Disentangling relationships between habitat conditions, disturbance history, plant diversity, and American black cherry (Prunus serotina Ehrh.) invasion in a European temperate forest:

DOI: 10.1111/j.1472-4642.2007.00453.x © 2007 The Authors

204

Journal compilation © 2007 Blackwell Publishing Ltd www.blackwellpublishing.com/ddi

Diversity and Distributions, (Diversity Distrib.)

(2008)

14

, 204–212

BIODIVERSITYRESEARCH

ABSTRACT

Whether non-native plant invasions are causes, consequences, or independent of thelow species diversity in recipient ecosystems remains a debated question. We tried totest these three hypotheses in the special case of the American black cherry (

Prunusserotina

Ehrh.), a gap-dependent tree species, which is invading European temperateforests. We compared plant communities, soil properties, and disturbance historybetween

P. serotina

-invaded and uninvaded paired-stands in a managed mixed forest.Relationships between invasion, disturbances, plant communities, and environmentalconditions were investigated using redundancy analyses with variation partitioning.Several soil characteristics differed between paired stands, but were rather componentsof stand invasibility than invasion effects, except for topsoil available phosphorus.The disturbance history was similar among paired stands except for the amount ofstorm-induced tree falls, which correlated with the invader’s density. Wild boar-disturbed soil areas were more important beneath

P. serotina

canopies, suggesting apositive feedback on its own establishment. Overall, species assemblages in invadedand uninvaded stands were similar; their ecological inconsistency suggested amanagement-sustained non-equilibrium. Habitat conditions and disturbancesexplained most of the variation in both plant diversity and

P. serotina

density, the lasttwo factors exhibiting a weak direct association. We conclude that in managed forestecosystems where plant communities are mainly driven by non-interactive factorsand immigration processes, non-native plant species can naturalize without beingdirectly influenced by measured features of the plant community in the receivingenvironment on the short term.

Keywords

Alien plant invasion, biological invasions, disturbance, forest management,

invasibility, non-equilibrium community,

Prunus serotina

.

INTRODUCTION

Understanding the causes and consequences of non-native plant

invasions is increasingly attracting attention, given their ecological,

economical, and societal deleterious effects (Vitousek, 1990;

Levine

et al

., 2003; Dietz & Steinlein, 2004), and their relevance

for testing theories in community ecology (Shea & Chesson,

2002). Invasive species are widely accepted as one of the leading

causes of biodiversity loss (Cronk & Fuller, 1995; Chapin

et al

.,

2000; Sax & Gaines, 2003). Yet, the degree to which non-native

invaders influence vegetation composition remains poorly

explored. Much of the evidence is based on field observations of

contemporary non-native dominance with low native species

richness. Those are often used to infer causal relationships

(Gurevitch & Padilla, 2004; Didham

et al

., 2005), and to speculate

about competitive exclusion of natives (‘driver’ model

sensu

MacDougall & Turkington, 2005; Fig. 1a). The impact of an

invader is expected to correlate with its own population density,

since any biomass (or space or energy) controlled by the invader

constitutes resources no longer available to other species (Parker

et al

., 1999). Hence, when an invader becomes the durably

dominant species of a community, long lasting shifts in species

composition are likely to occur. Such invaders have been labelled

‘transformers’ (Richardson

et al

., 2000), ‘invasive engineers’

(Cuddington & Hastings, 2004), or ‘strong invaders’ (Ortega &

Pearson, 2005), since they can change ecosystem functioning

1

Dynamiques des Systèmes Anthropisés,

Université de Picardie Jules Verne, 1, rue des

Louvels, F-80037 Amiens Cedex, France,

2

Laboratory of Forestry, Ghent University,

Geraardsbergsesteenweg 267, B-9090 Melle-

Gontrode, Belgium

*Correspondence: Guillaume Decocq, Laboratoire de Biodiversité végétale et fongique, Université de Picardie Jules Verne, 1, rue des Louvels, F-80037 Amiens Cedex, France. Tel./Fax:

+

33 (0)3 22 82 77 61; E-mail: [email protected]

Blackwell Publishing Ltd

Disentangling relationships between habitat conditions, disturbance history, plant diversity, and American black cherry (

Prunus serotina

Ehrh.) invasion in a European temperate forest

Olivier Chabrerie

1

, Kris Verheyen

2

, Robert Saguez

1

and Guillaume Decocq

1

*

Invasion–disturbance–ecosystem interactions

© 2007 The Authors

Diversity and Distributions

,

14

, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd

205

over a substantial area (e.g. excessive use of resources, disturbance

promotion/suppression, litter accumulation, soil nutrient flow

changes; Richardson

et al

., 2000; Davis, 2003; Vanderhoeven

et al

., 2005).

Disturbance has been early recognized as an extrinsic factor

promoting biological invasions, so that the most disturbed

habitats – or, more generally, those exhibiting the largest

resource fluctuations – are also the most susceptible to invasion

(Crawley, 1987; Hobbs & Huenneke, 1992; Levine & D’Antonio,

1999; Williamson, 1999; Alpert

et al

., 2000; Davis

et al

., 2000).

This has led to the ‘passenger’ model (Fig. 1c), which predicts

that non-natives are less dispersal-limited than natives and thus

are able to preempt space (and energy) before natives can

reestablish (MacDougall & Turkington, 2005). Here, there is no

direct relationship between the invader’s abundance and the

native species loss, both being consequences of the disturbance.

Recent advances in invasion biology have also shown that the

classic diversity resistance hypothesis, which argues that, all else

being equal, diverse native communities readily resist invasion

(e.g. Elton, 1958; Levine & D’Antonio, 1999), is more likely to

apply at small (local) scales. In this case, a species-rich plant

community, which is mainly driven by competitive interactions

(niche theory), is less susceptible to invasion than a species-poor

one, which is mainly driven by immigration processes (neutral

theory) (Tilman, 1997; Naeem

et al

., 2000; Kennedy

et al

., 2002;

Brown & Peet, 2003). Conversely, at larger (regional) scales, a

positive relationship between native and non-native species

richness is expected, as heterogeneity is increasing (Brown &

Peet, 2003). The environmental heterogeneity hypothesis

recently emerged as a heuristic generalization: environmental

heterogeneity both increases invasibility and reduces the impact

on native species in the community, by promoting invasion and

coexistence mechanisms that are not possible in homogeneous

environments (Melbourne

et al

., 2007). Hence, the degree to

which a native community is unsaturated, or lacks diversity

due to a limited regional pool of species, highly accounts for its

invasibility (Tilman, 1997; Gilbert & Lechowicz, 2005). According

to this third scenario, which we will call the ‘opportunist’ model

(Fig. 1b), invasion is the consequence of the low native species

diversity and not vice-versa, and can even occur without covarying

extrinsic factors.

In this study, we tested these three models in the special case of

the American black cherry (

Prunus serotina

Ehrh.).

Prunus

serotina

is a gap-dependent tree species native to North America

which is largely spreading throughout temperate forests of

Western and Central Europe, especially on well-drained, nutrient-

poor soils (Starfinger, 1997; Chabrerie

et al

., 2007a). For this

purpose, we compared plant communities, soil properties and

disturbance history between invaded and uninvaded paired

stands, in a managed mixed forest. More specifically, the

following research questions were tackled:

1

Do stand structure, vegetation, soil properties, and disturbance

history differ with respect to the presence of

P. serotina

in the

canopy?

2

Do those differences correlate with the invader’s density?

3

Once the collinearity among variables has been reduced,

can we explain the observed ecosystem–disturbance–invasion

interplay by one of the three models outlined above?

METHODS

Study site

This study was carried out in the temperate deciduous forest of

Compiègne, located in northern France (49

°

22

′

N; 2

°

54

′

E; 32–

148 m altitude) and covering 14,417 ha. This forest was chosen

because it contains a wide range of habitat conditions and is the

most invaded by

P. serotina

in France (Chabrerie

et al

., 2007b).

The climate is of oceanic type with a mean annual temperature of

10.3

°

C and annual rainfall of 677 mm. The geological substrate

mainly consists of palaeogeneous sands (

c

. 60% of total area) and

cretaceous chalks (

c

. 20%), locally covered by quaternary loess

and alluvial deposits.

The forest is currently managed as an even-aged plantation of

common beech (

Fagus sylvatica

), oaks (

Quercus robur

,

Quercus

petraea

), and Scots pine (

Pinus sylvestris

). The silvicultural cycle

lasts 130 years for

Q. robur

, 180 years for

Q. petraea

, 110 years for

F. sylvatica

, and 100 years for

P. sylvestris

, and always starts by

a large clear-cut. During this time interval, thinnings are

conducted every 4–10 years. Natural disturbances consist of

windthrows; in the past 30 years, two severe storms occurred in

Figure 1 Three hypothetic models accounting for the invasion–diversity–disturbance relationships: (a) ‘driver model’: the invasion causes diversity change; (b) ‘opportunist model’: the low diversity of the recipient community allows the invasion; (c) ‘passenger’ model: disturbance causes both diversity change and invasion, simultaneously but independently. Dashed arrows indicate implicit relationships, whereas full arrows indicate explicit relationships.

O. Chabrerie

et al.

© 2007 The Authors

206

Diversity and Distributions

,

14

, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd

1984 (364 m

3

ha

–1

, i.e. 129,075 trunks over 933 ha) and 1990

(227 m

3

ha

–1

, i.e. 134,451 trunks over 1199 ha).

Prunus serotina

has probably been introduced directly in the forest around 1850,

but has been reported as an invader only in 1970. It has now

spread over more than half of the total forest area.

Sampling design and data collection

Following a first field survey and by using available soil maps, we

first stratified forest stands according to soil type, dominant tree

species, and abundance/absence of

P. serotina

. Only closed-

canopy stands were retained. Then, we randomly selected 32

invaded stands, defined as stands where

P. serotina

has become

the dominant species in tree and shrub layers. For each invaded

stand, we paired an uninvaded stand (no

P. serotina

at all) sharing

the same soil (based on the visual aspect of soil cores) and canopy

(based on the respective proportion and mean diameter at breast

height (d.b.h.) of dominant tree species) conditions, and being as

close as possible to the invaded stand. This paired-site sampling

approach reduces the impact of possibly confounding environ-

mental variables on the observed differences in response variables

between invaded and uninvaded stands, while representing the

full range of

P. serotina

-invaded forest stands in the study area.

In each stand, a temporary 25 m

×

25 m plot (625 m

2

) was

disposed so that soil profile, stand structure, and relief were

uniform; the local environment was thus considered as homoge-

neous. Total vegetation cover was estimated for each vegetation

layers: tree layer:

>

6 m; shrub layer: 2–6 m; undershrub layer

>

0.5 m; and herb layer

<

0.5 m. Vascular plant species were

recorded during periods of peak vegetation cover (June to July

2004), and their cover abundance within each layer was scored

using Braun-Blanquet’s scale (Braun-Blanquet, 1964). Species

nomenclature follows Lambinon

et al

. (2004). For invaded

stands, we also described the

P. serotina

population in a

40 m

×

40 m quadrat which was centred on the smaller 625 m

2

one: canopy height (m), d.b.h. (cm), and canopy cover (%) of all

P. serotina

individuals with a d.b.h.

>

5 cm. To estimate the

population age we took a wood core at breast height from the

individual(s) exhibiting the largest d.b.h. using a Pressler’s auger,

for further tree ring counting (WinDENDRO® version 2004b;

Gagnon & Morin, 2004).

At the same time, in the 625 m

2

plots, we recorded the canopy

gap number, fallen tree cover (%), bare soil cover (%), stump

number, and wild boar-disturbed area (%). We measured litter

(L horizon) and upper organic horizon (OF and OH horizons)

thickness (cm) following the nomenclature of Brethes

et al

.

(1995). We also randomly collected soil litter in three 0.25 m

×

0.25 m quadrats, and took six 15-cm depth soil cores to count

the number of Fe

3

+

spots before merging them into a composite

soil sample for further laboratory analyses. Litter biomass was

weighted after drying (3

×

1 h at 70

°

C). Soil samples were dried

(65

°

C for 48 h) and sifted (2 mm sieve) prior to analyses.

All chemical analyses followed Aubert (1978): organic matter

(loss on ignition at 850

°

C for 15 h), pH H

2

O, total nitrogen,

N-NO

3–

and N-NH

4

+

(Kjeldahl method), organic carbon (Anne’s

method), and plant available phosphorus.

Information about natural (i.e. storm-induced tree falls) and

management-related disturbances were extracted from the local

forest service databases for the last 50 years (1955–2005). The

volume of logs that has been collected after each storm event

(TF) or cut down (CT) was cumulated into 10-year classes, from

1955 to 2005, prior to data analyses.

Data analysis

Comparison between invaded and uninvaded stands

Paired comparisons of vegetation and environmental characteristics

between invaded and uninvaded stands were conducted using

Wilcoxon rank tests (

P

<

0.05). For the vegetation, we compared

species richness (S) and Shannon’s evenness index (H

′

) for the

whole plant community and for each vegetation layer separately.

Further analyses were restricted to the herb layer since it contains

the major part of diversity in temperate forests, it is less affected

by forest management and is expected to respond more quickly

to environmental changes (e.g.

P. serotina

invasion) than woody

layers do (Halpern & Spies, 1995; Decocq, 2000).

Prunus serotina

was not included in the species matrix but used as an environ-

mental variable.

Species composition difference between invaded and

uninvaded stands was described using indicator species analysis

(ISA; Dufrêne & Legendre, 1997). Monte Carlo test of significance

was based on 1000 randomizations (

P

<

0.05). Multiresponse

permutation procedures (MRPP; Zimmerman

et al.

, 1985) were

used to test whether invaded and uninvaded stands exhibited

distinct species assemblages. ISA and MRPP were carried out using

PC-ORD® version 4.25 software (McCune & Mefford, 1999).

Analysis of the invasion gradient

To search whether an increasing

P. serotina

density was associated

with increasing environmental, vegetation, and disturbance

changes, we conducted Spearman rank correlations (

P

<

0.05)

between

P. serotina

density (i.e. cumulated cover values in the

tree and shrub layers and stem density) and the ‘delta variables’ of

the paired stands. Uninvaded stands being taken as the reference,

‘delta variables’ quantify the deviation of each variable from its

reference value, which is expected to be the consequence (or the

cause) of the invasion. They were obtained for each variable by

subtracting the value recorded in the uninvaded stand by the one

of its paired invaded stands.

Analysis of the ecosystem–disturbance–invasion interplay

We pooled the ‘delta variables’ into three sets (Table 1): diversity

variables (D), environmental variables (E), and disturbance

variables (P). A fourth set included invasion variables (I):

P. serotina

cover in each vegetation layer, basal area, stem density,

basal area/stem density ratio, canopy height, diameter, and

population age. To reduce collinearity among variables and to

provide synthetic variables, we performed a Principal Com-

ponents Analysis (PCA) on each of these sets, and retained the

Invasion–disturbance–ecosystem interactions

© 2007 The Authors

Diversity and Distributions

,

14

, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd

207

Table 1 Mean (± standard error) values of diversity, environmental, and disturbance variables in pairs of invaded and uninvaded stands (n = 32) and correlation between the delta values and cumulated cover value of Prunus serotina in the tree and shrub layers (PCTS) or P. serotina stem density (PSD). Z is the value of the Wilcoxon’s rank test. *P < 0.05, **P < 0.01, ***P < 0.001.

Invaded Uninvaded Z

Spearman rank

correlations with

PCTS PSD

Diversity variables (D)

Species richness

Tree layer 3.8 ± 0.2 3.7 ± 0.3 −0.332 −0.053 −0.043

Shrub layer 3.2 ± 0.3 2.5 ± 0.3 −1.886 −0.127 −0.021

Undershrub layer 5 ± 0.5 2.9 ± 0.4 −3.066** −0.358* −0.326

Herb layer 28.6 ± 2.2 33.5 ± 3 −1.029 −0.165 −0.049

Total 29.6 ± 2.2 34.5 ± 3 −1.019 −0.167 −0.059

Environment variables (E)

Sum of species covers without P. serotina in vegetation layers (%)

Tree layer 72.5 ± 4.8 104.7 ± 5.1 −3.586*** −0.263 −0.331

Shrub layer 19.9 ± 3.3 19.3 ± 4.1 −0.355 −0.068 0.003

Undershrub layer 19.4 ± 3.7 23.3 ± 5.6 −0.184 −0.153 −0.205

Herb layer 184.3 ± 18.2 191 ± 22.6 −0.159 −0.183 −0.108

Canopy tree species cover (%)

Fagus sylvatica 22 ± 4.6 36.9 ± 5.7 −2.471* 0.06 0.089

Quercus robur 17.9 ± 3.3 24.3 ± 4.5 −1.428 −0.191 −0.265

Carpinus betulus 19.7 ± 4.5 21.9 ± 4.1 −0.52 −0.438* −0.399*

Pinus sylvestris 6.7 ± 2.9 6.7 ± 3.1 −0.106 0.139 0.033

Soil properties

Litter thickness (cm) 3 ± 0. 4.4 ± 0.7 −3.09** −0.09 −0.101

Litter necromass (kg m–2) 0.9 ± 0.1 1 ± 0.1 −1.253 0.117 −0.103

Organic horizon thickness (cm) 2.2 ± 0.3 2.5 ± 0.4 −0.458 0.356* 0.103

Litter/organic horizon ratio 1.8 ± 0.2 3.5 ± 0.6 −2.332* −0.291 −0.223

Number of Fe3+ spots 0 ± 0 0.4 ± 0.2 −2.428* −0.17 −0.229

P2O5 (g/1000) 3.2 ± 0.31 2.1 ± 0.24 −2.489* 0.351* 0.163

pH-H20 4.7 ± 0.1 5.5 ± 0.2 −3.6*** −0.079 0.044

Organic matter (g/100) 10.4 ± 1 11.2 ± 0.8 −0.552 0.159 0.238

Total nitrogen (g/1000) 1.2 ± 0.1 1.4 ± 0.1 −1.725 0.143 0.261

NH4+ (g/1000) 0.02 ± 0.002 0.019 ± 0.001 −1.157 0.101 −0.028

NO3− (g/1000) 0.006 ± 0 0.006 ± 0 −0.103 −0.26 −0.411*

Organic C (g/100) 5.6 ± 0.7 6.4 ± 0.7 −0.823 0.407* 0.401*

C/N ratio 47.4 ± 5 48.8 ± 5.3 −0.131 0.245 0.258

Disturbance variables (P)

Volume of tree falls (m3 ha–1) collected per decade

1995–2004 5 ± 1 4.2 ± 1.3 −0.817 0.127 0.163

1985–1994 53.5 ± 15.6 26 ± 9.8 −1.968* 0.21 0.055

1975–1984 6.8 ± 2.4 5.7 ± 1.9 −0.057 0.351* 0.369*

1965–1974 1 ± 0.4 8.7 ± 3.3 −1.601 0.309 0.442*

1955–1964 16.7 ± 9.3 14.6 ± 5.8 −0.896 0.323 0.42*

Volume of cut trees (m3 ha–1) extracted per decade

1995–2004 28.8 ± 3.5 56.2 ± 9.9 −2.273* 0.012 0.072

1985–1994 44 ± 11 35.1 ± 10.8 −0.504 −0.02 −0.068

1975–1984 30.3 ± 12.65 24.2 ± 8.6 −0.156 0.365* 0.422*

1965–1974 42 ± 10.54 147.64 ± 30.9 −2.883** 0.086 0.17

1955–1964 34.9 ± 13.26 67.1 ± 21.1 −1.111 0.11 0.18

Bare soil cover (%) 8.6 ± 1.9 7.8 ± 1.4 −0.054 −0.16 0.02

Number of stumps 8.3 ± 1 9.5 ± 1 −1.076 −0.153 −0.121

Wild boar-disturbed area (%) 9.4 ± 3.7 1.7 ± 0.6 −2.405* −0.078 0.013

O. Chabrerie et al.

© 2007 The Authors208 Diversity and Distributions, 14, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd

first three PCA axes for subsequent analyses. Then, to help in

discriminating among the three hypothesized models (i.e. driver,

passenger, opportunist), the four PCA-derived groups of

synthetic variables were subjected to three consecutive redundancy

analyses with variation partitioning (RDA-VP): (1) P, E, and I as

explanatory variables and D as response variable; (2) P, D, and I

as explanatory variables and E as response variable; and (3) P, E,

and D as explanatory variables and I as response variable (Borcard

& Legendre, 2002; Økland, 2003). The explained variation of

each group was expressed as fractions of the total variation

explained by all groups of synthetic variables, after setting the

latter to 100%.

Results of RDA-VP were interpreted using the results pre-

viously obtained (ISA, MRPP, paired comparisons, correlations

with P. serotina density) and Hill’s criteria of causality (Hill,

1965): (1) statistical strength of the association (i.e. the stronger

the differences between paired stands, the higher the probability

of a causal relationship between P. serotina and the observed

differences); (2) reproductibility (i.e. our results should be

consistent with those previously reported in the literature for

P. serotina); (3) specificity (i.e. characteristics of invaded stands

should not be found in uninvaded stands); (4) temporality

(i.e. only the factors that undoubtfully occur before P. serotina

invasion can be a cause of the latter); (5) biological gradient

(i.e. the effect variables should strongly correlate with invasion

density); (6) plausibility (i.e. the derived relationships should not

be counterintuitive); (7) consistency (i.e. the derived relation-

ships should be consistent with biological invasion theory);

(8) experimentation (i.e. the derived relationships should be

supported by available experimental results); and (9) analogy

(i.e. our results should be similar to those reported for other case

studies of plant invasion).

RESULTS

Differences between paired stands

In invaded stands P. serotina cover reached (means ± SE)

29.7 ± 5.1%, 30.3 ± 3.5%, 32.1 ± 3.6%, and 41.4 ± 4.7%, in the

tree, shrub, undershrub, and herb layer, respectively. Basal area,

canopy height, and stem density averaged 3.25 ± 0.57 m2 ha–1,

10.18 ± 0.62 m, and 687 ± 198 ha–1, respectively. The oldest

individual was 24 ± 2 years old on average.

Paired comparisons revealed few differences between invaded

and uninvaded stands (Table 1). Uninvaded stands exhibited a

higher tree cover and proportion of F. sylvatica, a thicker litter

layer (and a lower L/OH ratio), a higher pH value and more iron

spots, and higher amounts of cut trees in both 1965–74 and

1995–2004 decades. Invaded stands differed by a species-richer

undershrub layer, higher contents of available phosphorus, a

higher quantity of tree falls in the 1985–94 decade, and more

decaying wood and wild boar-associated bare soil covers.

Only 24 (14%) of the 170 species recorded in the herb layer

showed a preferential distribution among stands, but indicator

values were always low (< 64), and no species was exclusive of

one stand type (Table 2). Results of the MRPP (see Methods)

confirmed that species assemblages only slightly differed between

invaded and uninvaded stands (T = –3.41; P = 0.00899; A = 0.012).

Correlations with the invasion gradient

Species richness of the undershrub layer and Carpinus betulus

cover were significantly decreasing with P. serotina density, while

OH thickness, available phosphorus, organic carbon were

increasing (Table 1). The quantity of tree falls in the 1955–84

period, as well as the amount of cut trees in the 1975–84 decade,

positively correlated with the invasion gradient.

Analysis of the ecosystem–disturbance–invasion interplay

When plant diversity (D) was taken as the response variable,

RDA-VP (see Methods) indicated that most of the variation is

explained by environmental variables and, to a lesser degree, to

disturbance variables (Fig. 2). Invasion variables alone explained

only 12% of the total variation. Similarly, disturbance variables

alone explained 10% of this variation, but shared 24 and 11% of

the variation with environment and invasion, respectively.

When invasion (I) was taken as the response variable, RDA-VP

Table 2 Results of indicator species analysis conducted on the 170 species of the herb layer in pairs of invaded and uninvaded stands (n = 32) (Monte Carlo test: *P < 0.05, **P < 0.01, ***P ≤ 0.001).

Indicator species of invaded plots

Indicator value

Sorbus aucuparia 63.7 ***

Milium effusum 62.7 *

Convallaria maialis 60.2 **

Hyacinthoides non-scripta 59.7 **

Dryopteris carthusiana 58.6 *

Moehringia trinervia 57.0 ***

Calystegia sepium 45.7 **

Indicator species of uninvaded plots

Indicator value

Carex sylvatica 63.6 ***

Fraxinus excelsior 57.0 ***

Prunus avium 52.8 **

Brachypodium sylvaticum 52.7 **

Circaea lutetiana 46.5 *

Euphorbia amygdaloides 44.1 *

Lamium galeobdolon 42.8 **

Acer campestre 42.3 **

Arum maculatum 40.0 ***

Carex remota 35.7 *

Rosa arvensis 33.9 **

Vicia sepium 27.6 *

Carex pendula 24.8 **

Festuca gigantea 21.9 **

Crataegus laevigata 21.9 *

Corylus avellana 20.3 *

Invasion–disturbance–ecosystem interactions

© 2007 The AuthorsDiversity and Distributions, 14, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd 209

provided similar fractions of explained variation for the three

other sets of variables, but the total reached only 30%. When

environment (E) was taken as the response variable, the propor-

tion of the variation explained strongly decreased from diversity

to invasion, through disturbance variables. In the last two analyses,

a significant part of the variation was shared between disturbance

and diversity variables.

DISCUSSION

Herb species assemblages only differed slightly between invaded

and uninvaded stands, and the two sets of indicator species were

composed of species with very contrasting ecological require-

ments, life forms, and strategies. This indicates non-equilibrium

species assemblages (Gigon & Leutert, 1996) that are typical

of disturbance-driven ecosystems (Decocq, 2002). Invaded

stands were rather characterized by true forest species (e.g.

Hyacinthoides non-scripta, Convallaria majalis, Dryopteris

carthusiana) and uninvaded stands by edge species (e.g. Brachy-

podium sylvaticum, Vicia sepium), hygrophilous sedges (e.g.

Carex remota, C. pendula), and tree or shrub regenerations

(e.g. Fraxinus excelsior, Prunus avium). This suggests that

microclimate is closer to a true forest microclimate (including

low light availability at the forest floor, high hygrometry, and low

thermic variations) in invaded stands (Chen et al., 1999). This

would also explain the increase of both OH thickness and topsoil

organic carbon content along the invasion gradient.

MRPP confirmed that differences in species assemblages are of

low ecological significance (see the weak chance-corrected

within-group agreement A), as already reported for P. serotina-

invaded forests in Belgium (Godefroid et al., 2005; Verheyen

et al., 2007). However, these tenuous differences may be due to a

certain time-lag in the response of plant communities to the

invader’s dominance. Changes can be cumulative and slow,

taking many years to play out, and preventing the full effects of

an invader from appearing for many years (Crooks, 2005; Strayer

et al., 2006), especially in forest ecosystems. Hence, our short-

term study may have failed to capture such long-term changes.

Furthermore, it should be noted that the herb layer cover was

always low, suggesting that herbaceous communities maybe

unsaturated, probably because of the recurrent disturbances

that characterize managed forests. Using a similar approach,

Hejda & PyÍek (2006) also failed to find any effect of the invasive

Impatiens glandulifera on species diversity in species-poor

communities that were regularly flooding-disturbed.

RDA-VP shows that invasion variables alone poorly explain

plant diversity at the whole stand scale and vice-versa, indicating

weak direct relationships between P. serotina dominance and

diversity. Conversely, environmental variables provide the highest

explanatory power for plant diversity, even if there was considerable

overlap in the variation accounted for by disturbance variables.

The latter alone also explained a significant part of the total

variation in environmental variables (c. 17%), suggesting that

diversity is mainly driven by both direct and disturbance-

mediated environmental factors. Environmental and disturbance

variables explained a similar amount of variation in the invasion

data set (35 and 31%, respectively), but diversity alone and diversity

together with disturbances accounted for a significant part of

the variation, too. This indicates that invasion is controlled by

a host of factors at the stand scale, including soil properties,

disturbances, and resident plant community diversity.

Among the environmental variables differing between

invaded and uninvaded stands, topsoil phosphorus content was

the only one correlating with P. serotina density, suggesting a

causal relationship (Hill, 1965). This is consistent with Lorenz

et al. (2004) and was expected since the species has often been

planted for soil amelioration purposes, especially on nutrient-

poor soils (Starfinger, 1997). Changes in soil functions following

Figure 2 Venn diagrams showing the relative importance of the three sets of explanatory variables as summarized from redundancy analyses with variation partitioning for (a) the delta diversity matrix D, (b) the delta invasion matrix I, and (c) the delta environmental matrix E, successively used as the set of response variables. Both unique and shared components are shown, given in percentage, expressing the part of the total explained variation for each set of explanatory variables (i.e. values add up to 100%).

O. Chabrerie et al.

© 2007 The Authors210 Diversity and Distributions, 14, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd

non-native invasion have even been recently recognized as one

of the most widespread mechanisms for altering ecosystem

processes (Ehrenfeld et al., 2001; Levine et al., 2003; Vanderhoeven

et al., 2005). The other variables (litter thickness, L/OH ratio, soil

pH, number of iron spots, tree cover, beech cover) should rather

be considered as components of stand resistance to invasion. For

example, P. serotina establishment is known to be restricted by a

closed canopy, especially when the latter is composed of shade

providers, e.g. F. sylvatica (Starfinger, 1997). Those also accumulate

a thick litter layer able to impair seedling emergence (Sydes &

Grime, 1981).

As the average age of the oldest P. serotina trees was 24 years

old, invading populations would have established around 1980

the latest in the surveyed stands. This chronology is consistent

with the increase of both fallen trees from 1955 to 1984, and cut

trees during the 1975–84 decade. Surprisingly, wild boar-

disturbed soil areas were much more widespread in invaded

stands. Wild boars probably forage preferentially in invaded

areas that provide them more food (P. serotina fruits) and more

shelter than is provided by uninvaded stands. This suggests a

positive feedback on P. serotina invasion through the creation of

bare soil areas that promote seedling emergence and banking,

easing their spread and constraining potential competitors

(Mack et al., 2000; Crooks, 2002).

Among the diversity variables, only species richness of the

undershrub layer decreased with P. serotina density. This suggests

that most species establishing in disturbance-induced canopy

gaps are subsequently shaded out by the invader (Closset-Kopp

et al., 2007), as confirmed by the high proportion of tree and shrub

regenerations among indicator species of uninvaded stands.

Overall, consistently with Gurevitch & Padilla (2004), our

findings did not support a direct causal relationship between

plant invader dominance and native species diversity loss. This

link seems mostly mediated by environmental factors (including

soil properties and stand structure), that are partly controlled by

past disturbances. The latter directly and indirectly control stand

invasibility. Similar results have been reported with the invasive

tree Syzygium jambos in Puerto Rico (Brown et al., 2006). However,

we should distinguish between irregular, unpredictable (natural)

and cyclic, predictable (anthropogenic) disturbances. In

managed forests, management typically represents a disturbance

regime in which frequency, intensity, and spatial extent of

thinnings/clearcuttings are planned and thus, periodically affect

vegetation. Such ‘chronic’ disturbances are part of the history of

managed ecosystems and as such, should be better regarded as a

permanent constraint on plant communities rather than an

external disturbance (Briske et al., 2003; Decocq et al., 2004).

Conversely, storms occur unpredictably in both space and time,

and act as a real disturbance. Consequently, such ecosystems are

artificially maintained at ‘non-equilibrium’ and plant communities

are mainly driven by non-interactive factors and immigration

processes, as shown by the ecological inconsistency of species

assemblages. In such communities any species – native or

non-native – has a chance to find a vacant niche and to fill it in

(Levine, 2000; Brown & Peet, 2003). Consistent with the

concepts of ‘niche opportunity’ (Shea & Chesson, 2002) and

‘invasion window’ (Johnstone, 1986), disturbances free space

and release resources, providing opportunities for the invader

that just needs to be present (e.g. as seeds or suppressed

seedlings) in the right place at the right time to successfully

invade a recipient ecosystem.

To sum up our results, disturbances directly influence the

three ‘passengers’ of the system: (1) plant diversity, by disrupting

the resident plant community, (2) invasion spread, by releasing

seedlings from suppression, and (3) environmental conditions,

by altering stand structure and freeing resources. In turn, the new

environment affects both the invasion (i.e. the invader captures

released resources and thus increases its density) and the recipient

community diversity (i.e. by changing the relative abundance of

species). Subsequently, a series of feedbacks is likely to occur, in

which the invader may ‘drive’ plant composition, directly, by

shading out light-demanding species (especially seedlings of

other tree species), and/or indirectly, through the environment

(by modifying topsoil chemical properties, especially phosphorus

content) and/or disturbances (by promoting wild boar-induced

soil disturbances). Similarly, species composition changes may

provide ‘opportunities’ for the non-natives (i.e. free space can

host ‘opportunists’).

In this study we found weak short-term, direct effects of the

invader on the recipient plant community and topsoil chemical

properties. Consistently with MacDougall & Turkington (2005),

P. serotina was primarily passenger to disturbances, together with

diversity and environmental changes, and naturalized with little

impact on the recipient ecosystem. However, the disturbance–

ecosystem–invasion interplay cannot be reduced to a simple

‘driver/passenger/opportunist’ model, but rather refers to

an interacting array of factors. After Didham et al. (2005), we

conclude that driver and passenger – but also opportunist – models

should be considered as extreme cases of a general model into

which the respective importance of each between-component

relationship depends on both invasive species and recipient

ecosystem under consideration, and varies over time.

ACKNOWLEDGEMENTS

We are very grateful to associate editor Marcel Rejmánek and the

four anonymous referees for their helpful comments on the

initial manuscript. We are indebted to Sidonie Perrin, Marie

Deville, and Justin Kassi for their help in data collection. We

thank the French ‘Office National des Forêts’ for their facilities

during fieldworks. This study was financially supported by the

French ‘Ministère de l’Écologie et du Développement Durable’

(INVABIO II program, CR no. 09-D/2003).

REFERENCES

Alpert, P., Bone, E. & Holzapfel, C. (2000) Invasiveness, invasibility

and the role of environmental stress in the spread of non-

native plants. Perspectives in Plant Ecology, Evolution and

Systematics, 3, 52–66.

Aubert, G. (1978) Méthodes d’analyses des sols. Centre Régional

de Documentation Pédagogique (CRDP), Marseille, France.

Invasion–disturbance–ecosystem interactions

© 2007 The AuthorsDiversity and Distributions, 14, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd 211

Borcard, D. & Legendre, P. (2002) All-scale spatial analysis of

ecological data by means of principal coordinates of neighbour

matrices. Ecological Modelling, 153, 51–68.

Braun-Blanquet, J. (1964) Pflanzensoziologie grünzüge der vegeta-

tionsgrunde. Springer, Berlin.

Brethes, A., Brun, J.J., Jabiol, B., Ponge, J. & Toutain, F. (1995)

Classification of forest humus forms: a French proposal.

Annales des Sciences Forestières, 52, 535–546.

Briske, D.D., Fuhlendorf, S.D. & Smeins, F.E. (2003) Vegetation

dynamics on rangelands: a critique of the current paradigms.

Journal of Applied Ecology, 40, 601–614.

Brown, R.L. & Peet, R.K. (2003) Diversity and invasibility of

southern Appalachian plant communities. Ecology, 84, 32–

39.

Brown, K.A., Scatena, F.N. & Gurevitch, J. (2006) Effects of an

invasive tree on community structure and diversity in a tropical

forest in Puerto Rico. Forest Ecology and Management, 226,

145–152.

Chabrerie, O., Hoeblich, H. & Decocq, G. (2007a) Déterminisme

et conséquences écologiques de la dynamique invasive du

cerisier tardif (Prunus serotina Ehrh.) sur les communautés

végétales de la forêt de Compiègne. Acta Botanica Gallica, 154,

383–394.

Chabrerie, O., Roulier, F., Hoeblich, H., Sebert-Cuvillier, E.,

Closset-Kopp, D., Leblanc, I., Jaminon, J. & Decocq, G.

(2007b) Defining patch mosaic functional types to predict

invasion patterns in a forest landscape. Ecological Applications,

17, 464–481.

Chapin, F.S., Zaveleta, E.S., Eviner, V.T., Naylor, R.L.,

Vitousek, P.M., Lavorel, S., Reynolds, H.L., Hooper, D.U.,

Sala, O.E., Hobbie, S.E., Mack, M.C. & Diaz, S. (2000) Con-

sequences of changing biotic diversity. Nature, 405, 234–242.

Chen, J., Saunders, S.C., Crow, T.R., Naiman, R.J., Brosofske, K.D.,

Mroz, G.D., Brookshire, B.L. & Franklin, J.F. (1999) Microclimate

in forest ecosystem and landscape ecology. Bioscience, 49,

288–297.

Closset-Kopp, D., Chabrerie, O., Valentin, B., Delachapelle, H. &

Decocq, G. (2007) When Oskar meets Alice: lack of trade-off

in r/K-strategies makes Prunus serotina Ehrh. a successful

invader of European forests. Forest Ecology and Management,

247, 12–130.

Crawley, M.J. (1987) What makes a community invasible?

Colonization, succession and stability (ed. by A.J. Gray, M.J.

Crawley and P.J. Edwards), pp. 429–453. Blackwell Scientific

Publications, Oxford, UK.

Cronk, Q.C.B. & Fuller, J. (1995) Plant invaders: the threat to

natural ecosystems. World Wide Fund for Nature, Chapman &

Hall, London.

Crooks, J.A. (2002) Characterizing ecosystem-level consequences

of biological invasions: the role of ecosystem engineers. Oikos,

97, 153–166.

Crooks, J.A. (2005) Lag times and exotic species: the ecology and

management of biological invasions in slow-motion.

Ecoscience, 12, 316–329.

Cuddington, K. & Hastings, A. (2004) Invasive engineers. Ecological

Modelling, 178, 335–347.

Davis, M.A. (2003) Biotic globalization: does competition from

introduced species threaten biodiversity? Bioscience, 53, 481–

489.

Davis, M.A., Grime, J.P. & Thompson, K. (2000) Fluctuating

resources in plant communities: a general theory of invasibility.

Journal of Ecology, 88, 528–534.

Decocq, G. (2000) The masking effect of silviculture on substrate-

induced plant diversity in oak-hornbeam forests from northern

France. Biodiversity and Conservation, 9, 1467–1491.

Decocq, G. (2002) Patterns of plant species and community

diversity at different organization levels in a riparian landscape.

Journal of Vegetation Science, 13, 91–106.

Decocq, G., Aubert, M., Dupont, F., Alard, D., Saguez, R.,

Wattez-Franger, A., de Foucault, B., Delelis-Dusollier, A. &

Bardat, J. (2004) Plant diversity in a managed temperate forest:

understorey response to two silvicultural systems. Journal of

Applied Ecology, 41, 1065–1079.

Didham, R.K., Tylianakis, J.M., Hutchison, M.A., Ewers, R.M. &

Gemmell, N.J. (2005) Are invasive species the drivers of

ecological changes? Trends in Ecology and Evolution, 20, 470–474.

Dietz, H. & Steinlein, T. (2004) Recent advances in understanding

plant invasions. Progress in Botany, 65, 539–573.

Dufrêne, M. & Legendre, P. (1997) Species assemblages and

indicator species: the need for a flexible asymmetrical approach.

Ecological Monographs, 67, 345–366.

Ehrenfeld, J.G., Kourtev, P. & Huang, W. (2001) Changes in soil

functions following invasions of exotic understory plants in

deciduous forests. Ecological Applications, 11, 1287–1300.

Elton, C.S. (1958) The ecology of invasion by animals and plants.

Chapman & Hall, London.

Gagnon, R. & Morin, H. (2004) WinDENDRO, version 2004b for

tree ring analysis. Régent Instruments Inc., Chicontimi, QC,

Canada.

Gigon, A. & Leutert, A.G. (1996) The dynamic keyhole-key

model of coexistence to explain diversity of plants in limestone

and other grasslands. Journal of Vegetation Science, 7, 29–40.

Gilbert, B. & Lechowicz, M.J. (2005) Invasibility and abiotic

gradients: the positive correlation between native and exotic

plant diversity. Ecology, 87, 1848–1855.

Godefroid, S., Phartyal, S.S., Weyembergh, G. & Koedam, N.

(2005) Ecological factors controlling the abundance of non-

native invasive black cherry (Prunus serotina) in deciduous

forest understory in Belgium. Forest Ecology and Management,

210, 91–105.

Gurevitch, J. & Padilla, D.K. (2004) Are invasive species a major

cause of extinctions? Trends in Ecology and Evolution, 19, 470–

474.

Halpern, C.B. & Spies, T.A. (1995) Plant species diversity in

natural and managed forests of the Pacific Northwest. Ecological

Applications, 5, 913–934.

Hejda, M. & PyÍek, P. (2006) What is the impact of Impatiens

glandulifera on species diversity of invaded riparian vegetation?

Biological Conservation, 132, 143–152.

Hill, A.B. (1965) The environment and disease: association or

causation? Proceedings of the Royal Society of Medicine, 58,

295–300.

O. Chabrerie et al.

© 2007 The Authors212 Diversity and Distributions, 14, 204–212, Journal compilation © 2007 Blackwell Publishing Ltd

Hobbs, R.J. & Huenneke, L.F. (1992) Disturbance, diversity and

invasion: implications for conservation. Conservation Biology,

6, 324–337.

Johnstone, I.M. (1986) Plant invasion windows: a time-based

classification of invasion potential. Biological Review, 61, 369–

394.

Kennedy, T.A., Naeem, S., Howe, K.M., Knops, J.M.H.,

Tilman, D. & Reich, P. (2002) Biodiversity as a barrier to

ecological invasion. Nature, 417, 636–638.

Lambinon, J., De Langhe, J.-E., Delvosalle, L. & Duvigneaud, J.

(2004) Nouvelle flore de la Belgique, du Grand-Duché de

Luxembourg, du Nord de la France et des régions voisines, 5th

edn. Patrimoine du Jardin botanique national de Belgique,

Meise, Belgium.

Levine, J.E. (2000) Species diversity and biological invasions:

relating local process to community pattern. Science, 288, 852–

854.

Levine, J.M. & D’Antonio, C.M. (1999) Elton revisited: a review

of evidence linking diversity and invasibility. Oikos, 87, 15–26.

Levine, J.M., Vilà, M., D’Antonio, C.M., Dukes, J.S., Grigulis, K.

& Lavorel, S. (2003) Mechanisms underlying the impacts of

exotic plant invasions. Proceeding of the Royal Society of London

Series B, Biological Sciences, 270, 775–781.

Lorenz, K., Preston, C.M., Krumrei, S. & Feger, K.-H. (2004)

Decomposition of needle/leaf litter from Scots pine, black

cherry, common oak and European beech at a conurbation

forest site. European Journal of Forest Research, 123, 177–188.

MacDougall, A.S. & Turkington, R. (2005) Are exotic species the

drivers or passengers of ecological change in highly disturbed

plant communities? Ecology, 86, 42–55.

Mack, R.N., Simberloff, D., Lonsdale, W.M., Evans, H., Clout, M.

& Bazzaz, F.A. (2000) Biotic invasions: causes, epidemiology,

global consequences, and control. Ecological Applications, 10,

689–710.

McCune, B. & Mefford, M.J. (1999) PC-ORD. Multivariate analysis

of ecological data, Version 4, 1st edn. MjM Software Design,

Gleneden Beach, Oregon, USA.

Melbourne, B.A., Cornell, H.V., Davies, K., Dugaw, C.,

Elmendorf, S., Freestone, A., Hall, R., Harrison, S., Hastings, A.,

Holland, M., Holyoak, M., Lambrinos, J., Moore, K. &

Yokomizo, H. (2007) Invasion in a heterogeneous world:

resistance, coexistence or hostile takeover? Ecology Letters, 10,

77–94.

Naeem, S., Knops, J.M.H., Tilman, D., Howe, K.M., Kennedy, T.

& Gale, S. (2000) Plant diversity increases resistance to

invasion in the absence of covarying extrinsic factors. Oikos,

91, 97–108.

Økland, R.H. (2003) Partitioning the variation in a plot-by-

species data matrix that is related to n sets of explanatory

variables. Journal of Vegetation Science, 14, 693–700.

Ortega, Y.K. & Pearson, D.E. (2005) Weak versus strong invaders

of natural plant communities: assessing invasibility and

impact. Ecological Applications, 15, 651–661.

Parker, I.M., Simberloff, D., Lonsdale, W.M., Goodell, K.,

Wonham, M., Kareiva, P.M., Williamson, M.H., Von Holle, B.,

Moyle, P.B., Byers, J.E. & Goldwasser, L. (1999) Impact: toward

a framework for understanding the ecological effects of invaders.

Biological Invasions, 1, 3–19.

Richardson, D.M., PyÍek, P., Rejmánek, M., Barbour, M.G.,

Panetta, F.D. & West, C.J. (2000) Naturalization and invasion

of alien plants: concepts and definitions. Diversity and Distribu-

tions, 6, 93–107.

Sax, D.F. & Gaines, S.D. (2003) Species diversity: from global

decreases to local increases. Trends in Ecology and Evolution,

18, 561–566.

Shea, K. & Chesson, P. (2002) Community ecology theory as a

framework for biological invasions. Trends in Ecology and

Evolution, 17, 170–176.

Starfinger, U. (1997) Introduction and naturalization of Prunus

serotina in central Europe. Plant invasions: studies from North

America and Europe (ed. by J.H. Brock, M. Wade, P. PyÍek and

D. Green), pp. 161–171. Backhuys, Leiden, the Netherlands.

Strayer, D.L., Eviner, V.T., Jeschke, J.M. & Pace, M.L. (2006)

Understanding the long-term effects of species invasions.

Trends in Ecology and Evolution, 21, 645–651.

Sydes, C. & Grime, J.P. (1981) Effects of tree leaf litter on

herbaceous vegetation in deciduous woodland. II. Experimental

investigation. Journal of Ecology, 69, 249–262.

Tilman, D. (1997) Community invasibility, recruitment limitation,

and grassland biodiversity. Ecology, 78, 81–92.

Vanderhoeven, S., Dassonville, N. & Meerts, P. (2005) Increased

topsoil mineral nutrient concentrations under exotic invasive

plants in Belgium. Plant and Soil, 275, 169–179.

Verheyen, K., Vanhellemont, M., Stock, T. & Hermy, M. (2007)

Predicting patterns of invasion by Prunus serotina Ehrh. in

Flanders (Belgium) and its impact on the forest understory

community. Diversity and Distributions, 13, 487–497.

Vitousek, P.M. (1990) Biological invasions and ecosystem

processes: toward an integration of population biology and

ecosystem studies. Oikos, 57, 7–13.

Williamson, M. (1999) Invasions. Ecography, 22, 5–12.

Zimmerman, G.M., Goetz, H. & Mielke, P.W.J. (1985) Use of an

improved statistical method for group comparisons to study

the effects of prairie fire. Ecology, 66, 606–611.

Editor: Marcel Rejmánek