The goal for the National Ecological Observatory Network (NEON) as it ends construction and enters into full operation is to provide freely available, continental-scale data for ecologists to address the Grand Challenges of Environmental Science described by the National Research Council in 2001. Accurately documenting trends in populations and communities is paramount to understanding the impacts of climate change, land-use change, and invasive species on biodiversity across the ecosystems in the observatory. NEON biomonitoring efforts focus on sentinel taxa (listed at https://www.neonscience.org/data-collection/observational-sampling). The data products are designed to characterize the structure of each assemblage at the scale of a NEON site (34 aquatic and 47 terrestrial), and the sites are distributed among 20 eco-climatic domains to capture continental scale variation. Here, we present a site-scale assessment of NEON data (https://data.neonscience.org) collected during the construction phase of the observatory. Specifically, we characterized (1) the spatial scales at which NEON captures variation in measures of population and community structure and (2) the effect sizes that are detectable given the levels of spatial and temporal variation that have been observed to date. Our assessment used generalized linear mixed effects models, with zero-inflation terms when appropriate, and a power analysis based on numerical simulations.
Results/Conclusions
Approximately 80% of sites had between one and four years of data available for this assessment, which will provide guidance to the ecological community about where and how the observatory captures variation in each data product. For example, total herbaceous above ground biomass varied most among NLCD class at 10 sites, among plots at 13 sites, among subplots at four sites, and among years at 11 sites (38 sites analyzed). When modeled as a fixed effect, significant inter-annual variation was detected in 15 (out of 36) sites. The power analysis showed that given current levels of variation, we can detect a 20% year-to-year change with a power of 0.80 at the majority of sites. We have also applied this approach to assess our ability to detect change in species counts, relative abundances, percent cover, or other metrics used to quantify population and community structure for sentinel taxonomic groups across the observatory. Overall, this analysis demonstrates that effect sizes that are typically reported in the ecological literature fall within the range of effect sizes we can expect to detect across the observatory given current sampling efforts and the levels of variability that have been observed during the construction phase of NEON.