polar.ord {asbio}R Documentation

Polar ordinations

Description

The function currently creates two or three dimensional Bray-Curtis (polar) ordinations.

Usage


polar.ord(data, index = "steinhaus", endpoint = c("BC.original","PC_ORD.original", "var.reg"), get.resid.dist = FALSE)

Arguments

data A multivariate dataset, e.g. a community site x species matrix.
index The dissimilarity or distance measure to be used. Any method from get.dist can be used.
endpoint Endpoint selection is accomplished using either the Bray-Curtis original "BC.original" method (i.e. using the sites which are furthest apart), the PC-ORD original method "PC_ORD.original" (the first endpoint has the highest sum of distances from other sites, the second endpoint has the highest distance from the first endpoint), or using the variance regression "var.reg" method as described by McCune and Grace (2002). Note that the variance regression method in PC-ORD does not appear to be doing what it is supposed to be doing. The variance regression method in polar.ord is in accordance with in McCune and Grace (2002).
get.resid.dist Logical. Allows one to see the residual distance matrices.

Details

The Bray-Curtis method is a relatively easy to understand ordination procedure that is also known as polar ordination because it arranges points in reference to endpoints or poles. The method is strongly favored by ecologists from the University of Wisconsin (probably because this is the school John Curtis graduated from and taught at). Curtis developed this method with the Canadian ecologist James Bray in the early 50s.

While polar ordination seems to work reasonably well for recovering community patterns, it has recently been very difficult to publish papers using this technique for two reasons: 1) it has been judged "outmoded" compared to more recently developed matrix decompostion methods, i.e. DCA and CCA, and 2) the endpoint selection techniqes for axes can be arbitrary (more on this later). Edward Beals (1984) wrote a scathing critique of ecologists who undersold polar ordination. Here is an excerpt:

"While ordination as an approach to data analysis gained acceptance in the 1960's, the Bray-Curtis method came under attack beginning with Austin and Orloci (1966), and it quickly fell into disfavor among ecologists as new methods of ordination were introduced and championed. Only ecologists trained at the University of Wisconsin persisted in using Bray-Curtis, not out of blind loyalty, but because it generally gave more ecologically interpretable results than did newer or more sophisticated methods"

Endpoint selection is the most crucial step in the polar ordination process, because all other points will be placed in relation to the endpoints. The original Bray-Curtis method used the two most dissimilar points as endpoints. Two other methods, "PC_ORD.original" and "var.reg" are also allowed by polar.ord.

Value

Ouput includes scores, the amount of variance explained by axes, and, if requested, the residual distance matrices.

Author(s)

Ken Aho

References

Beals, E. W. 1984. Bray-Curtis ordination: an effective strategy for analysis of multivariate ecological data. Advances in Ecological Research. 14: 1-55.

McCune, B., and J.B. Grace. 2002. Analysis of ecological communities. MjM Software design. Gelenden Beach OR.

McCune, B. and E. W. Beals. 1993. History of the development Bray-Curtis ordination. Pp. 67-72 in J. S. Fralish, R. P. McIntosh, and O. L. Loucks eds. John Curtis: Fifty Years of Wisconsin Plant Ecology. Wisconsin Academy of Science, Arts and Letters, Madison WI.

See Also

get.dist

Examples

demodat<-matrix(ncol=3,nrow=5,data=c(2,3,5,7,9,1,4,0,6,2,5,5,10,2,2),
byrow=FALSE)
po.orig <-polar.ord(demodat,endpoint="BC.original",get.resid.dist=FALSE)

[Package asbio version 0.1 Index]