Open Earth Monitor — Global Workshop 2023

You are currently viewing an older schedule version. You can find the current version here.

To see our schedule in grid from, please enable JavaScript.

Wednesday, Oct. 4, 2023

Thursday, Oct. 5, 2023

Friday, Oct. 6, 2023

14:00

15min

Opening Session

EURAC Auditorium

14:15

30min

Open Earth Observation - Shaping the future by understanding the past

Julia Wagemann (PhD)

The era of open Earth Observation (EO) data started 2008 when the United States Geological Survey (USGS) made the Landsat archive available free-of-charge. Since then, the amount of open EO data has increased exponentially, also due to Copernicus, the European Union’s Earth Observation programme. The current wealth of open EO data available is unprecedented. This leads to the current situation that a considerable fraction of the open EO data produced and disseminated on a daily basis is not used as users cannot access, process and analyse the data. Questions on how EO data can be utilised to better support the Green New Deal and related communities, such as Renewable Energies, lead the discussion now and will do so in the coming years. A better understanding of the needs and requirements of different users (from EO data users to policy- and decision-makers) will be vital in shaping the future of open EO data.

“To create the future, we must understand the past” is a famous quote stated by astrophysicist Dr. Carl Sagan. Hence, in this keynote talk, I would like to take the audience on a time travel through the era of open EO data. We will take the perspective of an EO data user and first identify key milestones and developments since 2008 before we draw a more detailed picture of what it is like at the moment to discover, access, process and retrieve knowledge from open EO data. After we shed a light on the past and the present, time travel continues to the year 2030 and together with the audience, I’d like to develop a wish list on how open EO data shall be of value for different stakeholders in the future and extract the key requirements that are needed to achieve this.

EURAC Auditorium

14:45

30min

Earth Observation Open Science and Innovation

Anca Anghelea

Open Science is increasingly recognized as a catalyst for innovation. Back in 2016, the EC's DG-RTD laid a vision for European R&D which acknowledged that “the way that science works is fundamentally changing, and an equally important transformation is taking place in how companies and societies innovate. The advent of digital technologies is making science and innovation more open, collaborative, and global”. The concept of Open Science and Innovation is embraced by the European Space Agency in
its Agenda 2025, recognizing the value that such principles of innovation can bring for the space sector in terms of optimizing development cycles, accelerating time to market, and reducing cost.
In this talk we will present ways in which Open Science and Innovation are addressed in ESA's Earth Observation Programme, and how such principles are transferred into R&D and Scientific activities. We will dive into the "WHYs" and "WHENs" of adopting openness in the Earth Observation value chain, share lessons learned from community consultations and finally look into the "HOWs" of implementing Open Science and Innovation in EO building on elements of agility, sustainability and scalability provided by incorporation of new digital technologies and design-driven product development and science for society.

EURAC Auditorium

15:15

30min

Coffee Break

EURAC Auditorium

15:15

30min

Coffee Break

EURAC Conference Hall

15:15

30min

Coffee Break

EURAC Seminar room 2 & 3

15:45

20min

GEOSS Platform Plus: European efforts to reinforce the use of Earth Observations globally

Alessandro Scremin

The Group on Earth Observations (GEO) envisions a future where decisions and actions for the benefit of humankind are informed by coordinated, comprehensive and sustained Earth observations. A central part of GEO’s Mission is to build a Global Earth Observation System of Systems (GEOSS), a set of coordinated, independent Earth observation, information and processing systems that interact and provide access to diverse information for a broad range of users. The amounts of Earth observation data are increasing drastically over the last years, as they include different data sources from satellites to model outputs, from airborne sensors to ground stations and in-situ data. The main challenge is how to find and make use of these resources by users worldwide.
The European Space Agency, together with the Italian National Research Council and the University of Geneva are contributing to the implementation of GEOSS via the EU H2020 co-funded project GPP (GEOSS Platform Plus). This European contribution is reinforcing the use of Earth Observations globally focusing on provisioning of actionable information for climate change research, monitoring, and development of mitigation and adaptation actions. A user centric approach helps to focus on real user needs, listening and co-designing new implementations in close coordination with users. Examples of applications targeted include SDG 15.3.1 on Land Degradation, Climate Change impact on Norovirus Pandemic Risk, and SDG11.7 that relates to climate change, urban sustainability and health. Access to data products, services and information, and the possibility to generate actionable information to derive results as input to decision makers are main objectives. Relevant contributions to an evolved overarching GEOSS architecture that copes with changes in the GEO landscape, including general changes in the Science/Policy landscape in technological innovation are foreseen as well. In this context, GPP is playing an important role in connecting application developers (scientists, developers) to providers (of data, platforms, services, etc.) for enabling the generation of actionable information usable by the end users (decision makers, institutions, citizens). It will be then presented some real cases of how GEO communities and providers could contribute to GEOSS and how users can benefit from the tools and applications available in the GEOSS Ecosystem to support the different actors in different GEO incubator areas (also known as societal benefit areas).

EURAC Auditorium

15:45

20min

Global Analysis-Ready and Cloud-Optimized (ARCO) Landsat Mosaics: Challenges and implementation strategies

Leandro Leal Parente

Landsat is the longest running program to provide space-based data for Earth’s land surface. Based on nine satellites, the program has been monitoring the planet since 1972, consistently providing multi spectral images for several applications. Due to technology differences among the satellites / image sensors, the reflectance values may have significant variations across the entire time-series. Data gaps, due cloud cover, and stripe artifacts, caused by the Scan Line Corrector failure (Landsat 7), add an additional level of complexity for the users interested in perform long-term time series analysis and machine learning on this data. Considering these challenges and the potential of usability of the Landsat imagery, here we presented a workflow to produce analysis-ready and cloud-optimized (ARCO) global mosaics including: 1) data harmonization, 2) cloud and artifact screening, 3) temporal aggregation, 4) gapfilling, and 5) mosaicking. Relying on de-facto standards (Cloud-Optimized GeoTIFF - COG and SpatioTemporal Asset Catalog - STAC), the Landsat global ARCO mosaics have potential to boost the access of Landsat data and contribute with the monitoring of land use conversion, food production, biodiversity, climate change and land productivity.

EURAC Conference Hall

15:45

45min

Reproducible and Reusable Remote Sensing Workflows

Edzer Pebesma

Reproducibility and Reusability of workflows are increasingly important topics in Remote Sensing research when moving towards FAIR and open data science. This workshop discusses the current status quo, and how we can improve this with future activities.

OEMC project workshop

EURAC Seminar room 2 & 3

16:10

20min

COPERNICUS DATA SPACE ECOSYSTEM: IMMEDIATELY AVAILABLE DATA AND ASSOCIATED SERVICES

Jędrzej Bojanowski

The Copernicus Data Space Ecosystem (CDSE) represents a key milestone in access to Copernicus satellite data. First and foremost, the novelty relates to the paradigm shift that all Copernicus data (except for some of the raw data) is immediately available online - global coverage and the entire time range including the archive - at no cost, for any user. The product list includes Copernicus satellite imagery (Sentinel-1, Sentinel-2, Sentinel-3, Sentinel-5p), Copernicus Services and other satellite data missions (e.g. Landsat, SMOS, Envisat). In the orderable mode, historical Sentinel-1 RAW data and processing of Sentinel 1/2/3 data using official ESA processors is available. So-called Sentinel Engineering Data (mostly Level-0 data) will be available in the rolling 2-week archive. Moreover, CDSE will offer access to commercial satellite data.

The main advantage of immediately available data is that the user does not have to order and wait for the data. Direct access also allows bulk data processing and streaming, e.g. via OGC services (WMS/WFS). In this respect, only online data access provides sufficient capacity to visualise data online. Another advantage of immediately available data is the ability to partially read large data files if they are stored in an optimised chunked format such as Cloud Optimised GeoTIFF (COG) or Zarr for rasters, or GeoParquet for vectors. Partial reading is essential for parallel computing, which allows small chunks of data to be processed in parallel so that there is no need to wait for the data to be fully loaded before processing.

Another novelty of CDSE are the various interfaces where the data are available: from old-fashioned download to various interfaces providing capability to search the catalog connected to the same database to guarantee consistency. First interface is OData - an ESA-adopted standard, which is based on https RESTful Application Programming Interfaces. It enables resources, which are identified by URLs and defined in a data model, to be created and edited using simple HTTP messages. Another interface which is foreseen is the STAC catalogue and API that become a standard in the EO community, also being onboarded to OGC. The CDSE also provides Jupyter Hub - a very suitable tool for prototyping, developing, and testing applications for Earth Observation data processing. This is an open-source, online, interactive web application which gives access to computational environments and resources without burdening the users with installation and maintenance tasks.

The vast majority of the described capabilities are available free-of-charge for the individual's use - personal, research or commercial. For those interested in larger scale processing, there are practically unlimited resources available under commercial terms. The first one of these is CREODIAS, which allows user to access and process the data directly from federated cloud environment, order serverless processing of the EO products and access EO-dedicated services. Yet, additional third-party providers are joining CDSE to offer a variety of additional services (free and commercial) as members of the Data Space.

EURAC Auditorium

16:10

20min

Monitoring the Earth System with xcube

Gunnar Brandt, Tonio Fincke

The available information for monitoring the Earth has never been as abundant and accessible as today. Data volumes from Earth Observation, mathematical models and in situ measurements will continue to grow in the future. While this data richness opens unprecedented possibilities for monitoring and developing a holistic understanding of our planet, it poses challenges with respect to efficient data exploitation and fosters novel technological approaches for the joint exploitation of the different data streams. Despite considerable efforts for standardisation by different institutions in the past, data formats and models as well as interfaces for data access remain diverse, resulting in costly, bespoke solutions for finding, accessing, and processing heterogeneous input data. The xcube open-source Python package addresses such requirements and offers a suite of comprehensive tools for transforming data sets into analysis-ready data cubes. As it is designed as a framework, it is continuously extended to cater to changing needs. It integrates seamlessly into Python’s data science stack, as advocated by Pangeo, and extends it by specific data stores for the access to relevant data sets for Earth System Science and tools for data provisioning and exploitation.

EURAC Conference Hall

16:30

45min

Distributed computing on large geodata from multiple sources using the Julia Programming language

Felix Cremer, Daniel Loos

Spatiotemporal data cubes are becoming ever more abundant and are a widely used tool in the Earth System Science community to handle geospatial raster data.
Sophisticated frameworks in high-level programming languages like R and python allow scientists to draft and run their data analysis pipelines and to scale them in HPC or cloud environments.

While many data cube frameworks can handle harmonized analysis-ready data cubes very well, we repeatedly experienced problems when running complex analyses on multi-source data that was not homogenized. The problems arise when different datasets need to be resampled on the fly to a common resolution and have non-aligning chunk boundaries, which leads to very complex and often unresolvable task graphs in frameworks like xarray+dask.

In this workshop we present the emerging ecosystem of large-scale geodata processing in the Julia programming language under the JuliaDataCubes github umbrella.
Julia is an interactive scientific programming language, designed for HPC applications with primitives for Multi-threaded and Distributed computations built into the language.
We will demonstrate an example analysis where data from different sources (global fields of daily MODIS, hourly ERA5, high-resolution land cover), summing to multiple TBs of data, can interoperate on-the-fly and scale well when run on different computing environments.

OEMC project workshop

EURAC Seminar room 2 & 3

16:35

20min

European platforms to support the Green Deal

Benjamin Schumacher

The European Union's Green Deal is a transformative initiative aimed at achieving climate neutrality and sustainable economic growth by 2050. To address the complex challenges posed by climate change and environmental degradation, a robust infrastructure of platform services has emerged to support the Green Deal. At EODC, we implement and develop key components of this infrastructure, including the openEO Platform, ESA GTIF Austria, STAC, Pangeo, and the GREAT project. These services and projects play a pivotal role in gathering, processing, and disseminating critical environmental data and insights that drive policy formulation and sustainable practices.

The openEO Platform serves as a cornerstone for the Green Deal by providing standardized and open interfaces for accessing and processing Earth Observation (EO) data. Leveraging cloud computing and distributed resources, openEO enables researchers, policymakers, and businesses to harness the power of petabytes of environmental data. By fostering collaboration and interoperability, openEO supports the development of innovative applications and services that contribute to the Green Deal's goals, such as carbon monitoring and land-use planning.

The SpatioTemporal Asset Catalog (STAC) standardises the organisation and discovery of geospatial assets, making it easier to find and use EO data. By embracing STAC, the Green Deal ecosystem and platforms ensure that valuable environmental information, such as satellite imagery and climate models, can be efficiently located and employed in various applications, including environmental monitoring and disaster risk management.

Pangeo, an open-source platform for scalable Earth science, empowers researchers with tools and resources for analysing massive climate and environmental datasets. By offering Jupyter notebooks, cloud computing, and data catalogues, Pangeo facilitates collaborative research and data-driven decision-making. Pangeo's contributions are vital to the scientific underpinnings of the Green Deal, assisting in climate modelling, biodiversity assessment, and ecological forecasting.

ESA's Green Transition Information Factory (GTIF) demonstrates in Austria how to enhance the visibility and access to EO-derived datasets and knowledge for decision-makers, policymakers and citizens. The platform additionally showcases how to efficiently disseminate information for Austria's green transition initiatives allowing users to discover the potentials of transitioning to carbon neutrality by 2050 through EO data and cloud computing.

The GREAT project is focused on building a robust data infrastructure for the Green Deal. It aims to integrate data from various sources, including EO, environmental sensors, and socioeconomic indicators, into a comprehensive data space. This unified data repository enhances data sharing and accessibility, supporting the monitoring, evaluation, and adaptation of Green Deal policies and actions.

In conclusion, these projects form a cohesive and adaptable infrastructure that underpins the Green Deal's ambition to combat climate change and promote sustainability. By providing open and standardized interfaces, advanced geospatial capabilities, efficient data organization, scalable computing, and comprehensive data integration, these platforms facilitate data-driven decision-making, innovation, and drive progress toward the EU's environmental goals. In the face of the global climate crisis, these services are central tools in advancing the Green Deal's mission to build a greener and more sustainable future for all.

EURAC Auditorium

16:35

20min

The Hydrography90m dataset: present offering and future scope

Tushar Sethi

In tandem with the monumental increase in geo-data availability from remote sensors, field sensors and various publicly available environmental datasets, state-of-the-art geoinformatics algorithms have evolved to harness earth science data as never before. In the field of computational hydrology, these processes have yielded global information in fine detail, and of exceptional precision.

Hydrography90m is one such data product that pushes the boundaries of computational hydrology in several ways. It is a globally standardised and seamless hydrographic dataset that allows the mapping of headwaters in unprecedented density and detail. With the minimum upstream contributing area set at 0.05km^2, it comprises the highest density of headwaters compared to leading global hydrographic assessments. The dataset contains 1.6 million drainage basins and 726 million stream segments and sub-catchments. It is also designed to overcome the spatial and accessibility constraints of gauged locations and address the limitations of spectral analyses.

As for applications in scientific research, Hydrography90m is well-suited for both global and comparative area-of-interest studies. The dataset contains many essential stream features, such as stream slope, stream distance, types of stream order and flow indices. Hydrography90m thus offers significant utility in the assessment of freshwater quantity and quality, inundation risk, biodiversity and conservation, and resource management objectives, all in a globally comprehensive and standardised manner.

In terms of the underlying computational approach, Hydrography90m is based on a drainage flow algorithm that distributes downhill water flow in a realistic manner, following the concavity and convexity of terrain. Additionally, programming in a variety of open
source software provides unmatched computational power, and the implementation of different scripting procedures allows for bench-marking strategies to check for potential errors. Software employed includes GDAL, Pktools, and GRASS GIS.

The novel computational approach of Hydrography90m broadens the scope for using various remote and field sensor technologies, and the scripting procedure lays the foundation for more complex Machine Learning-based discharge assessments. Its design is a pivotal development for addressing the challenges of overfitting and universal coverage in hydrological modelling. Machine Learning can now enable the massive data integration that is vital for global scale hydrological studies; and hydrographic data with fine detail of headwaters provides an excellent foundation for interbasin connectivity and high-resolution discharge predictions. Meanwhile, other data-driven and ensemble methods that have emerged recently to address these technical challenges still remain limited as tools for basin-specific studies.

Given the multitude of resource and conservation applications, Hydrography90m can be a vital toolkit for achieving several UN Sustainable development goals. Additional uses of the dataset are relevant to freshwater flows and sediment transport computations, pollutant and nutrient concentration assessments, public health, and geopolitical and resource challenges. To date, Hydrography90m has been used in species distribution modelling for aquaculture, vector-borne disease mapping, and various ecological studies. Institutions engaged in water resource management, transnational security and environmental crime monitoring are also starting to derive value from the dataset’s attributes.

EURAC Conference Hall

17:00

45min

General Assembly

EURAC Auditorium

18:00

150min

social event

EURAC Auditorium

09:00

30min

DestinE Core Service Platform

Inés Sanz Morère

DestinE Core Service Platform integrates and operates an open ecosystem of services (also referred to as DESP Framework) to support DestinE-data exploitation and information sharing for the benefit of DestinE users and Third-Party entities.
DESP includes key essential services such as user management service; infrastructure as a service with storage, network, and CPU/GPU capabilities; data access and retrieval service, in particular from the DestinE Data Lake operated by EUMETSAT, as it is the backbone for the data generated by the ECMWF’s Digital Twin Engine; data traceability and harmonization services; basic software suite service for local data exploitation; data and software catalogue services; and 2D/3D data visualization service.
DESP additionally provides onboarding support for integrating external services and resources, making the ecosystem flexible, scalable, and easily adapted to the community needs. DestinE ecosystem aims to support the needs of a large and diverse community of users including general citizens, scientists and academics, commercial entities, or policy makers. DESP Framework defines the required conditions for a service to be part of the ecosystem, and therefore benefit of the available resources, and of the potential to engage with all DestinE users.

EURAC Auditorium

09:30

30min

Analysis Ready Data

Peter Strobl

The concept of “Analysis Ready Data” (ARD) was initially developed around 2015 within the Committee on Earth Observation Satellite (CEOS). CEOS defines ARD as “satellite data that have been processed to a minimum set of requirements and organized into a form that allows immediate analysis with a minimum of additional user effort and interoperability both through time and with other datasets”. Over the course of the past few years CEOS has issued a number of so-called ‘Product Family Specifications’ (PFS) which cover a variety of different sensing methods and observed parameters. Institutional and commercial satellite data providers have accepted these specifications and by now a broad variety of satellite image products are available as “CEOS-ARD certified”.
The popularity of the concept and the expectation for the ARD ‘label’ in terms of simplified usability and interoperability of a wealth of EO data has spurred the desire to expand ARD beyond classical EO parameters by including higher level products and to cover geospatial data more generically. This prompted a discussion on which categories, levels, or classes of analysis ready data are needed and how they could be defined and distinguished. These considerations are now taken up by a formal ISO/OGC Standard Working Group (SWG) which was launched recently.
In the meantime, the amount of available geospatial data increases exponentially and many of these are available free&open and calling themselves ‘ARD’. However, users relying on their interoperability are often overwhelmed by their diversity and remaining inconsistencies which often require considerable effort before appropriate data can be selected and joined sensibly. Access to proper reference data and benchmarking methods is therefore an important factor for leveraging ‘ARD’ in the future.

EURAC Auditorium

10:00

30min

Aggregating multiple emissions data for actors at all levels through a common schema in OpenClimate

Joaquin van Peborgh

The OpenClimate Network is an open source nested accounting platform allows users to navigate emissions inventories and climate pledges of different actors at every level, aggregating data from various public sources for countries, regions, cities and companies. Through this aggregation, it enables the comparison of how different data sources report emissions of certain actors, by harmonizing the way data is reported and identifying the different methodologies used.

Additionally, by nesting actors into their respective jurisdictions it facilitates the comparison between the pledges these actors have committed to, and to see if they are aligned towards the same climate targets, and how these compare to the goals of the Paris Agreement.

By aggregating data and exploring it in this nested manner, it also allows for the effective identification of data gaps for these actors, suggesting where efforts are needed to identify existing data sources or help produce new inventories. When data gaps are identified, the platform also prompts users to contribute data based on the open and standardized data model used to aggregate emissions and pledges data.

EURAC Auditorium

10:30

45min

Coffee Break

EURAC Auditorium

10:30

45min

Coffee break

EURAC Conference Hall

10:30

45min

Coffee Break

EURAC Seminar room 2 & 3

10:35

5min

Integrating Sentinel-1 Data into FORCE for Large-Area Analysis Ready SAR Data Cubes

Felix Lobert, Tom Broeg

The use of synthetic aperture radar (SAR) data has become increasingly important in remote sensing for environmental monitoring. SAR data provides valuable information on surface characteristics and changes, such as land cover and land use change, soil condition, and vegetation growth, making it a powerful tool for various applications, including agriculture, forestry, and climate change studies. However, processing and integrating SAR data into analysis-ready formats can be complex and time-consuming, requiring specialized knowledge and tools.
In this contribution, we propose a software project called force-sar, which aims to integrate Sentinel-1 data into the Framework for Operational Radiometric Correction for Environmental monitoring (FORCE). FORCE is a widely used data cube framework for processing and analyzing optical remote sensing data and the integration of SAR data into FORCE increases its capabilities for large-scale, multi-modal analysis.
Force-sar automatically queries available Sentinel-1 Ground Range Detected (GRD) imagery covering the spatial and temporal dimensions of your area of interest. The scenes are then directly accessed from satellite data repositories on cloud environments like Creodias or the Copernicus Data and Exploitation Platform Germany (CODE-DE). When no connection to such repositories is available, the data can also be downloaded from data centers like the Alaskan Satellite Facility (ASF). The scenes are processed to radiometrically calibrated gamma-naught backscatter data using a pre-built but customizable ESA SNAP processing graph. After resampling, reprojecting, and tiling the data, they are ready for ingestion into a FORCE data cube.
The integration of Sentinel-1 data into FORCE allows for the creation of SAR data cubes with consistent radiometric and geometric properties, covering large regions and spanning multiple time periods. This enables users to perform multi-sensor and multi-temporal analyses like change detection, compositing, and classification and regression tasks for environmental monitoring at large scales, thereby supporting decision-making and policy evaluation in frameworks like the EUs Green Deal or the Common Agricultural Policy.
Force-sar is already used in the ongoing Mowing Detection Intercomparison Exercise (MODCiX), where more than ten teams compare their algorithms for grassland mowing detection on a consistent and harmonized remote sensing and reference data set. A consistent data cube holding optical and SAR data covering test regions in eight European countries was created and is used for the study.
Force-sar provides a streamlined and fully containerized workflow for preprocessing and integrating SAR data into an existing data cube framework without the need for the installation of any external tools or dependencies. This opens up new possibilities for utilizing SAR data in large-scale environmental monitoring applications, particularly when working in cloud environments.

Poster presentation

10:40

5min

A Cloud-native Data Platform for Management and Processing of Big Geospatial Data

Monia Santini, Melissa Latella

Earth System Models and Earth Observations are crucial for studying the Earth, providing scientific insight into fundamental dynamics and valuable predictions about Earth’s future. However, they generate huge amounts of data, at different temporal and spatial scales, so it becomes of paramount importance to access them in a seamless and efficient way for scientific analysis. Usually Earth Science datasets are represented with hundreds or thousands of files that can introduce a lot of burden to the user in terms of management, since it requires the user to set up computational and storage resources for accessing and retrieving data and writing code to load and prepare data into in-memory data structures for analysis.

In this talk, we describe in detail the architectural design, implementation and deployment of a data management and analytics system in order to facilitate cataloguing, accessing and processing Earth Science data. The system has been designed using a cloud-native architecture, based on containerized microservices, that facilitates the development, deployment and maintenance of the system itself. It has been implemented by integrating different open source frameworks, tools and libraries and has been deployed using the Kubernetes platform and related tools such as kubectl and kustomize.

The Data Platform consists of different components that will be introduced and described together with the related technologies adopted: (a) the Catalog, based on Intake and MongoDB for cataloguing and indexing the datasets published and managed in the system, (b) the Analytics Engine, based on the geokube and dask Python libraries: geokube is used for specialised geospatial operations (such as extracting a bounding box or a multipolygon) according to different types of geoscientific datasets and dask for parallel and distributed processing; (c) the Broker implemented using RabbitMQ framework for managing the user workload requests; finally, (d) the Rest APIs and the OGC standard interfaces (i.e., WPS) to access data and to submit analytics workflows.

An instance of the Data Platform has been deployed in production at Euro-Mediterranean Centre on Climate Change (CMCC) for the delivery and analysis of data produced by the CMCC Research Divisions. In this talk, we will showcase different Use Cases, related to sectors such as climate change and wildfire management, that demonstrate how the system has been used at CMCC, within different projects and initiatives, for building downstream products and services that need to access, analyse and process Earth Science data.

Poster presentation

10:45

5min

AquaINFRA for blue research interlinking EOSC ecosystem with Green Deal data space

Kaori Otsu

The establishment of the European Open Science Cloud (EOSC) is one of the eight priorities of the European Open Science Agenda (2018), with the ambition of enabling to federate multidisciplinary research infrastructures. Among ‘Enabling an operational, open and FAIR EOSC ecosystem (INFRAEOSC)’ projects contributing to this priority, Blue-Cloud 2026 and AquaINFRA have been funded to protect oceans, seas, coastal and inland waters, in contribution to achieve the goals of the EU Mission "Restore our Ocean and Waters” by 2030.

Furthermore, the European data strategy (2020) identifies data spaces as the instruments to achieve a single market for data across sectors and countries, through a common and interoperable framework. In particular, the Green Deal Data Space will be interlinked with the EOSC ecosystem demonstrated by the Blue-Cloud 2026 and AquaINFRA projects, involving several research communities and data infrastructures that are contributing to enabling the European Digital Twin of the Ocean.

The main objective of the AquaINFRA project is to develop a research infrastructure equipped with FAIR multi-disciplinary data and services, allowing seamless data discovery and processing through an AquaINFRA Interaction Platform (AIP) in order to support marine and freshwater scientists and stakeholders and interact with EOSC seamlessly. More specifically, the AIP will include developing a cross-domain and cross-country search and discovery mechanism as well as building services for spatio-temporal analysis and modelling through Virtual Research Environments (VREs) where regional case studies are demonstrated highlighting the Mediterranean use cases in this presentation.

Poster presentation

10:50

5min

Advancing Earth Observation knowledge management through machine learning and semantic interoperability for EU policy support

Dominik Weckmüller, Iacopo Ferrario

The Knowledge Centre on Earth Observation activity is grounded in sound knowledge management practices and cutting-edge NLP technologies. It aims to create a common scaffolding for research projects on the one hand and policy needs on the other.

The User Requirement Database (URDB) stores and validates Core Copernicus Users requirements for Earth Observation (EO) products and applications. The URDB facilitates automated gap analysis and screenings across diverse data spaces in pursuit of the optimal matching pre-existing solution, initially examining Copernicus Services product catalogues, followed by a subsequent exploration of research findings from the EU Horizon programme. The Text Mining Application (TMA) leverages innovative advancements in machine learning utilizing Transformers to facilitate precise semantic document retrieval within an EO-specific subset of research outcomes financially supported by the European Union's research and innovation framework programmes. These programmes and the respective EO project data span from FP1 in 1984 to the most current initiative, Horizon Europe. The primary TMA objective is to empower users with rapid access to research findings for highly specific queries, while simultaneously offering a user-friendly database, an internal microservice as an API, and a GUI interface for more advanced metrics and visualizations. In the future, the URDB and TMA will be closely interlinked and integrated, enabling users of either platform to benefit from rapid access to the actual Copernicus datasets, as well as enhanced meta-information metrics and insight into research outcomes.

In addition to supporting gap and fit for purpose analysis, the main scope of the URDB is to enable requirement retracing across the components of the EO value chain, from policy needs to observations, and therefore supporting and tracking the evolution of the Copernicus Programme. The URDB's records are technology-agnostic quantitative requirements, expressed by verifiable, unambiguous and actionable technical specifications (horizontal resolution, measurement uncertainty, tasking time, etc.). The URDB's data model builds on the experience of existing requirement databases from Copernicus Core Services and international partners (e.g., USGS and NASA). One of the URDB’s and TMA’s core design principles is semantic interoperability: entities, relationships and attributes are clearly defined in a terminology and, when applicable, they follow international standards (ISO, OGC), recommendation and best practices (CEOS, GEO).
From a technical perspective, both, the URDB and TMA are self-hosted open-source databases with a GUI and an application layer for querying and performing analysis. While the URDB is based on PostgreSQL, the TMA utilizes a novel vector database known as Qdrant, which fulfils highly specific AI requirements and offers a user-friendly API.
The vision for both databases is to link them through web APIs to online data catalogues and tightly integrate them into the existing (meta-) data Copernicus landscape.

Poster presentation

10:55

5min

The ICOS ecosystem in-situ measurements portfolio

Simone Sabbatini

In-situ data collection is a fundamental part in the domain of ecosystem observations and monitoring. Continuous measurements of energy and matter exchanges at the ecosystems/atmosphere boundary by means of the eddy covariance (EC) technique are fundamental observations across the wide range of in-situ measurements, in particular concerning carbon and other greenhouse gases and water balances. Monitoring stations based on this technique and organised in networks at different scales, from national to global, providing precious insights to different users, are a standard in the environmental sector. The Integrated Carbon Observation System (ICOS) is one of such research infrastructures, working at the European scale. The ecosystem domain of ICOS, dealing with terrestrial observations in natural and anthropic ecosystems, is not only providing EC datasets, but also numerous meteorological and other auxiliary variables to support the activity of the network. In the present work we describe the portfolio of the main products included in a typical ICOS ecosystem station: from continuous measurements of CO2 and H2O exchanges, to the above- and below- ground meteorological parameters, to the discontinuous ancillary measurements of different vegetation characteristics – spanning from tree height to above-ground biomass, from soil characteristics to plant area index, from species distribution to litter mass. The continuous datasets are provided at different scales, from half-hourly to yearly. All the datasets, supplemented by a detailed set of metadata ensuring the consistency with the FAIR principles (Findability, Accessibility, Interoperability, Reusability), are stored on a safe repository and both openly and freely distributed (with license CC-BY 4.0) to researchers, modelers, and any user that requests them.

Poster presentation

11:15

90min

Build and visualize your own raster data cube

Leandro Parente, Murat Sahin

Raster data cube is a four-dimensional array with dimensions x (longitude / easting), y (latitude /northing), time, and bands sharing a: (1) single spatial reference system, (2) constant spatial cell size, (3) constant temporal duration, (4) temporal reference defined by a simple start and end date / time, resulting in a single attribute value for every dimension combination. Building a data cube consists basically in converting raw irregular raster data into a regular and dense structure, which may include information loss and needs to consider user definitions and application restrictions.

OEMC project workshop

EURAC Conference Hall

11:15

20min

Estimation of urban heat island integrating IoT sensors and EO thermal open data

Raniero Beber

Heat waves are more and more heavily affecting population and this is even more enhanced in cities rather than in the countryside where the urban heat island (UHI) effect worsen their duration and intensity. Current research on the estimation of the UHI effect adopts 3 main approaches: i) observational studies describing its driving processes, spatial patterns and/or magnitude; ii) Earth observation (EO) studies focusing on the surface urban heat island (SUHI) determined from land surface temperature (LST) retrieved from satellite thermal sensors (e.g. Landsat-TIRS, Sentinel-SLSTR); iii) modeling via mesoscale meteorological models like the Weather Research and Forecasting (WRF) or microscale models (i.e. ENVI-met) that require large computational effort and/or fine tuning of parameters. Municipalities need actionable data to support their decisions. It is thus crucial to develop intermediate approaches for the estimation of the UHI intensity and spatial extension without the need of advanced expertise to tune parameters or run complex meteorological models but, at the same time, able to provide reliable insight into the urban air temperature. The work presented in this contribution is performed in the framework of the USAGE project activities and focuses on providing a pipeline for the development of UHI maps in urban areas utilizing open data like EO, IoT ground sensor data, surface properties and a hybrid model based on machine learning and geostatistics. We present a pipeline that can be deployed with minor adaptation (i.e STAC end points) within GIS software environments. The ground sensor data are accessed via OGC SensorThings API and fed into the analysis. Pre-loaded 'semi-static' layers, like DTM, DSM, LU/LC, vegetation fraction, urban building morphology and shade maps are accessed via OGC Feature API and utilized to spatialize the air temperature at each time stamp received from the IoT sensors. Based on the revisit time of EO thermal data and its cloud-coverage level, LST observations are integrated to help the spatialization of the ground sensor's temperature, performed using a hybrid model combining machine learning and geostatistics. This allows for faster computation compared to classical geostatistics but, at the same time, to explicitly handle spatial correlation of data and errors. The aforementioned pipeline is suitable to derive UHI maps from given IoT ground sensor data. On the other end, to forecast the UHI effect up to 48 h, the pipeline ingest 2-m air temperature, relative humidity as well as wind speed and direction from the meteorological models (WRF or AROME) the open-meteo API.
The proposed pipeline is applied in the Alpine valley and city of Trento (Italy) and is then validated against high-resolution simulations with the WRF model, offline coupled with an urban parameterization scheme to reach a resolution of 100 m. The proposed pipeline can be used not only as a forecasting tool, but also as a UHI mitigation and planning tool by changing the ‘semi-static’ layers that involve the study area. This allows municipalities to predict the effects of their decisions.

EURAC Auditorium

11:15

45min

GEO Knowledge Hub to preserve and share EO applications: Introduction and practice

Felipe Carlos

Earth Observation (EO) applications enable decision-makers, researchers, and specialists to understand the phenomena of our planet, allowing global changes to be made from local actions taken by the public and private sectors. With the dissemination and use of Open Data practices, the EO applications have been enhanced, allowing numerous works to be developed, ranging from the analysis of anthropic actions on inland waters to the temporal analysis of land use and land cover changes. These advances and improvements in EO applications have made their development complex, requiring several materials to be used together with the data to compose the results. Consequently, organizing, sharing, and preserving these applications and the knowledge within them to enable reproduction and replication has become a challenge. Often these activities require specific expertise from researchers and specialists and technical infrastructure.

The Group on Earth Observations (GEO) and its community promote Open Data practices, being responsible for defining guidelines and developing the Global Earth Observation System of Systems (GEOSS) that enhances access to EO data. Recently, GEO started the development of a new component of the GEOSS ecosystem, the GEO Knowledge Hub (GKH), to foster the reproduction and replication of EO applications. Created based on the GEO Data Sharing Principles and the GEO Data Management Principles, the GKH allows users to share their EO applications and the underlying resources (e.g., processing scripts, datasets, and description notes), enabling people to understand, reproduce and replicate the shared EO application. In the GKH, the resources of an application can have files (e.g., satellite imagery datasets, in-situ data files) and metadata (e.g., title, authors, spatial location). Furthermore, each resource can be associated with an individual persistent identifier (DOI) created by the GKH, enhancing dissemination and citation.

Applications shared on the GKH can be found and used, making their knowledge accessible. For this, the GKH provides high-level features for organizing the application materials and facilitating their sharing. In addition, the GKH has a powerful search engine that enables textual, thematic (e.g., Sustainable Development Goal-oriented search), and spatial-temporal searches. In addition to share and search capabilities, the GKH provides features that facilitate community engagement, such as discussion sections (Real-time Q&A) and a feedback system. All these features are accessible through high-level web interfaces and Rest APIs, allowing the integration of various tools to use the digital repository.

The GKH is already being used to share and preserve many EO applications. For example, several GEO Work Programme Activities store their applications in the GKH (e.g., GEOGLAM, GEOVENER, Digital Earth Africa, and many others). These and other use cases have shown positive results, indicating that the GKH can assist in organizing, sharing, and preserving the knowledge generated in EO applications. Therefore, in this workshop, we will introduce the main concepts of the GKH, guidelines, and practices in how users can use it to share and preserve EO applications.

OEMC project workshop

EURAC Seminar room 2 & 3

11:45

20min

From low-cost AirQuality stations to open standard (OGC SensorThings) and open data with open source solutions (FROST + QGIS plugin for sensors)

Luca Giovannini

This is the story of 2 twin projects (namely AIR-BREAK and USAGE) undertaken by Deda Next on dynamic sensor-based data, from self-built air quality stations to the implementation of OGC standard compliant client solution.
In the first half of 2022, within AIR-BREAK project (https://www.uia-initiative.eu/en/uia-cities/ferrara), we involved 10 local high schools to self-build 40 low-cost stations (ca. 200€ each, with off-the-shelf sensors and electronic equipment) for measuring air quality (PM10, PM2.5, CO2) and climate (temperature, humidity).
After completing the assembling, the stations were provided to high schools, private households, private companies and local associations. Measurements are collected every 20 seconds and pushed to RMAP server (Rete Monitoraggio Ambientale Partecipativo = Partecipatory Environmental Monitoring Network - https://rmap.cc/).
Hourly average values are then ingested with Apache NiFi into the OGC’s SensorThings API (aka STA) compliant server of the Municipality of Ferrara (https://iot.comune.fe.it/FROST-Server/v1.1/) based on the open source FROST solution by Fraunhofer Institute (https://github.com/FraunhoferIOSB/FROST-Server). STA provides an open, geospatial-enabled and unified way to interconnect Internet of Things (IoT) devices, data and applications over the Web (https://www.ogc.org/standard/sensorthings/).
In second half of 2022, within USAGE project (https://www.usage-project.eu/), we released the v1 of a QGIS plugin for STA protocol.
The plugin enables QGIS to access dynamic data from heterogeneous domains and different sensor/IoT platforms, using the same standard data model and API. Among others, dynamic data collected by the Municipality of Ferrara will be CC-BY licensed and made accessible from municipal open data portal (https://dati.comune.fe.it/).

EURAC Auditorium

12:00

45min

Let's co-create the Green Deal Data Space

Joan Maso

The Green Deal Data Space (GDDS) will interconnect current fragmented and dispersed data from various ecosystems, both from the private and public sectors to facilitate evidence-based decisions and expand the capacity to understand and tackle environmental challenges, for example, for monitoring and reaching environmental objectives in biodiversity, resilience to climate change, circular economy and zero pollution strategies.

This workshop will be partaken by projects EuroGEO Action Group for the Green Deal Data Space in their quest to push the boundaries of data provision, and ensure a FAIR and TRUSTworthy data is available for building a more sustainable future. Some outcomes of the workshop may contribute to the new adhoc ISO TC211 working group on data spaces. Within this Workshop, the following projects will present their current approaches towards enabling the GDDS:

AD4GD: The aim is Integrate standard data sources (e.g. Insitu, RS, CitSci, IoT, AI) in the GDDS, improve semantic interperability, and demonstrate with concrete examples that climate change zero pollution, biodiversity general problems can be solved.

FAIRiCUBE: The core objective is to enable players from beyond classic Earth Observation domains to provide, access, process, and share gridded data and algorithms in a FAIR and TRUSTable manner. We are creating the FAIRiCUBE HUB, a crosscutting platform and framework for data ingestion, provision, analysis, processing, and dissemination, to unleash the potential of environmental, biodiversity and climate data through dedicated European data spaces.

USAGE (Urban Data Space for Green Deal) will provide solutions for making city-level data (Earth Observation, Internet of Things, authoritative and crowdsourced data) available, based on FAIR principles: innovative governance mechanisms, standard-based structures and services, AI-based tools, semantics-based solutions, and data analytics. It will provide decision makers with effective, interoperable tools to address environmental and climate changes-related challenges.

B³ - Global biodiversity is changing under multiple pressures including climate change, invasive species and land-use change. Yet biodiversity data are complex and heterogeneous, making it difficult to understand what is happening fast enough for decision makers to react with evidence-based policies. To solve this B³ will create Open workflows in a cloud computing environment to rapidly and repeatedly generate policy relevant indicators and models of biodiversity change.

GREAT: Funded by the Digital Europe program, aims to establish the Green Deal Data Space Foundation and its Community of Practice which builds on both the European Green Deal and the EU’s Strategy for Data. The project will deliver a roadmap for implementing and deploying the Green Deal Data Space, an infrastructure that will allow data providers and initiatives to openly share their data to tackle climate change in a multidisciplinary manner.

The Open Earth Monitor Consortium is working to contribute infrastructure to the GDDS.

OEMC project workshop

EURAC Seminar room 2 & 3

12:15

20min

Towards a better estimation of GPP for greenhouse gas accounting through the combined use of SIF, spherical grids and knowledge-guided AI

Gregory Duveiller

Some remote sensing signals provide valuable information only at spatial scales that are too coarse for comfort for various applications. Examples include sun-induced chlorophyll fluorescence (SIF), whose signal is related to gross primary productivity (GPP) of vegetation and how this is impacted by stress, or vegetation optical depth (VOD), which is related to how water content is distributed within a canopy, which itself is informative on forest biomass and structure. Within OEMC WP6, we are developing an EO spatial downscaling framework that will be specifically tailored towards improving carbon flux estimations. The spatial downscaling will employ finer resolution EO variables to achieve this super-resolution, but this will not be done merely numerically. It will instead combine our process-based knowledge of how these variables relate to each other to develop a hybrid modelling approach with knowledge-guided AI. This will further be implemented using a moving window adaptative approach over a spherical grid, which will be made possible by novel developments from OEMC WP3. The use case within OEMC is to develop SIF-based 1km spatial resolution GPP flux estimations based on measurements from the TROPOMI sensor on Sentinel-5P, which has a spatial resolution that is larger than 5 km. The main stakeholder for this product will be the Global Carbon Project (GCP), and in particular RECCAP, which aims to establish the greenhouse gas (GHG) budgets of large regions covering the entire globe at the scale of continents. Such endeavor would greatly benefit from the fine-level spatialized GPP fluxes we aim to provide. This use-case will largely leverage on in-situ data provided in OEMC WP4 for validation and calibration, particularly after specific developments to optimize the matching between in-situ points and remote sensing grids is achieved. This presentation will detail the blueprint for this task, which will combine synergistic efforts across various elements within the OEMC project.

EURAC Auditorium

12:45

60min

Lunch Break

EURAC Auditorium

12:45

60min

Lunch Break

EURAC Conference Hall

12:45

60min

Lunch Break

EURAC Seminar room 2 & 3

13:45

45min

Building a Digital Twin Earth for the Water Cycle: State of the Art and Challenges

Luca Brocca

Climate change is profoundly affecting the global water cycle, increasing the likelihood and severity of extreme water-related events. Better decision support systems are essential to accurately predict and monitor water-related environmental disasters and to manage water resources optimally. These will need to integrate advances in remote sensing, in-situ and citizen observations with high-resolution Earth system modelling, artificial intelligence, information and communication technologies and high-performance computing.
The Digital Twin of the Earth (DTE) for the water cycle is a breakthrough solution that provides digital replicas to monitor and simulate Earth processes with unprecedented spatial-temporal resolution and explicitly including the human component into the system. To get the target, advances in observation technology (satellite and in situ) and modelling are pivotal. The workshop will serve the community to assess the state of the art of these technologies and to identify challenges to be addressed in the near future.

OEMC project workshop

EURAC Seminar room 2 & 3

13:45

20min

Mapping land potential and tracking land degradation using EO data

Julia Hackländer

While land is increasingly degrading, robust monitoring approaches are required to identify land degradation processes and to ultimately tackle those. Land degradation is commonly assessed by comparison with the immediate past or surrounding. Comparison with its natural potential, however, i.e. a state of minimal human impact, could give further insights into the full degree of degradation, accounting for the “shifting baseline syndrome”, the gradual change in perception of what is considered the reference. Primary production is one of the key indicators for determining impacts of land degradation, which can be approximated by FAPAR, the fraction of absorbed photosynthetic active radiation, a metric directly related to primary productivity. Here, we present a novel methodology to assess land degradation in reference to its natural potential. Using a machine learning model approach, global time series maps spanning 2000 - 2022+ will be generated by simulating potential natural FAPAR in the hypothetical space of minimal human impact. This will allow performing gap analyses of actual and potential natural FAPAR to monitor impacts of land degradation and restoration efforts through time. Use-case scenarios on country level and project investment level will be demonstrated in the context of supporting UNCCD targets for land degradation neutrality (LDN). This research is carried out within the Open Earth Monitor Cyberinfrastructure project (OEMC) and received funding via the European Union's Horizon Europe programme under grant agreement No.101059548.

EURAC Auditorium

13:45

90min

OpenEO in action. Learn how to get started with openEO via the Web Editor, Python and R.

Peter Zellner, Michele Claus

openEO develops an open API to connect R, Python, JavaScript and other clients to big Earth observation cloud back-ends in a simple and unified way.
openEO Platform implements the openEO API in an federated cloud platform. Hence, it allows to process a wide variety of earth observation datasets in the cloud.
Users interact with the API through clients. This demonstration shows the usage and capabilities of the main clients: The Web Editor, the Python Client and the R-Client.
The Web Editor is a webtool to interactively build processing chains by connecting the openEO processes visually. This is the most intuitive way to get in touch with openEO Platform.
The Python Client and the R-Client are the openEO Platform entry point for programmers. The are available via Comprehensive R Archive Network (CRAN). They facilitate the interaction with the openEO API within the respective programming languages and integrate the advantages of the available geospatial packages and typical IDEs.
The classroom training teaches users how to accomplish their first a round trip through a typical openEO Platform workflow: login to openEO Platform, data and process discovery, process graph building adapted to common use cases, processing of data and the visualization of results
By combining the approaches of the visually interactive Web Editor and the programming based clients users are introduced stepwise to the concepts of openEO Platform and will gradually understand the logic behind openEO.

OEMC project workshop

EURAC Conference Hall

14:05

20min

Use of airborne hyperspectral images in support to Alpine forest managers

Thomas Maffei

Extreme weather and outbakes phenomena are occurring more and more frequently and are affecting forests throughout the Alpine region. Efficient and targeted forest management on a municipal, provincial or regional scale requires high quality, information-rich remote sensing data. Airborne hyperspectral imagery enables the acquisition of high-resolution data on entire portions of land in a short space of time, whereby the high amount of spectral information provides efficient tools for forest managers, such as mapping forest species, identifying invasive species, mapping bark beetle damage, calculating narrowband vegetation indices and analyzing health status. AVT Airborne Sensing Italia (AVT-ASI) uses the Specim AisaFenix sensor for hyperspectral image acquisition. The sensor works in the VNIR and SWIR spectral ranges and acquires 384 bands in pushbroom mode. At the beginning of October 2022, AVT-ASI acquired hyperspectral images over a forest area with size 350 km² near Bruneck, in the South Tyrol province, Italy, (Figure 1 a and b) to support the local forestry inspectorate in the evaluation of the forest health status. Indeed the area of interest has been strongly affected by the spread of the bark beetle, probably due to the damages caused by Vaia storm in 2018, followed by dry periods. The AisaFENIX images were preprocessed to correct atmospheric, radiometric and geometric effect, and then the most frequent tree species were mapped with machine learning algorithms. For the Picea abies (Norway spruce) class, further analysis were conducted using multiple narrowband vegetation indices (Figure 1 c to h), in order to assess the health status of the trees (Figure 2) and detect the effects of the presence of the bark beetle. The results have been validated by the forestry inspectorate with ground surveys. The georeferenced thematic product obtained by the hyperspectral aerial images resulted to be very useful for optimal forest management, in particular for the identification of possible infected trees at an early stage (green-attack) and the implementation of mitigation measures. The information was available at a degree of accuracy that is not achievable by VNIR + SWIR images acquired by satellite platforms, due to the low spatial resolution. However the availability of regular and frequent satellite images has the potential to allow for temporal analysis and change monitoring, starting from the detailed as-is situation obtained from the aerial hyperspectral images.
The presentation will show the scientific approach of the work done and critically discuss the achieved accuracy in the intermediate and final products.

EURAC Auditorium

14:25

20min

Machine learning ensembles for scalable crop yield modelling: Leveraging earth observation and a cloud-integrated spatial data infrastructure to support agricultural statistics

Patric Brandt

Estimating crop yields timely, is pivotal for official statistics on agricultural productivity to inform policy-making on sustainable food production. Existing approaches of collecting yield data annually from a large number of farms are resource-intense, though. Official crop yield statistics in Germany, for instance, relies heavily on extensive and time-consuming farm surveys and on-farm measurements.

The EU’s Copernicus earth observation (EO) program provides a plethora of satellite data, enabling the remotely sensed monitoring of agricultural land at high spatio-temporal resolution. EO imagery, open geospatial data on meteorological conditions and soil properties as well as advances in machine learning (ML) provide huge opportunities for model based crop yield estimation, covering large spatial scales with unprecedented granularity. Managing vast amounts of multi-source data required for yield modelling remains a challenge, though, particularly for public authorities. We present a model-based approach to estimate yields of multiple major crops cultivated in Germany, by employing ML ensembles, using a cloud-integrated spatial data infrastructure (SDI). Our SDI is built on interconnected components linking EO cloud computation and data storage, using the CODE-DE platform, with internal data cubes through web services.

Our model-based yield estimation approach integrates a number of dynamic and static predictors. Analysis ready data of multi-spectral Sentinel-2 imagery is used for space-borne retrieval of crop traits such as leaf area index and above ground biomass. Geospatial data on meteorological time-series are queried from our data cube, providing daily variables such as temperature, precipitation, and global radiation. External geospatial data on soil moisture and physicochemical soil properties are obtained from the Copernicus Global Land Service and SoilGrids 2.0 data portals, respectively. Crop-specific ML models are trained on multi-annual data (2018 - 2022) collected at agricultural parcel level for three crops, i.e. winter wheat, winter barley, and winter rape. The ensemble of ML regressors employed, includes Gradient boosted trees (CatBoost, LightGBM, XGBoost), Partial Least Squares, RandomForest, and Support Vector Machines. Parcel geometries obtained from the Integrated Administration and Control System (IACS) enable the spatially scaled application of trained yield models, covering larger administrative regions represented by two federal states.

RSQ values of best performing models, inferred from cross validations at parcel level, range between 0.67 - 0.74. Related normalized RMSE (nRMSE) values range between 12 - 19%. Aggregated yield estimates at district level compared against mean yields at district level obtained from official yield statistics for 2020 and 2021 show RSQ values for best performing models, ranging between 0.57 - 0.85. Related nRMSE values range between 5 - 10%.

Preliminary results are promising, suggesting several advantages compared to traditional yield estimation approaches, regarding area coverage, cost effectiveness, and timeliness. Our cloud-integrated SDI used as backbone enables full scalability for crop yield estimation at national scale. However, high quality training data inferred from a representative sampling across the country and open data access for IACS parcel geometries are required to lift current scalability barriers.

EURAC Auditorium

14:30

45min

Tools for Long Time-Series Processing For Alpine Environments Monitoring

Daniele Marinelli, Davide Andreatta

In recent years, several new satellite constellations have been put into service. This, together with the new policies for open data distribution, dramatically increased the availability of time-series with high temporal resolution.

The new widespread availability of high temporal resolution imagery has led to paradigm shift from change detection techniques where pairs of images are compared searching for abrupt changes (e.g. forest fires, forest cuts), to methods capable of tracking changes continuously in time. In particular, time-series allows for the monitoring of subtle and gradual changes for which the definition of a pre and post event date is not straightforward (e.g., vegetation stress caused by drought, bark beetle outbreaks) and anthropogenic processes happening at a finer timescale (e.g. mowing events).
Such data availability, together with increasing ease of access to both offline computing power and to cloud based computing platforms and new tools for data processing, is leading to the development of a wide variety of applications for near real-time monitoring using Earth Observation (EO) data intended to be used in decision making processes (e.g., forest management) by stakeholders such as government agencies. In this context, we present monitoring tools, implemented on the Google Earth Engine platform, that exploit spaceborne EO data to support decision making in Alpine environments affected by two threats connected to global change: pests outbreaks and land use intensification.
After the Vaia storm in 2018, bark beetle outbreaks have become more frequent in the Alps with estimates, at the end of 2022, of 8000 hectares infested by the pests only in the Trento province. Such phenomena must be monitored by detecting both past and new outbreaks. This is critical for the definition of recovery strategies for the affected areas and mitigation strategies to limit the spread of new outbreaks. The developed tool analyzes long Sentinel-2 time-series for bark beetle outbreaks mapping, generating a product that identifies the area hit by an attack and the first year and month of the detection. By processing new images as they are acquired, it performs a near real-time monitoring highlighting new attacks as soon as they are visible from the satellite data. This tool is currently being used by the Forest Service of the Province of Trento that is providing the generated products to the local stations.

The second tool we present uses vegetation indices time-series derived from Sentinel-2 imagery to estimate grassland mowing frequency. Grasslands in Europe are facing management intensification in accessible areas and abandonment in marginal ones, with significant consequences not only for grassland productivity, but also for fodder quality, nitrogen leaching, animal and plant diversity and grassland recreational value. For these reasons the availability of grassland mowing frequency data can contribute to the development of more targeted conservation and management measures. The model is now being used in several research and management contexts, including CAP subsidies conditionality monitoring and habitat suitability for ground nesting endangered bird identification.

OEMC project workshop

EURAC Seminar room 2 & 3

14:45

20min

Towards a large-scale tool for estimating potential land-cover impacts from Remote Sensing

Daniel E. Pabon-Moreno

Land cover changes affect the climate system at local, regional, and global scales. Previous studies have indicated that the changes in vegetation distribution have an impact on the land surface temperature and the energy balance at local and global scales. Assessing the effect of land cover change on climate variables is a fundamental step in understanding how deforestation and reforestation processes will impact the climate dynamics. At the same time, this knowledge can be of utmost importance for the design of reforestation or afforestation plans, such as those envisaged within the European Union’s Green New Deal, but also more generally, in any part of the world. In this talk, we will present some preliminary results on the effect of land cover change on climate variables for Europe and Africa. To develop the studies, we will develop a first technical implementation of the space for time technique in the programming language Julia. The space for time technique estimates the average change of local climate if the land cover changes from one class to another. For example, savannas are usually hotter than neighboring forest, then contrasting the local climate conditions we can evaluate the potential effect of the transition from forest to savannas. In the current state of Earth observation using satellite imagery, downloading large amounts of information is no longer feasible because the amount of information is many times larger than the infrastructure of research institutions and organizations. In this context, the development of software compatible with cloud computing infrastructure is more important than ever. Julia is a dynamic programming language focused on high-performance computation and easy scalability, following the philosophy of “write like python, run like C”. Despite being a relatively new programming language (11 years old), the use of Julia in science and cloud computing has grown exponentially recently, as more and more institutes and companies adopt it. For these reasons, our implementation of the space-for-time technique on the Julia programming language, will allow scientists and organizations to efficiently perform the analysis from laptops to remote servers and platforms.

EURAC Auditorium