| From: | "L. Jay Roberts" <lja...@gmail.com> |
| To: | unma...@googlegroups.com |
| Date: | 09/11/2012 01:58 PM |
| Subject: | [unmarked] advice for dataset with no recaptures |
| Sent by: | unma...@googlegroups.com |
We have a dataset of Yellow-billed Cuckoo surveys (4 visits at 1400 points in riparian habitat), but a seemingly fatal issue is that we have very few detections and zero repeat detections. We'd like to run a standard single season occupancy model to get a rough estimate of the number of territories in our study area. So that the occupancy metric is on scale with territory size we planned on aggregating adjacent point count locations at a range of sizes approximating the territory size of the species.
YBCU is an episodic and cryptic species (may or may not call when close to field observers). There were a total of 9 detections. But even with the largest aggregation of point count locations that we identified a priori, none of the sampling units had any repeat detections. The detections are just too far apart and we feel they are in fact different individuals so it would not really be appropriate to aggregate at that scale. Can we still fit an occupancy model without any repeat detections?
One possibility in these data is that all nine detections were from birds not associated with territories, the could simply all be 'drifters' that are un-paired and not breeding?