Objectives

  • Develop a sampling scheme.
  • Estimate total counts of desert shrubs using your sampling scheme.
  • Compare sampling methods.

Instructions

Collaboration tips:

  • Your group will receive one overall grade for your submission.
  • You will receive an individual peer evaluation grade.
  • Your individual final grade is the product of the group grade and your peer evaluation grade.
  • Be an active participant!

Record Data in Excel

One member of the group needs to create a data sheet using Excel in OneDrive:

  1. Log into your OneDrive
  2. Create an Excel spreadsheet
  3. Share the spreadsheet with the group members

Background

When we want to estimate populations of animals or plants in a area, we would like to be able to exhaustively survey the entire are of interest.

However, that is usually not possible due to constraints such as time, money, accessibility, etc.

How could we estimate the number of desert shrubs in a vacant block in Cathedral City, California?

  • Count the shrubs in smaller samples of the entire region!

The aerial satellite imagery for this exercise is from the empty block near the corner of Dinah Shore Drive and Da Vall Drive in Cathehdral City CA.

Sampling

In your group, you’ll decide upon a sampling strategy to estimate the total count of desert shrubs.

The region has been divided into quadrats of different sizes:

  • Large: 16 total quadrats
  • Medium: 64 total quadrats
  • Small: 128 total quadrats

As a group you’ll decide on a sampling strategy from those listed in Gardner:

  • completely random sampling
  • stratified random sampling
  • transects
  • haphazard sampling

For each of the quadrat sizes, you’ll randomly choose to count a number of quadrats (\(n\)) that cover 25%, 50%, and 75% of the total area as follows:

quadrat size total quadrats (N) quadrats to sample (n) approx. coverage
large 16 4 25%
large 16 8 50%
large 16 12 75%
medium 64 16 25%
medium 64 32 50%
medium 64 48 75%
small 128 32 25%
small 128 64 50%
small 128 96 75%

Note that for transect sampling, your final number of sampled transects may not exactly match the numbers in the above table.

NOTES:

  • Recall the terms precision and accuracy.
  • In real-life studies, we make imperfect observations and different observers, or replicated observations of the same quadrat may come up with slightly different counts.
  • For each coverage level (25%, 50%, 75%) you need to choose a different set of quadrats.
  • You will probably count some of the quadrats multiple times. This is OK, you should repeat the counting process each time whether you’ve already counted the shrubs in that quadrat or not. This method will capture some of the variability and human error that occurs in real-life sampling efforts.
  • Does this remind you of biased sampling?

Quadrat Grids

You can find the quadrat figures here:

Random Quadrat Sampling

You can use the script quadrat_sampling_script.R to select random transects. The file is located in the Course Files tab in the Course Materials section of the GitHub page.

Data Files

The first row in your file needs to contain the following column names:

  • quadrat_size
  • quadrat_ID
  • coverage
  • shrubs

An example data sheet might look like:

quadrat_size quadrat_ID coverage shrubs
medium A1 25% 12
medium A1 50% 11
medium C2 75% 9
small L7 50% 5
... ... ... ...

When your group has finished counting, make sure to export your data from Excel to a csv file

Name this file desert_shrubs.csv and save it in the data subdirectory of your main course RProject folder.

Estimating Shrub Population

To estimate the total numbers of shrubs, you’ll need to:

  • Count individuals in the quadrats you chose: \(x_1, x_2, x_3, ... , x_n\)
  • Calculate the mean counts in your chosen quadrats: \(\bar{x}\)
  • Use your mean quadrat counts to estimate the total counts: \(N\times\bar{x}\)

To calculate the mean quadrat counts you can use the formula for mean:

\(\bar{x} = \frac{x_1 + x_2 + x_3 + ... + x_n}{n}\)

You could calculate this manually, or… you could use R!

  • Import your data into R using read.csv() and here().
  • Remember that you have to have your course RProject open for this to work!
  • Do you remember how to create a vector and calculate its mean and standard deviation?

Report

Your group will submit a short report, in a pdf file, via Moodle.

Report Instructions

Your report must address the following questions:

  1. Which sampling scheme did you select?
  2. Which sampling scheme do you think is the best for estimating counts from quadrats and why?
  3. Which sampling scheme do you think is the worst for estimating counts from quadrats and why?

You need to include a table showing:

  • Mean quadrat counts.
  • Standard deviation of quadrat counts.
  • Estimate of total number of shrubs in the empty block.

for each quadrat size and sample coverage:

quadrat size covarage mean shrubs per quadrat st dev. shrubs per quadrat estimated total shrubs
large 25%
large 50%
large 75%
medium 25%
medium 50%
medium 75%
small 25%
small 50%
small 75%