September 2025
The two sessions of the meeting have been blended into a single report wherever possible for clarity and completeness. Each session featured a different presentation from a data product user. The text of this summary was edited with the help of AI.
The OBIS Products Coordination Group met in two sessions, chaired by Jon Pye and Stephen Formel, with Secretariat support from Silas Principe and Laurent Chmiel. The agenda covered updates from the Secretariat, presentations from two data product users (Artificial Light at Night in the first session, and Fisheries and Oceans Canada in the second), and progress on the Products Catalogue.
OBIS on AWS Open Data. The full OBIS dataset is now available on AWS Open Data as GeoParquet, with all extensions, including all extended measurement or fact (eMoF) data, giving access to measurements. It can be queried spatially and is a cloud-performant data source, updated weekly (https://github.com/iobis/obis-open-data; full documentation at https://github.com/awslabs/open-data-registry/blob/main/datasets/obis.yaml). Questions on how to use it can be sent to the OBIS Secretariat. OBIS species distribution maps will also soon be hosted on AWS, in cloud-performant formats; an agreement has already been signed.
A paper using the full OBIS dataset. Following the discussion at a previous meeting about a paper that used the full OBIS dataset without appropriate citation, OBIS met with the authors. The conversation went well: OBIS minted a DOI for the dataset, which will be published with a revision of the paper, and the authors plan to bring their product to the OBIS catalogue.
Citing OBIS data. There is now clearer guidance on how to cite OBIS data, using a new DOI system that combines the datasets used in an analysis under a single DOI and tracks the full provenance of the data. The service will soon be available to Node managers, and can be mobilized on request in the meantime.
Journals requiring data publication. Stephen Formel had seen a recommendation from an ICES journal that data supporting submitted papers be published to GBIF as part of the submission process. This can also be done through OBIS, and the Secretariat will try to contact a few journals to extend this kind of requirement and recognize OBIS’ ability to fulfil it.
→ Clara Baringo noted that the journal Neotropica already does this, and asked how researchers can best be directed to the right place; ideally, they should publish through their national Node. John Nicholls suggested that a template statement to share with journals would be useful. Katherine Tattersall suggested this is an interesting topic for the DCG and proposed raising it at a future meeting.
Presented by Tim Smyth (session 1)
Tim Smyth described growing awareness of light pollution, which has increased significantly in less than a hundred years. To estimate how widespread marine ALAN is, the models use a critical depth based on a light threshold of 0.1 µW m⁻². Globally, around 22% of coasts are impacted, an area of about 2 million km². A wide variety of studies have demonstrated impacts of artificial light, to varying degrees, on diverse taxonomic groups and at different scales. The team has modelled the depth below which the light threshold is reached, together with depth intensity categories from 1 to 4, which can be explored in an online viewer (https://goalann.eofrom.space/?state=dacfe9). The team also has its own in-house data portal, PIER, for the atlas (https://pml.ac.uk/science/External-data-products/). The slides can be shared.
→ Asked how often the model is recalculated and how long this takes, Tim Smyth explained that the model is now on version 2, which is hoped to be the stable version for the foreseeable future. A single snapshot does not capture all the variation: clear and cloudy nights, for example, reflect different amounts of light.
→ Asked for examples of decision makers using the maps, he described traction in several areas. The Great Barrier Reef Authority in Australia was keen to explore the issue, but the reef proved to be too far offshore to be much affected, and the authority was able to use the data convincingly to address that concern. A version of the portal is also part of GAZAM, a UN Ocean Decade activity, and there has been good traction with OSPAR and Blue Flag International. The initial idea of approaching OBIS came through the Inter-American Convention for the Protection and Conservation of Sea Turtles: if the information could be combined with OBIS habitat maps, the convention could bring it all together. It was noted that much of the relevant sea turtle data probably comes through OBIS-SEAMAP.
→ Asked whether the model has been validated in the field, he explained that much of the input comes from ESA ocean colour data, and the model includes sediment, CDOM (coloured dissolved organic matter) and phytoplankton. Validation is hard, because many radiometric instruments lack the sensitivity needed at these light thresholds.
Follow-up. The ALAN team had a follow-up conversation about the potential for maps of sea turtles and sea turtle nesting sites, for example.
Presented by Rémi Daigle and Jaimie Harbin, DFO (session 2)
Rémi Daigle and Jaimie Harbin presented an application that applies an Ecosystem-Based Management Framework, currently covering only the ecological side but with the hope of expanding. Canada has multiple conservation frameworks. The application calculates indicators for all sites, presents them as flower plots, and can incorporate any relevant data. OBIS is one of many data sources, used for OTN data, unique species, and richness and representation, with functional traits, native and non-native species, and sample coverage planned for the future. All data is currently used at once, but richness will now be calculated per project in OBIS. OTN data is also used to support ProtConn connectivity analysis. Jaimie Harbin gave a demonstration of the application.
The code is on GitHub under the group’s organization (https://github.com/dfo-mar-odis). The public-facing application is at https://github.com/dfo-mar-mpas/MarConsNetApp, as DFO datasets are often not public; the analysis lives in dfo-mar-mpas/MarConsNetAnalysis, and the code for pulling datasets in https://github.com/dfo-mar-mpas/MarConsNetData.
The aim of the catalogue is to bring together, in one place, the products developed by the Secretariat and the OBIS community, making them easier to access and discover. Development, which has been the Secretariat’s main focus, is slow but steady, and current work is devoted to updating the schema. The metadata model will make it possible to connect with ODIS and other systems. A more in-depth demonstration is planned for the meeting in Colombia.
Structure. CKAN has three basic parts: datasets, organizations and groups. Organizations correspond to OBIS Nodes and the Secretariat, and each dataset can belong to only one organization. Groups correspond to the institutions in the OBIS database, and a dataset can belong to any number of them. Groups and organizations are pulled directly from OBIS sources (the OBIS API and OceanExpert).
→ Ward Appeltans stressed the importance of tagging products by EOV, since the plan is to create products that can be used for CBD reporting and similar purposes.
Naming. Asked whether organizations are publishers, it was explained that the name is flexible and could be changed to “publisher” if more appropriate. Clara Baringo gave the example of traditional groups that publish as a collective rather than an organization, for which “publisher” would be a suitable alternative. Silas Principe noted that “datasets” might also need a different name, to avoid confusion with OBIS datasets; the discussion had started previously but needed to go further. Katherine Tattersall agreed, suggesting “products” as a good alternative, and Jon Pye suggested “data products” as the more precise term, which avoids the problem.
→ Clara Baringo also shared an example of a tool that could be a product: Mycena, which her team uses to organize data in the MoF extension (https://mycena.sibbr.gov.br/).