Breeding bird response to riparian buffer width in managed Pacific Northwest Douglas-fir forests

We examined the relative importance of riparian vs. upland habitats to bree ding birds by comparing species abundance, richness, and similarity of bird communities in managed Douglas-fir forests in western Washington State, US A. We also examined whether forested buffer strips along second- and third- order streams effectively maintain the pre-logging riparian breeding bird c ommunity by comparing species abundance, richness, and turnover among three treatments: (1) unharvested controls; (2) sites that were clear-cut, leavi ng a narrow (similar to 14 m) forested buffer on both sides of the stream; and (3) sites that were clear-cut, leaving a wide (similar to 31 m) foreste d buffer along both sides of the stream. Deciduous trees, berry-producing shrubs, and other deciduous shrubs less co mmon in adjacent upland forest characterized streamside zones. Despite diff erent vegetation features, riparian and upland habitats did not differ in a ny measures of bird species richness and composition. No species or species group was more abundant in the upland. Neotropical migrants, resident spec ies, and species associated with deciduous trees and shrubs in forested hab itats were more abundant in riparian habitats than in adjacent uplands. Tot al bird abundance and abundance of four species (American Robin [Turdus mig ratorius], Pacific-slope Flycatcher [Empinonax difficilis], Black-throated Gray Warbler [Dendroica nigrescens], and Winter Wren [Troglodytes troglodyt es]) were higher in riparian habitats. Abundance of these riparian associat es was correlated with percent cover of berry-producing shrubs and the numb er of deciduous trees in the canopy. We found that the number of breeding bird species on sites with narrow buff ers increased from slightly fewer than controls before harvest to an averag e of 10 more species than controls after harvest, a change reflected in an average 20% increase in species turnover on narrow-buffer sites relative to controls. Total bird abundance did not differ between treatments and contr ols. Resident species, those species associated with shrubs in forested hab itats and conifer trees, declined on both buffer treatments. Species associ ated with upland and riparian forests (Black-throated Gray Warble;, Golden- crowned Kinglet [Regulus satrapa], and Brown Creeper [Certhia americana]) d ecreased in abundance on riparian buffer treatments relative to controls, w hereas species associated with open, shrubby habitats (Dark-eyed Junco [Jun co hyemalis], Cedar Waxwing [Bombycilla cedrorum], and Song Sparrow [Melosp iza melodia]) increased in abundance on one or both riparian buffer treatme nts. High species turnover on narrow-buffer treatments indicated that buffers < 14 m on each side of the stream did not maintain the pre-logging bird commu nity. There was little difference in species turnover or species richness b etween the wide-buffer treatment and the control, indicating that a 30-m bu ffer on both sides of second-order and third-order streams maintains most o f the pre-logging bird community in the first two years postharvest. The Bl ack-throated Gray Warbler was the only riparian associate to decline on bot h the narrow- and wide-buffer treatments; its abundance was positively corr elated with buffer width, and a buffer greater than or equal to 45 m wide o n each side of second- and third-order streams was needed to support popula tions at densities found on unharvested controls. To maintain the entire br eeding bird community associated with forested riparian habitats in the coa stal Northwest, we recommend a minimum buffer of 45 m along both sides of s econd- and third-order streams. Habitat features such as deciduous trees (A lnus rubra and Acer macrophyllum) and berry-producing shrubs (especially Ru bus spectabilis) appear to be important and should be maintained within for ested riparian buffer strips. This study documents short-term effects of riparian treatments on the breed ing bird community, which may take several years to respond to habitat mani pulations. Thus, we recommend continued monitoring to assess long-term effe cts of buffer width reduction.