Using iNaturalist Observations to Evaluate the Extrapolation of Herbaceous Critical Load Models
Emma Censky1, *, Michael D. Bell1, Emmi Felker-Quinn1, and Christopher M. Clark2
Herb species critical loads (CLs) for Nitrogen (N) and Sulfur (S) take into account local climate and soil conditions, but the surveys used to develop these CLs were unequally distributed across the continental US (Clark et al 2019). A separate project developed an environmental hull for each herb species CL: i.e., the niche of soil, climate, and deposition conditions similar to those in which each herb occurred (Clark et al unpublished), but since the models were created with limited data it is unknown if the input data covers a species range. Our project tests the accuracy of hulls and the extent of CL exceedances using species occurrence data from iNaturalist, a community science platform of species observations. We used iNaturalist data for 12 common herbaceous species with CLs in national parks and compared their location to each species' hull.
There were more than 150,000 verified research-grade observations from iNaturalist for our 12 species of interest. Observations per species ranged from 1,088 to 20,164. Seven species had over half of their points within the environmental hull. 10 species had 80% of their points within 1 standard deviation of the hull. The conditions of Pteridium aquilinum's range were most aligned with the hull values and Bouteloua curtipendula's conditions were the furthest from the hull range. These comparisons provide relative confidence in which species CLs can be broadly applied across the continental US (CONUS) and which CLs should be restricted in their spatial application.
We used the 2019-2021 TDep total N and total S to determine exceedances of critical loads for all species locations. There were similar rates of exceedance of N CLs and S CLs across the original survey locations and across the iNaturalist locations. Using verified citizen science observations in combination with environmental hulls allows us to better evaluate how representative the hulls are to a species' actual CONUS range. These points also give more information to land managers and policy makers about how broadly the CLs should be applied and the number of species in exceedance of their critical load.
1 National Park Service, Air Resources Division, Denver, CO
2 U.S. Environmental Protection Agency, Office of Research and Development, Center for Public Health and Environmental Assessment, Washington, DC
* Corresponding Author: ecensky@nps.gov