Community science data fill biodiversity knowledge gaps for assessing the status of bees globally

Wednesday, November 16, 2022

1:54 PM – 2:06 PM PT

Location: Vancouver Convention Centre, Meeting Room 118

Presenting Author(s)

Keng-Lou James Hung

Assistant Professor
University of Oklahoma
Norman, Oklahoma

Co-Author(s)

PC

Paige R. Chesshire

Northern Arizona University
Flagstaff, Arizona
Michael C. Orr

Group Leader
Staatliches Museum für Naturkunde Stuttgart
Bethesda, Maryland
AH

Alice C. Hughes

University of Hong Kong
Hong Kong, Guangdong, China (People's Republic)

Co-Author(s)

Jessica L. Mullins

PhD Student
University of California
La Jolla, California

Co-Author(s)

KP

Katherine A. Parys

Research Entomologist
USDA - ARS
Stoneville, Mississippi
PS

Patricia Simpson

Independent Researcher
San Diego, California
Lindsie M. McCabe (she/her/hers)

Research Entomologist
USDA-ARS Pollinating Insect Research Unit
Logan, Utah
NC

Neil S. Cobb

BON-EARTH
Flagstaff, Arizona
JA

John S. Ascher

National University of Singapore
Singapore, Singapore

Reports of bee declines globally are of great concern, but available digitized specimen records may not be sufficient for assessing the distribution and conservation status of bees. In recent years, image-based community science (CS) records have emerged as a powerful data source for a variety of taxa. Here, we mobilize nearly one million expert-validated CS records for nearly 3000 bee species from the leading CS platform, iNaturalist, to examine the degree to which these records increase spatiotemporal, taxonomic, and ecological coverage of bees globally. We found that CS records of bees were globally distributed and growing exponentially, with particularly strong contributions in regions with sparse digitized specimen data. Despite a dearth of image-based identification resources for bees, the majority of global CS records have been identified to ecologically relevant levels (i.e., at least to subgenus) that enable assigning putative life history traits. Higher-resolution analysis of data from the contiguous United States revealed numerous species occurrence “gaps” in contemporary digitized specimens that are filled by CS records. Additionally, CS records filled temporal data gaps during under-sampled months, as well as ecological data gaps with respect to host plant associations. Taxa routinely identified by experts from CS images are functionally and taxonomically diverse, and a subset of such species are shown to be better monitored through iNaturalist than through digitized specimens. When appropriately mobilized, CS records can greatly enhance ecological and distributional modeling of bee taxa and enable monitoring of sizable portions of bee assemblages in near real-time.