weightedCor: Weighted correlation test of WA reconstruction

In analogue: Analogue and Weighted Averaging Methods for Palaeoecology

Weighted correlation test of WA reconstruction

Description

Weighted correlation between WA optima from training set and axis 1 scores of constrained ordination fitted to fossil data with WA model predictions for fossil samples as constraints.

Usage

## Default S3 method:
weightedCor(x, env, fossil, method = c("rda", "cca"),
            test = TRUE, type = c("simulate", "permute"), sim = 999,
            verbose = TRUE, ...)

## S3 method for class 'weightedCor'
plot(x,
     type = c("bubble", "null"),
     weighted = TRUE,
     size = 0.25,
     xlab = paste(x$env, "WA Optima"),
     ylab = "Axis 1 Score",
     xlim,
     main = "",
     sub = NULL,
     border = "gray75",
     col = "gray75",
     obscol = "red",
     fg = "black", ...)

Arguments

x	training set covariates, a matrix-like object usually of species/proxy data. For the plot method, an object of class "weightedCor", the result of a call to weightedCor.
env	training set response, a vector usually of environmental data.
fossil	matrix of fossil/core species/proxy data for which a reconstruction is sought.
method	constrained ordination method. One of "rda" and "cca". Currently only "rda" is supported.
test	logical; should the observed correlation be tested?
type	the type of test to apply. One of "simulate" or "permute". The latter is currently not implemented. For the plot method, the type of plot to produce.
sim	numeric; number of simulations or permutations to permform as part of the test
verbose	logical; should the progress of the test be shown via a progress bar?
...	arguments passed to other methods. In the case of the plot method, additional graphical parameters can be supplied.
weighted	logical; should the null distribution plotted be of the weighted or normal correlation.
size	numeric; the size of the largest bubble in inches. See symbols and argument inches for details.
xlim, xlab, ylab, main, sub	graphical parameters with their usual meaning.
border, col	The border and fill colours for the histogram bars.
fg	The colour of the bubbles drawn on the bubble plot.
obscol	The colour of the indicator for the observed correlation.

Value

The plot method produces a plot on the current device. weightedCor() returns a list with the following components:

wtdCorrel, Correl	numeric; the observed weighted and standard correlation.
data	data frame; containing the training set WA Optima, axis 1 species scores, and mean abundance for each species.
ord	the fitted constrained ordination.
model	the fitted WA model.
method	the ordination method used.
ndist	the null distribution produced. NULL if argument test was FALSE.
sim	numeric; the number of simulations or permutations used to test the observed correlations.
type	the type of test performed.
env	the deparsed version of env argument. Used for plotting.
call	the matched function call.

Author(s)

Gavin L. Simpson

References

Telford R.J. and Birks, H.J.B. (2011) A novel method for assessing the statistical significance of quantitative reconstructions inferred from biotic assemblages. Quanternary Science Reviews 30:1272-1278.

See Also

wa for details on fitting weighted average models.

Examples

data(ImbrieKipp, SumSST, V12.122)

Cor <- weightedCor(ImbrieKipp, env = SumSST,
                   fossil = V12.122, type = "simulate", sim = 49)
Cor

plot(Cor)
plot(Cor, type = "null")

analogue documentation built on Sept. 30, 2024, 9:41 a.m.