Neighboring Sites on Presence 4.9

announcements (new versions, changes, bugs, installation problems...) related to program PRESENCE

Neighboring Sites on Presence 4.9

Postby sam.garvin » Thu Sep 27, 2012 9:45 am

I'm working on Presence 4.9. I have observations for a bird species at 70 points. I put together a neighboring points file, and selected that text file for the "site neighbor file" from the main menu. Now how do I apply the neighboring points to the group presence analysis? The program still says the group 1, constant p model has already been run.
sam.garvin
 
Posts: 2
Joined: Wed Sep 19, 2012 11:01 am

Re: Neighboring Sites on Presence 4.9

Postby jhines » Mon Oct 01, 2012 7:52 am

The site-neighbor file is intended to be used for a special case in the multi-season model. It allows you to model colonization or extinction in one season as a function of occupancy in neighboring sites in the previous season. This file is not needed for the single-season analysis.

If you're trying to get occupancy estimates for groups of sites, then the thing to do is build indicator covariates which indicate which group each site belongs to. For example, if the first 10 sites are all 'urban' sites and the rest are 'rural' sites, then you would create a site covariate named 'urban' which contains a '1' if the site is a urban site, and '0' if it is rural. Then build a 'custom' model with an intercept ('1' in the design matrix) and urban ('urban' in design matrix) effect. There is an example of this sort of model in the help file.

Jim
jhines
 
Posts: 599
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: Neighboring Sites on Presence 4.9

Postby sam.garvin » Wed Oct 03, 2012 9:32 am

Thank you for your response! In my study, I am assuming that there is no significant difference between the types of sites ie. urban/rural, but I do want to account for the distance between sites. So is there no way to account for this proximity of sites in a single season model? If sites are neighbors, then the detection probability for the species may be over-estimated as a whole.
sam.garvin
 
Posts: 2
Joined: Wed Sep 19, 2012 11:01 am

Re: Neighboring Sites on Presence 4.9

Postby bacollier » Wed Oct 03, 2012 9:55 am

sam.garvin wrote:Thank you for your response! In my study, I am assuming that there is no significant difference between the types of sites ie. urban/rural, but I do want to account for the distance between sites. So is there no way to account for this proximity of sites in a single season model? If sites are neighbors, then the detection probability for the species may be over-estimated as a whole.



Sam,
If your looking for a method that incorporates distance between sites into the occupancy estimation procedure, I worked up some WinBUGS code that relies on penalized splines which works for a single season occupancy model that treats the matrix of distances between and among all sites and a mesh of fixed locations as a random effects term for space. I used the spatial process on the occurrence predictions, but using it on the detection prediction would be fairly straightforward.

The paper is: Collier et al. 2012: Diversity and Distributions 18: 158-167;

you can download the WinBUGS code example from
http://irnr.tamu.edu/bret/BretWebSiteDocs/WinBUGSCode.txt

bret
bacollier
 
Posts: 230
Joined: Fri Nov 26, 2004 10:33 am
Location: Louisiana State University


Return to software problems/news

Who is online

Users browsing this forum: No registered users and 8 guests

cron