Surveys with no detections

[dbuhl]

I have three years of data with three surveys each year. I am trying to run a multi-season model. However, in 3 of the surveys (2 the first year and 1 the third year) there were no detections. So three of my columns in the presence/absence sheet are all zeros. I am still having some problems with convergence and am getting some standard errors equal to zero or very large standard errors.

My question is, is it a problem to have surveys with no detections and could this be the cause of myconvergence issues and poor standard error estimates?

Thanks for your assistance.

[jhines]

The model with detection constant across years and surveys (p(.)) should converge. I'd start with that model, then work on more complicated models, like p(year), or p(survey) or p(survey*year). The p(year) model will require an additional constraint as you only have one survey with detections that year. I'd suggest constraining that parameter (p(yr1,survey1)) equal to one of the other p's, perhaps p(yr2,survey1). This needs to be done for either the p(year) or p(year*survey) models. Additionally, I'd constrain p(yr1,survey2)=p(yr1,survey3)=p(yr3,survey3)=0 if those are the years/surveys with no detections.

It takes at least 2 surveys with detections per season in order to estimate occupancy or detection probabilities. Since the 3rd year has two seasons with detections, these models should converge.

If you have covariates, you won't be able to model the surveys with no detections as a function of the covariates as there is no variation in detection to model. If there is very little variation in the covariates, or sparse data, models with covariates might not converge.

I hope this helps.

Jim