Error Propagation When Averaging N-mix Predictions

24 views
Skip to first unread message

Tyler Hodges

unread,
Aug 9, 2024, 10:10:57 AMAug 9
to unmarked
Hello everyone,

I am currently predicting abundances across four sites with an N-mixture model calibrated with two raster covariates. At each of these sites, I plotted a large number of points (~10,000), extracted the raster values, and then produced an expected abundance prediction at each point. I now wish to average these predictions within sites. However, I am unsure of the most statistically rigorous way to do this to ensure the SEs propagate appropriately, and unfortunately, none of the discussions I've found online have been very helpful. I was hoping that this group could provide some good insight and discussion (as it always does).

Thanks in advance for your help, and I'm looking forward to your suggestions.

Best,
Tyler

Tyler Hodges

unread,
Aug 21, 2024, 1:35:29 PMAug 21
to unmarked
Good afternoon, everyone,

I just wanted to provide an update on this in case anyone is following along or runs into the same problem. I was able to find a bit of nifty code and instruction in AHM v.1 (pg. 275-277) that shows how to use a parametric bootstrap to solve just this issue, and that is ultimately what I decided to use.

Thanks!

Best,
Tyler

Reply all
Reply to author
Forward
0 new messages