[AniMov] Brownian Bridge home range in Adehabitat - getvolumeUD issue

[Clément Calenge]

I am not sure, without the code describing what has been done... 
However, it seems that one of the two smoothing parameters of kernelbb 
are too large...
Probably sig1, actually, given that similar peaks are located with the 
classical kernelUD.  This short simulation illustrates the behavior:

## simulates a Brownian motion:
move <- simm.brown( seq(1,1000000, length=100) )

## show the trajectory:
plot(move)

## Use a too large sig1 for the estimation
## of the brownian bridge kernel:
ku <- kernelbb(move, sig1=500, sig2=50)
image(ku[[1]]$UD)
contour(getvolumeUD(ku)[[1]]$UD, levels=95, add=TRUE)

## The problem could be solved by diminishing sig1
ku <- kernelbb(move, sig1=5, sig2=50)
image(ku[[1]]$UD)
contour(getvolumeUD(ku)[[1]]$UD, levels=95, add=TRUE)


Did you use liker to find the value of lik1?
Regards,

Clément Calenge.


Fernucht wrote:
> Greetings,
>
> My question might be very simple: when using the Brownian Bridge
> (adehabitat) method to define the home ranges, the command getvolumeUD
> (lev=95) gives me a huge area, impossible to be true. Only if I restrict to
> ~lev=35 in getvolumeUD, I get a close estimate to classic kernel method
> (kernelUD).
>
> As you know, in the Horne et al. 2007 paper, top pag.4 , they compare both
> methods, and in both figures it's represented the 99% contour, so it
> shouldn't be because of the method.  Has this happened to anyone of you?
>
> Thanks
>
> I send you an image comparing both methods:
> http://old.nabble.com/file/p26335986/example.jpg 
>   


-- 
Clément CALENGE
Cellule d'appui à l'analyse de données
Office national de la chasse et de la faune sauvage
Saint Benoist - 78610 Auffargis
tel. (33) 01.30.46.54.14