Woodpeckers in the city: abundances are highest in large green spaces with complex understories

We documented occurrence and relative abundance (henceforth abundance) of five woodpecker species in 36 forested green spaces in Portland, Oregon (USA), to estimate minimum area requirements (MAR; 0.5 probability of occurrence) and to test for potential influences of green space area and type (heavily treed manicured green spaces versus undeveloped natural forests), habitat features, and landscape composition on the abundance of woodpeckers. Abundances of all species increased with green space area. Manicured and natural green spaces did not differ in canopy cover or total basal area, but woodpecker abundances were greater in natural than manicured green spaces for all species except Northern Flickers (Colaptes auratus). Flickers and Downy Woodpeckers (Picoides pubescens) were most abundant; MAR of both were under 1.5 ha, and both were found in essentially all green spaces. Hairy (P. villosus) and Pileated (Dryocopus pileatus) woodpeckers were least abundant, and with MAR of 34 and 51 h..., This dataset includes abundance and occurrence data for 5 woodpecker species, as well as vegetation data and basic landscape and urbanization measurements taken throughout 36 urban greenspaces in Portland, OR during 2015-2016. Standard passive point counts and audio broadcasts were used to survey woodpeckers, and circular plot drop-point vegetation surveys were used to capture habitat data. Highest woodpecker abundances were found in large green spaces with complex understories that were embedded in relatively undeveloped treed landscapes. We calculated minimum area requirements and found that the two most area-sensitive species (Hairy and Pileated woodpecker) had area requirements of 34 and 51 ha, respectively. We propose that management policies should be driven by the needs of species with the largest area requirements and greatest dependence on structurally complex vegetation structure.,
Descriptive statistics were performed using Statistix (v. 9; Statistix 2009)
Logistic regressions were performed using R 3.2.3
Generalized linear models were constructed using JMP (1989-2019)