|CALOVI, MARTINA - Pennsylvania State University|
|MILLER, DOUG - Pennsylvania State University|
Submitted to: US-International Association for Landscape Ecology
Publication Type: Abstract Only
Publication Acceptance Date: 3/2/2021
Publication Date: 4/12/2021
Citation: Goslee, S.C., Calovi, M., Miller, D. 2021. Modeling potential species abundance for management applications[abstract]. US-International Association for Landscape Ecology. P. 1.
Interpretive Summary: No Interpretive Summary is required for this Abstract Only. JLB.
Technical Abstract: The species distribution modeling approach provides tools and techniques for creating range maps that can be used for conservation planning and projecting climate change scenarios. However, the most common models are based only on presence, which does not correlate well to abundance at broad spatial scales. Abundance models are needed for managing provisioning ecosystem services, as in forestry and agricultural applications. We have developed a set of tools in R for modeling abundance of agronomic species using quantile Random Forest to predict potential maximum abundance based on biophysical factors. As niche theory explains, there are many possible reasons a species is absent from a site that are unrelated to site suitability. The use of quantile methods better fits this conceptual understanding by fitting the maximum (or as here, 90th quantile), and produces predictions that more closely match measured abundance values. We demonstrate the utility of this approach by modeling a set of agriculturally-important forage species in the northeastern United States, and show how not just range but abundance may shift with projected future climates. A hybrid approach incorporating both presence data and abundance data may provide the most useful models, since presence data are often far more readily available. The citizen science tool iNaturalist is an excellent source for presence data, and can augment research studies of abundance. The combined models are more effective than either alone as components of decision support tools for developing regional and long-term management strategies.