Table 1 Candidate covariates used to develop piping plover breeding density models and their data sources.

Category	Source	Temporal resolution	Candidate covariates
Site accessibility	Gravel and paved roads⁵⁰	Static	Road density within 5 km
Checklist effort	eBird²⁶	Annual	Checklist duration; Checklist distance
Anthropogenic	Gravel and paved roads⁵⁰	Static	Distance to roads (m); Binary classification: > 100 m from roads = 1, < 100 m from roads = 0
Anthropogenic	Settlements⁵⁰	Static	Distance to human settlements (m); Binary classification: > 1 km from settlements = 1, < 1 km from settlements = 0
Topography	DEM (U.S. Geological Survey National Map)	Static	Slope^a
Hydrology	DSWE⁴³	Annual	Percent surface water^a,b
Hydrology	NWI	Static	Distance to permanent lake features (m); Distance to permanent lake features—exponential function
Vegetation	NDVI (Landsat)	Annual	NDVI^a,b
Land cover	NLCD⁴⁹	Every 2–3 years	Percent crop and hay pasture^a; Distance to trees (m)

Site accessibility and checklist effort covariates were only applied to the eBird observation process and remaining covariates were applied to the latent species distribution process. Temporal resolution refers to the frequency at which data layers were produced across our study period. DEM Digital elevation model; DSWE Dynamic surface water extent; NWI National wetlands inventory; NDVI Normalized difference vegetation index; NLCD National land cover database.
^aThese covariates were considered at multiple spatial extents: 30 m, all adjacent pixels, 90 m, 150 m, 300 m, and 750 m.
^bThese covariates included a candidate quadratic effect.

Quick links

Search