How to calculate a cross-correlogram based on Morans I

  Hello again,

okay, i found the ncf package and the correlog function.

# correlog(x, y, z, w = NULL, increment, resamp = 100, latlon = FALSE, na.rm = 
FALSE, quiet = FALSE)

my short question is:

*if latlon = T and the positions are "degree, decimal-minutes" then increment 
has the unit of "km" and distance bins are in "km" as well - is that correct ?*

I found it in the sourcecode :

#then generating geographic distances
if(latlon){
#these are geographic distances from lat-lon coordinates
dmat <- matrix(0, nrow = n, ncol = n)
for(i in 1:(n-1)) {
for(j in (i+1):n) {
dmat[j, i] <- gcdist(x[i], y[i], x[j], y[j])
dmat[i, j] <- dmat[j, i]

with gcdist =

r <- 360/(2 * pi)
lon1 <- x1 / r
lat1 <- y1 / r
lon2 <- x2 / r
lat2 <- y2 / r
dlon <- lon2 - lon1
dlat <- lat2 - lat1
a <- (sin(dlat/2))^2 + cos(lat1) * cos(lat2) * (sin(dlon/2))^2
c <- 2 * atan2( sqrt(a), sqrt(1-a) )
return(6370 * c)
}

but i cannot the answer from this....does gcdist deliver km units ?

Cheers!
jens





Am 05.10.2010 12:57, schrieb Jens Floeter:

Dear list members,

i need a hint - i am stuck on my search for a function.

I would like to calculate cross-correlograms between two variables (predator 
and prey densities) and use this as an index of spatial overlap.

I used the spdep pakcage and use e.g.

sp.correlogram(xy.data.nb, prey, order = 5 , method = "I", zero.policy=TRUE, 
style = "W")
sp.correlogram(xy.data.nb, predator, order = 5 , method = "I", 
zero.policy=TRUE, style = "W")

to calculate the (auto-) correlograms for predator and prey individually.

But how do i combine them to calculate the cross-correlogram ?

Any hint would be highly appreciated...

Best wishes !
jens