Worldwide, riparian areas are considered among the most biologically productive and species-rich habitats on the landscape and provide important breeding areas for many bird species. In the Boreal Plain ecozone of western Canada, forests adjacent to riparian areas are generally protected from forest harvesting through the retention of treed buffer strips. <p>Riparian buffer strips are expected to provide habitat for wildlife including many passerine bird species. Recently, non-conventional methods of riparian management have been implemented in parts of the Boreal Plain with the intent of aligning forestry more closely with natural disturbance processes. How bird communities associated with these management scenarios diverge from natural disturbances and how riparian birds interact with disturbances in the adjacent upland habitat are key questions in the conservation of boreal riparian bird communities. To answer these questions, I surveyed birds inhabiting riparian areas with adjacent naturally disturbed (burned) and harvested forest to determine how bird communities differ early (1-5 years) post-disturbance and, separately, in a before-and-after harvesting study. <p>Riparian species associated with burned merchantable shoreline forests and riparian areas included Common Yellowthroat (<i>Geothlypis trichas</i>) and Eastern Kingbird (<i>Tyrannus tyrannus</i>). Le Contes Sparrow (<i>Ammodramus leconteii</i>) was associated with burned riparian habitats adjacent to non-merchantable forests (e.g., bog, fen), while Alder Flycatcher (<i>Empidonax alnorum</i>) and Wilsons Warbler (<i>Wilsonia pusilla</i>) were indicative of harvested sites with larger buffers (30 m). Riparian species richness was highest in burned non-merchantable sites. Multivariate Redundancy Analysis of post-disturbance bird communities showed greater divergence in overall (riparian and upland) community composition than one with only riparian species. This suggests reduced sensitivity of riparian birds to disturbances in forested areas compared to upland bird communities. However, a higher natural range of variability was exhibited in riparian bird community composition in post-fire sites than in post-harvested sites. This emphasizes that forest management practices do not currently fully approximate natural disturbance for boreal riparian birds. <p>To assess the response of bird communities in riparian habitats to forestry, I studied bird communities one year (2004) prior to forest harvest and two years (2005 and 2006) after harvest. One of three treatments, 1) 5-35% retention (0 m buffer), 2) 35-75% retention (10 m buffer with variable retention in the next 30 m), 3) 75-100% retention (50 m buffer) and unharvested reference sites, was randomly assigned to 34 wetlands. Treatments were designed to represent buffer management strategies currently applied in the Boreal Plain. Eight of 22 species showed a significant response (p<0.1) to treatment, year or year*treatment effects including two riparian species, the Common Yellowthroat and Song Sparrow (<i>Melospiza melodia</i>) that increased in abundance in harvested sites. Overall pre-disturbance communities diverged (p<0.05) over the three-year study period as shown using Multiple-response Permutation Procedures (MRPP). However, riparian bird communities did not diverge from pre-disturbance or from reference sites providing further evidence that riparian bird communities are less impacted by forestry in the adjacent upland habitats than overall bird communities. Therefore, alternative forest harvesting methods should be explored that encompass landscape-scale management including total buffer removal to maximize conservation objectives for boreal forest bird communities while attempting to maintain natural disturbance processes.

In an attempt to manage values other than timber production, forestry companies have sought a new paradigm to manage forest resources. Based on the hypothesis that wildlife in the boreal forest has adapted to habitat structures created by natural disturbances, some forest harvests have been modified to approximate patterns left after natural disturbance. Attempts at approximating natural disturbance have included retention of patches of live trees within cutblock boundaries, cutting to natural stand boundaries and application of harvest plans with spatio-temporally aggregated cutblocks (single-pass harvests). Single-pass harvesting is a recent attempt to better approximate natural disturbance in the boreal and has not been evaluated for its potential to sustain wildlife. I therefore contrasted residual patch pattern and composition, as well as landscape-scale avian abundance and composition in 1) single-pass; 2) multi-pass; and 3) salvage logged post-fire harvests, and contrasted these with unsalvaged post-fire sites. Post-fire sites were used to define the Natural Range of Variation (NRV). Seventy-two plots (12 post-fire, 15 post-salvage harvest, 16 single-pass harvest, and 29 multi-pass harvest) were surveyed for avian community composition and abundance one to five years post disturbance. <p>I contrasted the composition of remaining live forest stands at the landscape-scale and in residual patches by pairwise comparison of pre- and post-disturbance composition. At the landscape-scale, non-metric mutlidimensional scaling suggested post-disturbance landscape composition of post-fire and salvage-logged plots was similar to pre-disturbance landscape composition, with a tendency toward greater survival of hardwoods and lower survival of jack pine (Pinus banksiana) or black spruce (Picea mariana). However, harvesting of hardwoods and mixedwood stand types in single and multi-pass harvests led to landscapes with more bog and swamp habitats. <p>Comparison of residual patch composition with pre-disturbance composition was made using blocked multi-response permutation procedures. Post-fire plots (i.e. NRV) had residual patches that were representative of pre-disturbance composition, but with slightly more hardwoods and less black spruce/jack pine than expected by chance. All harvested treatments had similar biases among residuals to those left by fire, except that multi-pass harvests tended to leave less mixedwood habitat than expected. Multi-pass harvests also had less area in residual patches, and patches were smaller, more isolated and less complex in shape. Single-pass harvests had residual patches that were more representative of the size, shape, complexity, and change in composition seen post-fire. Multi-pass harvests only had 14% of the residual patch area in patches at least 5 ha in size, whereas this proportion was higher in fire (83%), salvage-logged areas (42%), and single-pass harvests (57%). Old-growth associated species might only persist in patches 5 ha or larger, and so multi-pass harvesting may have negative consequences for these forest birds.<p>Redundancy analysis indicated that bird communities differed from the NRV in all harvest treatments. However, single-pass harvests provided a slightly better fit to NRV than did multi-pass harvesting. Community similarity was influenced by non-linear responses to area harvested, residual retention, residual composition and pre-disturbance forest composition. An optimization routine was used to select harvest characteristics that would maximize community similarity to NRV. Optimization suggested that community similarity to NRV can be maximized by using single-pass harvests over multi-pass harvests, harvesting 66-88% of of a planning unit, and retaining 5-19% of the harvest area as live residual patches.<p>My results suggest that single-pass harvesting may be ecologically more sustainable than multi-pass harvests. Future studies are required to determine whether both harvesting treatments converge with NRV through time. Greater overlap of salvage-logged avian communities with NRV suggests that experimentation with prescribed fire as a post-harvest treatment may be the best method to bring harvests ecologically closer to NRV, and highlights the need to conserve early post-fire habitats.

Worldwide, riparian areas are considered among the most biologically productive and species-rich habitats on the landscape and provide important breeding areas for many bird species. In the Boreal Plain ecozone of western Canada, forests adjacent to riparian areas are generally protected from forest harvesting through the retention of treed buffer strips. <p>Riparian buffer strips are expected to provide habitat for wildlife including many passerine bird species. Recently, non-conventional methods of riparian management have been implemented in parts of the Boreal Plain with the intent of aligning forestry more closely with natural disturbance processes. How bird communities associated with these management scenarios diverge from natural disturbances and how riparian birds interact with disturbances in the adjacent upland habitat are key questions in the conservation of boreal riparian bird communities. To answer these questions, I surveyed birds inhabiting riparian areas with adjacent naturally disturbed (burned) and harvested forest to determine how bird communities differ early (1-5 years) post-disturbance and, separately, in a before-and-after harvesting study. <p>Riparian species associated with burned merchantable shoreline forests and riparian areas included Common Yellowthroat (<i>Geothlypis trichas</i>) and Eastern Kingbird (<i>Tyrannus tyrannus</i>). Le Contes Sparrow (<i>Ammodramus leconteii</i>) was associated with burned riparian habitats adjacent to non-merchantable forests (e.g., bog, fen), while Alder Flycatcher (<i>Empidonax alnorum</i>) and Wilsons Warbler (<i>Wilsonia pusilla</i>) were indicative of harvested sites with larger buffers (30 m). Riparian species richness was highest in burned non-merchantable sites. Multivariate Redundancy Analysis of post-disturbance bird communities showed greater divergence in overall (riparian and upland) community composition than one with only riparian species. This suggests reduced sensitivity of riparian birds to disturbances in forested areas compared to upland bird communities. However, a higher natural range of variability was exhibited in riparian bird community composition in post-fire sites than in post-harvested sites. This emphasizes that forest management practices do not currently fully approximate natural disturbance for boreal riparian birds. <p>To assess the response of bird communities in riparian habitats to forestry, I studied bird communities one year (2004) prior to forest harvest and two years (2005 and 2006) after harvest. One of three treatments, 1) 5-35% retention (0 m buffer), 2) 35-75% retention (10 m buffer with variable retention in the next 30 m), 3) 75-100% retention (50 m buffer) and unharvested reference sites, was randomly assigned to 34 wetlands. Treatments were designed to represent buffer management strategies currently applied in the Boreal Plain. Eight of 22 species showed a significant response (p<0.1) to treatment, year or year*treatment effects including two riparian species, the Common Yellowthroat and Song Sparrow (<i>Melospiza melodia</i>) that increased in abundance in harvested sites. Overall pre-disturbance communities diverged (p<0.05) over the three-year study period as shown using Multiple-response Permutation Procedures (MRPP). However, riparian bird communities did not diverge from pre-disturbance or from reference sites providing further evidence that riparian bird communities are less impacted by forestry in the adjacent upland habitats than overall bird communities. Therefore, alternative forest harvesting methods should be explored that encompass landscape-scale management including total buffer removal to maximize conservation objectives for boreal forest bird communities while attempting to maintain natural disturbance processes.

In an attempt to manage values other than timber production, forestry companies have sought a new paradigm to manage forest resources. Based on the hypothesis that wildlife in the boreal forest has adapted to habitat structures created by natural disturbances, some forest harvests have been modified to approximate patterns left after natural disturbance. Attempts at approximating natural disturbance have included retention of patches of live trees within cutblock boundaries, cutting to natural stand boundaries and application of harvest plans with spatio-temporally aggregated cutblocks (single-pass harvests). Single-pass harvesting is a recent attempt to better approximate natural disturbance in the boreal and has not been evaluated for its potential to sustain wildlife. I therefore contrasted residual patch pattern and composition, as well as landscape-scale avian abundance and composition in 1) single-pass; 2) multi-pass; and 3) salvage logged post-fire harvests, and contrasted these with unsalvaged post-fire sites. Post-fire sites were used to define the Natural Range of Variation (NRV). Seventy-two plots (12 post-fire, 15 post-salvage harvest, 16 single-pass harvest, and 29 multi-pass harvest) were surveyed for avian community composition and abundance one to five years post disturbance. <p>I contrasted the composition of remaining live forest stands at the landscape-scale and in residual patches by pairwise comparison of pre- and post-disturbance composition. At the landscape-scale, non-metric mutlidimensional scaling suggested post-disturbance landscape composition of post-fire and salvage-logged plots was similar to pre-disturbance landscape composition, with a tendency toward greater survival of hardwoods and lower survival of jack pine (Pinus banksiana) or black spruce (Picea mariana). However, harvesting of hardwoods and mixedwood stand types in single and multi-pass harvests led to landscapes with more bog and swamp habitats. <p>Comparison of residual patch composition with pre-disturbance composition was made using blocked multi-response permutation procedures. Post-fire plots (i.e. NRV) had residual patches that were representative of pre-disturbance composition, but with slightly more hardwoods and less black spruce/jack pine than expected by chance. All harvested treatments had similar biases among residuals to those left by fire, except that multi-pass harvests tended to leave less mixedwood habitat than expected. Multi-pass harvests also had less area in residual patches, and patches were smaller, more isolated and less complex in shape. Single-pass harvests had residual patches that were more representative of the size, shape, complexity, and change in composition seen post-fire. Multi-pass harvests only had 14% of the residual patch area in patches at least 5 ha in size, whereas this proportion was higher in fire (83%), salvage-logged areas (42%), and single-pass harvests (57%). Old-growth associated species might only persist in patches 5 ha or larger, and so multi-pass harvesting may have negative consequences for these forest birds.<p>Redundancy analysis indicated that bird communities differed from the NRV in all harvest treatments. However, single-pass harvests provided a slightly better fit to NRV than did multi-pass harvesting. Community similarity was influenced by non-linear responses to area harvested, residual retention, residual composition and pre-disturbance forest composition. An optimization routine was used to select harvest characteristics that would maximize community similarity to NRV. Optimization suggested that community similarity to NRV can be maximized by using single-pass harvests over multi-pass harvests, harvesting 66-88% of of a planning unit, and retaining 5-19% of the harvest area as live residual patches.<p>My results suggest that single-pass harvesting may be ecologically more sustainable than multi-pass harvests. Future studies are required to determine whether both harvesting treatments converge with NRV through time. Greater overlap of salvage-logged avian communities with NRV suggests that experimentation with prescribed fire as a post-harvest treatment may be the best method to bring harvests ecologically closer to NRV, and highlights the need to conserve early post-fire habitats.