Tutorial–ducknest analysis in R ✏️

Practical 2 – Duck nest analysis

This has been your first experience using the Distance package. There are only a handful of functions you need to successfully complete a distance sampling analysis. This practical gave you experience with these functions. The data are familiar to you because they the same data you used when trying to fit a detection function by hand in Exercise 1. This exercise lets the computer do the work. Compare the estimate of duck nest density produced by Distance with the estimate you manually produced.

• What is the estimate value of \(\hat{P}_a\)?
  0.869 0.039 614.25 0.933
• Given your estimate of \(\hat{P}_a\), what is its meaning?
  distance to which nests were detected area under the curve proportion of nests detected within 2.4m of transects number of iterations to estimate detection function parameters probability of detecting a nest within 2.4m of transect

Density converted to abundance

In the ducknest data frame, I failed to specify the size of the study area in the Area field. As a consequence, the ds function is only able to provide a density estimate. Given the size of the Monte Vista refuge is 47.7 \(km^2\), and the estimate of nest density from the output provided, complete the code fragment to provide an estimate of the abundance of nests.

• What is the estimated number of nests on the refuge?
• What assumption does our estimated abundance of nests on the refuge rest?
  Nests are uniformly distributed over refuge Ducks like water Transects representatively sample the refuge Detection function model fits the data

• How should the plot be interpreted?
  model that fits has all points near the 45° line this plot has nothing to do with model fit a rapid rise in the graphed points implies good model fit
• The numerical output from gof_ds is from the Cramer-von Mises test. Interpret the output.
  P-value is <1, therefore the model fit is poor The test statistic is smaller than the P-value, therefore the fit is adequate The test statistic is close to 0, therefore the fit is poor P-value is close to one, indicating there is support for proposition the model fits the data

To reinforce the idea that similar results are produced (for good data) from the different key functions, write a few lines of code to explore this, using R as a calculator. Value entered to one decimal place.

• What is the percentage difference between smallest and greatest density estimates using these three models?

References