Home / Summaries / Class notes - Ecological Methods I / species-abundance-pca

Eigenanalysis-based ordination

Q: When do you use euclidean distances?

Any type of data (if same units, else scale first) Best measure when using data other than species abundances For species abundances: be careful with many zeros , and with large values

Q: When do you use Chi-square distances?

For species counts or abundances , be careful with r are species that are not well sampled.

Q: When do you use Bray-Curtis distances?

For species counts, abundances or compositional data, be careful with large values.

Q: What is unconstrained ordination?

Simplifying complex datasets into a low-dimensional representation.

Q: What is the PCA: Eigenvalue?

Variance. Variance of each axis = SS/(n-1) = Eigenvalue of that axis

Q: What does the projections of samples onto the species vectors give?

An indication of species abundance .

Q: What does the projections of samples onto the species vectors give?

An indication of species abundance .

Q: Why are CA's not sensitive to the abundance paradox and double-zero problem?

Chi-square distances are weighted by site totals and 1/species total. However, rare species have a large weight in the ordination!

Q: Why are species plotted as points (and not arrows as in PCA)?

Abundance decreases away from the optimum of the species (in any direction). unimodal response: thus suitable for longgradients

Q: Why are species plotted as points (and not arrows as in PCA)?

Abundance decreases away from the optimum of the species (in any direction). unimodal response: thus suitable for longgradients

15 important questions on Eigenanalysis-based ordination

When do you use euclidean distances?

Any type of data (if same units, else scale first)
Best measure when using data other than species abundances
For species abundances: be careful with many zeros, and with large values