Abstract Detail

Floristics & Taxonomy

Webb, Campbell O [1], Ickert-Bond, Steffi [2].

Integrating taxonomic resources for Alaska plants.

Underlying any new taxonomic work, whether a regional checklist or a taxonomic revision, is the integration and review of preexisting information in publications, specimens, and images. Traditionally this integration required painstaking reading, transcribing and collating by a human researcher. As more and more information is available in digital form, much of this integration can now be performed as batch operations, either using existing cross-linkages among data resources, or by writing bespoke text processing programs to map identifiers between isolated resources. This latter process represents a form of digital curation that is fundamental to making efficient use of existing taxonomic work. As part of an ongoing, multi-stakeholder initiative to produce a new Flora of Alaska, we have integrated (i) taxonomic information in: Hultén’s (1968) Flora of Alaska and Neighboring Territories, the Flora of North America, Kew’s World Checklist of Selected Plant Families, the Panarctic Flora checklist, Alaska checklists from ALA (University of Alaska, Fairbanks) and the Alaska Center for Conservation Science, (ii) occurrence and image information from the ALA herbarium database (in Arctos), the Global Biodiversity Information Facility, and iNaturalist, and (iii) a new compilation of morphological character data. Scientific names are the identifiers in each resource that permit this integration, but the same name is often represented in different resources but a different character string due to transcription errors in the epithets and variation in how the name authority elements are presented. We developed a software tool (‘matchnames’) to compare two lists of names and find matches by using (i) approximate (“fuzzy”) regular expressions, (ii) a set of rules about the construction of the botanical authority elements, and (iii) decisions by a human operator. This tool has greatly sped up the integration of taxonomic resources. In addition to matching names, we also aligned taxonomic concepts for a subset of complicated taxa; this process still requires intensive human comprehension and inference. The outcome of this integration can be thought of as a giant network, with nodes being name usages (in documents and specimen labels), and edges of different types joining orthographically mapped names, taxonomic synonyms, and taxonomic concepts. This network can be queried to produce web-based applications, or readable, text-based compendia, and can be serialized in various ways for transfer and archiving (e.g., as SQL, XML, or RDF). We will discuss various choices we made and a web app we built, and present the generic steps of our pipeline as a template for checklist-making. For the larger Flora of Alaska project there will still be subsequent steps of choosing accepted names and writing new descriptions. However our integrated review of taxonomic names, concepts and occurrences also stands alone as a statement of work to date.

1 - University of Alaska Fairbanks, Museum of the North, Fairbanks, AK, 99775, USA
2 - University of Alaska Fairbanks, Herbarium (ALA) and Dept of Biology and Wildlife, Fairbanks, AK, 99775, USA

Keywords:
Biodiversity Informatics
Alaska
taxonomic names
taxonomic concepts.

Presentation Type: Oral Paper
Number: F&T II002
Abstract ID:642
Candidate for Awards:None