decostand: Standardization Methods
In vegsoup: Classes and Methods for Phytosociology

Description Usage Arguments Details Value Note Author(s) References See Also Examples

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`	`Vegsoup*` object.
`value`	`character`. Any method accepted by `vegan::decostand` plus the additional choice `"cap"` or `NULL`.

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]