thta0pi in spatial dependence model

[mrodtka]

Anyone else struggling with thta0pi in the spatial dependence model?

When I attempt to fix to 'eq.' as outlined in the help menu or fix it to some seemingly reasonable value (e.g. 0.3) the model fails to converge (3-4 sig digits) and returns an estimate of 0 SE 10 but when I attempt to let PRESENCE estimate the parameter it no longer returns a convergence warning but results in an estimate of 22 SE 78000!

I'm not sure if this a user issue, software issue, data issue or some combination thereof.

On a related note, should the Beta parameter in the design matrix for thta0pi delete automatically when thta0pi is fixed to eq. (it does not appear to), or should I manually be deleting it?

I'm working with PRESENCE ver 5.5

Thanks in advance.

Update: The model seems to 'work' (i.e. no convergence issues, reasonable looking Beta's and SE's) if I fix thta0pi to 0.8 or a similarly high value. In my dataset, of the 18 sites with detections (out of 93 sites surveyed total), 13 have detections occurring in the first segment. Could the relatively high proportion of early detections along each transect in my dataset be affecting the model's ability to estimate thta0pi?

Mike

[jhines]

Hi Mike,

I've corrected an error in PRESENCE where it wasn't doing the 'equilibrium' option when you entered 'eq.' as the fixed value for thta0pi. Please upload the most recent version and try again.

Note that you do need to delete the column from the 3rd design matrix (for thta0pi) when you use this option.

Jim

[mrodtka]

That worked. Thanks!