The reading group will meet once again this Friday (9th October) at 1pm. We're going to switch topics to Capture-recapture using modern technology with a paper by Prof Paul Sunnucks (Monash) and colleagues. Paul uses capture-recapture models together with non-invasive remotely collected samples to estimate population size and dynamics.
One question is what is different here to more conventional capture-recapture problems?
ReplyDelete- the way individuals are "captured" - in this case it was hairs that stick on tape outside their burrow, instead of traps or encounters. This for example means that the juvenile wombats might be under-sampled (or completely missed) because they aren't tall enough to hit the sticky tape, as acknowledged in the paper.
- possible misidentification. Seems low here but is different to tags, where if you recapture a tagged individual you know for sure it is the same one (up to human error). Replicate samples might reduce the chance of this, you can use a model that incorporates misidentification to try to handle this.
- more data. This is less field-intensive (easier to set up sticky tape or collect scats than to capture individuals) so you can sample more individuals per unit effort. Margin of error on CIs became a third as big as previous survey efforts indicating this can lead to better precision.