Blog Post #5: Design Reflections

After arriving at the field site and setting up equipment, I deemed it necessary to make a few changes, create standardizations, and record a few additional variables. They are as follows:

1. My initial plan was to measure all of the C. aurantiacus trichome colonies within the quadrat at each location. However, there were hundreds in some locations and hardly any in others, making it impractical to count. Instead, photos were taken directly above the quadrat, and thresholding was used to measure the percent of the quadrat they took up. This is a faster technique and relatively accurate as the contrast between the bacteria and the bottom of the reservoir is high. 
2. The area of bacteria in the quadrat depends on the height above the water and the distance from the bottom. Thus, I standardized the height above the water to 5cm and the depth of each reading was recorded.
3. Due to the high amount of variation of bacterial growth, I decided to take all readings in triplicate after my initial 5 replicates. This will help in later analysis to increase statistical power and accuracy. 
4. Additional variables were recorded as from viewing the arrangement of bacterial growth and the water source, there appeared to be more factors at play. Thus, I considered other variables and recorded them at each replicate location. They may or may not have an influence but better to have too much data than a lack of data.

Aside from having to make a few revisions to my protocol before beginning any sampling, the sampling process went very well. All the data collected appeared to be precise in reference to replicates and neighboring locations. Thresholding each image has turned out to be a fast and highly quantitative process allowing for a high level of analysis from each location. Because of the quick data collection and analysis protocol created, I have been able to return and collect a large magnitude of samples which will optimistically have more accuracy and a higher degree of statistical power in the analysis. The data collection was smooth and gave me data that can be analyzed in many different ways and thus I will continue to collect data in this fashion. Though systematic spacing of samples works well there are a few gaps that are necessary as part of the reservoir is frozen and cannot be accessed. 

The data collected initially was surprising, though it had nothing to do with the method of collection. With part of the reservoir being frozen, I expected temperatures to be extremely high at the source (maybe 15°C) and very cold near the start of the ice (0°C) but instead saw that the change in temperature was at most 2-3°C. The changes in bacteria growth did correlate with the water temperature as expected but the temperature swing was far more subtle than anticipated.