[R-sig-Geo] combining UD's in kernel denisty analysis

[Clement Calenge]

Dear Louise,



[snip]
>> data<-read.csv("three.csv")
>> library(adehabitatHR)
>> gmt<-(data$gmt)
>> data$gmt<-as.POSIXct(strptime(gmt, "%d/%m/%Y%H:%M:%S"))
>> tr=as.ltraj(xy, data$gmt, id, burst = data$Burst, typeII = TRUE, slsp ="remove")
>> lik<-liker(tr, sig2=1, rangesig1=c(10, 100))
>> hrBB<- kernelbb(tr, sig1 =  6, sig2 = 1, grid = 100, byburst=TRUE)
>> hrBB
>> image(hrBB)
>> plot(getverticeshr(hrBB, 95), add=TRUE, lwd=2)
>> plot(getverticeshr(hrBB, 50), add=TRUE, lwd=2)
> Error in plot(getverticeshr(hrBB, 50), add = TRUE, lwd = 2) :
>    error in evaluating the argument 'x' in selecting a method for function 'plot': Error in re[[i]] : subscript out of bounds

It is hard to understand this particular error without any reproducible 
example. I don't know the origin of this particular error, but there are 
other errors in your code. There are several animals/bursts in tr, so 
that if you want to display the home range contour for each animal, you 
should not use plot(…, add=TRUE) after image(hrBB). Rather, you should use:

par(mfrow=n2mfrow(length(hrBB)), mar=c(0,0,2,0))
lapply(1:length(hrBB), function(i) {
     image(hrBB[[i]], main=names(hrBB)[i])
     plot(getverticeshr(hrBB[[i]], 95), add=TRUE, lwd=2)
     plot(getverticeshr(hrBB[[i]], 50), add=TRUE, lwd=2)
})

I am not sure that this will correct your error, but as indicated, it is 
difficult to debug a code without any reproducible example.


[snip]
> # I then try to combine these 4 individual UD’s using the following code
>
> UDs<- lapply(1:length(hrBB), function(i) hrBB[[i]]$UD)

There is a confusion here. There is no component UD in the object of 
class estUD. This kind of code would have been correct with the "old" 
package adehabitat (no longer developed). The newer package adehabitatHR 
uses quite different classes of objects. You should read the vignette of 
adehabitatHR, by typing:

vignette("adehabitatHR")

Also, you should read documentations related to the classes of objects 
of class "SpatialPixelsDataFrame" used by the package sp, on which the 
package relies (the class estUD inherits from this class). Understanding 
these classes may help you to write correct code.


> UDend<- liUD[[1]]
> for (i in 2:length(liUD))
> UDend<- UDend + liUD[[i]]
> nlocs<- unlist(lapply(1:length(tr), function(i) nrow(tr[[i]])))
> wi<- nlocs/sum(nlocs)
> UDt<- UDs[[1]]
> for (i in 2:length(UDs))
> UDt<- UDt+wi[i]*UDs[[i]]
>
> # And also tried:
>
> liUD<- lapply(1:length(tr), function(i) {
> hrBB[[i]]$UD * (nrow(tr[[i]]) / sum(sapply(tr, nrow)))
> })
> UDend<- liUD[[1]]
> for (i in 2:length(liUD))
> UDend<- UDend + liUD[[i]]
>
> But get this error message:
>
> Error in image.default(UDend) : argument must be matrix-like

These errors are caused by the incorrect use of hrBB[[i]]$UD above.
This code would be correct if you had used the package adehabitat 
instead of adehabitatHR. For adehabitatHR, this code should work (not 
tested):

UDend<- hrBB[[1]]
hrBB <- estUDm2spixdf(hrBB)
nlocs <- unlist(lapply(1:length(tr), function(i) nrow(tr[[i]])))
wi <- nlocs/sum(nlocs)
UDend[[1]] <- apply(hrBB at data, 1, function(x) sum(x*nlocs))

However, be careful that the above code averages the UDs of all bursts 
and all animals. It may be sensible to modify this code so that the 
average is calculated for each animal, depending on your aim.
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