The results of the virtual forest sampling tutorial are presented in Table 1. The tutorial was completed using distance based sampling for all three sampling techniques. The most accurate sampling method for the more common species (density > 100) appeared to be the systematic sampling method. Results were mixed – with the haphazard sampling method providing the most accurate results overall for the most common species Eastern Hemlock; although the overall absolute percentage error for the four more common species was more than double that of the systematic sampling method.
The most accurate method for the rarer species (density < 100) appeared to be the haphazard sampling method, although results were mixed again as the systematic sampling method gave the more accurate results for Chestnut Oak. The overall accuracy for the rarer species appeared to decline from that of the most common. The most rare species (White Pine) was missed completely in both the systematic and random sampling methods. The haphazard sampling method returned a relatively accurate result for the White Pine. The haphazard method in this study had been applied by selecting 5 points from anywhere within each forest zone with no predetermined notions of where to place the marker when selecting with the mouse. White Pine was found in only one of the five sample locations within the zone, therefore the high accuracy of this method for White Pine is somewhat subjective – a different selection point could have missed the species completely.
All three methods took between 4-5 hours to complete the sampling with the fastest method being systematic sampling. Haphazard sampling took 27 minutes longer yet had the least sampling error of the three methods. The random sampling method took the longest amount of time, and also returned the greatest sampling error for all species together.
The overall conclusion is that the sample size was not sufficient to estimate the abundance of the different species within the virtual forest, or to determine which of the three sampling methods was preferable.
Table 1: Absolute percentage error for three different sampling strategies for species abundance in the Snyder-Middleswarth Natural Area (see attached file)