The natural history domain is rich in information. For hundreds of years, biodiversity researchers have collected specimens and samples, and meticulously recorded the how, what, and where of these objects of research. To retrace this information, however, deep knowledge of the collection and patience is necessary. Whereas traditional access methods (e.g., analysing paper logs of specimen finds) can be used for smaller collections, the sheer size of most current natural history collections prohibits this. At the same time, information technology has advanced to the point where it is able to capture the intricacies of biodiversity collection information and provide the first steps towards full digital access. The need for collection information access is dire, as lack of access impairs our ability to answer questions about species biodiversity, diversity and change through time (Scoble, 2010). Examples from the young field of biodiversity informatics stress that in order to assess and tackle problems such as predicting a species’ reaction to changing environment or prioritisation of preservation policies, digitisation of and access to (large) collection databases is imperative (Guralnick & Hill, 2009; Johnson, 2007; Raes, 2009; Soberón & Peterson, 2004). Although much progress has been made, for example with the Global Biodiversity Data Portal (Berendsohn et al., 2010), many collections have not yet been (fully) digitised. In this contribution, we first present a new approach to collection digitisation, as well as a novel collection registration management system (CRS) as implemented at the Netherlands Centre for Biodiversity (NCB Naturalis). The new approach to digitisation at NCB Naturalis implements a cascaded digitisation approach: in parts of the collection that have not yet been digitised, first a shelf or drawer is assigned a unique ID in the CRS, along with a description of the specimens contained within it. Whenever the shelf or drawer is revisited, the new policy dictates that specimens that are taken and used from this set be recorded in the CRS. Furthermore, the CRS is linked to taxonomic resources, which enable integration with reference sources. We present two use cases that illustrate the benefits for smarter collection information management systems, employing natural language processing techniques.

Contains fulltext : 94566.pdf (Publisher’s version ) (Open Access)