Periodic Reporting for period 2 - AD4GD (All Data 4 Green Deal - An Integrated, FAIR Approach for the Common European Data Space)
Okres sprawozdawczy: 2024-03-01 do 2025-08-31
AD4GD (All Data for Green Deal)’s mission is to conduct research and propose components that shape the GDDS as a hub for FAIR (Findable, Accessible, Interoperable, Reusable) data and standards-based services supporting key priorities of the Green Deal: reducing pollution, preserving biodiversity and mitigating climate change. On the one hand, this means bringing together data from multiple domains and sources. AD4GD provides meaning to environmental data by adding links to well-known vocabularies, concepts and units of measurement enabling the sharing and exchange of data in an understandable way. In this respect, AD4GD leverages the Essential Variables framework from the Group on Earth Observations (GEO). The project demonstrates that it is possible to integrate remote sensing data from satellites, in-situ sensors from consolidate research infrastructures (e.g. GBIF), sensors from the Internet of Things (IoT) and Citizen Science, as well as socioeconomic data in an interoperable, scalable and reliable manner. The performance and interoperability between the developed components are demonstrated through three pilot cases focused on water quantity and quality, biodiversity connectivity and air quality.
AD4GD demonstrated how to create semantically enriched formats in an automatic way applying vocabularies and data model schemas. When relevant concepts cannot be found in existing vocabularies, the OGC RAINBOW is used to import and organize these concepts, as done in the Essential Biodiversity Variables. Other domains scientists can read the enriched data, interpret the variable definitions, and units of measure added to the numerical or categorical values of the observations and know the meaning of the data. In-situ data can subsequently be exported automatically as a Sensor Things API (STA) service. Searching in a semantically enriched data catalogue coming from the official sources as well as Citizen Science, Artificial Intelligence models can find datasets that are useful for training data. AD4GD has adopted the OGC API Coverage standard to represent datasets organized as multidimensional data cubes.
The results were exposed in standards from the International Data Space Association (also implemented by the SIMPL platform). We deployed the current version of the Eclipse Dataspace Connector (EDC) to create a prototype of the GDDS. AD4GD, USAGE and FAIRiCUBE projects, connected 3 instances allowing secured file exchange of data among the 3 projects, including static files and OGC APIs query results. It is expected that these 3 EDC nodes, and the semantic uplift mechanism, constitute the embryo of a future GDDS the SAGE project will deploy in the next 2.5 years. AD4GD explored alternative solutions for the exchange of restricted data in a data space, such as FACTS smart certificates for immutable assets.
To demonstrate approaches to zero pollution in water reservoirs, AD4GD selected Berlin’s small lakes. This pilot tested strategies for measuring water quality using remote sensing, deployed new in-situ oxygen sensors, experimented with algorithms that detect lakes in need of urgent interventions and implemented a ML model to predict water quantity based on three different scenarios of rain intensity. Terrestial habitat connectivity affects the distribution of species within ecosystems. AD4GD conducted a pilot to accelerate the computation of connectivity indices by combining data coming from remote sensing maps, species occurrence observations (some contributed by citizens) and sensor data to enhance regional and local actions. Poor urban air quality causes serious health issues across Europe. The results demonstrate that the resolution of the Copernicus Atmosphere Monitoring Service (CAMS) can be improved in cities by including Citizen Science air quality sensors collected by communities such as the Sensor.Community network.
The results were exposed in standards from the International Data Space Association (also implemented by the SIMPL platform). We deployed the current version of the Eclipse Dataspace Connector (EDC) to create a prototype of the GDDS. AD4GD, USAGE and FAIRiCUBE projects, connected 3 instances allowing secured file exchange of data among the 3 projects, including static files and OGC APIs query results. It is expected that these 3 EDC nodes, and the semantic uplift mechanism, constitute the embryo of a future GDDS the SAGE project will deploy in the next 2.5 years. AD4GD explored alternative solutions for the exchange of restricted data in a data space, such as FACTS smart certificates for immutable assets.
To demonstrate approaches to zero pollution in water reservoirs, AD4GD selected Berlin’s small lakes. This pilot tested strategies for measuring water quality using remote sensing, deployed new in-situ oxygen sensors, experimented with algorithms that detect lakes in need of urgent interventions and implemented a ML model to predict water quantity based on three different scenarios of rain intensity. Terrestial habitat connectivity affects the distribution of species within ecosystems. AD4GD conducted a pilot to accelerate the computation of connectivity indices by combining data coming from remote sensing maps, species occurrence observations (some contributed by citizens) and sensor data to enhance regional and local actions. Poor urban air quality causes serious health issues across Europe. The results demonstrate that the resolution of the Copernicus Atmosphere Monitoring Service (CAMS) can be improved in cities by including Citizen Science air quality sensors collected by communities such as the Sensor.Community network.
The most practical components resulting from the research done in AD4GD are:
* Automated ingestion routines of air quality observations coming from diverse low-cost sensors. This integration of several data sources was applied in the air quality pilot.
* A Data Trustworthiness Framework to validate the results of the 3 pilots: water availability prediction, connectivity calculation, and air quality distribution.
* A Green Deal Information Model that combines common vocabularies as a basis for the Green Deal Data Space.
* A Semantic Uplift pipeline that generates semantically tagged datasets based on RDF and dynamic Sensor Things API services.
* An OGC RAINBOW server ingested with the Green Deal Information Model and the Essential Biodiversity Variables framework.
* TAPIS, "Tables from APIs ", as a client application developed in HTML and JavaScript to support interoperability testing, data visualization and analysis.
* An OGC API Coverages connector to the Open Data Cube that provides easy access to the data contained.
Scientific results and tools for scientists include:
* Evaluation of the Eclipse Data Connector solution for the exchange of files and OGC Web APIs outputs between two internal nodes and with the FAIRiCUBE and USAGE project connectors.
* Acceleration of habitat connectivity calculations with optimized classical algorithms but also with machine learning trained models.
* New Open Graph oriented workflows for habitat connectivity computation.
* Water modeling and prediction of the status an evolution of water quantity on Berlin lakes.
* Water assessment of water quality on Berlin lakes with remote sensing
* An air quality routine that enhances the resolution of CAMS air quality distribution products in cities.
* A metadata catalogue based on GeoNetwork for the FAIR exploitation of the project results and a transformation of this metadata to the EDC catalogue.
* Automated ingestion routines of air quality observations coming from diverse low-cost sensors. This integration of several data sources was applied in the air quality pilot.
* A Data Trustworthiness Framework to validate the results of the 3 pilots: water availability prediction, connectivity calculation, and air quality distribution.
* A Green Deal Information Model that combines common vocabularies as a basis for the Green Deal Data Space.
* A Semantic Uplift pipeline that generates semantically tagged datasets based on RDF and dynamic Sensor Things API services.
* An OGC RAINBOW server ingested with the Green Deal Information Model and the Essential Biodiversity Variables framework.
* TAPIS, "Tables from APIs ", as a client application developed in HTML and JavaScript to support interoperability testing, data visualization and analysis.
* An OGC API Coverages connector to the Open Data Cube that provides easy access to the data contained.
Scientific results and tools for scientists include:
* Evaluation of the Eclipse Data Connector solution for the exchange of files and OGC Web APIs outputs between two internal nodes and with the FAIRiCUBE and USAGE project connectors.
* Acceleration of habitat connectivity calculations with optimized classical algorithms but also with machine learning trained models.
* New Open Graph oriented workflows for habitat connectivity computation.
* Water modeling and prediction of the status an evolution of water quantity on Berlin lakes.
* Water assessment of water quality on Berlin lakes with remote sensing
* An air quality routine that enhances the resolution of CAMS air quality distribution products in cities.
* A metadata catalogue based on GeoNetwork for the FAIR exploitation of the project results and a transformation of this metadata to the EDC catalogue.