top of page

Part 1. Lecture: Distance dependent detection

In this part of the workshop we describe about how your ability to detect animals depends on how far the animals are from you. This is a very intuitive concept and makes an easy introduction into statistical models describing detection probabilities and the parameters that describe this model.

â€‹

• We use an example of 16 detectors placed in a square configuration within our study area.

â€‹

• Five animals are within the study area, each making 10 calls which are detected (or not) by the detectors. We pretend for now that we know  the location of each animal and that they make 10 calls.

â€‹

• Measuring the distances between the location of the animals and each detector for each call detection allows us to generate a histogram of distances. We do this for:

1. all calls (pretending we detected all calls at each detector).

2. calls actually detected by the individual detectors where detection depended on distance to the animal.

â€‹

• A comparison of the two histograms reveals the shape of the detection model which describes how detection probabilities decrease with increasing distance.

Part 2. Lecture: Mark-recapture concepts

Mark-recapture (also called capture-recapture) concepts help us estimate the parameters of the detection model from part 1 of this chapter.

â€‹

• We first illustrate the concept using samples taken by two detectors out of a population of N animals.

• Step 1: Animals detected ("captured") by detector 1 are marked so that we can re-identify them when detected by detector 2. This way, detector 1 sets up trials for detector 2: will detector 2 detect any of the animals that detector 1 detected and marked?

• Step 2: We then look at how many marked animals were detected by detector 2. These are the trials with a successful outcome.

• Step 3: The ratio of numbers from step 2 and step 1 gives us the probability of detection.

â€‹

• Spatial capture-recapture includes information on where the detectors were.

â€‹

• In a two detector scenario we show how both detectors can set up trials for each other.

â€‹

• Using 16 detectors we show how for each detector, the other 15 set up trials.

â€‹

• Pretending we know the location of the detected animals, we know at which distance these trials were set up. This alleviates the need to know how many animals there were (or how many calls they made) from part 1.

bottom of page