Community Sampling Exercise — Results Summary — All Sampling Techniques
Blog Post 4: Sampling Strategies
For the Virtual Forest Tutorial, the three sampling methods that I used are as follows:
- Area based systematic sampling along a topographic gradient, using a belt transect with subsamples. Estimated time to sample: 12hours 36 minutes.
- Random Sampling of quadrats. Estimated time to sample: 12hours 42 minutes.
- Stratified haphazard sampling (quadrats were subjectively chosen without preconceived bias). Estimated time to sample: 13hours 8 minutes.
The technique that had the fastest estimated sampling time was the area based systematic sampling along a belt transect, with a time of 12hours 36minutes. Although this time was the fastest, the slowest at 13hours 8minutes only took 32 minutes longer. I believe this is due to the longer travel distance between quadrats in both the Random sampling and Stratified Haphazard sampling methods.
| True Density Value | Belt Transect | % Error | Random Sampling | % Error | Haphazard
Stratified |
% Error | |
| Eastern Hemlock | 469.9 | 388.0 | 17.4% | 645.8 | 37.4% | 456.0 | 2.9% |
| Sweet Birch | 117.5 | 72.2 | 38.7% | 170.8 | 45.4% | 144.0 | 22.5% |
| Striped Maple | 17.5 | 28.0 | 60% | 16.7 | 4.5% | 28.0 | 60% |
| White Pine | 8.4 | 0.0 | 100% | 16.7 | 98.8% | 0.0 | 100% |
For the most part, accuracy of sampling increased (% Error dropped) the more abundant the species was in the virtual forest. The only exception to this rule is the 4.5 % error from the Random Sampling of Striped Maples. Even though this exception exists it is likely due to random chance. The Haphazard Stratified sampling method also had the smallest margin of error when sampling the two most abundant species.
I believe that the Haphazard Stratified method encompassed all 5 topographical categories which increased the accuracy of the sampling. The Belt Transect method, although sampling through all 5 categories, still seems to be less accurate because it only takes samples in a straight line.
In doing this Virtual Forst Tutorial, I conclude that the Haphazard Stratified method ended up being the most accurate when compared to the true density values of the most abundant tree species. Accuracy was lost with all methods when sampling the two least abundant species. Not only was the Haphazard Stratified method more accurate, it did not take a significantly greater amount of time compared to the other methods.