decay.model: Fitting distance decay models to pair-wise assemblage similarity

Description

Fits a negative exponential or power law function (via GLM) describing (i) the decay of assemblage similarity with spatial (or any other) distance, or, equivalently, (ii) the increase of assemblage dissimilarity with distance.

Usage

decay.model(y, x, model.type="exponential", y.type="similarities", perm=100)

Arguments

codedist object, either containing similarities or dissimilarities between pairs of assemblages.

codedist object, containing distances (spatial or other) between pairs of assemblages.

model.type

functional form of the model, either negative exponential or power law, partial match of "exponential" or "power".

y.type

polarity of the codedist object (i.e. 1 means total similarity or total dissimilarity), partial match of "similarities" or "dissimilarities".

perm

number of randomizations to assess significance.

Value

The function returns a list with:

data

dataframe containing distances (spatial or other) and similarities (or dissimilarities).

model

the fitted GLM.

model.type

functional form of the model, either negative exponential or power law.

y.type

similarties or dissimilarities.

pseudo.r.squared

similarties or dissimilarities.

a.intercept

intercept of the model, i.e. similarity or dissimilarity at distance=0.

b.slope

slope of the model, i.e. rate at which similarity decreases with distance, or dissimilarity increases with distance in a negative exponential or power law model.

p.value

significance of the model, as estimated from a randomization test.

References

Nekola, J.C. & McGill, B.J. 2014. Scale dependency in the functional form of the distance decay relationship. Ecography, 37, 309-320

G<U+00F3>mez-Rodr<U+00ED>guez, C. & Baselga, A. 2018. Variation among European beetle taxa in patterns of distance decay of similarity suggests a major role of dispersal processes. Ecography, in press

Examples

Run this code

# NOT RUN {
require(vegan)

data(BCI)
## UTM Coordinates (in metres)
UTM.EW <- rep(seq(625754, 626654, by=100), each=5)
UTM.NS <- rep(seq(1011569,  1011969, by=100), len=50)

spat.dist<-dist(data.frame(UTM.EW, UTM.NS))

dissim.BCI<-beta.pair.abund(BCI)$beta.bray.bal

plot(spat.dist, dissim.BCI, ylim=c(0,1), xlim=c(0, max(spat.dist)))

BCI.decay.exp<-decay.model(dissim.BCI, spat.dist, y.type="dissim", model.type="exp", perm=100)

BCI.decay.pow<-decay.model(dissim.BCI, spat.dist, y.type="dissim", model.type="pow", perm=100)

plot.decay(BCI.decay.exp, col=rgb(0,0,0,0.5))
plot.decay(BCI.decay.exp, col="red", remove.dots=TRUE, add=TRUE)
plot.decay(BCI.decay.pow, col="blue", remove.dots=TRUE, add=TRUE)
# }