Species indicator values as an important tool in applied plant ecology Session: Databases and information systems 46th IAVS Symposium 8-14 June 2003 Naples.

Species indicator values as an important tool in applied plant ecology

Session: Databases and information systems

46th IAVS Symposium 8-14 June 2003Naples (Napoli), Italy

Martin Diekmann

Vegetation Ecology andNature Conservation Biology, IFOE

http://www.uni-bremen.de/

46th IAVS Symposium 8-14 June 2003, Naples

Contents

• Biological indication: Introduction

• Species responses and indicator values

• Applications of indicator values

• Indicator values and measurements

• Calibration of indicator values

• Conclusions


Examples

Indicator values for climatic and soil variables(Ellenberg et al. 1992) Indicators of habitat continuity (e.g. ancient forests)(Wulf 1997)

Indicators of management (e.g. mowing sensitivity)(Briemle & Ellenberg 1994)

Grime’s classification of species into habitat and strategy types(Grime et al. 1988)

Biological indication –making use of the specific reactions of organisms to their environment


Example: Reaction – soil or water pH (R)

1 Indicator of extreme acidity (Andromeda polifolia)2 Between 1 and 33 Acidity indicator (Pteridium aquilinum)4 Between 3 and 55 Indicator of moderately acid soils (Milium effusum) 6 Between 5 and 77 Indicator of weakly acid to weakly basic conditions (Hepatica nobilis) 8 Between 7 and 99 Indicator of calcareous or other high-pH soils (Primula farinosa)

Ellenberg values


Species responses and indicator values

Criticism: Species have no distinct ecological optima.

Gausscurve

Ecological gradient (e.g. pH)

Speciesresponse,

e.g. biomassproduction


A Gaussian response is preferable, but it is sufficient that the response is unimodal or linearly increasing/decreasing (skewed).

Most species response curves look that way, if

• The environmental variable in question is important.

• The total variation is large enough.

• The distribution of measured values along a gradient is more or less regular.


Response curves of forest vascular plants in Sweden(Diekmann, unpubl.)

Anemone nemorosa

Anemone ranunculoides

Melica nutans

Actaea spicata


Weighted site averages

Site score = (x1y1 + x2y2 + ... + xnyn) / (x1 + x2 + ... + xn)

x1, x2, ..., xn frequencies of species 1 to n;

y1, y2, ..., yn indicator values of species 1 to n.

Applications of indicator values


Weighted site averages serve to:

• Obtain information about the environment of relevés, habitat patches, grid squares, etc., and to make environmental comparisons between sites.

• Interpret ordination diagrams, by correlating the axis scores of sites with the corresponding values of measurements.


• ... calculate the size of the species pool of communities.

0 100 200 300 400

Coniferousforest

Halophytic

Tall herbvegetation

Freshwaterand bogs

Deciduousforest and

scrub

Talus andalpine

meadow

Anthropogenicheaths andmeadows

Ruderal

# species R > 6

0 20 40 60 80 100

Coniferousforest

Anthropogenicheaths andmeadows

Freshwaterand bogs

Deciduousforest and

scrub

Tall herbvegetation

Ruderal

Talus andalpine

meadow

Halophytic

% species R > 6

(Ewald, in press)


• ... predict the occurrence of species.

(Dupré &Diekmann

1998)[based on logistic regression]


• examine changes in environmental conditions and reconstruct historical site conditions on the basis of former species assemblages.

(Diekmann et al. 1999)


Weighted averages may be unreliable, if

• The plots studied are spatially highly heterogeneous.

• The environmental gradient is too short.

• The habitat is strongly affected by management (e.g. heavy grazing).

• There is a sudden environmental shift.

• The vegetation mainly responds to extremes and not to mean values.


Indicator spectra

(Diekmann 2003)

1 2 3 4 5 6 7 8 9

0

5

10

R indicator value

No.

of

spec

ies

Elm-ash forest


(Diekmann 2003)

1 2 3 4 5 6 7 8 9

0

1

2

3

4

5

6

7

8

9

R indicator value

No.

of

spec

ies

Basicolouspine forest


Indicator values and measurements

If weighted averages and measurements are

- highly correlated,We conclude that the indicator values for the variable in question work well.

- weakly correlated, there are two possible conclusions:1) The indicator values for that variable are unreliable.2) We have measured the wrong variable!

Criticism: Indicator values are not systematically derived from measurements,but mainly inferred from field experience. They may prevent us from conducting measurements!


Parameter Correlation Soil reactionpH (H2O), pH (KCl) + (Persson 1980, Degorski 1982,

or pH (CaCl2) Lawesson & Mark 2000, Wamelink

et al. 2002)+ (Rodenkirchen 1982)+ (Ellenberg 1992, Herzberger & Karrer 1992,

Seidling & Rohner 1993, Diekmann 1995, Hill & Carey 1997, Ertsen et al. 1998, Schaffers & Sýkora 2000, Diekmann, unpubl.)

Base saturation ++ (Degorski 1982, Seidling & Rohner 1993, Schaffers & Sýkora 2000)

Al/Ca ratio ++ (Seidling & Rohner 1993)Ca2+ saturation ++ (Schaffers & Sýkora 2000)Ca2+ amount ++ (Degorski 1982, Schaffers & Sýkora 2000)Total calcium ++ (Schaffers & Sýkora 2000)


Correlation measured values – weighted site scores

(Diekmann 2003)

Swedish forests


Parameter Correlation

NitrogenN PARAMETERSTotal N + (Vevle & Aase 1980, Schaffers & Sýkora 2000)

- (Hill & Carey 1997, Ertsen et al. 1998)Total C/Organic matter - (Hill & Carey 1997)

+ (Diekmann, unpubl.)C/N ratio + (Ellenberg 1992, Schaffers & Sýkora 2000)

- (Diekmann, unpubl.)Mineral N (NO3

- + NH4+) + (Schaffers & Sýkora 2000)

- (Lawesson, unpubl.)Mineral NO3

- + (Schaffers & Sýkora 2000)

% NO3- + (Schaffers & Sýkora 2000)

Total mineralisation + (Rodenkirchen 1982, Ellenberg 1992)- (Schaffers & Sýkora 2000,

Ammonification rate + (Diekmann, unpubl.)Nitrification rate + (Diekmann, unpubl.)Nitrification ratio ++ (Diekmann, unpubl.)


Correlation measured values – weighted site scores

(Diekmann 2003)

Swedish forests


Parameter Correlation

NitrogenOTHER NUTRIENTSTotal P + (Hill & Carey 1997, Ertsen et al. 1998)Available PO4

3- + (Ertsen et al. 1998, Schaffers & Sýkora 2000)

- (Lawesson, unpubl.)Total K + (Ertsen et al. 1998)Available K+ + (Hill & Carey 1997, Schaffers & Sýkora 2000)

- (Lawesson, unpubl.)Exchangeable Mg - (Lawesson, unpubl.)

VEGETATION-DERIVED PARAMETERSBiomass/”standing crop” + (Boller-Elmer 1977, Briemle 1986)

++ (Melman et al. 1988, Hill & Carey 1997, Ertsen et al. 1998, Schaffers & Sýkora 2000)

N accumulation of plants ++ (Ertsen et al. 1998, Schaffers & Sýkora 2000)Foliar N concentration ++ (Thompson et al. 1993)Tissue N concentration + (Schaffers & Sýkora 2000)


Calibration of indicator values


Three calibration approaches:

1) Observation of single species.

2) Indirect calibration on the basis of large phytosociological data sets.

3) Measurements of environmental parameters.

Criticism: Species may shift in their responses, especially across geographical gradients.


(Hill et al. 2000)


-3 -2 -1 0 1 2 3 4 5

-3

-2

-1

0

1

2

3

4

5

Calibrated R indicatorvalues in The Netherlands

Cal

ibra

ted

R in

dica

tor

valu

es in

Gre

at B

rita

in

R-squ adj. = 0.850

Galiumpumilum

(Diekmann 2003; data from Hill et al. 1999 and Schaffers & Sýkora 2000)


• Indicator values are subjective, but if they coincide with the ecological optima of species, they are very useful for many applications.

• Indicator values need to be verified by measurements and calibrated.

• Whenever possible, the use of indicator values should be accompanied by measurements.

Conclusions

Species indicator values as an important tool in applied plant ecology Session: Databases and information systems 46th IAVS Symposium 8-14 June 2003 Naples.

Documents

species indicator values

ifoe46th iavs symposium

environment46th iavs

press46th iavs symposium

yn indicator values

indicator valuescriticism

indicator of calcareous

indicator of weakly