COS 185-1 - An overview of variation in biodiversity across the National Ecological Observatory Network (NEON)

Wednesday, August 9, 2023

1:30 PM – 1:45 PM PDT

Location: B110

Presenting Author(s)

Eric R. Sokol

Ecologist
National Ecological Observatory Network (NEON), Battelle
Boulder, Colorado, United States

Coauthor(s)

SP

Stephanie M. Parker

National Ecological Observatory Network (NEON), Battelle, Colorado, United States
SR

Sydne Record

Associate Professor
Department of Wildlife, Fisheries, and Conservation Biology, University of Maine, Maine, United States

Abstract: Large, open-access datasets can facilitate ecological synthesis and promote the democratization of science by providing access to freely available high quality ecological data. The National Ecological Observatory Network (NEON) is funded by the US National Science Foundation to collect and publish such data from 81 sites across the US for 30 years. Here, our goal is to summarize the variation in biodiversity data across the observatory for nine sentinel taxonomic groups, including algae, carabid beetles, birds, fishes, aquatic macroinvertebrates, mosquitoes, plants, small mammals, and ticks. We used the ecocomDP R package to access NEON organismal data using a standardized data model in which the data have been prepared for community ecology analyses (e.g., abundances standardized to sampling effort). As the observatory collects and publishes more data, the growing size of NEON datasets will present a challenge to researchers. Thus, we present a workflow that uses the arrow package for R to select and analyze manageable chunks of data. We used the vegan package for R to carry out a variation partitioning analysis for each taxonomic group at each NEON site by fitting distance-based RDA models to calculate the adjusted R² coefficients for each variable that was a candidate to explain variation in community composition. We partitioned variation into inter-annual, intra-annual (seasonal), and intra-site spatial (e.g., variation in community composition among habitats, plots, etc.) components. Data used in this analysis include the NEON 2023 Data Release (including observations from 2013 through 2021) and provisionally published 2022 data. Total site-level variation explained by temporal and spatial variables (adjusted R²) ranged from 24% (11% SD) for carabid beetles to 44% (6% SD) for zooplankton. Temporal components were largest in aquatic taxa, with zooplankton showing the largest inter-annual (25%, 9.6% SD) and seasonal (12%, 2.9% SD) fractions. Spatial variability was more prevalent in terrestrial taxa, where plants showed the most spatial variability with 11% (6.7% SD) explained by National Land Cover Database (NLCD) class and 36% (10% SD) explained by plot. Our assessment shows that a quarter to half of the site-level variability in community composition can be attributed to spatial (e.g., habitat type) and temporal (e.g., month, year) sampling design variables. We hope that this overview of sources of variation in NEON biodiversity data can inform potential data users and facilitate the use of these data in future cross-network synthesis efforts.