The data I have collected for my experiment includes measuring each of the sunflower plants shoot height and the maximum basal diameter (mm). The systematic sampling strategy that I used allows for easy data collection and I plan to continue to use this approach. However, my three-sunflower plant/ sampling units that are located in each of the pots do not have enough distance between each other to be considered independent from one another. It would have been better if I had used individual plant pots for each of the sunflower plant replicates, so that they are considered independent from one another.
Category: Post 5: Design Reflections
Blog Post #5 – Experimental Design/Data Collection Reflections
The original study area had to be re-evaluated as area size was noted to be a possible confounding variable, and also because one of the city parks staff completely destroyed one of the proposed study sites in early June. As a result, I decided to take a closer look at the main study area, and in doing so I noticed that there was a gradient in vegetation species richness already present in several areas of the Dufferin Wetlands Park. I decided to divide the main study area into four quadrants to conduct my data collection, and before collecting bird species data, I surveyed each quadrant and counted the number of vegetation species present. Bird species data was collected from June 5-12, 2021 using the point counts sampling method.
The most pronounced difficulty in data collection that I noticed was the initial identifications of the bird species in the area. This was easily the most time-consuming task as some birds were extremely active and harder to identify than others. Once identifications were complete, I did not find the actual data collection replicates to be a difficult task, and found it to be relatively straight-forward.
At this point in time, I plan to make a few small changes but will ultimately stick to using this technique to collect data for my experiment. To explain, I will be adding a control site to the data collection, which will be a nearby parking lot. I will also attempt to control for the time of day by sampling data at the same times for each replicate, and also plan to standardize the point count sampling times for each area to 5 minutes. Cumulatively, I feel that these changes will help to produce more consistent data.
Blog Post 5: Design Reflections
For the collection of my initial data, I used stratified random sampling to sample the number of Western sword ferns in the marsh area, upland transition areas and the upland area. My hypothesis is that Western sword ferns prefer the upland area compared to the marsh and upland transition as a result of drier soil. The difficulties I encountered were that I didn’t realize how dense some of the areas of the forest were, with lots of bushes and shrubs. I had to really bushwhack and force my way to some of the locations which were very challenging. The data I did collect was in line with my hypothesis so it was not that surprising. Although my sampling technique was very challenging I plan to keep using this technique since I feel it will yield more precise results although I do plan to increase the size of my sampling unit as a felt 1meter squared was too small for this study.
Blog Post 5 – Design Reflections
The most challenging part of implementing my sampling strategy was spacing my sampling sites appropriately so they maintained independence. Although I would have liked to have included more sampling sites, it was not feasible given the limited area of the shrub habitat type. This has also made randomization challenging, although I have introduced it by randomizing which site is sampled first during each survey round. So far, the data I have collected has not been overly surprising. Bird species which are more associated with forest habitat, such as Townsend’s warbler (Setophaga townsendii) and red-breasted nuthatch (Sitta canadensis), have not been recorded within the shrub habitat sampling sites. I plan to continue collecting my data using the breeding bird point count methodology.
Post 5: Design Reflections
I collected my initial data on May 20th, 2021 around 1:15pm. I decided on using quadrats along transects for my study. I brought with me a pre-measured 36m string, four 4m strings, and a soil moisture meter. There is a trail that goes along the top of the bank, which is where I chose to place the transect. I used a random number generator to choose the number of footsteps I would take from the entrance of the trail and placed the 36m string there, along the top of the bank. I placed my quadrats along the transect about 4m from each other and collected my data. I used the soil moisture meter directly in the centre of the quadrat to measure how moist the soil was and then I counted how many ferns were in that quadrat. In order for a fern to be counted, more than 50% of its total size had to be within the quadrat.
One of the struggles I had was that the bank was quite steep at some parts which made it difficult to place the quadrats safely. I was afraid I would slip and hurt myself. I think I might bring somebody with me to help me navigate the terrain safely next time. I also realized later that I should have made sure that the soil moisture meter was inserted into the soil at the same depth in every quadrat in order to reduce any inaccuracies in the moisture readings. I will mark the meter with a Sharpie next time and make sure I push the meter up to that point during each of the readings. Overall, I was quite happy with my study design as it is not overly difficult for me to conduct and is very interesting.
I did realize I will have to revise my predictions about my hypothesis. I had originally predicted that as soil moisture increases, fern abundance would decrease. However, after also choosing a quadrat by the creek that had a lower soil moisture level and observing a lower number of ferns, I will now predict that as soil moisture increases, fern abundance also increases.
Blog Post 5: Design Reflections
The hypothesis for my research project is the length of time a Robin spends foraging in the meadow will differ from in the dog park. I predicted that the length of time a Robin spends foraging in the meadow location will be greater than in the dog park location due to the greater number of dogs present within the dog park than the meadow. The sampling method I have chosen will be the Point Count method. I visited the Meadow 5 times, approximately the same time between 5:00-7:00 PM and observed the Robins in the meadow for 30 minutes on May 3, 4, 5, 7 and 8, 2021. The sampling strategy was relatively simple. I sat at a picnic bench with my binoculars and timed the presence or absence of Robins foraging. Considering the time of day, I was surprised by the number of Robins actively foraging and how easily they would return from the meadow once a dog had left the area. I also observed some competitive behaviour between Robins, which I also thought was very interesting. Given the simplicity and lack of resources available, I have chosen to continue using the same sampling method.
The excerpt from the field journal can be found here: https://drive.google.com/drive/folders/1xbUUajwJ8BOtYKSs8DY5s1f7doOIsc72?usp=sharing. Observations were recorded on the left-hand pages, including weather and activity within the meadow. The right-hand pages document when the foraging time (FT) began, which is defined when at least one Robin was present in the meadow. The foraging time stopped (FTS) when no Robins were present in the meadow. Time was recorded as minutes and seconds. The number of birds and dogs was recorded during one complete cycle of FT to FTS. The FT would begin again when at least one Robin returned to the vacant meadow. Please refer to table one for a list of acronyms.
Table 1. List of Acronyms
|
Acronym |
Meaning |
| FT | Forage Time |
| #B | Number of Birds |
| #D | Number of Dogs |
| FTS |
Forage Time Stopped |
Percy Herbert, Blog Post 5: Design Reflections
For my sampling strategy I opted for haphazard sampling of individual rose plants. For the initial data collection in Module 3 I selected one plant in each of the following height ranges: 1-50cm, 51-100cm, 101-150cm, 151-200cm, and above 201cm. The reason why I have opted for haphazard sampling over random or systematic is that the area where the roses are located is specific and not large. Dividing the land into quadrats would also be exceedingly difficult due to the thickness of the roses and surrounding underbrush. There are many unbranched wild rose plants that fall within each of the height ranges outlined above. Therefore it was easy to find one from each height range to observe for the initial data collection. The only difficulties in sample collection are the thickness of the underbrush making it tough to access the roses, and that the vegetative buds are beginning to form into leaves and small branches.
The initial data was not overly surprising. The sample size was much to small to derive any meaningful conclusions, however, the initial data supported my theory that the spacing of the vegetative buds is not related to the height of the plant. I also collected data on the number of vegetative buds on each individual plant and the distance from the apical bud to the lowest vegetative bud. These additional pieces of information did not provide any interesting information and I do not believe that I will continue to collect these data as I move forward with this experiment.
I plan to continue to use the same sampling method in an equal number of individual plants will be observed in each height range. I believe that by simply haphazardly observing many more individual plants in each height range I will be able to have enough measurements to perform an ANNOVA analysis.
Blog Post 5: Design Reflections
I had two main difficulties when establishing the sampling of my research area. The first issue was establishing quadrats of equal size due to the terrain. I initially wasn’t sure how I was going to measure out this distance given the uneven terrain and thick tree cover. However, I decided that using a string pre-measured at 20m worked better than trying to navigate through the forest with a tape measure. Using this method I did manage to space out several 400m2 quadrats and will create more to further my research. Using this method I found it easy to search for my response variable (presence of bracket fungi on trees) and will continue to with this method.
The second significantly bigger challenge that occurred was the annual spring break up of the river which flooded out a large section of my study area. I initially established my quadrats starting from the bridge and ending near the high school. However, about a 500m stretch of park was entirely flooded due to record-high water levels. I therefore had to adjust my study area slightly and anticipate being unable to establish quadrats in the flooded areas. This was a shame because I had visualized bracket fungi on a few of the trees in these now flooded areas but did not get a chance to study them before the water level rose. It is unlikely I will be able to access these areas again for the duration of the course.
For my sampling method I opted to use the quadrat method, using the Kiwanis trail as a makeshift transect. This allowed me to study areas on both sides of the trail, which vary a fair bit in terms of species abundance and moisture content. However, at present, most areas on the east side of the trail (closer to the river) are inaccessible, therefore for my initial study, I only included counts from quadrats on the west side of the trail which weren’t waterlogged. I suspect the water in the east areas will recede within the coming weeks so I may be able to resume sampling in these areas later. However, if the water level does not recede, then I will adjust my study method and create a transect west of the trail rather than studying quadrats on the east side that may continue to be inaccessible. The key downside of this modification to my research methods is that it cuts out the trees nearer the river that sit on a steeper slope and are subjected to different environmental conditions to those trees inland. I was hoping to see if any key differences could be found between trees based on their proximity to the river
On a final note, I am still struggling to determine what to study for the predictor variable(s) and how to go about measuring them.
Reudink, Blog Post 5: Design Reflections
I collected data in the forested area on my property off of Clandeboye rd, MB, on April 15th, 2021. There was snow on the ground (up to a foot in height in certain areas) after a snowstorm. The day’s weather was 5 °C with clear skies. My study area is approximately 200m in length and 40m in width, so I made three transects and systematically sampled every 20 paces along each transect. The data I collected were soil samples and measurements of tree circumference at breast height (CBH) for trees over 25cm in CBH. Initially, I was planning on doing square plots along each transect but changed my mind during the first sample. It was very difficult to mark down a square plot with the tools I had and the highly dense forest, so I decided to do circle plots (5m radius) instead. By doing this, I was able to stand in the middle of a plot, collect my soil sample (10cm depth), and measure whether any of the trees of interest were within 5m from me. If they were within 5m, they would be within my plot so I included them in my sample. The only logistical difficulties I encountered was having to walk through snow in a dense forest (and dig through it) and making sure that there was no organic matter in my soil samples. My cat kept me company during the expedition and I made sure to remove any gross organic matter from my samples the next day before I weighed them. I plan to continue this approach for my study.
The data I collected was surprising because I was expecting both the EAST and WEST perimeter transects to have a greater mean CBH, but instead it was rather an increasing gradient from EAST to WEST for CBH (i.e., EAST mean CBH 40.4cm, CENTRE mean CBH 51.7cm, and WEST mean CBH 82.8cm). I am currently dehydrating my soil samples after measuring their initial mass so I will not know the moisture content associated with each transect for a few more days.
Blog Post 5: Design Reflections
For my initial data collection, I implemented stratified random sampling using Google Earth and QGIS. I created polygons based on Google Earth satellite images for each predictor zone, exported them as a KML file and used QGIS to generate random points within one of the polygons to collect sample replicates. I then exported these points as a GPX file, put them on my GPS and located them in the park to take samples. I used discrete classes to represent percent coverage as outlined in the sampling design tutorial, ranging from 1-6.
I think my method for generating worked fairly well, but I fear that my areas my be too small to justify stratification. I’m also unsure if statification is the best approach, given that the basis of the stratification is also the predictor variable (dominant tree species as an indirect measure of soil moisture). The zones are fairly distinct in the park, but there are some interspersed wetter/drier sites, leading me to think that perhaps I should use a non-stratified approach and just record the predictor variable with each individual sample. The number of random points which land in one of the smaller zones (arbutus/ garry oak, alder) may be smaller, but since frequency of predictor variable is not a measure of concern it may be ok if I have more samples from the doug-fir/grand fir zone.
