Landscape and vegetation effects on breeding songbirds in the mixedwood forests of southern Ontario

grantor: University of TorontoMixedwood forests in southern Ontario were studied to evaluate the relative impact of landscape influences and vegetation structure on the composition of the avian community. Data from central point censuses, vegetation surveys, and a GIS database were used in Redundancy Analyses, and revealed that local forest cover and edge distance had the most important landscape effects on the avian community but edge distance was a less powerful predictor than forest cover. Edge effects were complex and variable in their impact on birds through changes in the abundance of resources, predators, and parasites. The present study demonstrates that efforts to develop a general theory of edge effects, even if limited to birds in agricultural landscapes, may be ineffectual because of the confounding effect of forest cover. Vegetation structure explained more variance in the avian community than landscape effects alone, contrary to some earlier reports. As a result, conservation of North American avifauna requires management efforts to maintain forest cover plus reforestation plans that mimic the diversity of forests that have been lost, in addition to the current focus on protecting large forests and minimizing the amount of edge created by harvesting. The abundance of food resources may affect the distribution of insectivorous songbirds in fragmented landscapes, and the biomass of flying insects consumed by flycatchers was greater in less fragmented forests, but neither of the flycatching species was landscape-sensitive, so no link can be established between resource abundance and landscape effects on birds. The detectability of birds varied with date during the breeding season, time of day, and with temperature at the time of the census for 15 of 43 common species. In addition, evidence for observer bias ranged from four times to nine times the level that would be expected by chance. These biases can be minimized by planning multiple visits to each census point.Ph.D