Category: Post 4: Sampling Strategies

Blog Post 4

TRU Open Learning Leave a comment

Random sampling had the fastest estimated sampling time of 12 hours and 25 minutes. The slowest estimated sampling time was systematic taking 12 hours and 38 minutes. Haphazard sampling was in between taking 12 hours and 36 minutes.

Percentage Error respectively (Eastern Hemlock, Sweet Birch, Striped Maple, White Pine):  *note that Eastern Hemlock and Sweet Birch are the two most common species, while the Striped Maple and White Pine are the rarest species

Systematic: 10.7%, 5.5%, n/a (none found), 52%

Random: 20.6%, 8.1%, n/a (none found), n/a (none found)

Haphazard: 5.3%, 11.3%, 42.9%, 1.2%

Based on these results it is hard to determine which strategy is the most accurate since the random sampling did not find any Striped Maple or White Pine and systematic sampling did not find Striped Maple either. For common species, systematic and haphazard sampling were about equally accurate. My results suggest that 24 was not a large enough sample size to capture species in the rare communities thus insufficient in accurately estimating the abundance of these species.

 

 

 

 

Post 4: Sampling Strategies

TRU Open Learning Leave a comment
Systematic had the fastest estimated sampling time at 12 hours and 35 minutes.

Red maple – 8.25%

White Oak – 37.91%

Striped Maple – 0%

White Pine: (12.8-36.0)/36.0*100 = -64.44%

 

Random

Red Maple – -8.84%

White Oak – 16.40%

Striped Maple – -57.5%

White Pine- 0%

 

Haphazard

Red Maple – 5.68%

White Oak – 37.92%

Striped Maple – 0%

White Pine– 60%

 

The accuracy decreased as the species abundance increased, and I think systematic and haphazard are the most accurate and actually had similar results for both.

 

Sampling Strategies

TRU Open Learning Leave a comment

In this exercise I used a virtual forest tutorial to test the accuracy of estimating species abundance using three different sampling techniques: area, random or systematic; distance, random or systematic and finally, Haphazard.

The distance, random or systematic would have been the most efficient model as its estimated time to sample is only 4 hours 43 minutes compared to the other two ranging around 12-13 hours. This is most likely influenced by how spread apart the sample sites are as there are equal sample sites in each method.

Percent error for each sampling technique for the most abundant and most rare species was:

  • Area, random or systematic: Eastern hemlock-  14.9% error. White Pine = 8.3% error
  • Distance, random or systematic: Eastern hemlock = 30.2% error and White pine= 100% (none found)
  • Haphazard, are: Eastern Hemlock= 4.7% and White pine = 90%

The most accurate for abundant species was the Haphazard method at 4.7% error rate. It estimated 448 out of the 469.9 total Eastern Hemlocks in the area. However, this technique over-estimated by double the amount of White Pine. The most effective at the rare species of White Pine was the area, random-sampling technique which had a small error rate of 8.3%. It estimated 7.7 of the 8.4 total White Pine trees. The accuracy trend between abundant and rare species was dependent on the method used. The most constant technique was area, random or systemic where the percent error decreased with the rare but only slightly. The least reliable was the distance, random or systematic which had a high error rate for both species; finally, the haphazard technique had an accurate estimation for abundant species but unreliable estimations for rare species.

24 sampling points allow fairly reliable sampling results for area, random or systematic sampling, but if using the other techniques it would be useful to add more sampling sites to improve estimations of rare species.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

After completing the virtual forest sampling tutorial, the data was as follows:

Sampling Time (hr:min): The fastest sampling methods was the systematic sampling methods at 12hrs 36mins estimated.

Systematic = (12:36)

Randomized = (13:12)

Haphazard = (13:02)

Percentage Error of density of species:

Eastern Hemlock – Systematic (1.30%) Randomized (6.41%) Haphazard (37.1%)

Red Maple –           Systematic (-12.5%) Randomized (-49.5%)   Haphazard (27.8%)

White Pine –           Systematic (42.9%) Randomized (-4.76%) Haphazard (-4.76%)

Striped Maple –     Systematic (37.1%) Randomized (82.9%) Haphazard (37.1%)

Sample Calculation: Eastern Hemlock (systematic) = (estimated (476.0) – actual (469.9)) / actual (469.9) x 100 = 1.30%

Populations with greater numbers of individuals had much more accurate estimations than the rare species.

The systematic methods was the quickest as well as the most accurate sampling method. Surprisingly the randomized methods seemed substantially less accurate than the haphazard method. I believe it may be limited to this one trial however. I would expect the haphazard to be the least accurate because of bias whether intentional or subconscious.

 

 

Blog Post 4

TRU Open Learning Leave a comment

The virtual forest tutorial allowed for the testing of three different sampling strategies in the Snyder-Middleswarth Natural Area: systemic sampling, random sampling, and haphazard sampling. The technique that had the fastest estimated sampling technique was systematic sampling with a time of 12 hours, 36 minutes followed by random sampling and then haphazard sampling.

Yes, the accuracy changes with species abundance for all three sampling techniques. The percent error for the common species was very low compared to the percent error for the rare species. The two most common species were the Eastern Hemlock and the Sweet birch and appeared to be best represented by systemic and haphazard sampling methods. The two most rare species were the Striped Maple and the White Pine and appeared to be best represented by the same sampling methods.

Based on the values of the Shannon Weiner diversity index, systemic and haphazard are the most accurate since they have the highest value at T=1.5 E=1.4.

 

  Tree Species Actual Density Systemic Density % Error Haphazard Density % Error Random Density % Error
Common Eastern hemlock 469.9 476 1.3 500 6.4 456 -3.0
Sweet birch 117.5 116 -1.3 108 -8.1 200 70.2
Rare Striped maple 17.5 4 -77.1 16 -8.6 0 -100.0
White pine 8.4 8 -4.8 4 -52.4 0 -100.0

 

 

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

The virtual forest tutorial allowed for the testing of three different sampling strategies: systematic sampling, random sampling and haphazard sampling. It was determined that the fastest sampling time was for systematic sampling at 12 hours, 34 minutes. Followed by haphazard sampling and random sampling, 12 hours and 37 minutes, and 12 hours and 47 minutes, respectively.

The two most common tree species were the Eastern Hemlock and Red Maple. The most accurate of the three techniques for the Eastern Hemlock was systematic sampling with a 1.3% percent error, and the most accurate for the Red Maple was Haphazard with a percent error of 5.4%. The least effective sampling strategy for Eastern Hemlock was Haphazard with a percent error of 45.4%, and the least effective for Red Maple was random sampling with a percent error of 68.4%. Additionally, systematic sampling had a percent error or 26.0% for Red Maple, and random sampling had a percent error or 7.3% for Eastern Hemlock.

The two least common tree species were the White Pine and the Striped Maple. Systematic sampling was the best technique for White Pine and the worst technique for Striped Maple, with percent errors of 4.8% and 174.3%, respectively. The next best sampling strategy for White pine was haphazard sampling with a percent error of 98.8%, followed by random sampling with a percent error of 147.6%. For Striped Maple, the best sampling strategy was random sampling with a percent error of 18.9%, followed by haphazard sampling with a percent error of 66.9%.

Overall, the accuracy of the sampling got worse with lower species abundance as shown with the difference with the two groups above, the most common and least common species. For example with a larger species abundance (the most common species group), the percent error did not go above 68.4%. However, with a smaller species abundance (the least common species group), the percent error went as high as 174.3%. Overall, there is not one sampling strategy that clearly stands out as the most accurate, however, Systematic sampling is the closest, it has the two lowest percent errors for Eastern Hemlock and White Pine. It is hard to tell which is more accurate between random sampling and haphazard sampling as their percent errors are all fairly similar.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

For my Virtual Forest sampling strategies I chose the Snyder-Middleswarth Natural Area. The fastest sampling method was systematic sampling requiring only 10 hours of sampling while random and haphazard sampling required ~12.5 hours each. This appears mainly due to the amount of travel time to and from sample locations.

Common species such as Eastern Hemlock, Sweet Birch, and Red Maple appeared to be best represented by random sampling and haphazard sampling. Both produced low and sometimes negative values, indicating the true value and estimated value were extremely similar, or slightly overestimating.

More rare species of lower abundance, such as Yellow Birch, Chestnut Oa, Striped Maple and White Pine were often over represented through sampling. The systematic sampling indicated no Striped Maple or White Pine at all in the area which completely removes species from the area analysis. The haphazard sampling method demonstrated lower percent errors indicating that it is the most useful while analyzing rare species. The accuracy of species density decreased when analyzing rare species, % errors or over 100 or -100 indicate the data is not representative.

The sample size of 24 functioned well for random and haphazard methods, however, with the systematic method completely missed two species types. These rare species may be included in the data more efficiently if the sample size was increased to 50. Therefore, 24 was not sufficient in estimating the total abundance of tree species.

Blog Post 4 – Virtual Forest

TRU Open Learning Leave a comment

According to the sampling strategies I conducted in the virtual forest tutorial the systematic sampling technique had the fastest estimated sampling time at 12 hours and 36 minutes and Haphazard had the slowest estimated sampling time at 13 hours and 2 minutes. All 3 strategies used the area based methods.

Species  Systematic  Random  Haphazard 
Eastern Hemlock  1  14  25% 
Sweet Birch  31  3  56% 
Striped Maple  37  100  242% 
White Pine  42  100  42% 

As species abundance increase sampling accuracy increased or in other words the percentage error is lower.

Based on the results I obtained during the haphazard sampling indicated that one of the rare species percentage of error was 242 percent. The results from the virtual forest had estimated 60 striped maple and the actual was only 17.5. It appears to me that the systematic sampling method produced more accurate estimates than the other two methods.

Blog Post 4

TRU Open Learning Leave a comment

Results using the three sampling strategies on Snyder-Middleswarth Natural Area

Random Sampling

Total time to sample: 12hrs, 12 mins

% Error for Dominance

Most Common

Eastern Hemlock: 107 individuals, 15.9%

Red maple: 28 individuals, 17.1%

Least Common

Sweet Birch: 19 individuals, 14.3%

Yellow Birch: 17 individuals,  4.8%

Missed the striped maple and white pine entirely

Shannon-Wiener’s index of diversity for the community: 13.3%

Systematic Sampling:

Total time to sample: 12 hrs 12 minutes

% Error for dominance

Most Common:

Eastern Hemlock: 119 Individuals, 1.5%

Sweet Birch: 31 Individuals, 9.4%

Least Common:

chestnut oak: 11 individuals’, 50%

white pine: 2 individuals, 11.1%

Shannon Wiener’s index of diversity for the community: 6.6%

Haphazard Sampling:

Total time to sample: 13hrs

% Error for Dominance

Most Common Species:

Eastern Hemlock: 134 individuals, 5.4%

Yellow Birch: 34 individuals, 2%

Least Common Species:

Striped Maple: 8 Individuals, 157%

Chestnut Oak: 19 Individuals, 1.8%

Shannon Wiener’s index of diversity for the community: 6.6%

 

Comparing these three strategies it would appear that there is little meaningful difference in which one is most efficient with the time required to collect samples. Both the random and systematic sampling techniques requiring 12hrs and 12 minutes while the haphazard sampling required 13 hrs. In regards to the Shannon Wiener’s index of diversity for the community the haphazard and systematic sampling techniques only had a 6.6% error while the random sampling technique had a 13.3% error. The random sampling had relatively low % errors in the most common species (15.9% & 17.1%) and same for the most common species (14.2% & 4.8%); however, it missed two whole species present in the community. Systemic Sampling had low % errors in the most common species (1.5% & 9.4%), but with the least common species one high percent error 50%. Using Haphazard sampling had all percent errors under 6% except for in one of the two least common species a huge percent error of 157%. Due to these results I would probably deme random sampling the least accurate sampling method despite the haphazard sampling having one sample with 157% error, as it missed whole species and had the highest error Shannon Wiener’s index, while the Systematic Sampling was the most accurate on both accords. It seemed that the less abundant a species the more chance for high error, as they could be missed entirely. I would add that a stratified random sampling technique would probably have been the most accurate in this community as it seemed like some species were only present in small patches of land.

Blog Post 4 – Sampling Strategies

TRU Open Learning Leave a comment

In the virtual forest tutorial, I tried sampling techniques to compare their accuracy.

Haphazard (Area-sample)

  • Total quadrats sampled: 5
  • Area sampled: 500 sq. m
  • Species sampled: 6
  • Total specimens sampled: 39

There was an estimated sampling time of 2 hours 42 minutes. The error percentage for the 2 most common species was 13.9% and 34.2%. The error percentage for the 2 rarest species was 344% and 57%.

Haphazard (Distance-sample)

  • total points sampled: 5
  • species sampled: 5
  • total specimens sampled: 20

There was an estimated sampling time of 59 minutes. The error percentage for the 2 most common species was 42.3% and 28.1%. The error percentage for the 2 rarest species was 72% and 44%.

Systematic (Area-sample)

  • Total quadrats sampled: 5
  • Area sampled: 500 sq. m
  • Species sampled: 6
  • Total specimens sampled: 34

There was an estimated sampling time of 2 hours 39 minutes. The error percentage for the 2 most common species was 4% and 43.6%. The error percentage for the 2 rarest species was 47.1% and 43.6%.

Systematic (Distance-sample)

  • Total points sampled: 5
  • Species sampled: 7
  • Total specimens sampled: 20

There was an estimated sampling time of 59 minutes. The error percentage for the 2 most common species was 13.2% and 12.9%. The error percentage for the 2 rarest species was 51.5% and 11.8%.

From these results I determined that the distance method is the fastest way to sample. When comparing the error percentage of the different strategies for the 2 most common and 2 rarest species, it seems the accuracy does change with species abundance. Generally speaking, the systematic distance sampling method was more accurate than the haphazard one.