Category: Post 4: Sampling Strategies

Blog Post 4: Sampling Stategies

TRU Open Learning Leave a comment

Blog Post 4: Sampling Strategies

 

I sampled the Snyder-Middleswarth Natural area using area-based systematic, random, and haphazard methods. The technique with the fastest sampling time was the area-based systematic approach with a sampling time of 12 hours, 5 minutes, while random and haphazard techniques had sampling times of 12 hours 34 minutes, and 12 hours 36 minutes, respectively. The least abundant species showed the most accuracy in results, White Pine had a 0% error for two different methods, random and haphazard sampling strategies. In general, the area-based random sampling technique was the most accurate for each species. Area based systematic sampling showed the least accuracy in results. Below is a list of percent error for the two most rare species, Red Maple and White Pine, as well as the most common, Eastern Hemlock and Sweet Birch, for each sampling strategy.

 

Area Based Systematic:

Eastern Hemlock-45.4%

Sweet Birch-20.6%

Red Maple-82.5%

White Pine-50.0%

 

Area Based Random:

Eastern Hemlock-9.1%

Sweet Birch-6.38%

Red Maple-1.85%

White Pine-0%

 

Area Based Haphzard:

Eastern Hemlock-25.0%

Sweet Birch-6.38%

Red Maple-22.6%

White Pine-0%

Blog Post #4

TRU Open Learning Leave a comment

Sampling Theory Using Virtual Forests

I completed the Community Sampling Exercise on the Snyder-Middleswarth Natural Area and received the results from 3 different types of surveys to compare.  The systematic-area method produced the shortest estimated sample time of 12 hours and 6 minutes.  A quick calculation comparing the estimated density data with the actual data revealed the systematic-area method had a percent error of 28.7%, while random sampling had a percent error of 11.8% and haphazard sampling demonstrated the greatest accuracy rate at 11.6%.  The percent error for the Striped Maple was 2.1%, and the White Pine was 14.4% both which are rare species.  The common species were Easter Hemlock 6.2% and the Sweet Birch had a percent error of 5.4 %.

The most accurate way to measure both the common species and the rare species was the haphazard method. The accuracy declined with the rare species as some of the rare species were not detected by some of the sample strategies. This leads me to believe 24 sample points did not cover enough ground to accurately represent the rare species.  While 24 was adequate to represent the common species, I would recommend increasing the sample points for greater accuracy of the rare species.

Blog Post 4: Sampling Strategies

TRU Open Learning Leave a comment

In the virtual forest tutorial, the three sampling techniques used to sample tree species in the Snyder-Middleswarth Natural Area were random/systematic sampling, random sampling and haphazard/subjective sampling. I found the technique with the fastest sampling time was the second exercise- random sampling. As opposed to the first and third exercises, the second exercise only required 24 locations to be sampled randomly, without any other restrictions. After reviewing the percentage errors for both the most common and rare species, it was found that the third exercise, haphazard/subjective sampling, was the most accurate technique, with a 3.75% error for Eastern Hemlock, and a 28.6% error for Striped Maple. The second most accurate technique was random/systematic sampling, with a 10.6% error and 52.4% error for Eastern Hemlock and White Pine, respectively. The random sampling technique had percentage errors of 14.7% for Eastern Hemlock and 54.8% for White Pine. These varying percentage errors indicate that with changing species abundance, the accuracy of each technique changes as well.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

Using the virtual forest, I sampled the Snyder-Middleswarth Natural area first systematically, then randomly, then haphazardly. Each of the different sampling strategies had a sample size of 24. The most efficient strategy was area-based systematic sampling, which was estimated to take 12 hours, 6 minutes to sample. Area-based random and haphazard sampling would take 12 hours, 48 minutes and 13 hours, 9 minutes respectively. Percentage error was calculated using the estimated and true values of species density.

Area-based systematic sampling:

Eastern Hemlock: 21.5%

Sweet Birch: 27.7%

Striped Maple: 100%

White Pine: 100%

Area-based random sampling:

Eastern Hemlock: 39.2%

Sweet Birch: 14.9%

Striped Maple: 100%

White Pine: 142.9%

Area-based haphazard sampling:

Eastern Hemlock: 13.2%

Sweet Birch: 19.2%

Striped Maple: 0.09%

White Pine: 100%

 

Based on these percentage errors, the most accurate sampling method for the Eastern Hemlock was haphazard sampling with only 13.2% error, and random sampling for Sweet Birch with 14.9% error. These are the two most common species of trees in the Snyder-Middleswarth Natural area. The two most rare species, Striped Maple and White Pine, were most accurately sampled using haphazard and systematic/haphazard sampling respectively. Haphazard sampling of Striped Maple resulted in only a 0.09% error and both systematic and haphazard sampling resulted in 100% error as there were no White Pines recorded. In random sampling, White Pines were over-represented and had an error of 142.9%.

Accuracy increased with abundance of a species, as seen by the significantly lower percentage errors in the more common species vs. the rare species. Although each strategy used 24 samples to gather data, White Pine was undetected in both systematic and haphazard sampling, suggesting that the total number of samples was insufficient to truly capture the number of species in the community and their abundance. Of the three strategies, haphazard sampling seemed to most accurately estimate the abundance of each species in the area, as the percentage errors for the common species were relatively low and Striped Maple (rare species) was present.

Blog Post 4: Sampling Strategies

TRU Open Learning Leave a comment

The first sampling technique I explored was area/haphazard. I sampled 27 quadrats, which was estimated to take 14 hours and 30 minutes. The percent error for the two most common species were 11.0% and 14.2% respectively. The two most rare species had percent errors of 62.6% and 79.7%. Accuracy changed drastically when abundance decreased and sample time is not optimal, therefore, this strategy is not the best choice for the Mohn Mills community.

The second method I tested was area/random. The most abundant species had percent errors of 9.2% and 14.5% while the two least abundant were 13.3% and 47.9%. I believe that accuracy only changed drastically due to an outlier. Otherwise, they might be very similar. Estimated sampling time for this method, also 27 quadrats, was 14 hours and 16 minutes. This is very similar to the first method’s sampling time.

The third method I looked at was distance/haphazard. The sampling time for 27 quadrats was only 5 hours and 15 minutes, making it much more reasonable than the area strategies. Percent error for the most common species was 13.2% and 13.2%, while the two rarest were 8.57% and 26.9%. Although the last percent error was higher than the others, these values are the most consistent out of all three sampling techniques. Along with the reasonable sampling time, this makes the distance/haphazard method the best choice for this community.

Post 4

TRU Open Learning Leave a comment
Tree Species Actual Density Distance Systematic Data Distance Random Data Distance Haphazard Data
Eastern Hemlock

(Most Common) 

 

469.9

  

308.6

 % Error

 

34.3

  

395.7

 % Error

 

15.8

  

427.4

 % Error

 

9.0
Red Maple 118.9 105.8  

11.0

 45.7  

61.6

 63.7  

46.4
Sweet Birch 117.5 167.5  

42.5

 60.9  

48.1

 127.3  

8.3
Yellow Birch 108.9 114.6  

5.2

 159.8  

46.7

 127.3  

16.9
Chestnut Oak 87.5 141.1  

61.3

 68.5  

21.7

 90.9  

3.9
Striped Maple 17.5 8.8  

49.7

 0.0  

100

 27.3  

56.0
White Pine

(Most Rare)

8.4 0.0  

100

 0.0  

100

 9.1  

8.3
Survey Time 4 hours, 5 min 4 hours, 44 min 4 hours, 11 min

All three sampling strategies appear to have little difference in time.  However, the Random and Haphazard sampling strategies should take longer overall if travel time between plots is considered.

The most accurate sampling strategy for the most common and most rare species was found using the Haphazard sampling strategy and the sampling error for both were similar.

The sampling error greatly increased from the most common to the rarest species in both the random and systematic sampling strategies.  However, no real pattern was observed.

The number of plots was sufficient in capturing the number of species in the community but to improve the accuracy of the data more plots should be added.

Blog Post #4 – Sampling Strategies

TRU Open Learning Leave a comment

For the Virtual Forests tutorial, I chose to use the area-based methods for my 3 samples. The fastest technique for sampling was the systematic technique along a topographic gradient with a time for 12 hours and 36 minutes.  What surprised me about the results was that the random and haphazard techniques, each taking 13 hours and 14 minutes, did not take much more time than the systematic approach.

The two most common species I found in my samples were the Eastern Hemlock and Sweet Birch.

Systematic Random Haphazard
Actual Density Measured Density Percentage error (%) Measured Density Percentage error (%) Measured Density Percentage error (%)
Eastern Hemlock 469.9 388.0 17.4 304.0 35.3 436.0 7.2
Sweet Birch 117.5 72.0 38.7 96.0 18.3 112.0 4.7

Analysis of the data collected for the 2 most common species indicates that the haphazard method of sampling was the most accurate strategy, with both common species having percentage errors in the single digits.

 

Systematic Random Haphazard
Actual Density Measured Density Percentage error (%) Measured Density Percentage error (%) Measured Density Percentage error (%)
Striped Maple 17.5 28.0 60 16.0 8.6 12.0 31.4
White Pine 8.4 0.0 100 0.0 100 4.0 52.4

Analysis of data collected for the 2 most rare species shows that all 3 sampling methods provided very inaccurate results. The second rarest species, the Striped Maple, was well sampled in the random method with a percentage error of only 8.6%. However, the rarest species, the White Pine was not found at all using this method. As species abundance decreased, percentage error of sampling using all 3 methods decreased.

 

Overall, 24 plots does not appear to be enough to get an accurate representation of species density across the range of species in the geographical area. I would predict that increasing the number of plots would increase the accuracy of all 3 sampling techniques.

To test this theory, I repeated both the haphazard and systematic techniques using 50 plots instead of 24. I found the same number of species (7) as before, however the haphazard method now yielded percentage errors of 0.02% for the most common species (Eastern Hemlock) and 19% for the rarest species (White pine). The systematic method now yielded a percentage error of 3% for the most common, and 19% for the rarest species.  I conclude, based on this observation, that more sampling plots, regardless of method, yield more accurate results.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

After completing the different sampling techniques and comparing the results, the fastest estimated technique was systematic with an estimated time of 12hr 5mins the next fastest method was haphazard with a sampling time of 12hr 44 minutes. The final and longest sampling method was random with an estimated time of 12hr 55minutes.

 

The most common species sampled was Eastern hemlock, systematically sampling gave a percentage error of -4.2%, randomized sampling gave a percentage error of -16.6% and haphazard sampling gave a percentage error of 27.7%. For the second most common species sweet birch, systematically sampling gave a percentage error of -14.9%, randomized sampling gave a percentage error of 2.8% and haphazard sampling gave a percentage error of -29.1%. When comparing these two data sets for the most common trees we can see that systematically sampling gave the lowest error percentage for eastern hemlock and randomizes sampling gave the lowest percentage for sweet birch.

 

The most least species sampled was White pine, systematically sampling gave a percentage error of 98.8%, randomized sampling gave a percentage error of 98.8% and haphazard sampling gave a percentage error of 98.8%. For the second least common species Striped maple, systematically sampling gave a percentage error of -100%, randomized sampling gave a percentage error of 18.9% and haphazard sampling gave a percentage error of -4.5%. When comparing these two data sets for the least common trees we can see that all sampling methods provide the same results for White pine and haphazard sampling gave the lowest percentage error for Striped maple.

 

Overall as species abundance decreased percentage error of all sampling methods increase dramatically, from this data set randomized sampling appears to be the most effective

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

Each technique took about the same amount of time sampling: systematic took 12.37H, random took 12.36H, and haphazard took 12.44H.

The most common species was Eastern Hemlock, next was Sweet Birch. Systematically sampling easter hemlock gave a 7.4% sample error and systematically sampling Sweet Birch gave 17.2%. Randomly sampling Eastern Hemlock gave 2.2% sample error and randomly sampling Sweet Birch gave 14.1%. Haphazardly sampling Eastern Hemlock gave 40% and 22% for Sweet Birch. Therefore the least amount of sample error came from the random sample and the most came from haphazardly sampling.

The lease common species was White Pine then Striped Maple was a bit less rare. Percent error of systematically testing White Pine was 70% and it was 100% for Striped Maple. Percent error of randomly sampling was 60% for White Pine and 100% for Striped Maple. Percent error for systematically sampling was 70% for White Pine and 120% for Striped Maple. None of these methods proved to be accurate enough to use in a proper data set. But randomly sampling had the least percent error for White Pine.

The more abundant the species it, the more accurate the sampling is. To increase accuracy, the number of samples taken can be increased.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

I used the distance-based method for the Sampling Theory Using Virtual Forests Tutorial. Below is a table summarizing the comparison between the actual and estimated densities of the seven tree species in the Snyder-Middleswarth Natural Area as well as the percentage error using each (distance-based) sampling method. Also included in the table is the estimated sampling time for each method.

Based on the results above, the systematic method had the fastest estimated sampling time.

The two most common species are the eastern hemlock and red maple while the two rarest species are the striped maple and white pine. Between the three different strategies, percentage error is generally inversely related to species abundance. This is more prominently seen in the random sampling strategy.

When we only look at the most common and rarest tree species, the haphazard approach is the most accurate. However, when we consider all tree species, the systematic approach is the most accurate with a maximum percentage error of 57.7% versus the 100% error with the haphazard approach for the striped maple.

Overall, the systematic approach is best because of its relatively lower sampling time and for its relatively higher accuracy.