User: | Open Learning Faculty Member:
| Systematic | Random | Haphazard | |
| Fastest estimated sampling time | 12 hrs 38 mins | 12 hrs 45 mins | 12 hrs 30 mins |
| Percentage error Eastern Hemlock |
((520-469.9)/469.9)) *100 10.7%
|
((479.2-469.9)/469.9)) *100
1.98% |
((583.3-469.9)/469.9)) *100
24.1% |
| Percentage error Sweet Birch |
((124-117.5)/117.5)) *100 5.5%
|
((120.8-117.5)/117.5)) *100
2.8% |
((166.7-117.5)/117.5)) *100
41.9% |
| Percentage error Striped Maple | ((0-17.5)/17.5)) *100
100% |
((12.5-17.5)/17.5)) *100 -28.6%
|
((20.8-17.5)/17.5)) *100
18.9% |
| Percentage error White Pine |
((4-8.4)/8.4)) *100 -52.4%
|
((0.0-8.4)/8.4)) *100
-100% |
((20.8-8.4)/8.4)) *100
147.6% |
| Accuracy | Moderate accuracy for common species
Poor accuracy for least common species |
High accuracy for common species Poor to very poor accuracy for least common species
|
Poor accuracy for common species
Moderate to very poor accuracy for least common species |
Of the three sampling types, haphazard had the fastest sampling time but only marginally. It was 8 minutes faster than systematic sampling and 15 minutes faster than random sampling. I would expect random sampling to have the longest estimated sampling time as this can be a difficult method to carry out under field conditions. None of these estimated sampling times would give me enough information to choose which sampling method would be appropriate for a given project to reduce time costs. Possibly with a larger sample size (>24) the estimated sampling times would show greater variance amongst the sampling methods, providing better information to make a decision for project cost.
To calculate percentage error, I based this calculation on the information provided for density (not frequency or dominance). The percentage error for the two most common species (Eastern hemlock & sweet birch), haphazard had the highest percentage error for the sampling method. This is to be expected because the plant community is more heterogeneous which tends to offer more biased, unrepresentative estimates. Random sampling offers reliable estimates with the least amount of bias and as a result, had the smallest percentage error for the two most common species.
The two least common species had large variations in accuracy of density. The percentage error for density may be showing such variations as White Pine and Striped Maple are typically small species in terms of basal area (or diameter at breast height). This would cause a disproportionate representation of density and it may have been a better idea to calculate percentage error on dominance and not density. Dominance provides the total basal area of a given species within the unit area of the community.
Species abundance does not appear to heavily influence percentage error accuracy in my findings, it only affects the result from overestimates and underestimates. Possibly with a larger sample size, this error would be greatly reduced. Using a species-area graph would have helped with sample size, ensuring that species richness is represented but once the graph starts to level off (no more addition of new species) no more additional samples are needed.
Random sampling had the greatest time estimate but it also had the highest accuracy for the density of common species. I also appreciate the lack of bias in this method and would tend toward this sampling method.
