• Aucun résultat trouvé

Forest edge responses present a variety of patterns in southwestern France

N/A
N/A
Protected

Academic year: 2021

Partager "Forest edge responses present a variety of patterns in southwestern France"

Copied!
23
0
0

Texte intégral

(1)

Open Archive TOULOUSE Archive Ouverte (OATAO) OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible.

This is an author-deposited version published in : http://oatao.univ-toulouse.fr/

Eprints ID : 16374

To cite this version : Alignier, Audrey and Deconchat, Marc Forest

edge responses present a variety of patterns in southwestern France.

(2009) In: 9th Annual Meeting of the British Ecological Society - BES 2009, 7 September 2009 - 10 September 2009 (Hertfordshire, United Kingdom). (Unpublished)

Any correspondance concerning this service should be sent to the repository administrator: staff-oatao@listes-diff.inp-toulouse.fr

(2)

F

OREST

E

DGE

R

ESPONSES PRESENT

A VARIETY OF PATTERNS IN

S

OUTHWESTERN

F

RANCE

Audrey Alignier & Marc Deconchat

(3)



Spatial dynamics of landscape

modified by human activities:

Fragmentation/ Defragmentation

- Loss of habitat (e.g. by forest cutting) - Isolation of patches

F

OREST

E

DGES IN

L

ANDSCAPE

:

A CRITICAL

ROLE FOR

V

EGETATION

B

IODIVERSITY

- Isolation of patches

Increase of discontinuity proportion in landscape

Ex. Progressive fragmentation of a woodlot since 1750, in southwestern France (Arrignon, 2003)

(4)

F

OREST

E

DGES IN

L

ANDSCAPE

:

A CRITICAL

ROLE FOR

V

EGETATION

B

IODIVERSITY



Edge = zone, in the forest, under

discontinuity influence

(Murcia, 1995)

Discontinuities influence

environmental conditions to which vegetation respond (richness,

Forman (1995)

Theoretical pattern of response to discontinuity, widely accepted.

Distance from border

V a ri a b le v a lu e

Field Forest edge Forest interior

vegetation respond (richness, abundance, composition).

(5)

O

UR

Q

UESTIONS



How do forest vegetation (richness, abundance and

composition) and abiotic variables respond to edge

effect ?

Comparing vegetation response curves to forest edges between

several transects

Method

several transects



Can we characterize a common pattern of response to

edges ?

Attributing a statistical model to responses of vegetation and

abiotic variables

(6)

S

TUDY

S

ITE

Nord

1km

Site LTER « Vallées et coteaux de Gascogne »

Subatlantic climate with mediterranean influence

(average annual T°C = 11°C; average annual rainfall = 800mm) Woodlots managed by private owners ; coppice with standards Dominant species: oaks (Quercus robur, Q.pubescens, Q.petreae), hornbeam (Carpinus betulus), wild cherry (Prunus avium)

(7)

S

AMPLING

D

ESIGN

 28 transects, extended from

the border to 40m into the forest interior

Border =line formed by the first tree trunks (Murcia,1995)

 40m away from clearcut or

40m Border Forest interior Adjacent land cover

2m

Crop or meadow One transect:

Localisation map of 28 transects studied (in red)

 40m away from clearcut or

(8)

D

ATA

A

NALYSIS

: M

ODEL

A

PPLICATION

PER

T

RANSECT

Vegetation data:

 Total species richness

 Total abundance (cover percent)  Composition

(Scores on Axis 1 of centered PCA on presence/absence matrix)

In relation to distance from border

Response curves

Models adjustment and

1 Comparison Environmental data:  Soil temperature  Soil moisture (RH)  Soil pH  Soil penetrability  Canopy openess (%)

2 Models adjustment and

best model selection

(Ewers & Didham, 2006)

Characterization of edge reponses by adjusted models Comparison of models

* 28 transects

Alignier – Response patterns to forest edge

(9)

D

ATA

A

NALYSIS

: B

EST MODEL

S

ELECTION

Distance from border

V a ri a b le v a lu e NULL

Distance from border

V a ri a b le v a lu e EXPONENTIAL v = tested variable d = distance from border

βx= constant

Selection on AIC

ε

+ =v v v=β0exp(β1*d)+ε LINEAR

Distance from border

V a ri a b le v a lu e

Distance from border

V a ri a b le v a lu e LOGISTIC Widely accepted theoretical model « Simple » models - Complexity +

Selection on AIC

criterion

ε β β + + = d v 0 1* ( ) ( ) (β β β ) ε β β + − + − + = 3 * 2 exp 1 0 1 0 d v

(10)

R

ESULTS

: V

EGETATION

123 species from 42 families ; 75 herbaceous

perennials, 42 woody species and 6 annual species

2 0 0 3 0 0 4 0 0 5 0 0 b u n d a n c e Hedera helix Rubus fructicosus gr. Tamus communis Lonicera per. H. helix R. fructicosus L.peryclimenum T.communis 1 5 2 0 ri c h n e s s Max = 15 Min = 6 Mean = 10.7

High variation in species richness between transects (similar results for abundance).

0 20 40 60 80 100 120 0 1 0 0 2 0 0 Species rank A b u n d a n c e 80% of species have an occurrence frequency <10%. 1 3 5 7 9 11 13 15 17 19 21 23 25 27 5 1 0 1 5 Transect T o ta l s p e c ie s ri c h n e s s

(11)

R

ESULTS

: A

BIOTIC

V

ARIABLES

High variation of soil temperature between transects:

variation between transect (up to 4°C) > variation within a transect (max 1.3°C)

pH was stable with distance from

border except for 7 transects with a slight decrease in forest interior as in Marchand & Houle (2006).

General decrease of % canopy

Soil moisture was stable with General decrease of % canopy

openess with distance from border.

High variation of abiotic variables between transects,

often > variation within transect.

Results consistent with previous studies.

Soil moisture was stable with

distance from border, except for 6 transects which present an increase of soil moisture in forest interior.

(12)

R

ESULTS

: V

EGETATION

R

ESPONSE

M

ODELS

8 1 0 6 0 8 0 1 0 0 0 .0 0 .5 1 .0 1 .5

Species richness Percent cover Composition

Example for transect n°19:

Different models between vegetation descriptors

0 10 20 30 40

4

6

Distance from border (m)

0 10 20 30 40

2

0

4

0

Distance from border (m)

0 10 20 30 40 -1 .5 -1 .0 -0 .5

Distance from border (m)

EXPONENTIAL NULL LINEAR

(13)

R

ESULTS

: A

BIOTIC

R

ESPONSE

M

ODELS

1 6 .8 1 7 .0 1 7 .2 1 7 .4 5 .8 6 .0 6 .2 1 5 1 6 1 7

Soil temperature pH Soil moisture

( Penetrability) (Canopy openess) Example for transect n°19:

Different models between abiotic descriptors

0 10 20 30 40 1 6 .2 1 6 .4 1 6 .6

Distance from border (m)

0 10 20 30 40 5 .2 5 .4 5 .6

Distance from border (m)

0 10 20 30 40 1 2 1 3 1 4

Distance from border (m)

(14)

R

ESULTS

: F

REQUENCY OF

M

ODELS FOR

ALL

TESTED VARIABLES

Model complexity 25 30 35 40 45 F r e q u e n c y (% )

No discernible

edge effect

0 5 10 15 20 1 2 3 4 F r e q u e n c y Adjusted models

Null Linear Exponential Logisitic

Model « null » dominant

BUT

Edge effects in the majority of cases

(15)

R

ESULTS

: M

ODELS

A

DJUSTMENT

10 15 20 25 30 N u m b e r o f tr a n se c ts Logistic Power

Abiotic variables Biotic variables

Complexity +

Logistic model

dominant

Simple models are dominant for the majority of variables.

0 5 10 N u m b e r o f tr a n se c ts Linear Null Complexity

(16)

D

ISCUSSION



How forest vegetation (richness, abundance and composition)

respond to edge effect ? And abiotic variables ?

No unique (general) model but different models between vegetation descriptors and abiotic variables.

(17)

D

ISCUSSION

V a ri a b le v a lu e

Field Forest edge Forest interior

Border

NULL

NO EDGE EFFECT

Distance from border Distance from border

EXPONENTIAL

LINEAR

(18)

D

ISCUSSION



How forest vegetation (richness, abundance and composition)

respond to edge effect ? And abiotic variables ?

No unique (general) model but different models between vegetation descriptors and abiotic variables.



Can we characterize a common response pattern to edges ?

High variability of edge responses between transects: - Different adjusted models according to tested variables

- Null model predominant for overall transects NO EDGE EFFECT - Selection among 4 models not always the best adjustment to data

(19)

D

ISCUSSION

How to explain the absence of edge effects ?



Small woodlots (from <0.05 to 5ha)



Human management

Frequent perturbation over years Frequent perturbation over years Management by private owners

Perspectives :

To integrate forest management (cuttings) and forest continuity as explicative variables in data analysis

(20)

This work was granted by the French Ministry of Research and Education. Thanks to:

L.Burnel, M.Fakorellis, J.Willm for their field assistance

A.Cabanettes, M.Goulard for their helpful comments on statistical analysis Private woodland farmers and owners

SEVAB for financial support in participation of this meeting

Thanks for your attention

Thanks for your attention

(21)

Hemispherical photography Hedera helix Rubus fructicosus Tamus communis Lonicera periclymenum 0 20 40 60 80 100 120 0 10 20 30 40 50 60 70 80 90 S p e c ie s n u m b e r Fréquence (%)

(22)

3 5 0 4 0 0 4 5 0 P e n e tr a b il it y 1 4 1 6 1 8 2 0 2 2 % C a n o p y o p e n e s s 0 10 20 30 40 3 0 0

Distance from border (m)

0 10 20 30 40

1

2

1

4

(23)

R

ESULTS

: M

ODELS

A

DJUSTMENT

For all transects:

Simple Complex 30 40 50 60 70 80 90 100 Composition Percent cover Species richness % Cnpy Openess pH Penetrability O c c u r r e n c e f r e q u e n c y

Null Linear Power Logisitic

Models complexity 0 10 20 30 1 2 3 4 Penetrability Soil moisture Soil T°C O c c u r r e n c e

Références

Documents relatifs

Wooded Savanna : Pericopsis laxiflora, Pterocarpus erinaceus, Pseudocedrela kotschyi, Parinari curatelifolia, Maranthes polyandra, Uapaca togoensis, Detarium microcarpum,

The purpose of this note is not to dwell on the localization problem, but only to stress the elfect of scatterer motion on the edge singularities. For a given history, the

By using the following theorem, we can reduce the disjoint paths problem to an equivalent smaller problem if the input graph has a large clique minor.. An outline of the proof

(a): Acoustic streaming tracing from fluorescence particles tracking, at zero flowrate; (b): Static mixing effect without acoustic streaming, (0 V) and (c-f) Active mixing

Statistical models were applied to mean plant indicator values (MIV) from Ellenberg and Ecoplant databases for soil pH, soil nitrogen (N), soil humidity (F), light (L) and

We used the same test to detect seasonal variations in the selectivity of animals (S i ) for the main vegetation components, one on deciduous-leaved species (broad-leaved trees

We quantitatively derived average surface properties of dune and interdune areas (texture and composition) from modeling of the microwave backscattered signal and Monte Carlo

When facing the need to model adsorption separation processes, or simply fixed bed adsorption/desorption cycles, in the presence of water one must give a mathematical expression