Is vegan::cca() treating each possible value of a categorical environmental variable as a separate factor?

user_16672975 0
Viewed 47

I am applying canonical correspondence analysis to a dataset of herbaceous plant species presence in 48 quadrats arranged across a large forested area, where environmental variables for each quadrat include both continuous (% cover of tree canopy above the quadrat) and categorical (microtopography, with possible values of bowl, slope, ridge, and level) variables. When plotting the CCA, I notice that all possible values of the categorical variable are plotted, while only the name of the single continuous variable is plotted (figure below). Does this indicate that function cca() is treating each possible value of the categorical variable as a distinct factor, rather than levels of a single factor?

cca1 <- cca(formula = quadrats_df ~ canopy_cover + microtopography, data = sitedesc_df, scale = -3)

plot(cca1, type = "text", display = c("sites","cn"))

Where quadrats_df contains quadrat-level vegetation data, and sitedesc_df contains quadrat-level environmental data.

Plotting cca1 yields the figure below.

Canonical correspondence analysis of vegetation data including continuous and categorical variables as factors

vegan r
1 Answers

Answer - no, there is no error and cca() is correctly treating possible values of microtopography as levels of a single factor.

From ter Braak and Verdonschot 1995:

  • "A CCA ordination diagram may consist of the following elements: points for species, sites, and classes of qualitative environmental variables, and arrows for quantitative environmental variables" (page 265).
  • "A qualitative environmental variable consists of a number of classes that partition the sites; each class is naturally represented by a point on the diagram, namely the centroid of the sites points belonging to the class" (page 268).

Reference: Ter Braak, C.J. and Verdonschot, P.F., 1995. Canonical correspondence analysis and related multivariate methods in aquatic ecology. Aquatic sciences, 57(3), pp.255-289.

user_166729750
Related