[R-sig-eco] adeHabitatLT trajectory partitioning

[Clement Calenge]

Dear Dylan,


[snip]
> In the gus example 10 models are generated with a mean dist from 0 to
> 130km. The max step length for gus is ~117km /day so I assumed the 130km
> is a rounded up approximation of this, to cover the whole spectrum of
> observed step lengths. The parameter that is confusing me is the 5km
> standard deviation of the models. The vignette says this was chosen
> after a visual exploration of the distribution of distances for gus, but
> I don't see it. Does this value perhaps come from the standard error of
> the distances, ~3.6km?

Actually, the value of 5 km was determined very, very, roughly on the 
basis of the graph displayed on page 43 of the vignette. If you look at 
the parts of the graph where the distances are stable (e.g. during 
March), you can see that, the observed distances vary approximately +/- 
5km both above and below the mean. This is a very very approximate 
value, chosen visually just to illustrate the approach; but I agree that 
there are probably much better ways to define the models used to 
partition the trajectory. (actually, I prefer the approach of Lavielle 
when I have to segment a trajectory for exploration purposes -- it 
relies far less on the correct specification of a model).
HTH,

Clément Calenge.

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