decostand {vegsoup} | R Documentation |
This method provides the standardizations implemented in function
decostand
of vegan plus one more.
Only the species matrix will be affected by standardization.
## S4 method for signature 'Vegsoup' decostand(x) ## S4 replacement method for signature 'Vegsoup' decostand(x) <- value
x |
|
value |
|
Standardization is often necessary before the species matrix is subject to
analysis by a multivariate method. This can easily be achieved by setting
slot 'decostand' using the method of the same name.
The standardization method will be applied when ever any method for that object
requests a species matrix (see as.matrix
). It is possible to
supply more than one method and these are applied sequentially to the data
(order matters). In fact, method wisconsion
invokes a double
standardization (max
followed by total
) as it is the case in
vegan. Currently only the method
argument of
vegan::decostand
is implemented, whereas argument MARGIN
is not.
Method cap
calculates cumulative abundance profiles (CAP, De
Cáceres et al. 2013). Currently any additional decostand
method will be ignored for method cap
. This standardization is
especially suited for forest vegetation when vertical stand structure
is available (vegetation layers).
decostand(x)
returns a character vector depending on the number of methods.
decostand(x) <- value
sets the standardization method(s).
If the decostand method is redefined for an object of class VegsoupPartion the object will be subject to revaluation as to honor the newly assigned decostand method. In this way it is possible to compare different standardizations with the same partitioning method.
Manipulation of the sites data has to be performed by modifying elements of
slot 'sites', which is represented as data.frame
. Use the
extract and replace operators "$"
and
names<-
method to modify columns of the sites data. See
Extract
for details.
Roland Kaiser
De Cáceres, M., Legendre, P., He, F. 2013 Dissimilarity measurements and the size structure of ecological communities. Methods in Ecology and Evolution 4, 1167–1177
as.matrix
,
as.dist
,
vegdist
coverscale
data(barmstein) x <- barmstein # no standardization by default decostand(x) # standardization divide by margin total decostand(x) <- "total" # retrieve a standardized species matrix range(as.numeric(x)) # double standardization decostand(x) <- c("hellinger", "standardize") range(as.numeric(x)) # remove any standardization decostand(x) <- NULL # replace decostand method for VegsoupPartition object # and recompute partitioning x <- barmstein decostand(x) <- "pa" x <- VegsoupPartition(x, k = 2) # revaluation decostand(x) <- NULL # revaluation again decostand(x) <- "wisconsin" # resetting with the same value invokes no recomputation decostand(x) <- "wisconsin" # cumulative abundance profiles (CAP) data(windsfeld) x <- windsfeld # subset forest plots x <- x[x$tcov1 != 0 | x$tcov2 != 0, ] # order layers layers(x) layers(x) <- layers(x)[c(2,1,3,5,4)] # select woody species j <- c("lari.deci", "pinu.mugo", "sorb.aucu", "loni.caer.caer", "sali.appe", "pice.abie") j <- unique(unlist(sapply(j, grep, colnames(x)))) x <- x[, j] taxon(x) # raw matrix m1 <- as.numeric(x) # cumulative abundance profiles decostand(x) <- c("cap") m2 <- as.numeric(x) m1[,grep("lari.deci", colnames(m1)), drop = FALSE] m2[,grep("lari.deci", colnames(m2)), drop = FALSE]