Abstract Detail

Biodiversity Informatics & Herbarium Digitization

Quantifying Error in Occurrence Data: A Framework for the Use of Citizen Science and Digitized Herbarium Data in Plant Families of the Southeastern United States.

iNaturalist is a database with over 50 million observations and it therefore has the potential to be a rich source of organismal occurrence data for years since its creation in 2008. From limited past studies assessing the taxonomic accuracy of citizen science databases such as iNaturalist, there is reason to believe that these records are accurately identified in certain geographic regions, particularly for charismatic organisms (e.g., birds). Few studies have aimed to quantify the accuracy of iNaturalist identifications across a broader scale of less readily “identifiable” organisms, such as vascular plants. iNaturalist serves as a contrast to herbarium data, which currently are the primary source of occurrence data for large-scale biogeographic studies/niche models in plants. Misidentification in herbarium specimens is assumed to be at a low level, under the assumption that curators are keeping up with updating and checking identifications of specimens, but the level at which there is a misidentification rate across varying groups of plants (e.g., different families of angiosperms) has not been quantified in herbarium or iNaturalist data. Quantifying the rate of misidentification could lead to a better understanding of the types of data being used and how they compare to one another. Preliminary results in families native to the Southeastern United States (Ericaceae, Gentianaceae, Cyperaceae) have shown a comparable rate of misidentification among digitized herbarium specimens and research grade iNaturalist observations within the study area, and found biases in plant observations towards more “common” species in iNaturalist data. This provides a framework for understanding how each of these data types can be improved upon, and how these types of data could be used together to build accurate models that can capture change in species ranges and requirements over time.

1 - University of Florida, Biology, 1659 Museum Rd, Gainesville, FL, 32611, USA

Keywords:
none specified

Presentation Type: Oral Paper
Number: BI&HD I006
Abstract ID:1013
Candidate for Awards:None