Simulating heterogeneity

[sixtystrat]

I am running some simulations in Density to look at different trap cluster designs for a bear DNA mark-recapture study. We have some earlier data and I have a density estimate of 0.0022 bears/ha, g0=0.21, and sigma = 2337 m. Capture heterogeneity is probably going to be a problem so I need to simulate that in the data. Is there a way to also estimate the distribution and CV of the capture heterogeneity as required by the simulator? What values should I use? Thanks!!!
Joe Clark

[murray.efford]

Hi Joe

The only way I know to estimate CV is to fit a particular mixture model (e.g. 2-class finite mixture) and then compute each CV from the fitted parameters. If you use a finite mixture, the simulator in Density offers a direct mechanism for simulating a mixture of classes differing in density detection parameters (hit the 'group' button instead of 'add' for the second and later classes). I suggest you play with this to see how it works before attempting SECR estimates.

There is an argument in a manuscript that I'll send you offline for taking a more relaxed approach: heterogeneity is not always as poisonous as you might expect. The ms also includes the formula for getting CV(theta) from a mixture of thetas.

Murray