Category: Post 5: Design Reflections

My sampling day went fairly well. I think I worked out some of the kinks in my mind beforehand, however once I was in the field collecting data I realized there were certainly things I needed to change. First of all, I was only recording presence/absence of the three lichen types on each aspect of the tree (NSEW). I have decided I want to record the percent cover of each type on each aspect of the tree, because sometimes all three types would be on one side but one was much more dominant and that information was being lost.

Secondly, I need to determine how far up and down the tree I want to sample so that there is consistency in the data. I could focus on the base of the trees, or perhaps one metre above and below breast height. I could even make transects on each tree.

Thirdly, I may need to only sample in the upper and lower forest strata, or focus my studies in the rocky outcrop only. The rocky outcrop, or Garry Oak ecosystem, at the top of Mount Douglas is quite a different community than the closed forests. All the trees were deciduous and the foliose lichen was present on every tree, whereas in the closed forest most of the trees were coniferous and the foliose lichen was only present on one of the ten trees which just so happened to be very close to the top of the mountain.

I think for my final data collection I will use the same sampling strategy that I used which is the distance based method. I used a random number generator on my phone to tell me how many paces to take on the trail, and how many paces to take perpendicular to the trail on the right or left side of the trail, then once I walked the paces I sampled the closest tree to me. I treaded carefully in the forest and was minimally invasive. Sometimes it was hard to walk a straight line because of downed logs or ditches which may have made it less random because I would sort of guess which way to walk and for that reason I may need to follow a compass bearing (I’ll think about this some more).

Finally, I would like to take replicates from more places in the park to get a more representative sample. I only sampled the southern portion for this preliminary data.

On January 23^rd from 2pm to 3pm (MST), I did my first field data collection. I chose to use the haphazard sampling method. It made it easy to select five distinct point counts across the area of study using Google maps. Point counts was an effective way of sampling bird population and human traces. However, because the haphazard method does not provide a true randomization, I am aware that the data collection might have been more effective and accurate if I had selected more than five point counts. In the future, I would either use a true random sampling method, or I would use the haphazard method while doubling the number of watch points. I think that will generate more accurate results.

Prior to collecting the data with four other observers, I pre-defined “human traces” as being any of the following: debris, food waste, a pair of footsteps, a passerby, pet or pet feces, a motorized vehicle, or any human-installed unit such as a garbage bin, picnic table, and what not. Bird feeders was one human-installed unit which was a category on its own since we assume it has a direct impact on bird population. After collecting the data, I reflected with the other observers, and we all agreed that the footsteps category was not useful, hard to count, and confusing since with the snow, there were many confounded footsteps. Counting hikers would have been sufficient. Moreover, we concluded that if this data collection had to be done again, we should count the number of nests observable at eyesight as well, and possibly the number of bird caches. I had not included those, because the species observed was the black-billed magpie exclusively, and I was worried that we would confuse other species’ nests or caches with the ones that actually belonged to the black-billed magpies. I would like to try including them next time around.

All in all, the results were not surprising as the point count which featured the most traces of human presence and also a bird feeder was the one with the highest black-billed magpie population, as expected. Still, for a second data collection, I would double the number of watch points, or use a true random sampling method, for more accurate results. Also, I would revise the pre-defined categories of human traces and keep track of nests and observed caches as well.

My sampling strategy was to use a 12m measuring rope to count the number of big sagebrush and green rabbitbrush along the hillsides coming up from Guerin Creek. I counted the number of plants within a 3m segment of rope, extending 0.5m on either side of the rope to give a 3m2 quadrat.

Sampling this number of quadrats was very doable, in total there were 8 unique lines measured and bushes counted and it took one hour. In implementing my sampling strategy, a few things went differently than planned. First, I misjudged the depth of the creek valley, so ended up with double the distance I was expecting to cover. My rope wasn’t designed for more than 12m of measurement, so I only measured the initial 12m as planned, and then counted bushes from an elevation of 12m to the top of the hillside. Additionally, there were between 15-30cm of snow cover, depending on the hillside. This prevented the counting of any young bushes, therefore only mature bushes grater than 40cm high were included in the count.

The data I collected was close to what I expected after doing some research (table 1). The intital 3m are likely flooded in the spring, and so there were few established brush plants right creekside. The number of plants actually then increased all the way to 12m, so clearly they are able to germinate and grow that far from the water table. However, there were 6.67 sagebrush plants per m between 9-12m; while from 12-45m there were only 1.13 bushes per m. When I take another sample I will include measures from 12-18m, 18-24m, 24-30m and 30-36m from the creekbed. This will help elucidate more information about how far above the water table the bushes thrive. Rabbitbrush is much less frequent on the hillsides, so it’s accuracy may not be as good as the big sagebrush measures.

I will need to sample around the entire creek, since there were significant differences between the topography of the two hillsides. The bushes in the second site were nearly double the size and the hillside was very steep compared to Site 1.

Table 1. Mean number of sagebrush and rabbitbrush bushes counted at various elevations from the Guerin Creek bed at Thompson Rivers University in Kamloops, BC.

My time collecting data at Lost Lagoon, Stanley Park, I had no difficulties in implementing my systematic sampling strategy. I focus the study on examining bird species presence and record other species that I have observed.  I’ve noticed a pattern of abundance at three different sites along an urbanized gradient.

Revised hypothesis: Bird species presence and abundance is impacted on how close their natural habitat is located near an urbanized city.

Revised predications are:

1. Bird species richness will be highest in areas closer to the city (south side of lagoon)
2. Bird species evenness will be highest in areas with the highest percent cover of natural habitat (west side of lagoon).
3. Bird species richness/evenness will be lowest in areas near the main highway/road because of lack of food (east side of lagoon).

My response variable will remain as bird species presence and abundance (continuous variable), and my explanatory variable of percent cover (categorical variable) of natural habitat vs. urbanized areas (roads, buildings, trails) at each site as a whole will also remain the same.

In addition, the followings changes will be made in the experimental design in regards to the way data is collected:

1. Initial sampling took place around 1:00pm and overall lots of activity and species present.
2. My new sampling strategy includes two survey sites (point count surveys) in each of the three habitats along the gradient. During the replicate exercise, I expanded the number of point count surveys in one of the sites to five. While it would be ideal to maintain this type of replication it is too time consuming to perform five replicate surveys in each of the three sites. As a result, I will perform two replicates per site, and will conduct surveys at each site on at least three different days.

During the next blog I will go into more detail on the sampling strategy, sample unit, how these changes have impacted the data collection, and any ancillary patterns that appear in the data to support the new hypothesis and predictions.

For my sampling, I witnessed a lady feeding crows bread which I never thought of as a treatment level for my experiment. At this point, the number of crows was much great than any other point confirming that as the anthropogenic sources increase the number of crows does as well. I will add anthropogenic sources to certain points for the rest of my field study. The point counts that I observed were as predicted for my experiment that as the anthropogenic sources in the area increased so did the density of the crows. With these added treatment levels I feel that it will present clear evidence that anthropogenic sources allow for a greater density of crows.

Blog 5: Design Reflections

My original idea was to sample deadwood (logs and stumps) and determine if a form of succession occurred on wood. I took samples from the forest but ran into several problems:

Location. The area is not too big. For this idea a larger area probably has to be sampled.

Human activity. I sampled a location that I knew had some human activity and areas that looked untouched. There seemed to be less wood on the ground than I expected. I since discovered that some people who living around the parameter of this property do go in on occasion and remove fallen trees for firewood. Since the forest isn’t that big and the area surrounding is over 40 years old too much of the wood may have been removed to capture samples of vegetation growing on logs.

Order of plant growth. I thought of treating fresh cut or broken wood like bare rock. Lichen would grow, followed by moss and then vascular plants. I saw some fresh stumps in this forest that had rotten wood in its cross-section and tiny tree seedlings growing in that. I did not take into account this factor, that some wood might already provide a habitable location for tree seeds to take root immediately after death. I also saw one stump beside the access road that had soil piled onto it during road maintenance in the last couple of years that was host to a tree seedling. The wood under the soil was still rock hard.

Wood that does host growth. The wood that did have growth was all conifer (cedar, hemlock, Douglas fir). None of the deciduous alder deadwood in my survey areas had growth. Logs outside of the survey did have some moss but nothing else. It appears, in this forest anyways that only the conifer trees are hosting new vegetation. Since the coniferous area accounts for only one third of this forest then the study area shrinks even more. There didn’t seem to be any rule that determines which deadwood plays host to vegetation. Other than type of tree that becomes nurse logs, there doesn’t seem to be any other rule about which ones will host vegetation. Growth occurred on a mix of wood, both new and old. There are stumps bearing more advanced vegetation that are clearly more recent than some other stumps that are very old and only bearing moss. One stump that does host a couple of fairly tall cedars still bears a loggers notch (part of a method of logging by hand use up to the 1930’s), but other stumps nearby even more decayed didn’t have even moss.

Time. The report in my Blog 2 was about a study on decaying wood in an experiment that had been ongoing for 65 years at the time of publication. Perhaps succession on dead wood would be best studied from fresh logs and stumps and the growth documented over time, as opposed to finding specimens with no know history and trying to piece together a pattern of succession.

Modification of idea. I visited another location in my region that has the same type of coast rainforest. This area is larger and is a regional park. There is also some known historical background. In 1913 an above ground concrete water pipe was laid through to supply Victoria. A strip of forest on either side was cleared so that trees wouldn’t fall on the pipe. The pipe was decommission many years ago but was left in place and the forest on either side is closing in to reclaim the gap. I thought this might be a better place to find samples of deadwood with successive stages of vegetation growth. I found a photograph of the pipe installation crew at work in 1913 with a giant stump prominent. I went out to the exact location and after much difficulty found the very rotted remains covered under thick shrub. Most of the other original stumps were also gone. There was also a lack of logs laying around. I did notice a pair of newer stumps in a clearing that showed something interesting. The ground in this small clearing was covered with shallon shrub too thick to even walk through. The stumps, which rose above the height of the shrubs, stood out because they bore several varieties of plants. None of the plants were shallon and they were the only plants in the clearing not shallon. On the way out of the park I noticed that the few other stumps and logs along the trail bore a greater variety of plants than the ground surrounding them. I went back to my own study area and took another look at my original samples. The logs and stumps with plants had more growth on them than on the ground around them. I am going to concentrate on determining if fallen wood is more habitable to new growth than the ground. The study area does see some human activity and there is a heavy concentration of deer that does feed in here. The park I came from also sees lots of activity in the form of hiking and mountain biking. I will locate an area to pursue this idea further.

Initial data collection for my backyard observations of Acer glabrum stand density was done using two types of sampling methods: random and systematic. I did not have any problem with the sampling, but I have decided to tweak my methods somewhat in order to best address my hypothesis.

The gist of my project is to quantify the density of maple trees relative to the large cedar tree in my yard. Hypothesis being that stand density increases the further the plot is from the cedar. For the random sample, I used a random number generator app on my phone to establish 10 plots within the site. For the systematic method, I measured stand density along 3 transects at 5, 10, and 15 meters from the base of the tree at designated bearings of 20°, 40°. And 60° for a total of 9 plots.

These methods both affirmed my hypotheses but for final data collection I think I will use the systematic method, but more comprehensively. The study site is located between 0°-100° from the base of the cedar tree. Using the systematic method, and to satisfy the rule of tens, I will establish transects at 10°, 20°, 30°, 40°, 50°, 60°, 70°, 80°, 90° and 100°with one meter plots located at 5, 10, 15, and 20m along each transect. Each plot will be assessed for the number of maple trees within it. So, my field notes will look like this:

This will give me a total of 50 replicate sites with enough distance between plots to be independent of each other. This would seem to be a method that will provide a robust quantification of the density of maple trees at the site, relative to the cedar tree.