region.N with shared buffer

[ndusit]

Hi Murray
I did estimated bear density of three study sites in one large national park. Mask.check was tested for appropriated buffer. However, the buffer was large and overlap among three sites. When I estimated abundance from region.N(), it does three estimated. I'm not sure that the total abundance is sum from three sites or it is wrong where buffer overlap across sites. I never recapture same individual bear among sites.

Thanks in advance,
Dusit


region.N(fit.site, region = ABB.mask)
$KET.ABB
estimate SE.estimate lcl ucl n
E.N 43.53288 13.87128 23.66579 80.07808 13
R.N 43.53288 12.20162 27.35910 77.92442 13

$KSP.ABB
estimate SE.estimate lcl ucl n
E.N 43.15794 16.91488 20.57283 90.53728 10
R.N 43.15794 15.58702 23.81176 89.60238 10

$KKP.ABB
estimate SE.estimate lcl ucl n
E.N 213.9764 61.63540 123.0406 372.1201 31
R.N 213.9764 59.87442 128.9295 372.8825 31

[murray.efford]

Hi Dusit
Good question! I think the answer depends on ABB.mask. If that is a list of three different masks, one for each of the sites, then for each session (site) the estimated N is the number in the buffered area (mask) for that site, whether or not it overlaps with other buffered areas. Of course, the sum may then be more than the actual number in the total (union) of the buffered areas.
What is the number that you want? If density varies by habitat within the park, and you want to both allow for this when estimating the total number in the park, I think you need to do this: start out by building one mask for the whole park with a covariate to indicate the different habitat areas; then read all your data as a single session, fit a model D~habclass (or whatever), and use region.N with that big (park) mask as 'region'. If a lot of your large mask is outside the buffered study areas then it is efficient to use a trimmed version of the big mask during fitting. You can make that using some code like this

Code: Select all

smallmask <- subset(bigmask, distancetotrap(bigmask, allcamerasites) < 5000)

for a 5-km buffer.
Murray

[ndusit]

Hi Murray
Thanks, now it look good. But, D~1 and D~x+y are identical.
I will try to work with habitat covariate in mask next step.

Best,
Dusit