For the virtual sampling exercise, I sampled from the Snyder-Middleswarth Natural Area. I utilized random, systematic, and haphazard sampling strategies. Of these, the systematic was the most time efficient clocking in at 4hrs and 5 minutes, though the haphazard was a close second at 4 hrs 27 minutes compared to the random strategy which took 12 hours and 34 minutes.
There was no method that I found to be particularly accurate in its estimation. I calculated error rates for each tree species rather than just the 2 most common and 2 rarest in an attempt to narrow down the “best” method but was still left with few answers.
My error calculations showed:
Eastern Hemlock (actual density 469.9) – Random – 3.3%; Systematic – 22.2%; Haphazard – 3.6%
Sweet Birch (actual density 117.5) – Random – 31.2%; Systematic – 11.7%; Haphazard – 60.9%
Yellow Birch (actual density 108.9) – Random – 57.9%; Systematic – 55.0%; Haphazard – 9.6%
Chestnut Oak (actual density 87.5) – Random – 52.3%; Systematic – 14.3%; Haphazard – 47.0%
Red Maple (actual density 118.9) – Random – 33.1%; Systematic – 5.3%; Haphazard – 8.1%
Striped Maple (actual density 17.5) – Random – 90.3%; Systematic – 60.6%; Haphazard – 47.4%
White Pine (actual density 8.4) – Random – 1.2%; Systematic – 123.8%; Haphazard – 100%
All in all, the accuracy was not great across the board so I’d generally recommend Systematic or Haphazard sampling strategies based solely on time savings compared to the Random sampling strategy.
