Figure 8. Study area within the boreal forest of northern Alberta. Sites in aspen stands are represented by dark grey dots (n=69) and sites in black spruce stands are represented by green dots (n=18).
Aspen
Ten-minute unlimited distance point counts were used to survey birds in trembling aspen-dominated stands in northern Alberta (Figure 8). In this study, the stations were clustered in groups of nine, with the groups systematically placed on the landscape with respect to a gradient of seismic line density (Figure 9). A total of 69 sites and 621 stations were sampled. Vegetation data, including percent canopy closure, canopy height, shrub height, and shrub and tree species composition, was collected at each station.
Figure 9. Sampling design at an aspen site, consisting of nine point count stations (represented by dark grey dots) in a grid with approximately 300 m between points. Seismic lines are shown by brown lines.
Black spruce
Songbird species abundance in black spruce-dominated stands in northeastern Alberta was also assessed using 10-minute unlimited distance point counts (Figure 8). Point count stations were located in areas which had: 1) no seismic lines within 200 m, 2) one seismic line within 200 m, and 3) two or more seismic lines within 200 m. The point count stations were clustered on the landscape in groups of twelve, with each group (i.e. site) containing four stations of each treatment type (Figure 10). In total, there were 18 sites and 216 stations surveyed. Local habitat characteristics, including stand age, stand height, and proportion of black spruce, were calculated for areas within a 200 m radius of each station using Alberta Vegetation Inventory (AVI) layers in ArcGIS.
Figure 10. Sampling design at a black spruce site, consisting of twelve point count stations divided between three treatments. Black dots represent stations with multiple seismic lines, green dots represent stations with no lines, and blue dots represent stations with single lines. Seismic lines are shown by brown lines.
Detectability
Using equation 1, counts were converted to densities (number of birds per hectare) to account for differences in detectability between species (Buckland et al. 2002).
Equation 1: D= C/((r^2)*pi)*10000 where D = density
C = mean count of a species per treatment group
r = effective detection radius
The effective detection radius was calculated using a large dataset of birds across boreal Canada, and is defined as the distance at which the probability of detecting an individual beyond a set distance (eg. 50 m) is the same as the probability of missing an individual within this distance.
The data obtained in the two studies were entered into Access databases. A flat file of data for each study was then created in Excel by querying the databases. In both cases, the sampling unit for which counts of bird species were recorded was the station, and stations were clustered on the landscape into sites.
Aspen data table
In this study, the key predictors were the line density within 200 m of the station and the line category (0 = no lines, 1 = >0 to 3 km per square km, 2 = >3 to 6 km per square km, 3 = >6 km per square km), which are shown in red in Table 1. Environmental predictor variables were also measured and are highlighted in green. The response variables are counts of the 20 species present at more than 10% of the sites, three of which are shown in yellow.
Table 1. Flat file table of aspen data with the ID variable in the first column. Locational variables are coloured grey, while predictor variables are red and green, and response variables are yellow.
Black spruce data table
In this dataset, the key predictor variables are the treatment (none, single, or double seismic) and the line density within a 200 m radius of the station, which are shown in red in Table 2. Environmental predictor variables (shown in green) were also measured. The response variables are counts of each species. Three species are highlighted in yellow in Table 1 as an example, although there were a total of 12 species in the actual dataset.
Table 2. Flat file table of black spruce data with the ID variable in the first column. Locational variables are coloured grey, predictor variables are in red and green, and the response variables (species) are highlighted in yellow.