Single-season occupancy
********************
1. For the BLGR data, construct the following models.
- detection a function of log(field size) and varying over occasions, occupancy constant
-detection varying over occasions, occupancy varying over sites in relation to log field size
-- detection a function of log(field size) and varying over occasions,occupancy varying over sites in relation to log field size
-- detection constant, occupancy constant
-- detection varying over occasions, occupancy constant
2. evaluate AIC for each model
3. Generate model averaged predictions for occupancy over combinations of bqi = 0 and 1 and field sizes of 1,5,25,and 30. Use the built in predict function
#######################
Sample size -single season
-For the constant psi and p model, evaluate precision (CV= se/estimate) of the occupancy estimate based on the analysis above.
- Assuming that psi and p remain constant in a future study, indicate combinations of number of sites and number of replicates that will achieve
double precision (0.5*CV) i.e., CV is measure of variance, precision is 1/CV, so half of a given CV gives twice the precision. For example, if CV goes from 0.2 to 0.1 precision has doubled, from 0.2 to 0.05 precision has increased by a factor of 4.
####################
Multi-season occupancy
For the Grand Skink example, create the following additional models
-- extinction as a function of pasture , initial occupancy constant, detection constant, colonization constant
-- colonization as a function of pasture , initial occupancy constant, detection constant, extinction constant
-- colonization and extinction both a function of pasture , initial occupancy constant, detection constant
Include these in the model set (you will have to add them to the list) and get model averaged estimates of extinction
and colonization in relation to pasture or not pasture
-- Take estimates of parameters under the no covariates (~1) model as pilot estimates. Investigate what the implications for precision of the estimates would be of a) cutting the number of study sites in half at the same number of study years and replicates per site b) keeping the site number the same and doubling replication