GEOLocate World: An Expanded Tool for Georeferencing Natural History Collections

Tulane University, New Orleans LA

Investigators

Abstract

Tulane University is awarded a grant to expand the GEOlocate tool to world-wide coverage, extending GEOLocate's architecture to include multilingual support, and develop a taxonomic footprint verification service to expedite visualization and correction of GEOLocate-determined coordinates. Natural History collections have traditionally served as a resource for research on the taxonomy and systematics of organisms. Increasingly, specimens and their data are being used in studies of ecology, conservation and environmental biology. A major hurdle in implementing such studies with museum data is that collection localities in most museum databases are not georeferenced (i.e., they lack latitude and longitude coordinates). In the vast majority of instances, the collection locality is recorded or digitized as a string of text describing the location as a position along a road some distance and compass direction from a town or other geographic point of reference. Traditional methods of georeferencing these data (capturing map coordinates) are tedious and time consuming, typically involving finding the locality on either hardcopy or digital maps, plotting the locality, and determining the coordinates by hand or with the aid of the computer. In prior support the PIs have developed a software tool called GEOLocate that uses a patternmatching algorithm to break down the locality text string into basic directional, distance and geographic place components. The algorithm uses place-name databases and other information sources to quickly return fairly accurate latitude and longitude coordinates. GEOLocate includes visualization, correction and error-estimation features that allow users to more precisely fix a position and return geo-coordinates to a database. The current version GEOLocate was developed for georeferencing North American collections (United State, Canada and Mexico) and is highly effective in this area. The tool has been distributed to over 140 users and is gaining wide acceptance in the natural history collection community. This award is about expanding coverage of the tool to the entire world and making it multilingual to serve international users and to allow North American users to georeference international collections. A world version of GEOLocate will be a valuable tool for georeferencing the estimated 3 billion specimen records in the world's natural history collections and would help facilitate uses of specimen data in the emerging field of biodiversity informatics. As such, GEOLocate World would be an important component of computing toolkits in the Global Biodiversity Infrastructure Facility (GBIF) and the proposed Legacy Infrastructure Network for Natural Environments. The taxonomic footprint verification service will allow experts to visualize taxonomic data in a geographic context. This project will provide cross-discipline training for two graduate students from computer sciences and biological sciences.