search

jclbrooks / MD_Stream_Salamanders

MIT License
0 stars 0 forks source link

Variation in occupancy, detection, colonization, extinction #15

Open djhocking opened 5 years ago

djhocking commented 5 years ago

------------ psi and p can't both have means that vary randomly ---------

vary p randomly

psi fixed mean prob (intercept) that only varies by region

Need to make hard decisions about what parameters vary randomly and which are stochastic processes from a common distribution (e.g. all transects in a region have same prob of occupancy, colonization, and extinction except for any fixed effects like forest cover and mean summer temperature).

For example

Z[i,t] ~ dbern(psi)
# make psi vary by region and covariates

as opposed to

logit(psi[i]) ~ dnorm(mu_psi, sd_psi)
Z[i,t] ~ dbern(psi[i])
djhocking commented 5 years ago

There are no years with data from all regions and so little data in any year. Maybe consider a different dataset for each region but using something in common like $p$ that will help across sites and years.