by jhines » Mon Sep 23, 2019 1:15 pm
Hi,
With only 1 survey per season, it's impossible to estimate occupancy and survey-specific detection probabilities as you have more parameters than data. If you assume detection probability constant over surveys, then you can get estimates of occupancy/colonization/extinction and detection. This is a pretty strong assumption and if not true, could lead to biased estimates of occupancy/col/ext, and/or incorrect model effects on those parameters. A slightly better model would assume detection varies by a survey covariate like "effort" (or anything which might be related to detection). This is still a strong assumption which would require justification if used.
My point is that it's possible to do something with this type of data, but it has the potential to give very misleading results. Having replicate surveys each season eliminates the need to make those strong assumptions.
Jim H.