Missouri Botanical Garden Open Conference Systems, TDWG 2016 ANNUAL CONFERENCE

Font Size: 
AMAZONFISH: collating, curating and publishing fish occurrence data for the Amazon river basin
Aaike De Wever, Céline Jézéquel, Koen Martens, Thierry Oberdorff

Last modified: 2016-09-29

Abstract


The Amazon basin is one of the regions on earth with the highest freshwater biodiversity. Although fish species are generally well studied, with an estimated 2400 species for the basin, the knowledge on the spatial distribution of these organisms is still greatly deficient and large taxonomic and sampling gaps prevent a comprehensive analysis and modeling. The AMAZONFISH project (http://amazon-fish.com) aims to build the largest freshwater fish biodiversity database for the entire Amazon basin, to fill these gaps in order to improve our understanding of this unique ecosystem.

To do so, we will mobilize and integrate existing data from different sources including museum and university collections, and on-line databases together with newly generated data from sampling campaigns and data extracted from the literature. The structure of the database is constructed to facilitate data export in the Darwin Core standard and to cover both point and basin level data (the latter using the HydroBASINS catchment delineation). Major sources of data include GBIF (the Gloabl Biodiversity Information Facility), FishNet, SpeciesLink and IABIN (the Inter-American Biodiversity Information Network). One of the project's challenges is to identify the original data sources from these data publishers, identify overlaps and avoid duplication. Integrated data is checked for systematic reliability and consistency, using the FishBase Consortium database and the California Academy of Science’s Catalog of Fishes as the nomenclature authority upon import, and will be subjected to further quality control as outlined in the project’s Data Management Plan. In order to ensure the public availability of the data, and to foster its reuse in comprehensive analyses and modelling work, we will both encourage the publication of original raw data, and work out a solution for releasing the curated occurrence data at either (sub)basin level and/or point locality level.