ALBERT

Description: Highlights: • Open-access online scholarly biodiversity databases are threatened by a lack of funding and institutional support. • Strategic approaches to aid sustainability are summarised. • Issues include database coverage, quality, uniqueness; clarity of Intellectual Property Rights, ownership and governance. • Long-term support from institutions and scientists is easier for high-quality, comprehensive, prestigious global databases. • Larger multi-partner governed databases are more sustainable; i.e. ‘bigger (multi-partner) databases are better’. Abstract: Scientists should ensure that high quality research information is readily available on the Internet so society is not dependant on less authoritative sources. Many scientific projects and initiatives publish information on species and biodiversity on the World Wide Web without users needing to pay for it. However, these resources often stagnate when project funding expired. Based on a large pool of experiences worldwide, this article discusses what measures will help such data resources develop beyond the project lifetime. Biodiversity data, just as data in many other disciplines, are often not generated automatically by machines or sensors. Data on for example species are based on human observations and interpretation. This requires continuous data curation to keep these up to date. Creators of online biodiversity databases should consider whether they have the resources to make their database of such value that other scientists and/or institutions would continue to finance its existence. To that end, it may be prudent to engage such partners in the development of the resource from an early stage. Managers of existing biodiversity databases should reflect on the factors being important for sustainability. These include the extent, scope, quality and uniqueness of database content; track record of development; support from scientists; support from institutions, and clarity of Intellectual Property Rights. Science funders should give special attention to the development of scholarly databases with expert-validated content. The science community has to appreciate the efforts of scientists in contributing to open-access databases, including by citing these resources in the Reference lists of publications that use them. Science culture must thus adapt its practices to support online databases as scholarly publications. To sustain such databases, we recommend they should (a) become integrated into larger collaborative databases or information systems with a consequently larger user community and pool of funding opportunities, and (b) be owned and curated by a science organisation, society, or institution with a suitable mandate. Good governance and proactive communication with contributors is important to maintain the team enthusiasm that launched the resource. Experience shows that ‘bigger is better’ in terms of database size because the resource will have more content, more potential and known uses and users of its content, more contributors, be more prestigious to contribute to, and have more funding options. Furthermore, most successful biodiversity databases are managed by a partnership of individuals and organisations.

Description: For the last 15 years, Biodiversity Information Standards (TDWG) has recognized two competing standards for organism occurrence data, ABCD (Access to Biological Collections Data; Holetschek et al. 2012) and DarwinCore (Wieczorek et al. 2012). These two representations emerged from contrasting strategies for mobilizing information about organism occurrences (also commonly called species occurrence data). ABCD was capable of representing details of more kinds of information, but was necessarily more complicated. DarwinCore, on the other hand, was simpler but more limited in its ability to represent data of different kinds and formats. TDWG endorsed both standards because the different projects and communities that generated them remained dedicated to their different strategies and tool sets, and the Global Biodiversity Information Facility (GBIF) developed the ability to integrate data published in either standard. Since their inceptions, DarwinCore and ABCD have become more similar. DarwinCore has gotten more complicated through the addition of terms and has begun to assign terms to classes. ABCD is now expressed in RDF (Resource Description Framework), potentially enabling re-use of terms with alternative structures among classes. At the same time, methodologies for conceptual modeling and representing complex scientific data have continued to evolve. In particular, a suite of modeling and data representation methods related to linked data and the semantic web, i.e., RDF, SKOS (Simple Knowledge Organization System), and OWL (web Ontology Language), promise to make it easier for us to reconcile shared concepts among different representations or schemas. A mapping between ABCD 2.1 and DarwinCore has existed since before 2005.*1 ABCD 3.0 and DarwinCore are both now represented in RDF. In addition, the BioCollections Ontology (BCO) covers many of the shared concepts and is derived from the Basic Formal Ontology (BFO), an upper level ontology that has oriented many other biomedical ontologies. Reconciling ABCD and DarwinCore through alignment with BCO (in the OBO Foundry; Smith et al. 2007) would better connect TDWG standards to other domains in biology. We appreciate that many working scientists and data managers perceive ontologies as overly complicated. To mitigate the steep learning curve associated with ontologies, we expect to create simpler application profiles or schemas to guide and serve narrower communities of practice within the wider biodiversity domain. We also plan to integrate the current work of the Taxonomic Names and Concepts Interest Group and thereby eliminate the redundancy between DarwinCore and Taxonomic Concepts Transfer Schema (TCS; Kennedy et al. 2006). At the time of this writing, we have only agreements from the authors (i.e., conveners of relevant TDWG Interest Groups and other key stakeholders) to collaborate in pursuit of these common goals. In this presentation we will give a more detailed description of our objectives and products, the methods we are using to achieve them, and our progress to date.