific workflows. Through their offerings and services, RDIs can shape research practices and are strongly connected with the communities of users that identify and associate themselves with them.

Naturally, the potential of RDIs faces many challenges. Even though it is clear that RDIs are indispensable for solving big societal problems, their wide adoption requires a cultural change within research communities. At the same time RDIs themselves must be developed further to serve user needs. And, also at the same time, the sustainability of RDIs must be improved, international cooperation increased, and duplication of development efforts must be avoided. To be able to provide a community of diverse career stages and backgrounds with a convincing infrastructure that is established beyond national and institutional boundaries, new collaboration patterns and funding approaches must be tested so that RDIs foster cultural change in academia and be a reliable foundation for FAIR and open research. This needs to happen while academia struggles with improving researcher evaluation, with a continuing digital disruption, with enhancing scholarly communication, and with diversity, equity, and inclusion.

In the Earth System Sciences (ESS), several research data infrastructures and components are currently developed on different regional and disciplinary scales, all of which face these challenges at some level. This session provides a forum to exchange methods, stories, and ideas to enable cultural change and international collaboration in scientific communities, to bridge the gap between user needs, and to build sustainable software solutions.

Cloud computing has emerged as the dominant paradigm, supporting practically all industrial applications and a significant number of academic and research projects. Since its introduction in the early 2010s and its widespread adoption thereafter, migration to cloud computing has been a considerable task for many organisations and companies. Processing of big data close to their physical location is a perfect use case for cloud technologies and cloud storage infrastructure which offer all the necessary infrastructure and tools, especially if cloud infrastructure is offered together with HPC resources.
Pangeo (pangeo.io) is a global community of researchers and developers that tackle big geoscience data challenges in a collaborative manner using HPC and Cloud infrastructure.
This session's aim is threefold:
(1) Focuses on Cloud/Fog/Edge computing use cases and aims to identify the status and the steps towards a wider cloud computing adoption in Earth Observation and Earth Modeling.
(2) to motivate researchers that are using or developing in the Pangeo ecosystem to share their endeavors with a broader community that can benefit from these new tools.
(3) to contribute to the Pangeo community in terms of potential new applications for the Pangeo ecosystem, containing the following core packages: Xarray, Iris, Dask, Jupyter, Zarr, Kerchunk and Intake.
We encourage contributions describing all kinds of Cloud/Fog/Edge computing efforts in Earth Observation and Earth Modeling domains, such as:
- Cloud Applications, Infrastructure and Platforms (IaaS, PaaS SaaS and XaaS).
- Cloud federations and cross domain integration
- Service-Oriented Architecture in Cloud Computing
- Cloud Storage, File Systems, Big Data storage and Management.
- Networks within Cloud systems, the Storage Area, and to the outside
- Fog and Edge Computing
- Operational systems on the cloud.
- Data lakes and warehouses on the cloud.
- Cloud computing and HPC convergence in EO data processing.
Also presentations using at least one of Pangeo’s core packages in any of the following domains:
- Atmosphere, Ocean and Land Models
- Satellite Observations
- Machine Learning
- And other related applications
We welcome any contributions in the above themes presented as science-based in other EGU sessions, but more focused on research, data management, software and/or infrastructure aspects. For instance, you can showcase your implementation through live executable notebooks.

In the big-data era, science attempts to address global societal problems (such as climate change, pandemics, environmentally sustainable exploitation of our resources) often with computationally expensive advanced methodologies that require machine-to-machine (m2m) interaction and interoperable data and services. At the same time, data science and technologies are rapidly evolving and offer new tools to address scientific demands, revealing yet more technological challenges as the complexity increases at all stages of the data and service development lifecycle.

To serve the complex, heterogeneous and diverse demands of their end-users, data providers try to work on federated solutions using FAIR enabling resources. Scientists and developers investigate the requirements for integrating systems that operate not necessarily using the same technologies but rather adopting technical solutions that facilitate the interoperability among them. Starting at the very low levels of data (e.g. near real-time, minimally processed, etc) and reaching higher-level derivative data collections, products and services, many issues are still open: metadata models, authentication and authorisation systems, ontologies, machine actionable licenses and PIDs that support the findability, accessibility and sustainable future of data (enabling proper citation and attribution to both creators and funders, usage tracking, etc.).

Groups facilitating global data sharing, networks and services are e.g. Federation of Digital Seismograph Networks (FDSN), OneGeology, OneGeochemistry, WorldFAIR, Earth Systems Grid Federation (ESFG), OGC, W3C, GEO, CODATA/DDI Cross-Domain Data Initiative), the European Open Science Cloud (EOSC) and the paneuropean cluster of environmental research infrastructures (ENVRI).

This session brings together diverse disciplines and a variety of experts (data centre architects, data stewards, developers, ontologists, scientists). We seek contributions that demonstrate scientific use cases (in the field of earth/environmental sciences), discuss ICT and data challenges and recommend best practices based on experience from interoperability frameworks. We welcome abstracts from small scale scientific methodologies developed by (multi)disciplinary groups aiming at multi/inter-disciplinary science to larger scale integrated platforms that offer interoperational data and services by multiple discipline-focused providers and/or cross-domain communities of providers

Destination Earth (DestinE) is an ambitious initiative of the European Union aiming to develop – on a global scale - a highly accurate digital model of the Earth that will help understand and simulate the evolution in the state of our planet, better predict impact on human system processes, ecosystem processes and their interaction.

DestinE will exploit state-of-the-art technologies, including high-performance computing, high-resolution Earth system models and novel approaches in analytics, including artificial intelligence, and offer unprecedented interactivity with the system for users.

A number of tangible outcomes are expected from these developments: Earth system simulations will become more skillful, the intensity and magnitude of natural disasters will be predicted more reliably, decision makers will have tools to tackle more efficiently the effect of climate change and much more.

Work is currently ongoing by the three implementing agencies (ESA, EUMETSAT, ECMWF) to develop the three components of the DestinE system: the Core Service Platform, the Data Lake and the Digital Twin Engine. This session aims at presenting progress towards the implementation of the DestinE system. It will also highlight opportunities to contribute to this challenging and ambitious endeavor and co-evolve the system together.

Need for Smart Solutions in earth, environmental and planetary sciences: Tackling data challenges and incorporating applied earth and planetary sciences into artificial intelligence (AI) models opened a new avenue for creating comprehensive methodologies and strategies to answer a wide variety of theoretical and practical questions from detecting, modelling, interpreting and predicting changes in the earth and environment’s ecosystems in response to climate change to understanding interactions among the ocean, atmosphere, and land in the climate system. Therefore, AI and Data Science (DS) in earth, environmental and planetary sciences are one of the fastest growing areas. The performance of the AI/DS models improves as it gains experience over time. Various mathematical and statistical models need to be investigated to determine the performance of AI models. Once the learning process is completed, then the model can then be used to make an assumption, classify and test data. This is achieved after gaining experience in the training process. This session aims to make available to the world community of earth, environment and planetary sciences-related professionals a collection of scientific papers on the current state of the art and recent developments of AI and DS applications in the field. This session will shed light on many recent research activities on applying AI/DS techniques into a single comprehensive document to address engineering, social, political, economic, safety, health, and technological issues of earth, environment and planetary sciences challenges and opportunities. The purpose of this session is to improve and facilitate the application of intelligent systems for the earth, environmental and planetary sciences to highlight new insight for creating comprehensive methodologies for analyzing/processing/predicting/management strategies in the fields of fundamental and applied sciences problems through the decision-making abilities of artificial intelligence and machine learning techniques.

Visualisation of scientific data is an integral part of scientific understanding and communication. Scientists have to make decisions about the most effective way to communicate their results everyday. How do we best visualise the data to understand it ourselves? How do we best visualise our results to communicate with others? Common pitfalls can be overcrowding, overcomplicated plot types or inaccessible color schemes. Scientists may also get overwhelmed by the graphics requirements of different publishers, for presentations, posters etc. This short course is designed to help scientists improve their data visualization skills in a way that the research outputs would be more accessible within their own scientific community and reach a wider audience.
Topics discussed include:
- Choosing a plot type – keeping it simple
- Color schemes – which ones to use or not to use
- Creativity vs simplicity – finding the right balance
- Producing your figures and maps – software and tools
- Figure files – publication ready resolutions
This course is co-organized by the Young Hydrologic Society (YHS), enabling networking and skill enhancement of early career researchers worldwide. Our goal is to help you make your figures more accessible by a wider audience, informative and beautiful. If you feel your graphs could be improved, we welcome you to join this short course.

Science impacts human society in many ways but of
particular importance is the application of scientific
results to the design of forecasting systems.
Forecasting systems are indispensable for making
informed decisions under risk. Informative and reliable
weather forecasts for instance help to better prepare
for or to reduce the exposure to adverse weather.
Therefore, there is a need for an objective and well
understood framework for ``forecast verification'',
i.e. qualitative and quantitative assessment of
forecast performance.

Statistical methods compare historical forecasts with
corresponding verifications, indicating whether the
forecasting system behaved significantly different (in
a statistical sense) from what was expected.

This short course will introduce the participants to
the fundamentals of statistical forecast verification.
Some necessary statistical theory will be discussed as well, and some hands-on numerical experiments will take place using freely available code. More specifically, the course will cover the following topics (more or less in that order)

* Forecast types and scoring rules
* Tests and p-values
* How to cope with dependent data
* How to cope with forecasts of spatial fields
* Code, literature, and further resources

Target audience are researchers (both from academic institutions and operational centres) who are either new to forecast verification or who have practical experience but want to know more about the theory. The course is NOT restricted to atmospheric forecasts, nor exclusively to the assessment of operational forecasting systems. The discussed methods are applicable in many other fields such as parameter estimation, data assimilation, model evaluation, and machine learning.

Humans have been successfully mapping the remotest and most inhospitable places on Earth, and the surfaces and interiors of other planets and their moons at highest resolution. The remaining blank spots are located in areas that are hardly accessible either through field surveys, geophysical methods or remote sensing due to technical and/or financial challenges. Some of these places are key areas that would help to reveal geologic history, or provide access to future exploration endeavors.

Such extreme and remote locations are commonly associated with the ocean floor, or planetary surfaces, but these extreme worlds might also be found in hot deserts, under the ice, in high-mountain ranges, in volcanic edifices, hidden underneath dense canopy cover, or located within the near-surface crust. All such locations are prime targets for remote sensing mapping in a wider sense. The methodological and technical repertoire to investigate extreme and remote locations is thus highly specialized and despite different contexts there are commonalities not only with respect to technical mapping approaches, but also in the way how knowledge is gathered and assessed, interpreted and vizualized regarding its scientific but also its economic value.

This session invites contributions to this field of geologic mapping and cartography of extreme (natural) environments with a focus on the scientific synthesis and extraction of information and knowledge.

A candidate contribution might cover, but is not limited to, topics such as:

- ocean mapping using manned and unmanned vehicles and devices,
- offshore exploration using remote sensing techniques,
- crustal investigation through drilling and sampling,
- subsurface investigation using radar techniques,
- planetary geologic and geophysical mapping,
- geologic investigation of desert environments,
- subglacial geologic mapping...

The aim of this session is to bring together researchers mapping geologically and geophysically inaccessible environments, thus relying on geophysical and remote sensing techniques as single source for collecting data and information. We would like to keep the focus on geologic and geophysical mapping of spots for which we have no or only very limited knowledge due to the harsh environmental conditions, and we would thus exclude areas that are inaccessible for political reasons.

The session aims to focus on the implementation of GEO data sharing and management and FAIR principles to In Situ Data for environmental and climate purposes. This effort also benefits to the implementation of the UN Sustainable Development Goals (SDG) and other global challenges.
In the context of Earth Observation, common understanding exists regarding the application of FAIR principles and GEO Data Sharing and Management principles for remote sensing satellite data, while in-situ data still faces issues related to its fragmentation, legal, organizational, technical aspects including a lack of standardization, harmonization, and integration to better solve the global challenges the planet and our society. It is then a major bottleneck to upscale from local to global. Global initiatives such as GEO, its work on Essential Variables and the European component EuroGEO; the Copernicus programme through the activities of the In-Situ Component; the activities conducive to the Common European Green Deal data space; and the work around the UN SDGs are providing important approaches to this aim.
In this context, topics as:
• Promoting free, full, open and timely access to in situ data; implementation of the GEO Data Sharing and Management principles;
• Methodological approaches to collect and manage in situ data requirements, including the application of the Essential Variables concept as a potential common framework;
• Pushing in situ data and its fit-for-purpose as a driver to define data collections: temporal and spatial resolution, thematic accuracy, model harmonisation and interoperability, formats, spatial and temporal coverage;
• Challenges in in situ data accessibility and re-usability (including both technical and licensing aspects);
• In-situ data gap analysis and best practices and recommendations

will be appreciated.

Research data repositories can store a variety of research outputs, ranging from raw observational data derived from monitoring infrastructures including satellites, drones, aircraft, remote sensors, in situ analytical laboratories, etc. to downstream derived products from scientific projects and the full suite of small and highly-variable data from the long-tail communities as well as sample descriptions. Repositories play an important part in the support of open data and open science, whilst their curation of datasets, together with operation and continuous development of access services, are key to making their data assets FAIR. Many data and sample repositories and their data provider communities are continually working on improving dataset identification and attribution, versioning control, as well as provenance recording and tracking, and improving and enriching relevant metadata records.

Traditionally, data centers monitored their usage by logging access to data sets (number, volume, success/fail rate) and used connection IP addresses for rough geographic differentiation. That usage data, together with user feedback received through other channels, supported data center governance and was sufficient for reporting to funders and other stakeholders. However, in recent times many repositories are being asked by their sponsors and funding agencies to provide information on what data and services are used by whom and for what purpose in greater detail than customary in the past: these more detailed requests can be in conflict with privacy legislation. More sophisticated systems now have to be implemented to enhance usage statistics, and the costs often compete with activities such as improving customer service and user experience.

This session will showcase a range of best practices in research data repositories that are working on making data and metadata Open, FAIR and accessible to both humans and machines. Contributions are welcome on unique identification of datasets and physical samples in a repository, attribution to ensure both funders and creators are credited (particularly of the raw observational datasets), provenance tracking and on making QA/QC assessments of the published datasets publicly available and FAIR. Methodologies and systems for implementing user identification and usage tracking are also invited, including protecting privacy and balancing scarce resources vs the need to continuously improve the user experience.

Awareness of the importance of the reproducibility of research results has increased considerably in recent years. Knowledge must be robust and reliable in order to serve as a foundation to build further progress on it. Reproducibility is a complex topic that spans technology, research communities, and research culture. In the narrow sense, reproducibility refers to the possibility of another researcher independently achieving the same result with the identical data and calculation methods. Put simply, one could say that research is either reproducible or not, but more practically there is a continuum of reproducibility where some factors weigh more heavily on influencing results. Replicability or replication, on the other hand, is a broader term and refers to one’s ability to replicate their own research. One problem, however, is that a large percentage of existing studies cannot be successfully reproduced or replicated. This endangers trust in science.

However, with the increasing complexity, volume and variety of Earth System Science (ESS) data - where data can be of multiple types like source code, entire workflows, observational or model output data - and the continuing push towards compliance with the FAIR data principles, achieving reproducibility is challenging. Dedicated solutions do exist only for a subset of implementation factors, but are mostly focused on single institutions or infrastructure providers. Current developments to establish the FDOs (FAIR Digital Objects) and corresponding frameworks go one step further to eventually enable a global interoperable data space to achieve scientific reproducibility. The adoption of Artificial Intelligence (AI), especially machine learning (ML), and other computational-intensive processes complicate this even further.

This session will explore current practices, methods and tools geared towards enabling reproducible results and workflows in ESS. We promote contributions from the areas of infrastructures, infrastructure requirements, workflow frameworks, software/tools, description of practices or other aspects (e.g. provenance tracking, quality information management, FDOs, AI/ML) that must be considered in order to achieve and enable reproducibility in Earth system sciences. These can be contributions that are generally valid and/or transferable or focus on certain areas of application. Finally, best practice examples (or as a counter-example bad practice) are also invited.