by jhines » Sat Sep 26, 2015 1:17 pm
Hi Lea,
Site number 3 has missing data for all surveys except survey 33, which contains a '0'. A couple of the site covariates have missing data for site 3, so the program is trying to warn you that a model which estimates occupancy or detection as a function of that covariate won't have a covariate value for that site. If all surveys for that site contained missing data, it wouldn't be a problem. So, you won't be able to use that covariate, unless you can enter a value for that site. The same holds true for a few of your survey covariates (you have presence/absence data for a site/survey, but no covariate value for the site/survey).
Some folks will replace the missing covariate data with the mean or median value of the other sites. Others will just delete that site. It's up to you. If you ignore the warning and run a model which uses those covariates, PRESENCE will plug in a very large negative number for the covariate.
Models which use no covariates, or covariates which don't containe missing data will be unaffected by this warning.
Jim