Distance-based ordination
16 important questions on Distance-based ordination
What are multivariate analyses (MVA's)?
- ecological phenomena are inherently complex
- often with many response variables
Why would you not do seperate univariates analyses?
- Many independent tests: not efficient
- Multiple testing: loose statistical power
- Multicollinearity: several species may response similarly to predictors (redundancy)
What are the 3 types of ordination plots?
- Scatterplot (sites)
- Biplot (sites + species)
- Triplot (sites + species + predictors)
- Higher grades + faster learning
- Never study anything twice
- 100% sure, 100% understanding
How are euclidean distances calculated?
What are the steps of polar ordination: axis 2?
- Calculate distance between sites in terms of species composition
- Select the two most different sites: these are placed on opposite ends (poles) of axis 1
- Order the remaining sites between these poles
- Choose sites for axis 2 perpendicular to axis 1: select 2 other sites which are close together on axis 1, but very dissimilar
- Order all sites between these poles on axis 2
What is an ordination diagram?
What is a short gradient?
What is the abundace paradox?
Two locations sharing all species could appear more different (thus higher distance) than two locations which do not share any species but which have low abundances!
What are the solutions for double zeros?
- log(1+x) (reducing the influence of large values)
- sqrt(x) (reducing the influence of large values)
- ‘Hellinger’ (computing sqrt of relative abundances per location)
And/or use an asymmetric distance measure
- measures that treat double zeros and double presences asymmetrically
What is Hellinger distance?
- less senstive to abundance paradox and double-zero problem
What are the disadvantages of Polar ordination?
- Heuristic rules rather than a formal model
- Axes are not statistically independent by design
- Different people will end up with different ordinations
What is the Principal Coordinates Analysis (PCoA)?
- Based on robust mathematics
- Axes are statistically independent by design
- The same solution every time (on the same dataset)
What is the aim of nonmetric multidimensional scaling?
Why can/will multiple people performing polar ordination on the same dataset arrive at different ordinations, especially if the dataset becomes larger and more complex?
What can the Manhattan distance on presence-absence data be interpreted as?
What are Hellinger distances?
The question on the page originate from the summary of the following study material:
- A unique study and practice tool
- Never study anything twice again
- Get the grades you hope for
- 100% sure, 100% understanding

















