Due to time constraints of my summer employment and narrow sampling window relative to seasonality (i.e., counting live fern fronds will be difficult as they die in fall), I knew that much of my data would need to be collected in one or two long field sessions. After a cursory review of the data from the second of my two most recent site visits, I am pleased that the long afternoon of sampling may suffice as enough content for analysis and subsequent report commentary. I am working on phrasing a specific hypothesis and predictions, but for now, it seems that maidenhair fern frond density as a proportion of overall frond density (including western sword-fern) is higher downslope of burn piles than in similar control areas of western redcedar understory.
This is a good thing, not just because of my time constraints (wildland firefighter), but because of one difficulty I encountered: sample disturbance. I had read about this prior to sampling, and quickly realized its reality. It is difficult to carefully lay out a quadrat and step around it in a transect so as not to snap maidenhair fronds, which are relatively fragile. I was able to easily identify broken-top fronds and include them in the count, but subsequent sampling would be slightly skewed. Speaking of the count, I likely re-counted one or two fronds, or missed fronds entirely, in some quadrats. That said, this would present a minor error relative to sample size. The difficulties of field sampling render it less an exact science than the laboratory setting; therefore, ecology focuses on multiple samples and statistical analyses. I believe the six total transects, three each of burn piles and controls, should provide enough data to be statistically significant.
Beyond the preference of maidenhair for burn piles, I was curious if more maidenhair grew immediately adjacent to burn piles than farther downslope in the pile transects. In this regard, I may be surprised; it did not seem like pile proximity had an effect on maidenhair frond count. Perhaps this could be due to local factors like time of sunlight on exact individual plants, or micro-scale soil moisture gradients. However, those observations are likely beyond the scope of the BIOL 3021 report. That said, I did try to isolate some variables with other observations: hourly relative shading percent of each transect, average slope across each transect, and azimuth of each transect. For this, the control transects were preferentially selected to roughly mimic the environmental conditions of neighboring pile transects, and the initial data review suggests that control transects were indeed similar to burn pile transects.
As I continue data analysis, literature review, and subsequent teaching modules, I may find that I will need more data before fronds die in the fall. But for now, this seemed to be an appropriate sampling technique.