Multiple sites but not all sites sampled in all years

[Fish_Boy]

OK, the subject outlines the problem. I have four locations that were sampled over 4 years, with the exception of one site that was skipped in one year. I need to estimate N for each location. Can I setup a robust design encounter with all the data using location as a covariate and still get an N estimate for each location? or do I have to generate multiple encounters, one for each location? There is some movement among locations across years, but it is overall rare.