Amy F. Darling, Carol I. Stefan, and Corey De La Mare
March 2013
Executive Summary
Canada’s boreal forest is important habitat for many North American birds, which play a vital role in the health of forests, and provide ecosystem services. Expanding human development in northern Alberta raises concerns that fragmentation and loss of habitat could be leading to declines in forest wildlife. Objectives of this boreal songbird analysis were to examine habitat and disturbance associations of boreal songbirds, and determine what habitat and disturbance variables had the greatest influence on mature forest songbird relative abundance and species richness, with the ultimate goal of developing an empirical habitat model to predict the influence of development on mature forest songbirds in northeastern Alberta.
Local habitat associations were assessed separately for all songbirds, species of concern, and mature forest songbirds using fine scale habitat designations made by observers in the field. Detailed vegetation and disturbance data with a broader coverage (i.e., up to 1000 m beyond the point count centre) were available for 1,386 point counts which were used for predictive modeling. Competing hypotheses (models) were compared using Akaike’s Information Criteria (AIC) to search for a parsimonious suite of factors useful for predicting variation in mature forest songbird relative abundance and species richness. Multiple spatial scales were considered to better delineate the scale(s) at which disturbance affects songbird population and community parameters.
Seventy-six species and 6,886 individual boreal songbirds, including 27 species and 2,840 individuals in the mature forest songbird guild were recorded during 2,302 point counts (50-m radius surveys) in the Athabasca and Cold Lake portions of the Oil Sands Region (OSR) (ESRD 2012). The six most commonly observed species comprised 50% of all observations, with Tennessee warbler as the most commonly detected species (17% of observations). Mean (± standard deviation) relative abundance (i.e., detections per point) or density, species richness and/or diversity of all boreal songbird species combined, songbird species of concern, and the mature forest songbird guild were positively associated with land cover types or habitat groups likely to provide high quantity and quality foraging and nesting microhabitats. Songbird parameters tended to be highest in the coniferous white spruce, mixedwood aspen-white spruce, treed swamp and deciduous aspen-balsam poplar habitat groups. Songbird parameters tended to be lowest in dry or very wet, nutrient poor, and/or open habitat dominated by jack pine and/or black spruce. The habitat most preferred by mature forest songbirds, mixedwood aspen-white spruce, was consistent with previous research. All boreal songbirds combined were also positively associated with non-treed shrubby wetlands, whereas mature forest songbirds were negatively associated with non-treed wetland habitat groups.
Empirical habitat models predicted the influence of development on mature forest songbird relative abundance and species richness in northeastern Alberta. Top models included variables for height of the tallest vegetation layer, landscape preferred habitat groups, landscape avoided habitat groups, and three landscape disturbance types (wide linear disturbance, narrow linear disturbance and non-linear disturbance). The direction (positive/negative) and strength of the influence of disturbance on mature forest songbird relative abundance and species richness depended on the type of disturbance. Examination of cumulative effects of disturbance on songbirds should consider and account for different relationships with different types of disturbance. Measures of fitness consequences associated with habitat selection are required to understand the adaptive significance of habitat use.
The primary drivers of mature forest songbird relative abundance are found both at the local scale and at greater spatial extents. Local habitat characteristics in the 50 m point count are important determinants of mature forest songbird relative abundance, as are habitat characteristics in the smallest landscape examined (250 m). However, when controlling for local habitat group in disturbance models, the broadest landscape scale examined, up to 1 km from the point count, had the strongest influence on songbird relative abundance. Disturbance variables had stronger influences at the coarser landscape scales examined, especially at the 500 m spatial scale, whereas preferred habitats had stronger influences at the finer scales (local habitat and at the 250 m landscape scale).
Assessment for songbirds often focuses on fine scale, local habitat use and subsequent mitigation attempts might focus on limiting the proximity of disturbance to sensitive elements. Though this research confirms that local habitat variables are important, considering only local scale variables would fail to account for important landscape level variation in songbird numbers. Examination of zones of influence should move beyond local, fine-scale examinations of distance from disturbance and consider the composition of surrounding natural habitats and cumulative effects of surrounding disturbance, which can influence songbird habitat choice in a hierarchical manner. Models combining local and landscape level variables (habitat composition and structure as well as disturbance types) are recommended to accurately predict development impacts on boreal songbirds.
# 09-9200-50