Re: [unmarked] simultaneous bird counts in three areas as dependent variables

114 views
Skip to first unread message

Andy Royle

unread,
Jul 31, 2012, 5:54:06 PM7/31/12
to unma...@googlegroups.com
hi Rhine,
 There are a number of classes of models in unmarked that should be helpful for your problem.
 I suggest starting with the N-mixture model, unmarked function "pcount" -- look at the help files, and also review the materials from some recent workshops
and also you might view the 5 hour webinar that I link to below.
regards,
andy
 
 
J. Andy Royle
Research Statistician
USGS Patuxent Wildlife Research Center
12100 Beech Forest Rd.
Laurel, MD 20708
http://profile.usgs.gov/professional/mypage.php?name=aroyle
andy_...@usgs.gov
phone: 301-497-5846
fax: 301-497-5545

Book: "Hierarchical Modeling and Inference in Ecology: The Analysis of Data from Populations, Metapopulations and Communities" by J. A. Royle and R.M. Dorazio.  

unmarked: A very useful R package for fitting certain hierarchical models using likelihood methods. Available from: http://cran.case.edu/web/packages/unmarked/index.html

A 5 hour "introduction to unmarked" Webinar can be found here: http://www.pwrc.usgs.gov/Royalvideo.cfm

-----unma...@googlegroups.com wrote: -----

To: unma...@googlegroups.com
From: Rhine <sin...@franklinpierce.edu>
Sent by: unma...@googlegroups.com
Date: 07/31/2012 09:49AM
Subject: [unmarked] simultaneous bird counts in three areas as dependent variables

Hello,

 

I'm analyzing a data set on bird counts (single species) and I'm looking for a statistical model that can include the spatial structure of the data.

 

The bird is fruit eating and we are interested in the effects of fruit availability on bird numbers. Birds were counted and fruiting was estimated for several different species of trees monthly in three different zones over the course of several years. The zones are geographically distinct but close enough that the birds can move freely between them. The birds could be in any of the zones in any month.

 

As a result, it's possible that bird counts within a given zone reflect fruit availability within that zone as well as the other two zones. For example, a low bird count in zone 1 could reflect low fruit availability in that zone and/or high fruit availability in one or both of the other two zones.

 

Does anyone know of a way to include in one statistical model bird counts from all three zones as the dependent variables and fruit estimates of trees within the three zones as the independent variables?

 

I'm also interested in suggestions for dealing with issues inherent to bird count data (issues of detection probability, though I don't think we can use occupancy modeling because we only have one bird count per zone per month) and how to handle the issue of repeated measures.

 

Many thanks for any suggestions!

 

Rhine


Rhine

unread,
Aug 1, 2012, 5:59:16 PM8/1/12
to unma...@googlegroups.com

Many Thanks Andy.

 

I reviewed the documents you suggested and have watched much of the webinar which I found quite helpful. However, I have a fundamental question that I need to find the answer to in order to figure out whether a hierarchical model is appropriate for our data.

 

Here's the key question at the moment: Is it possible to estimate detection probability when there is no replication of bird counts per sample period? I think the answer is no, but I may be missing something.

 

Our data set involves monthly bird counts over several years.  Each month, there is a count in each of three different geographic zones.

 

For example:

 

# birds

                june       july         aug         etc.

zone1    0              5              2

zone2    1              3              2

zone3    3              1              1

 

 

The birds are capable of moving between zones, so the bird counts each month are not replicates of an estimation of abundance in the zone. The number of birds in zone1 in june is the only estimate of abundance in that zone for that month.

 

I also have estimates of fruiting in each zone as covariates.

 

For this data structure, is there an unmarked function such as pcount() or pcountOpen() that will help with the issue of imperfect observation?

 

Or, as I suspect, am I restricted to using a poisson regression with glm since I do not have multiple visits per count?

 

Thanks,

Rhine

Andy Royle

unread,
Aug 1, 2012, 6:18:58 PM8/1/12
to unma...@googlegroups.com
hi Rhine,
 the pcountOpen function will fit a class of models for this type of data and also estimate encounter probability. This is the Dail-Madsen model (Biometrics 2011).
regards
andy

 
J. Andy Royle
Research Statistician
USGS Patuxent Wildlife Research Center
12100 Beech Forest Rd.
Laurel, MD 20708
http://profile.usgs.gov/professional/mypage.php?name=aroyle
andy_...@usgs.gov
phone: 301-497-5846
fax: 301-497-5545

Book: "Hierarchical Modeling and Inference in Ecology: The Analysis of Data from Populations, Metapopulations and Communities" by J. A. Royle and R.M. Dorazio.  

unmarked: A very useful R package for fitting certain hierarchical models using likelihood methods. Available from: http://cran.case.edu/web/packages/unmarked/index.html

A 5 hour "introduction to unmarked" Webinar can be found here: http://www.pwrc.usgs.gov/Royalvideo.cfm

-----unma...@googlegroups.com wrote: -----

Date: 08/01/2012 05:59PM
Subject: Re: [unmarked] simultaneous bird counts in three areas as dependent variables
Reply all
Reply to author
Forward
0 new messages