My hypothesis is that nest cavity density will significantly increase along with increasing successional stages, ranging from early to late. My prediction is this trend will occur due to forest harvesting practices affecting early regeneration (ie. elimination of remnant older trees) and cavity density will increase as the natural structural complexity of the forest increases with age, and conditions favour primary cavity nesters.
Null Hypothesis:
The null hypothesis is that there will be no significant difference between the average densities of nest size cavities across the sampled forest succession classes of Early, Mid, Mature, and Old.
Alternative Hypothesis:
There will be significant differences detected between the average densities of nest size cavities across the sampled forest succession classes of Early, Mid, Mature, and Old.
Design Reflections:
In general, my pre-survey design followed the framework necessary for effective experimental design, including steps to identify focus variables of interest, defining the population of the study, and detailing plans for data collection to be representative of the population.
I developed reference criterion to guide the design, using BCTS GIS data and other mapping tools online (Google Earth Pro 3.3.0) to delineate seral stage categories, BGC Zones, aspect/elevation/slope. My concern after visiting the site in person is that it will be necessary to ‘ground truth’ the GIS data to ensure the seral stage classifications match what actually is present at each site. I needed to clarify how to identify a nesting cavity, as opposed to a feeder or roosting cavity. I adapted my own criterion for this, based on existing guidelines for Pileated woodpecker cavity identification (Government of Canada, 2024). Since I was unable to bring a ladder into the field I am limited to inspecting cavities <3m height with a camera on a pole. Cavities higher on the trees were only possible to visibly assess using binoculars.
My goal in sampling was to have representative sample plots stratified across the different seral stage forest types, taking into account the need for consistent underlying site characteristics like:
- Same BEC zone classification: ICHmw5 (Interior Cedar Hemlock moist warm) = similar species/communities (see Figure 6)
- Close in elevation (1200 to 1400m elevation)
- Similar aspect (South)
- Similar slope (>15%)
- Minimum 300,000m2 area to accommodate 2 x 200m survey transects with widely spaced plots (20m apart alternating), allowing for buffering from edge effects (~30m buffer)
After selecting suitable polygons to represent all seral stage categories, I decided an efficient way to sample would be using systematic sampling along 2 independent transects from a random point at a random bearing, within each selected seral class polygon.
Using ArcGIS software, I generated one random UTM point per seral class polygon from which to commence systematic sampling of 400m2 plots (20m x 20m quadrats) along 2 separate 200m transects, both oriented at different random bearing directions. These seemed like appropriate plot dimensions to detect tree cavities via visual surveillance, with 10 replicate plots per seral stage polygon. For efficiency purposes, I planned to conduct sampling over 2 transects of 200m each to facilitate spacing of plots to maintain their independence (5 per transect). I thought by having 2 transects would be more representative and easier to access compared to 10 plots randomly distributed throughout the entire polygon area.
Heading out into the field for the first time, I was easily able to find my pre-determined random UTM points of commencement for each transect, and use the random compass bearing to establish the transect direction. Bushwacking with snowshoes through the early and mature forest types was challenging with plenty of windfall to contend with. I found the mature and old growth sampling to be much easier. I used a measuring rope that I pre-measured and flagged at 20m intervals to indicate the corners of each plot. I think it was hard enough to go 200m per transect than attempting a longer stretch!
I used an eslon tape to measure out perpendicular from the transect line and flagged the corners of the plots as I went, to give me a visible boundary reference. I took GPS points using my inreach at the centre of each plot. Unfortunately when I went to upload this data, the points were skewed on the map.
At each plot I then recorded:
- General notes about site information (UTM coordinates, Plot #, composition/cover%, slope/aspect/moisture)
- Cavity tree characteristics (species, decay class, height, DBH, photos)
- Specific cavity info (cavity size, height on tree, wildlife signs
- Other observations
I think the approach is working, and I will continue it for the remainder of the survey. It would be nice to have a way to better view the cavities to confirm them as nesting type, but for that I may need to get some specialized climbing gear – which would likely cause disturbance to anyone inside!