Webinar “PHIDIAS HPC – Building a prototype for Earth Science Data and HPC Services”
The first PHIDIAS HPC webinar took place on the 13th of February, addressing the HPC European community and conforming to the scope of PHIDIAS project, which aims at developing and implementing a set of interdisciplinary services and tools for Earth Sciences based on HPC.
The Webinar, titled “PHIDIAS HPC – Building a prototype for Earth Science Data and HPC Services”, was intended to introduce the PHIDIAS HPC initiative to the European HPC and Research community, targeting also Big data scientists, Earth observation researchers, Marine life experts and Research and academies. It gathered members of Research Centres, experts in Marketing and PR and professionals working in aerospace and defence fields.
The webinar has been an exhaustive and open opportunity to showcase how PHIDIAS aims at increasing the HPC and Data capacities of the European Data by pursuing the following objectives:
- Building a prototype for earth scientific data
- Enabling Open Access to HPC Services
- Strengthening FAIRisation
- Creating a framework combining computing, dissemination and archiving resources
The webinar was introduced by Boris Dintrans, the Director of National Computing Centre for Higher Education (CINES), a French public institution, located in Montpellier (France) and supervised by the French Ministry for Higher Education and Research. As the project coordinator of PHIDIAS, he presented what the EU-funded project can offer to the HPC community, highlighting the main challenges that will be addressed, which are the following:
- Handling the diversity of data coming from the Earth System Research Infrastructure.
- Developing and testing transversal methods and tools that can be applied to data coming from other scientific domains such as health and environment data.
- Dealing with the scalability of Data processing tools.
- Working on industrialization and strength development of HPC/HPDA/AI workflows.
The first key speaker was Pascal Prunet, Chief Executive Officer at SPASCIA, who brought his wide expertise in the field of interpretation of remotely sensed data to explain how it is possible to improve the efficiency of intelligent screening of environmental satellite data through two processing prototypes: PCA-based screening of L1 data (SWIR) for detection of extreme events, and new AI methods for objective/automatic detection of plumes from L2 products.
The second contribution was given by Jean-Christophe Desconnets, senior engineer at IRD, and geomatic scientist at MICADO research team of ESPACE-DEV. His intervention focused on on-demand image processing for environmental monitoring and showcased two examples of uses cases coming from environmental monitoring community: Sentinel-1/Sentinel-2-derived Soil Moisture product at Plot scale (S2MP) over agricultural areas, and Remote sensing images processing with artificial intelligence.
The last Use Case that PHIDIAS will be developing was introduced by Cecile Nys, assistant manager of Ocean Data cluster at Ifremer, who boosted the use of cloud services for marine data studies, highlighting the need to combine and collocate data from several data sources, both in situ and satellite, and to adopt new data structures based on big-data technologies.
The final presentation was given by Aleksi Kallio, development manager of data analytics at CSC – IT Centre of Science Ltd., who focused on AI services targeting the user communities.
The webinar led to some main takeaways, which are the following:
- It is extremely important to make use of the huge quantity of data offered by Earth observation from space which at the moment is not fully exploited: PHIDIAS will be working to provide real-time detection and analysis of extreme events, such as air pollution emission plumes.
- It is essential to merge scientific experimental algorithms and catalogues with community user-needs.
- Taking advantage of HPC architecture and big data infrastructure will facilitate selection and image analysis activities for environmental monitoring.
- Current data structures of in-situ marine data are not very efficient. This leads to the need of new ways to visualize data and to process these data, as well.
If you missed the webinar, you can find it recorded at this link.
