Category: Post 4: Sampling Strategies

Blog Post 4: Sampling Strategies

TRU Open Learning Leave a comment

For the virtual tree sampling tutorial, I selected Mohn Mill. The three sampling strategies used were the Area haphazard method, the distance-haphazard method and the distance random/systematic method. The distance-haphazard method had the fastest sampling time.
The two most common species were Eastern Hemlock and Sweet Birch, and the two rarest species were the Stripped maple and White pine.
The haphazard area method had the lowest percentage error, while the haphazard distance error had the highest percentage error when used to measure the most common species.
For the rare species, the distance random/systematic method had the lowest percentage error, but no Stripped maple trees were located. The haphazard method had the highest percentage error.
The accuracy appeared to be decreasing with a decrease in species abundance for the haphazard methods. However, for the distance random/systematic method, the accuracy increased with decreasing species abundance.

Instructor: Robyn Reudink

Blog Post 4

TRU Open Learning Leave a comment

The results of the virtual tree sampling tutorial showed that the haphazard sampling method is the most efficient in terms of time spent, at a total of 12 hours and 30 minutes. This is in contrast to systematic sampling which took 12 hours 37 minutes and random sampling at 12 hours and 51 minutes.

Haphazard sampling is the most accurate method for common species. When measured with the haphazard method, the two most common species, Eastern Hemlock and Red Maple, had percentage errors of 3.34% and 29.7% respectively. In contrast, systematic sampling was the least accurate method for both species (18.3% for Eastern Hemlock and 41.3% for Red Maple).

Systematic sampling is the most accurate for some rare species but not others. For Striped Maple systematic sampling had a percentage error of 14.3% and for White Pine 185%. Random and haphazard sampling had a percentage error of 100% for White Pine.

For common species all sampling methods were relatively accurate. Accuracy declined for all methods as species rarity increased. The most rare species, White Pine, had the lowest accuracy for all sampling methods. For rare species, 24 sampling units may not be enough to get an accurate representation of their abundance. Therefore, although time consuming, increasing the sampling unit size may be necessary when studying rare species of trees and other plant life. Increasing sampling unit quantity would increase accuracy of abundance for all species, both rare and common.

Blog Post 4

TRU Open Learning 1 Comment

For the virtual forest tutorial, I used the area-based model for my systematic, randomized, and haphazard sampling of vegetation. The most efficient method of sampling for my results was systemic sampling. On common species, it had a percentage error of 3.71% versus 18.75% for random and 9.48% for haphazard sampling. The technique which required the least amount of time was the random method. However, all sampling techniques had very similar sampling times with random being 12hrs 20mins, systemic was 12hrs 37mins and haphazard was 12hrs 25mins. The percent error for the rarest species was best using the haphazard sampling technique. It seems that the systematic approach is more useful for common species as it didn’t take up many rare species, and the haphazard sampling was more accurate for the rarer species.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

In the virtual forest tutorial, using the area-based method for samples taken in the Snyder-Middleswarth Natural Area, the data collected in the haphazard method had the fastest estimated sampling time of 12 hours, 26 minutes.

From the collected data set we can gather that the most common species is Eastern Hemlock and Red Maple. The actual density for each is 469.9 and 118.9, respectively. The data set for the systematic sampling method provided the lowest percent error for each species at 4.7% for Eastern Hemlock and 2.4% for Red Maple. Conversely, the random sampling method resulted in the largest percent error at 25.5% for Eastern Hemlock and 120.8% for Red Maple. Comparatively, the rarest species sampled, the White Pine and Striped Maple held different results. The data set for haphazard sampling resulted in the smallest perfect error at 1.2% for White Pine and 18.9% for Striped Maple. Interestingly, both White Pine and Striped Maple concluded to the same largest percent error through random sampling at 100%.

It is notable that overall, the higher the actual data of species and collected data of species presented results with lower percent error. The average percent error for Eastern Hemlock was 13.7%. However, the rarest species, White Pine, had the next lowest percent error at 41.2%

The data set collected resulted in the systematic sampling method having the smallest percent error overall at an average of 10.8%, which was closely followed by the haphazard sampling method at 11.6%.

 

Table 1: Summary Data Set of Snyder-Middleswarth Natural Area Comparing Density with Area-based Methods

 

Species Systematic Random Haphazard
Actual Density Data Percent Errora Data Percent Error Data Percent Error
Most Common:
    Eastern Hemlock 469.9 448.0 4.7% 350.0 25.5% 520.8 10.8%
  Red Maple 118.9 116.0 2.4% 262.5 120.8% 137.5 15.6%
Rarest:
  White Pine 8.4 8.0 4.8% 0.0 100.0% 8.3 1.2%
  Striped Maple 17.5 12.0 31.4% 0.0 100.0% 20.8 18.9%
Estimated Sampling Time 12 hours, 41 minutes 12 hours, 45 minutes 12 hours, 26 minutes

 

aThe calculation used in finding the value of percentage error for each category is:
(E – T)/T*100, where E = estimated value and T = true value

Table 1 contains the summary for my findings for each species, eastern hemlock, red maple, white pine, and striped maple through systematic, random, and haphazard sampling in the Snyder-Middleswarth Natural Area. The table compares actual density of each species to the gathered data I found during sampling and the amount of time it would take to produce these samples.

Blog#4

TRU Open Learning Leave a comment

Blog Post 4: Sampling Strategies
Create a blog post describing the results of the three sampling strategies you used in the virtual forest tutorial. Which technique had the fastest estimated sampling time? Compare the percentage error of the different strategies for the two most common and two rarest species. Did the accuracy change with species abundance? Was one sampling strategy more accurate than another?
I looked at the Mohn forest for the exercise. I found that the time spend was approximately the same when using the haphazard and random methods for area, with the random taking slightly less time. The haphazard strategy resulted in more species found (richness) and more plants counted. In this particular situation though, I think the random method is better because the forest is a quite patchy, having had many disturbances. If it was the Snyder forest, which is quite homogenous, randomness wouldn’t be as important.

Post #4 Sampling Strategies

TRU Open Learning 1 Comment

Overall, the most efficient method of sampling for my results was the random sampling. On common species, it had a percentage error of 7.5% versus 13.8% for systematic and 43.6% for haphazard sampling. It also had the best percentage error at 72.4% for the rare species, versus ? (did not account for the species) for systematic and 198.2% for haphazard sampling. Considering the fact that random and systematic had only a 27 minute difference in sampling times (12h38m vs 12h11m), it shows how much more efficient and accurate random sampling can be in both achieving results and saving time and resources in research. Haphazard sampling was the least accurate in terms of percentage error, but visiting the 5 sites versus the 24 of the other methods only took two hours and forty minutes. One notable thing I encountered while doing this activity is I found was that despite the 43.6% percentage error of common species for haphazard sample, this was due to skew because the two most common species having percentage errors of 86.25% and 0.925%, and that last value was the most accurate of any species using any method, the second smallest percentage error being 3.2%. Another interesting find was that systematic sampling had relatively similar accuracy for the common species, but it did did not register the two rare species that both random and haphazard sampling did. This is interesting because despite the fact that the haphazard sample sites were strategically chosen, but it only had a fifth of the quadrants as systematic sampling and a much higher percentage error of common species. What these results showcase is the strength of random sampling and what effect human bias can have on the results of sampling.

Blog Post 4: Sampling Strategies

TRU Open Learning Leave a comment

For the virtual forest tutorial, I used the area sampling technique.  The random sampling strategy had by far the fastest estimated sampling time at 4hours 51minutes compared to the systematic strategy (12 hours 37 minutes) and the haphazard sampling times (13 hours 2 minutes).

For the two most common tree species, the eastern hemlock and sweet birch, the percentage error was low for the systematic strategy (7% and 5%) and was highest for the random strategy (19% and 36%) (Figure 1).  The percent error from the haphazard strategy varied with the eastern hemlock having the lowest percent error at 6% and a high percent error of 21% for the sweet birch (Figure 1).  Based on the percent errors for the most common species, the systematic strategy seems to be the most accurate, the random being the least accurate and the haphazard being unpredictable with one of the percent errors being low and the other being high.

When looking at the percent error for the two most rare species, the striped maple and the white pine, all the sampling strategies had large percent errors for at least one of the species; this shows that accuracy increases with species abundance and decreases with species rarity.  The systematic strategy had large percent errors (178% and 100%) as well as the random strategy (185% and 48%) (Figure 1).  The haphazard strategy had a comparatively lower percent error for the striped maple at 14% but the white pine had a large percent error at 138% (Figure 1).  All the strategies for rare species were not able to accurately represent the population and the haphazard strategy having a high and low value again shows it is an unpredictable sampling strategy.

I think the systematic strategy was the most accurate when looking at the percent errors for all the tree species.  Surprisingly the haphazard strategy had lower percent errors overall than the random strategy which had the largest overall percent errors.

Post 4: Sampling Strategies

TRU Open Learning 1 Comment

Using the area-based sampling, these are my results from the visual forest tutorial:

Based on my results, all three sampling techniques took about the same time, but the random sampling was slightly faster at 12 hours and 15 minutes.The two most common species were the Eastern Hemlock and the Sweet Birch. Systematic sampling yielded the lowest percentage error for these species. The two least common species were the Striped Maple and the White Pine. As you can see, the percentage errors were extremely high for the Striped Maple using the systematic and haphazard sampling techniques and very low using random sampling. For the White Pine, all the sampling techniques yielded high percentage errors but haphazard sampling was the lowest at 50%. The accuracy seemed to be greater with the more abundant species as the percentage errors were lower compared to the less abundant species. Based off my results, I cannot say that one sampling strategy was more accurate than the others, as the results were varying. I would need a larger sample size in order to obtain more accurate results. 

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

When participating in the virtual forest tutorial activity, I used three sampling strategies, haphazard area sampling, stratified-random sampling and systematic sampling.

When looking at the haphazard area sampling method results, I found the sampling time to be the fastest at 18 hours 59 minutes. Following this was systematic sampling at  26hours 58 minutes and finally random at 63 hours 18 minutes. With haphazard, I found reducing bias selection was difficult when considering the topography and cluster images. I was drawn to selecting those areas first in hopes of greater concentration of results. Unfortunately this was not the case and haphazard proved to be the least accurate method of sampling for common and uncommon species. The most accurate method was also the most time consuming. When filling in the random sampling selections, I found the greatest variation in discovered species as well as the highest accuracy in percentages.

The two most common species were:

Red maple 

Haphazard-9.0%

Random-1.0%

Stratified-3.25%

Chestnut oak

Haphazard-34.0%

Random-3.25%

Stratified-3.75%

The two rarest species were:

Sweet birch

Haphazard-90.0%

Random-0.1%

Stratified-82.0%

 White ash

Haphazard-90.0%

Random-0%

Stratified-100%

When reviewing the effects of density on the accuracy of species sampling I found higher density species had a greater accuracy in recordings. The lower density species, when found have a very high accuracy, but often are not found which greatly reduced the accuracy.

Post 4: Sampling Strategies

TRU Open Learning Leave a comment

The area around the stormwater pond is called Marshall Springs. The type of sampling strategies that I am going to use is area-based sampling. I will be taking 16 samples. They will consist of 1m x 1m quadrats. I will have 4 quadrats in each zone that I have picked, natural wetland, riverine (lower elevation), riverine (higher elevation), and grassland. I will use random and systematic methods to determine the most abundant species of vegetation in each quadrat along with the soil moisture. I will only tabulate the species that are present for, let’s say, above 10% coverage in the quadrat. I will be looking at the percent coverage of vegetation in each quadrat and not the number of individuals.
Given the area of the park and region I am looking at I could have increased the size of the quadrats or the number of quadrats which could better reflect the actual percent coverage of vegetation in each region and more accurately display the soil moisture. The soil moisture is also only reflective of the current state. If we were to be experiencing a rainfall or drought then obviously the soil moisture would fluctuate.
I will also be trying to calculate the moisture content of the soil in each zone. I will collect 1 cup (250ml) of dry soil and weigh it. I will then collect samples from each quadrat and determine the moisture content present. My fiancée worked at a garden center through her degree and helped me with calculating moisture content as she said that different plants require different moisture content in order to optimize growing potential.