Bootstrapping and Royle-Nichols model

[janderson0723]

My occupancy study has 36 sites (transects) visited 3x each with 26 presences. Dataset is in my last post. I am using the Royle-Nichols model for analysis.

I'd like to bootstrap using my dataset to see how llambda, r, and psi would be with a few hundred sites and at least 80 presences. What is the recommended approach for doing this? Can this be done using PRESENCE? I selected the bootstrap option (although not completely sure what its doing in this case) when running the Royle-Nichols model in PRESENCE, but the output doesn't appear to be different in any way than when I don't select the bootstrap.

Also, what is the recommended number of presences and sites for good estimates with the Royle-Nichols model?

Thanks
Jim.

[jhines]

PRESENCE doesn't handle GOF for any model except the single-season model (at this time).

I've written a script to simulate data in order to test the parameter estimates and this could possibly be used in your case. Contact me directly (jhines@usgs.gov) if you're interested in obtaining this.

I can't answer your last question with anything but 'it depends'. It depends on your view of 'good estimates'. It depends on the population parameters. It depends on what you're going to do with the estimates. You're on the right track to simulate data with different input parameters to get an idea of what sort of estimates you'll get.

Jim