Classification of communities

1
Classification of communities
And now, for something completely different....
2
Classification of communities

• Data collection (sampling) covered in lab
  lectures
• Now, how use data to decide which stands belong
  to the same association?
• Recall.
• Association more formal and precise unit
• basic unit of plant community classification
• composed of many stands (stand a particular
  member of an association)
• taxonomic analogy association is species, stands
  are individuals

3
Classification of communities

• Data collection (sampling) covered in lab
  lectures
• Now, how use data to decide which stands belong
  to the same association?
• Recall.
• Stands comprising association have relatively
• 1) consistent floristic composition
• 2) uniform physiognomy (see notes later.)
• 3) consistent habitat

4
Classification of communities

• Data collection (sampling) covered in lab
  lectures
• Now, how use data to decide which stands belong
  to the same association?
• 1) Tabular methods
• 2) Cluster analysis
• 3) Association analysis
• 4) Ordination methods

5
Tabular methods

• American approach often classification based on
  dominance
• Express dominance by
• Relative cover, density, basal area, biomass
• Importance values

6
Tabular methods

• May classify association based on dominants in
  layers
• Ex, Pinus ponderosa/Agropyron spicatum woodland
• Dominant tree and dominant grass both listed
• Samples often placed with some degree of
  randomness (not very subjective)

7
Tabular methods

• European approach often classification based on
  entire flora
• Braun-Blanquet technique (relevé technique)
• Based on sample called a relevé
• 1) Get familiar with vegetation of area
• 2) Choose representative stand (very subjective)
• 3) Compile species list
• 4) Do species-area curve to determine minimum
  quadrat size
• 5) Place one minimal area quadrat (relevé)
  subjectively in stand
• 6) Collect data on species abundances and
  distribution

8
Tabular methods

• Classification use table method. Start with raw
  data table
• Sort through species and discard widespread and
  rare ones, also shuffle columns to make
  differentiated table

Cover, sociability (dispersion)
9
Tabular methods

• Differentiated table has differential species
  those characteristic of similar types of stands
• Have high fidelity few stands outside of
  association have them
• Have high constancy most stands w/in association
  have them

10
Tabular methods

• Define associations based upon those differential
  species
• Note
• 1) method very subjective (esp. relevé technique)
• 2) differential species are often NOT dominants
  in stands
• Example, may define association based on moss
  species on tree trunks

Olympic National Park, Washington
11
Cluster analysis

• Expresses similarity of stands graphically (in
  two dimensions)
• How express similarity? Coefficient of Community
  (CC)
• Mathematically shows how close stand composition
  is between two samples

12
Cluster analysis

• Two major indices
• Jaccards Index
• Sorensens Index

Picards Index
13
Cluster analysis

• Two major indices
• Jaccards Index
• Sorensens Index
• Each can use presence data or weight that by
  cover values

14
Cluster analysis

• Note possible values range from 100 (complete
  identity) to 0 (no species in common)

15
Cluster analysis

• Generate dendrogram
• Note Y axis represents resolving power 40 is
  great power and 0 is no power to distinguish
  differences between stands

16
Cluster analysis

• Decide level of similarity to use for defining
  association
• CC of 10 (threshold III) 2 associations
• CC of 20 (threshold II) 7 associations
• CC of 30 (threshold I) 15 associations

17
Cluster analysis

• Note this technique uses all data from stand to
  decide similarity Polythetic technique
• Also is divisive technique starts with all
  stands lumped together and splits them into
  smaller groups

18
Cluster analysis

• Note this technique uses all data from stand to
  decide similarity Polythetic technique
• Also is divisive technique starts with all
  stands lumped together and splits them into
  smaller groups
• Result associations based on floristic
  similarity. Dominant species fairly influential
  in defining associations.

19
Association analysis

• Also divisive technique
• But is monothetic

20
Association analysis

• Also is divisive technique
• But is monothetic uses presence or absence of
  single species to assign stands into groups
• Like table method uses differential species, but
  these selected based on their influence upon
  other species
• How document influence? Use contingency table
  analysis

21
Association analysis

• How document influence? Use contingency table
  analysis
• Calculate chi-square for every pair of species in
  all samples

22
Association analysis

• Will get matrix of chi-square values
• Species
• Species A B C D
• A X 57 12
  23
• B 57 X 17
  3
• C 12 17 X
  1
• D 23 3
  1 X
• Total 92 77 30 27

23
Association analysis

• Species
• Species A B C D
• A X 57 12
  23
• B 57 X 17
  3
• C 12 17 X
  1
• D 23 3
  1 X
• Total 92 77 30 27
• Select species with highest chi-square total
• Split stands into 2 groups with A and without A

24
Association analysis

• Then repeat procedure on each group
• One group will still have species A present
• Other group will lack species A
• Continue to divide groups until total chi-square
  values become less than some threshold number

25
Association analysis

• Example, 70 quadrats placed in salt marsh
• Association analysis used to define associations
  from this sample
• Procedure used until total chi-square of 7
  achieved

26
Association analysis

• Figure showing result (70 quadrats in salt marsh)
• Stands divided into groups (8) depending on
  presence/absence of particular species (32, 38
  is group 1, 32, -38 is group 2 etc.)
• Each group represents an association

27
Association analysis

• Note associations defined on presence/absence of
  most ecologically dominant (influential) species

28
Ordination Methods

• Distill data into graphical representation
• Stands mapped in ordination space. May be
  multidimensional

29
Ordination Methods

• Distill data into graphical representation
• Stands mapped in ordination space. May be
  multidimensional
• Generally, are polythetic techniques (use data
  from all species)
• Usually classify stands based on dominant species

30
Ordination Methods

• Can provide information about variables that
  correlate with axes and thus explain why stands
  located where they are in ordination space