Study Design and Using the Cluster() Command

[adamdillon]

Murray,

Thanks for your input on my past posts. In light of your last response on separate sessions and how SECR is "cheating" some, I thought I might ask you some advice about the general set up for my models.

I am working with Island Foxes on the California Channel Islands.

The trapping design consists of 18 separate grids (1000m x 200m) located across the island ( grids > 2 km apart). Each grid is trapped for 6 days and the trapping has been run for 7 years. Foxes are separated into 2 age classes (adults and pups), as well as two sex classes (male and female). I am looking to estimate the density of adults and pups for each year of the study.

So here had been my thought process so far, any advice or suggestions is greatly appreciated.
Since I am estimating the density of adults and pups annually, I could split the data into 2 files (adults and pups) or into 14 files (adults x 7 years, pups x 7 years), but it seems as though some parameters (g0 and sigma) may be shared across age classes and/or across years and therefore all the data should be in a single data file and analyzed all at once.

Currently I have 28 sessions (7 years x 2 age classes x 2 sexes), although I'm starting to rethink this approach. Alternatively I could set it up with 14 sessions (7 years x 2 age classes) and enter sex as an individual covariate. Since I want to estimate density annually for adults and pups, I have been using the full likelihood method. One thing I'm having trouble with is the cluster() command. It seems as though my study design is ideal for using the cluster command but I'm having trouble implementing it. In reading about trap.builder it seems as though a lot of the commands are used to delineate the polygon where the clusters
will be determined. In my study there are 18 designated grids or clusters, is there a way to assign the clusters more directly? Can I enter another column in the Detector File that has a "cluster" number associated with it and just assign the clusters by that column?

I'm sure I'll have more questions to follow, but it seems like I need to sort out my SECR design and cluster issue first.

Thanks for the help and your patience.

Adam

[murray.efford]

[I don't think there is a Cluster() command, but...]

Can I enter another column in the Detector File that has a "cluster" number associated with it and just assign the clusters by that column?

Yes, you can input one trap covariate for the cluster number (clusterID), and another trap covariate for the trap number within a cluster (clusterTrap), and use those subsequently to assign clusters. Remember the '/' on each line before any trap covariates. I haven't done this for a while, but I think something like this should work:

Code: Select all

clusterID(mytraps) <- covariates(mytraps)$myclustID
clusterTrap(mytraps) <- covariates(mytraps)$myclustTrap


Murray

[murray.efford]

that should be

Code: Select all

clustertrap(mytraps)