Lesson 8: Vegetation classification
•
•
Review of ordination vs. classification
Approaches to classification
–
Physiognomic, Dominance types, Synusial approaches,
Total flora, Numerical approaches
•
Braun-Blanquet approach
• Numerical approaches (dendrograms)
Classification vs. ordination
CLASSIFICATION
(1) Used for defining discrete vegetation types. Gives us a first order information about
vegetation, i.e. the primary qualitative differences between plant communities.
(2) Necessary for mapping vegetation
(3) Useful for ordering the variation we see in vegetation and for:
(a) communication,
(b) comparison of vegetation between regions
(c) habitat evaluation
(4) Works with homogeneous communities (homogeneous in all layers of the plant canopy).
(5) High sampling intensity because replicate samples are needed in all units.
(6) Species presence or absence is considered more important than minor variations in quantity.
Sampling is primarily QUALITATIVE.
ORDINATION
(1) Useful for examining the causes of vegetation patterns. Gives us second order information
regarding the fine structure within vegetation classes by sampling the variation throughout
each vegetation type.
(2) Based on ordering the vegetation samples according to their similarity to each other.
(3) Evolved primarily from timber survey methods. Often ignores details in pattern and
composition of undergrowth species during sampling.
(4) If sampling is random or systematic, communities can be large and heterogeneous in
undergrowth and habitat.
(5) Sampling is primarily QUANTITATIVE. Variation in species quantities is as important as
presence or absence of species.
(6) It is generally possible to use ordination techniques with samples used for classification.
However, samples collected for ordination can be used for classification only if they are
floristically complete and from homogeneous sample sites.
(7) Two primary approaches:
Direct ordination examines vegetation changes along known environmental gradients (e.g. a moisture
gradient or elevation gradient.
Indirect ordination examines the environmental causes of vegetation patterns, through correlation of
ordination axes with environmental variables.
.
Some common classification approaches
•
•
•
•
•
Physiognomic approaches (e.g., Raunkaier, Dansereau, Küchler, Beard,
PFTs)
Dominance types (Whittaker, Daubenmire, US Forest Service)
Synusial approaches (Gams, Du Reitz, Lippma)
Total flora (Association approach of Braun-Blanquet, Nature Conservancy)
Numerical approaches (Greig-Smith, Goodall, Curtis, McIntosh)
Physiognomic approach
• Based on the dominant growth forms that create the general
appearance of the vegetation.
• Useful at higher levels of classification and can contain many
plant associations or community types.
• The Nature Conservancy uses a physiognomic approach at
higher levels.
Example of a physiognomic legend
from the Circumpolar Arctic Vegetation Map (CAVM)
CAVM Team (2003)
CAVM Mapping Units:
The 400 plant community types were grouped into five broad physiognomic categories:
B – Barrens
G – Graminoid Tundras
P – Prostrate-shrub Tundras
S – Erect-shrub Tundras,
W – Wetlands.
And then subdivided into 15 major mapping units based primarily on dominant plant functional types.
Barrens
B1. Cryptogam, cushion-forb barrens
B2. Cryptogam-barren complexes
B3. Noncarbonate mountain complexes
B4 Carbonate mountain complexes
Graminoid tundras
G1. Grass, rush, forb tundra
G2. Graminoid, prostrate-shrub, forb tundra
G3. Non-tussock sedge, dwarf-shrub, moss tundra
G4. Tussock-sedge, dwarf-shrub tundra
Prostrate-shrub tundras
P1. Prostrate dwarf-shrub, herb tundra
P2. Hemiprostrate/Prostrate dwarf-shrub tundra
Low-shrub tundras
S1. Erect dwarf-shrub tundra
S2. Low-shrub tundra
Wetlands
W1. Sedge, grass, moss mire
W2. Sedge, moss, dwarf-shrub mire
W3. Sedge, moss, low-shrub mire
Barrens:
B1 – Cryptogam, herb barren
Eskimonaesset, North Greenland, C. Bay
Subzone A
B2 – Cryptogam barren complex (bedrock)
Daring Lake vicinity, Canada. D.A. Walker
Mostly shield areas in
subzones C and D.
Graminoid-dominated tundras:
High Arctic: mineral soils
Low Arctic: peaty soils
Subzone D
Subzones A and B
G1 – Rush/grass, cryptogam tundra
Amund Ringnes Island, Canad a. D.A. Walker
Subzone C
G2 – Graminoid, prostrate dwarf-shrub, forb tundra
Eureka vicinity, Ellesmere Island, Canada. D.A. Walker
G3 – Non-tussock sedge, dwarf-shrub, moss tundra
Ambarchik, Yakutia, Russia, D.A. Walker
Subzone E
G4 – Tussock sedge, dwarf-shrub, moss tundra
Imnavait Creek, Alaska. D. A. Walker
Shrub-dominated tundras
High Arctic: prostrate-shrubs
Subzone B and C (nonacidic)
P1 – Prostrate dwarf-shrub, herb tundra
Bund e Fjord, Axel Heiberg Island, Canad a. D.A . Walker
Subzone C (acidic)
P2 – Prostrate/hemiprostrate dwarf-shrub tundra
Zackenbe rg, East Greenland, H.H. Christiansen
Low Arctic: erect shrubs
Subzone D
S1 – Erect dwarf-shrub tundra
Daring Lake, Canada. D.A. Walker
Subzone E
S2 – Low-shrub tundra
Seward Peninsula, Alaska. D.A. Walker
Wetlands:
Subzone B and C
W1 – Sedge/grass, moss wetland
Resolute, Cornwallis Island, Canada. D.A. Walker
Subzone D
W2 – Sedge, moss, dwarf-shrub wetland
North Slope coastal plain, Alaska. D.A. Walker
Subzone E
W3 – Sedge, moss, low-shrub wetland
Yukon-Kuskokwim Delta, Alaska. S. S. Talbot
Detailed descriptions of CAVM mapping units
Example:
B1. Cryptogam, herb barren (Figure 7a)
Dry to wet barren desert-like landscapes mainly in Subzone A and on some coarse-grained, often calcareous
sediments in subzones B and C. Sparse (2-40%) horizontal plant cover, and very low vertical structure
(generally <2 cm tall) with a single layer of plants where they occur. Dry herb barrens composed of few
scattered vascular plants are present over much of the landscape. Snow-flush communities are often a
conspicuous component, forming dark streaks on the otherwise barren lands, composed largely of bryophytes
and cryptogamic crusts. In upland areas, vascular plant cover is generally very sparse (<2%), mainly scattered
individual plants often in crevices between stones or small (< 50 cm diameter) cryoturbated polygons. Sedges
(Cyperaceae), dwarf shrubs, and peaty mires are normally absent.
Dominant plants: The most common vascular plants are cushion forbs (Papaver dahlianum ssp. polare,
Draba, Potentilla hyparcticaa, Saxifraga oppositifolian) and graminoids (Alopecurus alpinus, Deschampsia
borealis/brevifolia, Poa abbreviata, Puccinellia angustata, Phippsia , Luzula nivalisa, L. confusaa), lichens
(Caloplaca, Lecanora, Ochrolechia, Pertusaria, Mycobilimbia, Collema, Thamnolia, Cetraria, Flavocetraria,
Cetrariella, Stereocaulon), mosses (Racomitrium, Schistidium, Orthotheciumn, Ditrichumn, Distichiumn,
Encalypta, Pohlia, Bryum, Polytrichum), liverworts (e.g., Gymnomitrion, Cephaloziella), and cyanobacteria.
Representative syntaxa: Communities of the classes Thlaspietea rotundifolii Br.-Bl. et al. 1947 (e.g.
Papaveretum dahliani Hofm. 1968) and Salicetea herbaceae Br.-Bl. et al. 1947 (e.g. Phippsietum algidaeconcinnae Nordh. 1943).
Alaska tundra map derived from the CAVM
• 15 major categories,
subdivided into 33
subunits.
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
• Back side of the
map contains
detailed
information
regarding specific
plant communities
in each bioclimate
subzone, floristic
province, and
topographic
position.
Dominance type approach to classification
•
•
Based on the most easily noticed species in the overstory and
understory.
Dominance types of Whittaker (1962)
– Broad groups of vegetation with similar dominance of the trees in the upper
plant canopy but often with heterogeneous understories.
– Similar to the “associations” of Clements, which were in marked contrast to
the narrowly defined associations of the Braun-Blanquet approach.
•
Sociations of Du Rietz (1921)
– Plant community types with dominant species in each layer of the
vegetation canopy.
– Basic unit of vegetation for the Northern European School of
phytosociology.
– Useful concept in species-poor areas such as the boreal regions, but not so
useful in species-rich areas.
•
“Associations” of the Daubenmire (1952) and the US Forest Service
– Names of units based on one dominant species in the overstory and one in
the understory (e.g., Pinus ponderosa/Purshia tridentata).
The Braun-Blanquet approach: Classification based on total floristic
composition
• Plant community types are conceived as units based purely on
their total floristic composition.
• Among the species that make up the community, some are
better indicators of a given community than others. The
approach seeks to use those species whose ecological
relationships make them most effective indicators; these are
diagnostic species.
• The diagnostic species are used to organize communities into a
hierarchical classification, with the association as the basic
unit. This step is not always followed in the U.S.
Plant communities in the Braun-Blanquet approach
• Plant communities are recognized by their total floristic
composition.
• Units have relatively consistent groups of taxa with several
diagnostic taxa that characterize the plant community.
• If the plant communities are placed within the Braun-Blanquet
hierarchical system and given association names, then to be
valid:
– The description must be published in English
– with complete tables and
– named according to international rules of nomenclature.
3 Phases of the Braun-Blanquet approach
• Phase I, Field phase: collection of the relevé data (we’ve done
this.)
• Phase II, Analytical phase: sorted-table analysis leading to the
differentiated table. (We will do this in the lab.)
• Phase III, Syntaxonomic phase: naming of the unit according
to international rules of nomenclature and placement within the
hierarchic system of syntaxa. (We won’t do this.)
Phase I, Field phase, collection of Relevé data
•
•
Covered in Lab 1 and 2.
Review of key points:
– Reconnaissance critical to develop preliminary concept of primary plant
communities and their landscape relationships
– Choice of sample sites to avoid mixed, heterogeneous stands (subjective
sampling)
– Minimal area
– Centralized, replicate sampling
– Braun-Blanquet cover-abundance estimates
•
Collection of site and soil data is often part of the process.
Sample Relevé protocol
Phase II. Analytical Phase
(Sorted-table analysis)
•
GOAL: To prepare a table that shows the relevé data in a well-organized form
so that important trends of species distribution between the samples can be
immediately recognized.
•
(1) Relevés are grouped according to pre-recognized units. (These groups will
likely change during the table analysis process.)
(2) Species that are common to several relevés are identified and grouped
together. These diagnostic or character species become the key to identification
of vegetation units.
•
Steps in sorted-table analysis leading to the differentiated table
1. Construct a raw table.
2. Reorder species (table rows) to form a constancy table.
3. Extract species of intermediate constancy and use these to determine
differential species based on a set of fixed rules.
4. Reorder the relevés (table columns) and species (table rows) in relation
to the presence or absence of groups of differentiating species into a
differentiated table.
Step 1: Production of the
raw table
•
GOAL: To organize the field data in
columns (relevé numbers) and rows
(species).
•
1. The table is arranged with the
columns being the relevés and the
species the rows.
2. Order of arranging the data is not
important, but it may be easier to
recognize differential species if the
relevés are arranged according to
preconceived groupings along a
gradient from dry to wet (this order
will undoubtedly change at later
stages of the table analysis).
Species are best transferred in the
order they appear on the data
sheets.
•
Example Raw Table
Step 2: Constancy
table
•
GOAL: To show
the species
arranged in order
of constancy
(frequency).
•
1. Add up the
number of
occurrences for
each species and
enter in the right
hand column.
•
2. Rearrange the
species from
highest frequency
to lowest
constancy.
First two steps in table analysis
2 Order plots by preliminary types and
species by frequency
1 Raw |Table
+
+
CARMEM
1
2
+
2
2
2
1
1
1
+
CARVAG
+
+
.
+
.
+
+
.
.
.
CASTET
1
1
.
.
.
1
+
.
+
.
CHRINT
.
.
+
.
+
r
.
1
+
CIRCIR
+
.
+
.
.
+
.
+
DACARC
.
+
+
.
.
+
+
DISINC
.
.
1
.
1
1
DRYINT
3
2
2
3
3
ERIVAG
2
3
.
1
FLACUC
+
+
1
+
LECEPI
.
.
+
.
MINARC
.
.
+
SALARC
1
2
TOMNIT
3
2
Circle
+
Circle
.
140 Const.
Circle
.
141 142 97
Circle
+
96
Circle
.
47
Tundra
+
20 46
Tundra
.
19
Tundra
r
18
Tundra
Circle
CARCAP
Rel.
Tundra
Circle
141 142
Circle
96
Tundra
47
Tundra
140 46
Circle
20
Tundra
97
Circle
19
Tundra
18
Tundra
Rel.
CARMEM
1
2
2
2
1
1
1
+
+
2
10
DRYINT
3
2
3
2
3
2
3
3
2
3
10
FLACUC
+
+
+
1
+
2
1
1
1
1
10
r
TOMNIT
3
2
3
2
3
1
2
2
1
1
10
+
.
SALARC
1
2
2
2
2
.
+
.
r
.
7
+
.
+
CARCAP
r
.
.
.
.
+
+
+
+
+
6
1
+
.
.
CHRINT
.
.
.
r
.
1
+
r
+
+
6
2
3
2
3
3
DACARC
.
+
.
+
+
+
.
+
+
.
6
+
1
1
.
.
.
ERIVAG
2
3
1
1
1
.
.
.
.
+
6
1
1
+
2
1
1
CARVAG
+
+
+
+
+
.
.
.
.
.
5
+
.
.
+
+
2
.
+
+
.
1
+
.
CASTET
1
1
.
1
+
.
+
.
.
.
5
r
2
.
2
2
.
+
.
CIRCIR
+
.
.
+
.
+
+
.
+
.
5
1
3
1
2
3
1
2
2
DISINC
.
.
.
1
1
+
.
.
1
1
5
LECEPI
.
.
.
.
.
+
+
2
+
+
5
MINARC
.
.
.
+
.
1
+
.
+
+
5
Courtesy of Anja Kade
Step 3: Ordinated partial table
•
GOAL: To identify groups of differential
species that characterize the various plant
communities.
1.
Groups of differential species are identified that co-occur in
each of the predetermined group of relevés and are relatively
rare elsewhere. (See the previous slide, red and blue
underlined species of species. ) These species are identified
by their constancy within groups (see next slides for
discussion of constancy).
2.
These differential species are then arranged together
according to their relevé groups to form the "partial table”
(Note: The partial table merely saves time so all the rare and
ubiquitous species do not have to be recopied with each
iteration of the table. With the advent of computers it is no
longer necessary to develop a partial table. Constancy is
calculated for all species in all the vegetation types.)
3.
The relevés within each vegetation type are ordered by the
number of differential species they contain (highest number of
differential species on the left.)
4.
The differential species within each vegetation type are
ordered by the number of relevés they occur (highest number
of plots at the top of the list.)
5.
The groups of differential species should be ordered
according their associated relevés so that they form a
diagonal from the upper left to lower right of the table.
6.
The process is repeated if new subgroups of differential taxa
can be recognized within the primary groups, as is done in the
table here with the red and blue groups within the larger
groups.
How to determine differential species
Calculate the constancy class of each species within each
vegetation type. Constancy is the percent of relevés that the
given species occurs within a vegetation type.
Constancy classes
Constancy
Class
Percent of plots in
which species occurs
I
II
III
IV
1-20%
21-40
41-60
61-80
V
81-100
Rules for deciding if a species is differential for a vegetation type when compared against another
vegetation type (from Daniels 1982)
•
•
A vegetation type must have at least two differential species. Their presence is
high in the vegetation type under consideration and relatively low in other types.
The rules for determining differential species for four situations are as follows:
Situation I. The vegetation type under consideration, A, is based on 5 or more
relevˇs, and the comparable vegetation type, B, is based on 5 or more relevˇs
Vegetation Type A
Vegetation Type B
Constancy
Cover
Constancy
Cover
V
Arbitrary
II, I, +, r or absent
As in Type A or less
IV
Arbitrary
I, +, r or absent
As in Type A or less
III
Arbitrary
I, +, r or absent
As in Type A or less _
Situation II. The vegetation type under consideration, C, is based on 2-4 relevˇs,
and the comparable vegetation type, D, is based on 5 or more relevˇs
Vegetation Type C
Vegetation Type D
Constancy
Cover
Constancy
Cover
In all records
Arbitrary
I, +, r, or absent
As in Type C or less
In 2 of 3,m or in
2 of 4 records
Arbitrary
Absent
Absent
Rules for deciding if a species is differential for a vegetation type (continued)
Situati on III. The vegetation type under consideratio n, E, is based on 5 or mo re
relev ˇ s, and th e comp arable vegetatio n typ e, F, is based on 2-4 relev ˇs
Vegetation Type E
Vegetation Type F
Constancy
Cover
Constancy
Cover
V
arbitrary
absent
_
IV
arbitrary
absent
_
III
arbitrary
absent
_
Situati on IV. T he vegetatio n typ e under consideration, G, is based on 2-4 relev ˇ s, and
the com parable vegetation ty pe, H, is based on 2-4 relevˇs
Vegetation Type G
Vegetation Type H
Constancy
Cover
Constancy
Cover
In at leas t 2
recor ds
arbitrary
absent
_
Step 4: Differentiated table
• The final order of the differential species
is listed with constant and rare species
at the bottom of the table in order of
their constancy.
• Environmental and other relevant
information is added at the top of the
table.
• Preliminary community type names are
assigned to each group of relevés. The
groups can be delineated with vertical
lines, and the associated species shown
in outlined blocks.
Sorted differential table: unranked plant community types
Example of a Differentiated Table
47 96
Circle
Circle
Circle
140 141 142
Circle
97
Circle
Tundra
Tundra
20 46
Tundra
Tundra
18 19
Tundra
Rel.
DRYINT
3
2
3
2
3
2
2
3
3
3
TOMNIT
3
2
3
2
3
1
1
1
2
2
CARMEM
1
2
2
2
1
1
+
2
1
+
FLACUC
+
+
+
1
+
2
1
1
1
1
Constant taxa for both communities
Differential taxa for "Tundra"
ERIVAG
2
3
1
1
1
.
.
+
.
.
SALARC
1
2
2
2
2
.
r
.
+
.
CARVAG
+
+
+
+
+
.
.
.
.
.
CASTET
1
1
.
1
+
.
.
.
+
.
Differential taxa for "Circle"
LECEPI
.
.
.
.
.
+
+
+
+
2
CHRINT
.
.
.
r
.
1
+
+
+
r
CARCAP
r
.
.
.
.
+
+
+
+
+
MINARC
.
.
.
+
.
1
+
+
+
.
DACARC
.
+
.
+
+
+
+
.
.
+
CIRCIR
+
.
.
+
.
+
+
.
+
.
DISINC
.
.
.
1
1
+
1
1
.
.
Companions
Courtesy of Anja Kade
Phase III. Syntaxonomic phase
• Determine the formal names for the vegetation units according
the international rules of nomenclature.
• Show the relationship to existing units within the hierarchy of
nomenclature.
The plant association concept
•
“An association is a plant community of definite floristic composition,
presenting a uniform physiognomy, and growing in uniform habitat
conditions. It is the fundamental unit of synecology.” (Pavillard 1935).
•
This is the term as it is accepted by the International Botanical Congress. It
is the same definition as that is used by the Braun-Blanquet approach.
•
In the U.S., “association” had several very different meanings. Clements
used the terms to indicate a subdivision of a formation, very broad units with
uniform physiognomy. The U.S. Forest Service uses it more in the sense of
a dominance type. These are very broad units dominated by a typical pair
of species. The term also developed different meanings in northern Europe
(DuRietz), England (Tansley), and Estonia (Lipmaa).
•
The U.S. Classification approach is evolving toward a definition similar to
that used in Europe, although there are still major important aspects of the
European approach that are not included, such as the nomenclature
system and the syntaxonomical approach.
Requirements for formally defined syntaxa
1. Must be published in a journal or book.
2.
Must contain an adequate description including:
a.
b.
Association tables with at least 3 relevés and specify a "type relevé”
Lists of character and differential taxa for higher units (alliance, order,
classes).
3. Author's name and year of publication are attached to the syntaxon
name.
Example: Clementsio rhodanthae-Rhodioletum integrifoliae Komárková
1976
Association tables
• The final table should show the differential/character species for
each subassociation within the association and
differential/character taxa for related alliances, orders, and
classes.
• A brief summary of the site information should be included at the
top of the table.
Diagnostic Species
• Differential taxa
– These species differentiate one community from the other local
plant communities.
– Species with high constancy in one community type and low
constancy in other communities.
• Character taxa
– Special type of differential taxa the differentiate the community
from all other described associations.
– Have ecological optimum in one community.
– These can only be determined after a thorough search of the
literature.
Example association table for Artemisietum maritimae in the SW Netherlands (W.G. Beeftink 1978)
Synoptic Syntaxon Table
• When several related associations from a region have been
described, the information can be summarized into a synoptic
(summary) table, showing the character taxa for all the
associations, alliances, orders, and classes.
Example synoptic syntaxon table
System of syntaxonomic nomenclature
Syntaxon
Class
Order
Alliance
Association
Subassociation
Suffix
Example
-etea
-etalia
-ion
-etum
-osum
Phragmitetea
Littorelletalia
Alnion glutinosae
Ericetum tetralicis
Artemisietum maritimae
agrostidetosum
Note: To convert a plant name to a syntaxon name, the suffixes are attached to the
plant genus names. The plant species names are converted to syntaxa names
with the use of Latin genitive case throughout. Names ending in -a becomes ae; -es becomes -is; -um becomes -um, and -us becomes-is.
Important points
• There are different approaches to classification (physiognomic,
dominance types, mixed ecological, pure floristic approaches).
• Braun-Blanquet is a pure floristic approach.
• 3 phases in the Braun-Blanquet approach (Field, analytic (table
analysis), syntaxonomic phases)
• Steps in the sorted table analysis procedure (raw table,
constancy table, sorted partial table, final differentiated table)
• Association concept (“...a plant community of definite floristic
composition, presenting a uniform physiognomy, and growing in
uniform habitat conditions. It is the fundamental unit of
synecology.” (Pavillard 1935))
• Syntaxonomic nomenclature of the B-B approach (hieararchical
system with specific rules for publication and naming of
communities)
Important references for classification
•
•
•
W estho ff, V ., and E . van de r M aare l. 1978 . T he B raun B lanquet app roach . Pages 287-399 in R . H . W hittake r,
ed ito r. C lass ification o f P lant Comm un it ie s . D r. W . Junk,
Den H aag .
E llenberg , H . 1988 . V egetation Eco logy o f Centra l Eu rope, 4
ed it ion . C am bridge U n ive rs ity P ress , C am bridge . 731 pp
D ie rschke , H . 1994 . P flanzensoz io log ie G rud lagen und
M ethoden . U lm er, S tu ttgart. 638pp