Analysis of double platform data
Analysis of double platform data
A fundamental assumption of “conventional” distance sampling is that we detect all animals at zero distance – i.e., that g(0)=1. In some circumstances this assumption is violated, for example in shipboard surveys of marine mammals where they may be underwater or obscured by waves when the observer looks for them even right on the transect line.
Here we introduce an extended type of distance sampling that can help deal with this issue: the use of two independent observation teams, sometimes called “double platform” (because on ships they operate from different physical platforms). The analysis approach used on double platform data is called “Mark Recapture Distance Sampling” (MRDS).
We introduce the topic with a lecture, and follow up with an exercise that covers two datasets: an artificial survey of a known population of golf tees in St Andrews (to illustrate the concepts), and a real-world survey of crabeater seals in Antarctica (Southwell et al. (2007)).
Photo by Luiz Felipe on Unsplash
Lecture materials
Exercise materials 
Supplemental materials
Further reading on MRDS
- See Burt et al. (2014) for a paper introducing MRDS methods. This is more accessible than the original MRDS reference (Laake & Borchers, 2004).
- Notes on interpreting output from Burt et al. (2014) faecal pellet data
- Borchers et al. (2006) develop point independence methods.
- Buckland et al. (2010) go one step further with “limiting independence” methods - but there is no accessible software to implement this.
- Laake et al. (2011) extend the methods to point transect surveys.
Other approaches for dealing with g(0)<1