Sightability models
For the moose example, fit additional models:
-- quadratic effect of groups size alone
--quadratic effect of groups size + vegetation cover
Compare these with the model with vegetation cover alone using AIC
###############
Distance sampling
-For the rabbit example, try some other key functions besides half normal (e.g., uniform, exponential) and compare in terms of model fit and AIC
-Compute model averaged estimates by habitat (poor=1, good=1) . Use the built in function modSel, for example:
#built in model selection
fL <- fitList(hn=mod1, hn_hab=mod2, haz=mod3,haz_hab=mod4)
ms<-modSel(fL)
predict(fL,type="state")
-- take the estimated CV for density (se/ estimate) as a starting estimate (it is around 0.15, regardless of habitat) . Using the formula (9.4) in Conroy and Carroll, find the total transect line lengths you would need to reduce this value to 0.10, and to 0.05.