Comparison of forest sampling strategies from virtual tutorial.

For the forest sampling tutorial, I used the distance-based method for all three sampling strategies (random, systematic, haphazard). The systematic sampling method was the fastest (4h 7 m) to survey, followed by haphazard (4h 38 m) then random sampling (4h 23 m). I expected systematic sampling to be the fastest, but I was surprised that haphazard was a bit quicker that random sampling. My sampling results indicated that Eastern Hemlock and Red Maple were the two most common species while White Pine and Striped Maple were the two least commons species.

For the two most common species the most accurate sampling method was systematic sampling which had the lowest percent error when compared to the actual densities. The percent errors increased when looking at random sampling and were even greater with haphazard sampling of the Eastern Hemlock. Notably though, the haphazard sampling for the Red Maple had a smaller percent error compared to random sampling which seems unusually, but both methods are not ideal with percent error over 20%.

For the two rarest species, random sampling appeared to be the best sampling strategy with percent errors of 11.4% (Striped Maple) and 7.1% (White Pine). Systematic sampling did not encounter either of these species leading to a percent error of 100%, while haphazard sampling overestimated the species density leading to 68% error (Stiped Maple) and 75% error (White Pine).

Overall, haphazard sampling does not seem to be a good sampling strategy. The percent errors for all species surveyed were higher than what seems reasonable (roughly under 10%). In the case of the tutorial both systematic and random sampling techniques each have their pros and cons. Random sampling seemed to better capture rare species and species in the mid range of densities. However systematic sampling became the more accurate method as species densities increased, although not in every case (e.g. Yellow Birch and Chestnut Oak had lower percent errors using random sampling, see Table 1 below).

Table 1: Forest tutorial experimental density results and percent error calculations