Category: Post 6: Data Collection

Blog Post 6

TRU Open Learning Leave a comment

Marshlands Conservation Area Kingston, Ontario- Rideau Trail 1040-1430

09.12.18, -3 °C, clear skies, chilly, no snow but snow-covered ground,

just North of Lake Ontario; wedged between Little Cataraqui Creek to the West of the trail

and Cataraqui Golf and Country Club to the East.

Sampling Strategy: Stratified sampling (all ash trees within plots A-E; 25X25 foot plots). I used the same plots as my original field data.

I altered my data collection protocol since last time. Previously I had been looking for larval galleries/bark fissures. This did not seem to give my data definitive patterns, so I decided to look at other factors of emerald ash borer (EAB) presence too. The other indicator of emerald ash borers I noted was woodpecker foraging/d-shaped exit holes in the trees. I included a category for trees which displayed both larval galleries/fissures/bark deformities and woodpecker foraging/D-shaped exit holes. To prevent double counting trees in a plot, I brought along a piece of yellow chalk to mark which trees had been observed already. The number of replicates in each plot varied from 20-38 ash trees. One problem with my sampling design has been noticing signs of EAB presence up in the canopy. I tried my best to overcome this problem using binoculars though. I have noticed that ash saplings do not tend to be infested by emerald ash borer. It makes me wonder if the age of a tree is more relevant to predicting EAB presence as opposed to the density of ash trees within a plot.

Results:

Plot A 13 out of 25 trees were infected. 52% infection rate.

Plot B 20 out of 29 trees were infected. 69% infection rate.

Plot C 16 out of 20 trees were infected. 80% infection rate.

Plot D 14 out of 38 trees were infected. 36.8% infection rate.

Plot E 12 out of 23 trees were infected. 52.2% infection rate

 

Figure 1: Ash tree with bark splitting and EAB exit holes/woodpecker foraging

Figure 2: Ash saplings unaffected by EAB

Figure 3: Rideau Trail

Figure 4: Measuring tape used to delineate 25 foot borders of each plot

Blog Post 6: Data Collection

TRU Open Learning Leave a comment

Blog Post 6: Data Collection

The field data collections at the View Royal Park have been occurring over multiple days at various times of the day. Squirrel abundance in response to dog walking has been measured. Data collection started on Saturday, February 23, 2019, while the weather was clear but chilly. following collections happened on Monday, Tuesday and Wednesday at 10:00, 1330, and 0700 respectively. The weather was clear and windy at approximately 4-5 degrees Celsius.

The predictive variable, dogs, will be organized into intervals of 0-1, 2-4 and 5+ dogs present at the park. The number of squirrels observed within these intervals will be measured and analyzed.

Each interval will be replicated at least 4 times. therefore data collection will continue until each interval (0-1, 2-4 and 5+ dogs at the park) has been noted and squirrel abundance has been measured 4 times within the intervals.

The most difficult part of implementing the sample design is the uncertainty of the number of dogs present at the park at a time. Without the ability to manipulate the number of dogs present data collection must be done over many days and times of the day.

Over four days of data collection patterns appear to support the hypothesis that squirrel abundance declines in presence of dogs as they pose a threat of predation. Although, The outlying intervals of 0-1 and 5+ have much higher replicate data than the median interval of 2-4. Number of humans without dogs in the park have also been collected as a control to show that the decrease in squirrels is not due to humans alone. The number of humans has shown no effect of squirrel abundance so far in data collection.

 

 

Post 6: Data Collection

TRU Open Learning Leave a comment

The four sites I selected each have a different tree density: site 1 has low density, site 2 has sparser density but denser than site 1, site 3 has a mixture of low and high density in different areas of plot, site 4 has high density. Sites 1-3 has an elevation of around 810masl and site 4 has an elevation of 860m. Each site was selected randomly by throwing a stick and choosing the site based on where the stick landed. Once the sites were chosen a 10m by 10m quadrat was set up using 4 wooden stakes and a 50 m rope to wrap around each stake to clearly outline the quadrat. GPS points were taken at each wooden stake in order to map the sites later. I then took field observation notes on the area and where the site was located based on the surrounding areas. Then I walked along transects within the quadrats in order to ensure that I counted all the trees in the site. When walking along the transects, I recorded how many spruce trees were dead and alive as well as the other trees and bushes in the site. One issue I had while implementing this sampling design was when I got to Site 4 with the higher tree density, it was hard to make a perfect 10m by 10m square and also difficult to get through the trees and ensure that I counted all of the ones I needed to. I noticed that the site with higher density had more dead trees than the sites with lower tree density, which provides evidence to prove my hypothesis as correct.

Post 6: Data Collection

TRU Open Learning Leave a comment

Data was collected January 16th 2019 at 9:00am at Piper’s Lagoon, Nanaimo, BC

I separated my area into six areas (A through F) in which I implemented a closest distance quadrant method. Each sample point I recorded the nearest four tree species, and each area contained 5 sample points resulting in 120 tree species recorded. I chose to separate the points enough that any tree would not be sampled twice. I recorded the location of points on a separate map. I decided to remove Alnus rubra from the analysis because only a single tree was observed (but not recorded). I found it difficult to choose representative sites without being biased. I wanted to record data that represented each area, therefore, chose to consistently follow the footpath.

The six areas studied had previously been divided into ocean exposure levels from 1-6, 1 being the most protected. Most protected E > F > A >B> D > C

The observed patterns during data collection were:

  • Quercus garryana and Arbutus menziesii occur in similar environments on rocky substrates with little soil
  • Pseudotsuga menziesii grow concentrated together forming forested environment
  • Quercus garryana was observed in every area in different abundances

These made me reflect on my hypothesis that ocean exposure affects angiosperm occurrence. Perhaps other factors such as substrate type are contributing to the pattern.

Post 6: Data Collection

TRU Open Learning Leave a comment

3 areas which are 10m x 10m in size were chosen at Papago Park around a man-made oasis. Area 1 was 5 meters south of the oasis. Area 2 was 20 meters south of the oasis. Area 3 was 50m south of the oasis. These 3 sites were chosen in a southern direction because south had the most unimpeded land (either by road or rock). The three sites were arranged in a line and did not veer substantially to the south-east or south-west. Each area was divided into grids of 25 subplots each measuring 2m x 2m. Each grid was assigned coordinates of x and y values. To pick the initial subplot in each area, the google number generator with n = 5 was rolled. I rolled it twice, with the first number acting as the x coordinate and the second number acting as the y coordinate. From there I used a systematic sampling method where each individual individual Larrea tridentata was counted and measured. From the first subplot, I would increase and decrease the x value until I have x = 1, 2, 3, 4 ,5. Y-values were y and y-1 (where y – 1 = 0, would wrap around back to around to y = 5).

5 subplots were selected in each of the three areas for a total of 15 replicates. The total time for the sampling and recording of data took approximately 1 hour and 35 minutes. Distance between each individual in the each chosen subplot to its closest neighbour was recorded, as well as the number of Larrea tridentata individuals and each plants height. There were few difficulties with this method of sampling. The only issue was that the number of individuals in area 1 was quite low and 2 of the chosen subplots ended up being empty which may skew the estimates of the height as empty plots had to be disregarded.

I noticed that the total number of plants (of any species)  increased dramatically in area 1 relative to the other 2 areas. However, the number of creosote bushes dropped substantially which was against my set out hypothesis. There was no ancillary pattern in the distribution of individuals but in terms of height, creosote closer the water were larger than those further away.

Blog Post 6

TRU Open Learning Leave a comment

The three elevation gradients that were chosen were 20m, 110m, and 200m. The point centred quarter sampling method was chosen for this study. Transect lines were set up horizontally across each of the three elevation gradients. Five points were set up along the transect lines, five meters apart from one another. As a result, five replicates were set up and randomization was established by placing them five meters apart. A compass was used to make sure that the lines were set up in the same direction. Each of the five points was split up into four quadrants. The distance of the closest black spruce species in each quadrant to the centre point was measured with a measuring tape and recorded. The diameter at breast height (DBH) was also measured for the closest trees by wrapping the measuring tape around the trunk of the tree at 4.5 feet above the ground and then dividing the measurement by 3.14.
I did experience a few problems while implementing my sampling design. I initially went out alone to collect the measurements but then had to return since I realized that I would need help from others to collect the measurements. It was also difficult to walk through the bushes and tree logs to measure the distance. I did notice ancillary patterns in that black spruce species prefer drier conditions as the elevation increases.

Blog Post 6: Data Collection

TRU Open Learning Leave a comment

I used the point count method to assess the abundance of waterfowl (dabbling ducks) within selected drainage channels within Colony Farm Regional Park. Eight point count locations were chosen along the existing dike network in the park based on channel visibility and varying cover of emergent vegetation (yellow pond lily, Nuphar sp.). Each point count was considered one replicate (sample unit). Selected locations (8) were visited on 6 separate occasions (sampling events), with 48 total replicates sampled. Percent cover of yellow pond-lily was visually estimated within each channel section, to a maximum distance of 80 metres. Channel sections were strategically selected on the basis that they varied in percent cover of yellow pond-lily (from 0% to approximately 75% cover) to provide a representative sample. Drainage channels chosen were fairly uniform in terms of their topography, hydrology, surrounding vegetation, and overall width.

I spent five minutes at each point count location observing dabbling duck species, specifically mallard (Anas platyrhynchos) and wood duck (Aix sponsa), within the selected channel sections.  Total number, species, and approximate life stage of observed waterfowl within the channels were recorded. Other information I recorded included the start and end time of each point count, weather conditions, date, start and end time of each sampling event, and other notes including the presence of other wildlife and/or predators in the area. All point counts were collected between the hours of 1800 – 2000 hours (prior to sunset at 2030 hours), as this was the time of day that waterfowl within the park were observed to be the most active. Randomization was incorporated into the study by using a random number generator (between numbers 1 to 8) to decide the order in which point count locations were visited during each sampling event.

One problem that I encountered while carrying out my sampling design was the low number of dabbling ducks observed within the channels for the duration of the data collection. To combat this, I visited the site more frequently (i.e. sampled more replicates) to ensure that I would have adequate data to analyze later on.  If I were to collect data in the future, I may choose to do so in the fall months when waterfowl occur in greater numbers. A second issue that arose had to do with the site conditions. Connectivity between channel sections led me to realize that not all replicates would be independent of one another. After reviewing relevant literature related to point count methodology, point count locations are suggested to be located greater than or equal to 250 metres apart. Based on the limitations of the study area, and in order to reduce variables between selected drainage channel sections, spacing of the point count locations at this minimum distance was simply not possible. Despite this, the layout of the study area allowed for ample visibility, which prevented double counting of birds and reduced bias among point count locations.

Initial summation of the total number of waterfowl within each channel has lead me to believe that my hypothesis may be rejected. The numbers did not indicate a strong preference of heavily vegetated channels by waterfowl, versus those channels with sparse emergent vegetation cover. This will be subject to further analysis.

Post 6: Data Collection

TRU Open Learning Leave a comment

My data involved collecting soil samples from each of my three locations. I randomly took 5 soil samples from within a 1m x 1m quadrant in each area. The quadrant was placed in an area which contained the main vegetation present, however, because in the main vegetation in location 1 was fir trees the quadrant was placed directly beside them. I took the soil samples from the top soil, meaning first I scraped off the top layer known as the humus and collected from the layer underneath. All my samples were placed in plastic bags to prevent moisture loss and I brought them to a lab at my University. There I placed them in individual foil cups, removing any rocks or other debris and pounding out the clumps. I then weighed them and placed them in an oven between 105-107 degrees celsius for 24 hours. Afterwards, I weighed the samples again and calculated the percent moisture of each.

I haven’t had any problems implementing my sampling design. However, I have noticed that there are far more factors than I initially considered that could also play a role in determining the vegetation present. This realization is not going to change how I go about my research, but I will take more consideration into this factor when I am writing my results.

Blog Post 6 – Data Collection

TRU Open Learning Leave a comment

I split each study zone in 5 quadrants and counted the trees/moss in each. I believe this means that I did 15 replicates in total. Although I am a bit unsure if replicates are defined the same way in a tabular experimental design. So far, I haven’t had any problems implementing my sampling design. The data seems consistent with the hypothesis. Perhaps I could find a relation between the amount of sunlight & soil moisture to moss growth/direction.

Blog Post 6: Data Collection

TRU Open Learning Leave a comment

I did my data collection on four separate days, first on September 12, then the 13th, 15th and 19th. I counted the number of ducks in each of the three locations at McArthur Island and categorizing them into each of the four states depending on if the ducks were on land or in water. I repeated this process at three times of the day (10am, 2pm, 6pm). After analyzing my data on the first day I realized that over the course of the day the temperature of the air and water increased, and so the number of ducks in the water increased substantially . At the pond and the river though, there was not much change over the three times I observed the ducks. After viewing this data my hypothesis still hasn’t changed and I will continue to analyze it and make some slight adjustments to my methods