Navigation – Plan du site

Channel vertical mobility, hydro-geomorphic disturbances and understory vegetation in floodplain forests of the Ain River (France)

Mobilité verticale du chenal, perturbations hydrogéomorphologiques et végétation dans les forêts alluviales de l’Ain (France)
Simon Dufour et Hervé Piégay
p. 371-386

Résumés

L’impact des perturbations liées aux crues dans les écosystèmes alluviaux a été étudié au cours des trois dernières décennies dans de nombreux contextes géographiques. Mais peu d’auteurs intègrent la mobilité du chenal (érosion latérale, incision/exhaussement) comme facteur de contrôle du régime de perturbations. Dans cette étude nous analysons 1) l’impact de la mobilité du chenal sur le régime de perturbations subit par unités végétales apparemment similaires au niveau de leur strate arborée et 2) l’influence des différences de régime de perturbations (fréquence et intensité) sur la végétation de la strate basse de ces unités. La position spatiale des unités forestières dans le corridor (distance au chenal) et la mobilité verticale du chenal au droit de ces unités contrôlent le régime de perturbations et entrainent des différences locales significatives. Dans des forêts apparemment similaires en termes de strate arborée, des différences fines au niveau de la fréquence des perturbations se traduisent par des différences dans la structure de la strate basse (diversité, fréquence des espèces à courte durée de vie, fréquence des plantules d’espèces arborées). Nos résultats montrent également que la fréquence des perturbations n’est pas suffisante pour expliquer les différences observées. Dans la gamme de perturbations observées, l’apport et le remaniement des sédiments sont également nécessaires pour affecter la strate basse. Ces processus sédimentaires maintiennent des espaces ouverts, limitent le développement des espèces qui dominent les sites non perturbés et accroissent la diversité locale (pour une richesse comparable). Enfin, nous observons que les effets d’une crue ne sont pas nécessairement les mêmes que ceux enregistrés à moyen terme, et l’impact à long terme reste difficile à évaluer sans un suivi approprié.

Haut de page

Notes de la rédaction

Article soumis le 25 janvier 2010, accepté le 13 juin 2010

Texte intégral

This project was funded by the Agence de l’Eau Rhône-Méditerranée-Corse, the Office National des Forêts, the Centre National de la Recherche Scientifique and the university of Lyons.We would like to thank Anne-Julia Rollet and Norbert Landon for providing help in field and topographical data. We are very grateful to John Stella and Alex Fremier for English improvement, Eric Tabacchi and anonymous reviewers for their useful suggestions and corrections.

Introduction

1Disturbance affects most ecosystems and an abundant literature has been produced within the last three decades (White and Jentsch, 2001). Usually, a disturbance is defined as any discrete event (e.g., fire, hurricane, grazing, floods) that modifies ecosystem, community and population structures by affecting the physical environment such as substrate availability (Picket and White, 1985). For plants, disturbance results in an initial decrease in biomass through destruction (Grime, 1977) and creates an increase in resource availability for organisms that survive or arrive within the community after disturbance (Van der Maarel, 1993). Whatever the spatial and temporal scale considered, disturbance events are one of the main causes of spatial heterogeneity in ecosystems, which influences biological diversity (see for example P.P. Sale, 1977; E.J. Chaneton and J.M. Facelli, 1991; G. Vivian-Smith, 1997; C. Syms and G.P. Jones, 2000). Because disturbances decrease competition for resources, they favour pioneer over competitive species. Disturbance tends to decrease competitive interactions and also competitive ability of species and thus modifies the dominance hierarchy within the community (Suding and Goldberg, 2001). Disturbance favours specific functional groups as R-selective species (Pianka, 1970; Whittaker and Goodman, 1979) and ruderal species (Rsensu Grime, 2001) with particular biological traits. For herbaceous vegetation in a riparian context, E.S. Menges and D.M. Waller (1983) observed that disturbance favours traits such as fecundity, short lifespan and tends to limit stoloniferous growth. At the species level, ecophysiological responses (such as an increase in photosynthetic efficiency and biomass) are also generated by a disturbance such as burial by sediment (Perumal and Maun, 2006). Concerning diversity, the Intermediate Disturbance Hypothesis (IDH), formalised by J.H. Connell (1978), predicts highest specific richness within habitats where disturbance is intermediate (see D.M. Wilkinson, 1999 for emergence of the IDH). This hypothesis has been widely studied and discussed in different contexts and systems (e.g., Fox, 1979; Huston, 1979; Wootton, 1998; Mackey and Currie, 2001; Roxburgh et al., 2004). In riparian corridors, the role of disturbance in maintaining high diversity has been confirmed at the river corridor scale (Tabacchi et al., 1996; Gilvear et al., 2000), as well as at the reach scale with a diversity of disturbance regimes between landforms (Pollock et al., 1998; Kamisako et al., 2007). At more local scales, for example one landform or one vegetation unit type, studies are less common and tend to focus on isolated former channels in floodplains (e.g., Bornette et al., 1998) or low dynamic systems (Vivian-Smith, 1997).

2In fluvial corridors, channel movement and sediment dynamics have a huge control over plant spatial complexity, community distribution, and biological diversity (Fonda, 1974; Pautou and Décamps, 1985; Hupp and Osterkamp, 1985; Salo et al., 1986; Bornette and Amoros, 1991; Hughes, 1997; Van Coller et al., 1997; Ward, 1998; Suzuki et al., 2002; Steiger et al., 2005; Bornette et al., 2008). Thus, in a floodplain and wetland contexts, studies mainly discuss disturbance consequences in terms of flow and/or sedimentation effects on vegetation (Menges and Waller, 1983; Nilsson, 1987; Bornette and Amoros, 1996; Pollock et al., 1998; Vervuren et al., 2003; Wintle and Kirkpatrick, 2007). Disturbance in riparian systems can be essentially characterised by a double gradient of flood frequency and intensity. On any given geomorphic surface, disturbance frequency is monotonically and negatively correlated to discharge required to inundate this surface; therefore inundation discharge is a close proxy for flood disturbance frequency. Disturbance intensity, the second gradient considered, is linked with flow velocity, flow depth and the grain size of sediment deposits, all of which increase under higher levels of scouring forces (i.e., near-bed shear stress) that affect plants. Despite the large body of literature linking plant community dynamics (and particularly forest and woodland communities) to physical processes, few authors have observed at the community scale the effect of slight differences in disturbance frequency and intensity, particularly for understory plants. In addition, only a few studies focused on understory plant communities studies that integrate the effects of non-equilibrium geomorphic processes such as channel migration (Wolfert et al., 2002 on levees in a lowland sand-bed river), embankment (Dufour et al., 2007 in pioneer habitats) or degradation (Bornette and Heiller, 1994; Bornette et al., 1996 for aquatic vegetation in former channels). Indeed, effects of channel degradation on terrestrial vegetation are well documented at the corridor scale and from a descriptive point of view (see J.-P. Bravard et al., 1997); however, more quantitative examples focused on communities, and in particular on understory communities, are still lacking.

3In this study we measured the local influence of flood disturbance frequency and intensity at the scale of a few square metres. We focused on the understory herbaceous and shrub vegetation of an alluvial forest with a homogeneous overstory strata but with variation in disturbance regime due to channel dynamics. We tested the influence of disturbance on vegetation and also bed vertical mobility control on disturbance regime. Our objective was to analyse how common flood events affect the understory community in terms of composition and regeneration of canopy species. We had main two hypotheses: (i) the spatial position of geomorphic landforms relative to an active river channel can be characterised by a double gradient of intensity and frequency of disturbance, with distance to channel scaling to intensity and vertical channel mobility reflecting disturbance frequency; (ii) these differences in disturbance regime influence understory community composition, structure and diversity, even in stands with equivalent overstory strata. Moreover, the magnitude and direction of changes in understory community structure following floods are predictable along these gradients of intensity and frequency.

Study sites, material and methods

Study sites and sampling strategy

4The Ain River is a sixth-order stream (Strahler, 1957) and the main tributary of the Rhône River upstream from Lyons (France) with a basin area of 3,640 km² (fig. 1A; Marston et al., 1995). The river’s mean annual discharge is 123 m3/s over the 1913-1977 record period; low flows (i.e., generally < 10 m3/s) occur in summer whereas annual peak and most high flows (Q1000 = 2450 m3/s) occur during winter months. In the lower valley (the last 40 km), the Ain River is a gravel bed river with sinuous pattern (sinuosity in 1991: 1.22) including some active meandering reaches. Within the lower river, the slope is 0.013 m/m(range 0.012-0.018 m/m) and the active channel width averages 120 m (fig. 1). Within the lower river, changes in channel elevation have occurred since the 1970s, and bed degradation has occurred in specific reaches due to sediment starvation by dams in the upper watershed and morphological adjustment in response to artificial cut-offs in the downstream reaches (Piégay et al., 2000).

Fig. 1 – Location of the Ain River (A) and the study sites (B), and sampling strategy (C): studied plots are located in geomorphically-stable or aggraded reaches or along river reaches with bed degradation over the 1976-1999 period (D).
Fig. 1 – Carte de localisation de l’Ain (A) et des sites d’étude (B) et stratégie d’échantillonnage (C) : les sites d'études sont localisés soit dans un contexte d'incision soit dans un contexte d'exhaussement/stabilité du chenal principal sur la période 1976-1999 (D).

Fig. 1 – Location of the Ain River (A) and the study sites (B), and sampling strategy (C): studied plots are located in geomorphically-stable or aggraded reaches or along river reaches with bed degradation over the 1976-1999 period (D).Fig. 1 – Carte de localisation de l’Ain (A) et des sites d’étude (B) et stratégie d’échantillonnage (C) : les sites d'études sont localisés soit dans un contexte d'incision soit dans un contexte d'exhaussement/stabilité du chenal principal sur la période 1976-1999 (D).

5Within the alluvial corridor, forest patches were identified on aerial photographs (2000 series provided by Institut Géographique National, scale 1:25,000). Sampling focused on floodplain surfaces with relatively homogeneous overstory vegetation in post-pioneer units (defined by R.A. Marston et al., 1995) with overstory species dominated by Populus nigra and Fraxinus excelsior. Initial field reconnaissance focused on defining floodplain areas in which to test a stratified design of two contrasting levels of the two disturbance attributes: disturbance frequency, and disturbance intensity. We used relative changes in bed elevation evident on long profile surveys from 1976 and 1999 to select areas of recent geomorphic change. Plots with high inundation frequency were typically located on eroded cutbanks in geomorphically-stable reaches, whereas less frequently-disturbed patches were located along river reaches with bed degradation, where floodplain forest plots are more elevated above the channel and require a higher (and less-frequently occurring) discharge for inundation. Areas of high and low intensity were distinguished by distance from the active channel. Using data from S. Dufour (2001) and H. Piégay et al. (2008), we identified a zone adjacent to the active channel where more intense and frequent sedimentation/erosion processes occur due to higher flow velocities during floods. Along the Ain River, this zone is between 40 m and 50 m wide and typically exhibits a coarser substrate (e.g., sand) compared to the inner floodplain areas with a finer (e.g., silt) substrate. According to this stratification system, we established 20 circular plots of 625 m² among the four disturbance conditions (i.e., frequency vs. intensity combinations): (i) near the channel (< 50 m) in a degraded reach (i.e., low frequency and high intensity; n = 4); (ii) within the floodplain (> 50 m) in a degraded reach (i.e., low frequency and low intensity; n = 5); (iii) near the channel in a stable or an aggraded reach (i.e., high frequency and high intensity; n = 6); and (iv) within the floodplain in a stable or an aggraded reach (i.e., high frequency and low intensity; n = 5). In order to control for effects of human alteration, edge with non-forested habitats and stand age (Bossuyt and Hermy, 2000, 2001; Jacquemyn et al., 2003; Verheyen et al., 2003), plots were chosen in large, similarly-aged floodplain stands of fluvial origin with no history of cultivation (field or aerial photos evidences).

Vegetation data

6Understory vegetation. In each plot, the herbaceous vegetation layer was measured in 4 systematically-positioned quadrats of 4 m2 each; a total of 80 quadrats were sampled in the study (fig. 1C). For each quadrat, species abundance was quantified as presence/absence of all species within 25 subquadrats; species names conform to the French flora database published by M. Kerguelen, 1999 (http://www.dijon.inra.fr/​flore-france). This quadrat subsampling takes into account within-quadrat heterogeneity (Palmer and Dixon, 1990). Germinants of woody species are counted when height is less than 50 cm. In order to take into account differing species’ phenologies, repeat surveys were carried out in spring, in summer and in autumn 2002 (fig. 2A). From the abundance data we calculated frequency of vegetation biological group (ruderal, exotic, annual) and life form (geophyte, chamaephyte, therophyte, hemicryptophyte and phanerophyte sensus C. Raunkiaër, 1934). For each species, abundance is given by the number of sub-quadrats that contains it. Flooding in winter 2002 (17 November: maximum flow 1160 m3/s at Chazey, daily mean flow 1000 m3/s = Q2.5), allowed us to evaluate the effect of a specific flood event on community response using post-flood surveys in 2003. A total of 40 flooded quadrats and 8 control (non-flooded) quadrats were resurveyed (respectively 4, 4, 20 and 20 plots in the four disturbance conditions described previously).

Fig. 2 – Daily discharge at Chazey gauging station (1998-2003) and floristic survey timing (A), and flow duration curve of daily discharge data (between 1963 and 2003; B).
Fig. 2 – Débits moyens journaliers à la station hydrologique de Chazey (période 1998-2003) et dates des relevés de végétation (A) et courbe des débits moyens journaliers classés (période 1963-2003 ; B).

Fig. 2 – Daily discharge at Chazey gauging station (1998-2003) and floristic survey timing (A), and flow duration curve of daily discharge data (between 1963 and 2003; B).Fig. 2 – Débits moyens journaliers à la station hydrologique de Chazey (période 1998-2003) et dates des relevés de végétation (A) et courbe des débits moyens journaliers classés (période 1963-2003 ; B).

7Control of plot-level environmental attributes. To control that plots were similar in term of overstory vegetation, all trees with a DBH greater than 7.5 cm were identified and their diametres measured. Community age was estimated in a relative way using dendrochronological measurement of the biggest ash trees in each plot (10 trees per plot, see S. Dufour and H. Piégay 2008). Organic matter content of the upper part of the sediment layer was evaluated by a loss on combustion method (Walkley and Black, 1934). In forest ecosystems, the quantity of available light in the understory layer is an important limiting factor (Emborg, 1998; Brosofske et al., 2001), and possibly interacts with the disturbance regime in flooded forests (Menges and Waller, 1983; Hall and Harcombe, 1998; Küßner, 2003). To measure the stress level generated by overstory light interception, canopy closure measurements were carried out for each quadrat by digital photos taken during summer 2002 at 1-m height with a wide-angle converter lens (Nikon WC-E63, angle of view 84 degrees, focal length 24 mm; Beckage et al., 2000). Image analysis was performed using threshold determination and pixel counting of sky vs. canopy (i.e., leaves, branches and trunks; ENVI software, ITT Company). This method gives a percentage of canopy closure as an estimation of light availability.

Disturbance measurement

8Flood frequency. We constructed water stage-discharge curves for each plot (i.e., 20) using topographic measurements and maximum flood recorders (level gauge with exchangeable color tape) for each sites. Maximum stage has been surveyed after 5 flood events (range of discharges: 150-1150 m3/s). The maximum discharge reached during the 5 events, for all the plots, come from the channel discharge recorded at the Chazey gauging station (fig. 1; there are no significant tributaries between the station and the plots). From the rating curves, we determined the threshold discharge corresponding to overbank flow for each plot (Dufour, 2005). Inundation frequency was then calculated for each plot from daily discharge data (reported at Chazey gauging station between 1963 and 2003) by analysing the frequency that daily flow equalled or exceeded the plot’s bankfull threshold discharge (fig. 2B).

9Flood intensity. The disturbance intensity gradient was evaluated via the effects of hydrogeomorphic processes on the plots, using the surface sediment grain size in the plots as a proxy for flow velocity during floods (Dufour, 2005). Sand proportion in a surface sediment core (sampled with a hand auger) was determined in the laboratory by sieving a single core for each quadrat taken from 0-10 cm depth. As floodplain substrate is usually a mix between sand, silt and clay, and mainly dominated by silt on the Ain River, a high percentage of sand expresses a relatively coarse grain size and therefore conditions of relative high velocity within the floodplain forest area during floods (Dufour, 2005). For the comparison before/after flood events (i.e., winter 2002 flood), we also used the amount of freshly deposited sediments (or the presence of vertical erosion) as a proxy of disturbance intensity with some sediment traps (flat tiles, see S. Dufour, 2005).

Statistical analyses

10Between-groups differences were tested using analysis of variance (ANOVA) for normally-distributed data, the Kruskal-Wallis and Mann-Whitney non-parametric tests for non-normal data and the Wilcoxon non-parametric test for non-normal, paired data. The threshold for statistical significance for each test was specified at p < 0.05. We analysed the vegetation pattern using De-trended Correspondence Analysis (DCA) and tested for significant differences in the floristic composition between groups using an analysis of similarity (ANOSIM). ANOSIM is a non-parametric test used to evaluate composition differences between two or more groups. It is based on a distance measure and expressed by the statistic R; a large positive R value indicates dissimilarity between groups (see Ø. Hammer et al., 2004). The Bray-Curtis index was used as the distance value and the significance computed by 5000 permutations of group membership (Hammer et al., 2004). The magnitude of change due to one flood event was evaluated by a Bray-Curtis Similarity index calculation for each quadrat (1 = no change; 0 = total change). We used multiple regression to model community composition change as a function of the explanatory factors disturbance frequency, disturbance intensity and other environmental attributes, including canopy closure, plot age, tree density, basal area and organic matter (tab. 1 and tab. 2). Model selection was performed using forward step-wise procedure in order to select the most important variables. All statistics were computed with Statview software (Baron and Vang, 1996), except multivariate analysis, which was computed using PAST software (Hammer et al., 2004).

Results

Disturbance quantification and plot characteristics

11As expected, the spatial position of forested plots within the floodplain (e.g., aggraded reaches vs. degraded and channel distance) created differences in local disturbance regime. Plots located on cutbanks (concave banks) in aggraded or stable reaches were lower in elevation, typically between 0.5 m and 1.5 m above the low-flow water level in the channel (fig. 3A). These plots were inundated at discharges < 750 m3/s (which is equivalent to a Q2 event), which corresponds to a higher relative flood frequency compared to plots in nearby degraded reaches. These latter plots are typically located > 2 m above the channel and inundated at discharges > 1100 m3/s, which is typical of a Q4 return frequency (fig. 3B). Thus quadrats in aggraded or stable reaches are hydrologically connected for a discharge exceeded approximately 2% of the time (over the1963-2003 period) versus < 0.0025% in degraded reaches (tab. 1 and fig. 2B). The substrate grain size analysis confirmed that sand content (30%) in the upper layer of the substrate is lower in the inner forest, i.e. farther than 50 m from the channel, than in near-channel plots (> 40% sand), and in particular compared to ones that are frequently flooded plots (70% sand). As we assumed that sand content is a proxy for flood intensity, our results indicate that the inner forest plots experienced lower flood intensity than those nearer the channel (tab. 1). In spite of these differences, the overstory vegetation indicated no significant variation in species, age, density and basal area among the four disturbance conditions (tab. 1). There was also no significant difference for organic matter content in sediments. Lastly, we observed a slightly higher canopy closure in the most frequently disturbed patches relative to less frequently disturbed plots (tab. 1).

Fig. 3 – Comparison of relative elevation between morphological contexts.
Fig. 3 – Comparaison entre l’altitude relative et le contexte géomorphologique.

Fig. 3 – Comparison of relative elevation between morphological contexts.Fig. 3 – Comparaison entre l’altitude relative et le contexte géomorphologique.

(mean values and 95% confidence interval, t test; A), and relation between elevation and overbank discharge (B).
(valeurs moyennes avec un intervalle de confiance à 95 %, comparaison par test t ; A) et relation entre l’altitude et le débit de débordement (B).

Tab. 1 – Characteristics of the sampling plots, mean values with standard deviation within the parentheses (letters give homogeneous groups from ANOVA and ANOSIM).
Tab. 1 – Caractéristiques des placettes étudiées, valeurs moyennes avec écart type indiqués entre parenthèses (les lettres indiquent la groupes homogènes issus des analyses ANOVA et ANOSIM).

Tab. 1 – Characteristics of the sampling plots, mean values with standard deviation within the parentheses (letters give homogeneous groups from ANOVA and ANOSIM).Tab. 1 – Caractéristiques des placettes étudiées, valeurs moyennes avec écart type indiqués entre parenthèses (les lettres indiquent la groupes homogènes issus des analyses ANOVA et ANOSIM).

1 measured at quadrat scale; 2 measured at plot scale; § evaluated over the 1963-2003 period at the Chazey-sur-Ain gauging station; # species< 10 % of the stand value are not indicated; Pn: Populus nigra, Fe: Fraxinus excelsior, Ap: Acer platanoïdes, Cm Crataegus monogyna; * when p is < 0.05 and *** < 0.0001 ; NS = not significant.

Impact of disturbance regime

12Community composition. Among the four disturbance conditions, the floristic composition of the understory strata was different only for plots that experienced both a high disturbance frequency and the highest intensity (i.e., close from the channel, stable or aggraded reaches; fig. 4A, black circle). Indeed, all of these quadrats are located on the right part of the species factorial map. The primary species that allow differentiation, and are characteristic of these plots, are Urtica dioica, Moehringia trinervia, Poa trivialis, Lamium maculatum, Polygonum lapathifolium, Calystegia sepium, Helianthus rigidus and Aegopodium podagraria(fig. 4B). All these species show a mean frequency lower than 0.05 per quadrat except for Aegopodium podagraria. This pool of species indicates particular ecological conditions, such as disturbed habitat, with many ruderal and short-lived species (Aegopodium podagraria, Alliariapetiolata, Galeopsistetrahit, Helianthus rigidus, Humuluslupulus, Impatiensglandulifera, Lamiumhybridum, Polygonum lapathifolium, Urticadioica), open canopy conditions (Poatrivialis, Calystegiasepium) and mesohygrophilic environments (Aegopodium podagraria, Alliariapetiolata, Calystegia sepium, Impatiensglandulifera, Lamiummaculatum, Urtica dioica).Beyond this pattern, the DCA shows a high internal variability between quadrats of a given context (fig. 4A). Analysis of similarity indicates significant differences between the groups (R = 0.646; p < 0.0001). When ANOSIM was run without the most disturbed plots, the results are not statistically significant (R = 0.109; p > 0.05), meaning that differences within groups are equivalent to those among groups.

Fig. 4 – Plots of samples (A) and species (B) with respect to the first two axes of a Detrended Correspondence Analysis (DCA).
Fig. 4 – Cartes factorielles des quadrats (A) et des espèces (B) selon les deux premiers axes de l’analyse des correspondances décentrée (DCA).

Fig. 4 – Plots of samples (A) and species (B) with respect to the first two axes of a Detrended Correspondence Analysis (DCA). Fig. 4 – Cartes factorielles des quadrats (A) et des espèces (B) selon les deux premiers axes de l’analyse des correspondances décentrée (DCA).

Axis limits are indicated in the corner of each plot.
La limite des axes est indiquée dans l’angle de chaque diagramme.

13Dominant species. The most frequent species are fairly consistent among the four disturbance conditions (fig. 5). However, among quadrats that underwent infrequent disturbance, two mesic forest species are dominant: Hedera helix (present in all the subquadrats) and Ligustrum vulgare. Other species in these two disturbance conditions are always < 25% (fig. 5). In plots that are flooded more frequently, these two species are less abundant. For plots near the channel with greater disturbance intensity, Hedera helix abundance exhibits its lowest ranking (4th) and there are no species with an absolute abundance > 40%.

Fig. 5 – Abundance of main species for each disturbance regime.
Fig. 5 – Abondance absolue des principales espèces de la strate basse en fonction du régime de perturbations.

Fig. 5 – Abundance of main species for each disturbance regime.Fig. 5 – Abondance absolue des principales espèces de la strate basse en fonction du régime de perturbations.

14Diversity and biological groups. We observed no significant differences in number of species by quadrat (tab. 2). However, species evenness is higher in plots that underwent frequent and intense floods (median = 0.87 vs. 0.81). Within-quadrat variability is also higher for this class; there is a higher variability of number of species by sub-quadrats and a higher frequency of empty sub-quadrats (tab. 2). Empty sub-quadrats imply that frequent and intense floods create more opens spaces at a few-metre scale.Non-native species frequency is not affected by disturbance regime (median for each group ≤ 5%), whereas ruderal species are more abundant (20%) in intensely disturbed plots, especially those with higher flood frequency (tab. 2). The disturbance effect was significant for only three biological forms: more frequently-disturbed plots had lower frequency geophytes and lianas, and higher frequency of woody seedlings (indeed ash germinants represent 96% of the “Tree germinant”).

Tab. 2 – Community diversity and composition. Median, inter-quartile range within the parentheses; letters give homogeneous groups from Mann-Whitney test.
Tab. 2 – Composition et diversité de la strate basse. Valeurs méd²ianes et intervalles inter-quartiles indiqués entre parenthèses ; les lettres indiquent les groupes homogènes issus du test de Mann-Whitney.

High frequency
High intensity

High frequency
Low intensity

Low frequency
High intensity

Low frequency
Low intensity

Significance level

Number of quadrats

24

20

16

20

Specific Richness (Rs)

9.5 (7)

9 (9)

9 (8)

8 (9)

NS

Evenness§

0.87 (0.23)
a

0.82 (0.26)
b

0.81 (0.19)
b

0.81 (0.22)
b

**

Coefficient of variance of Rs

0.48 (0.76)
a

0.31 (0.62)
b

0.34 (0.32)
b

0.28 (0.26)
b

**

Frequency of empty subquadrats

0.04 (0.28)
a

0 (0.16)
b

0 (0.08)
b

0 (0)
b

*

Exotic species

0.01 (0.33)

0 (0.31)

0 (0.33)

0.05 (0.39)

NS

Ruderal species

0.20 (0.66)
a

0.07 (0.56)
b

0.13 (0.52)
b

0.09 (0.21)
b

*

Geophyte

0 (0.05)
a

0.01 (0.1)
a

0.04 (0.14)
b

0.05 (0.02)
b

*

Therophyte

0 (0.10)

0 (0.23)

0 (0)

0 (0.01)

NS

Hemicryptophyte

0.11 (0.54)

0.09 (0.28)

0.07 (0.48)

0.05 (0.40)

NS

Chamephyte

0 (0.09)

0 (0.05)

0 (0.19)

0 (0.06)

NS

Shrub (phanerophyte)

0.35 (0.79)

0.31 (0.49)

0.38 (0.60)

0.39 (0.37)

NS

Tree germinant (phanerophytes)

0.16 (0.32)
a

0.11 (0.23)
b

0.10 (0.21)
b

0.05 (0.21)
c

**

Liana (phanerophyte)

0.22 (0.39)
a

0.41 (0.70)
b

0.34 (0.40)
b

0.36 (0.42)
b

**

Annual

0 (0.1)

0 (0.2)

0 (0)

0 (0)

NS

Biannual

0 (0.1)
a

0 (0)
b

0 (0)
b

0 (0)
b

*

Perennial

0.99 (0.2)
a

1 (0.2)
a

1 (0)
b

1 (0)
b

**

§  Shannon diversity divided by the logarithm of number of taxa (see equitability, Hammer et al., 2004); * when p is < 0.05 and ** < 0.01 ; NS = not significant

Impact of one flood event

15Comparison flooded/non-flooded. Pre- and post-flood comparison of floristic composition reveals that 18 species changed in frequency more than 1% in disturbed plots compared to 5 species in the unflooded plots. However, the change is statistically significant for only 6 species among the 18: increases for Ranonculusficaria, Galium aparineand Silenedioica, and decreases for Lamiumhybridum, Hedera helix and germinants of Fraxinusexcelsior. For Hedera helix, this result is consistent with the difference observed between plots measured in 2002 with different disturbance regimes (fig. 5) in showing that Hedera helix is sensitive to flooding. Germinants of Fraxinusexcelsior (i.e., 96% of “Tree germinant”) that are more abundant in disturbed plots (tab. 2) were negatively affected by one flood event (tab. 3). In the year that followed the flood, the number of species in disturbed plots increased. Other relevant changes observed include the decrease of tree germinants and perennial plants, and an increase in abundance of short-lived species (therophytes; tab. 3). Flooded plots also experienced an increase in geophytes species, mainly due to the high occurrence of Ranonculusficaria, and a decrease in liana forms.

Tab. 3 – Evolution of community characteristics after a flood (evolution in %, paired Wilconxon test).
Tab. 3 – Evolution de la strate basse après une crue (évolution en %, test apparié de Wilcoxon).

Flooded

Significance level

Non flooded

Significance level

Number of quadrats

40

8

Species richness (Rs)

+ 18.8

**

+ 0.1

NS

Evenness

+ 1.9

NS

+ 0.9

NS

Coefficient of variance of Rs)

- 1.6

NS

+ 44.0

NS

Frequency of empty subquadrats

- 13.6

NS

/

NC

Exotic species

0

NS

/

NC

Ruderal species

+ 11.7

NS

+ 14.8

NS

Geophyte

+ 393.3

***

+ 24.2

*

Therophyte

+ 24.1

*

/

NC

Hemicryptophyte

+ 5.6

NS

+ 15.9

NS

Chamephyte

+ 38.5

NS

+ 21.7

NS

Shrub (phanerophyte)

+ 1.2

NS

-1.2

NS

Tree germinant (phanerophytes)

-39.9

***

-7.5

NS

Liana (phanerophyte)

-7.5

*

-5.2

*

Annual and biannual species

100

*

/

NC

Perennial species

-1.9

*

/

NC

* when p is < 0.05, ** < 0.01 and *** < 0.0001 ; NS = not significant ; NC : not calculable

16Predicting community modification after disturbance. Flood-induced changes in community composition can be evaluated by comparing community similarity before and after the flood event. Pre-flood similarity index values were > 0.95 for control quadrats (those not flooded) whereas those for disturbed quadrats ranged between 0.20 and 0.97. Using sedimentation and erosion amount as a proxy for disturbance intensity, the greatest change was observed in plots that underwent erosion (implying a higher flow velocity) and high sedimentation amount of > 5 kg/m2 (fig. 6A); for these plots, mean similarity values after the flood were < 0.6. There was virtually no change in plots that experienced no sedimentation (similarity ~ 1), even in plots that were flooded and the change is moderate for low sedimentation rate.To develop a predictive relationship for the degree of community composition changes at the quadrat scale (e.g., similarity index) we performed a step-wise procedure to select the most relevant variables from the set of seven explanatory factors and environmental attributes: hydrological connection frequency of plots, grain size of overbank sediments, organic matter content, stand basal area, stand density, stand age and canopy closure. The resulting best model contained the two explanatory factors, connection frequency and grain size, which were proxies for disturbance frequency and intensity respectively; both factors were significant at p < 0.05. Using these disturbance attributes (e.g., disturbance frequency and sediment grain size), we can then predict degree of changes by a multiple regression model (fig. 6B); dissimilarity significantly increases in relation with flood frequency and intensity.

Fig. 6 – Degree of community changes after a flood (i.e., similarity index before/after flood).
Fig. 6 – Ampleur de la modification de la strate basse en termes de composition avant/après une crue.

Fig. 6 – Degree of community changes after a flood (i.e., similarity index before/after flood). Fig. 6 – Ampleur de la modification de la strate basse en termes de composition avant/après une crue.

A. Relation with sediment erosion and sedimentation processes (mean values and 95% confidence interval, letters give homogeneous groups from ANOVA). B. Comparison of similarity (before/after flood) observed and predicted by a multiple regression based on disturbance attributes (connection frequency and sediment grain size).
A. Relation avec les processus d’érosion et de sédimentation (valeurs moyennes avec un intervalle de confiance à 95 %, les lettres indiquant les groupes homogènes issus de l’ANOVA). B. Comparaison entre les valeurs observées et prédites des valeurs de la similarité avant/après crue, les valeurs prédites sont obtenues par une régression multiple basée sur le régime de perturbations (fréquence de connexion et granularité des sédiments).

Discussion and perspectives

Disturbance cause and effect

17Water flow and sediment reworking during floods are usually recognised as the main natural source of disturbance in floodplain ecosystems at the landscape scale (e.g., between different units; Pollock et al., 1998; Kamisako et al., 2007). Our results confirm studies that suggest that, at a finer scale, in floodplain forests these processes also influence community composition and structure (Bornette and Amoros, 1996; Wolfert et al., 2002; Dufour and Piégay, 2008). In this study, we observed forested plots flooded between Q1 and Q1000 events; however, substantial differences in understory community are observed at a threshold between Q2 and Q4 events. Our results also show that high disturbance frequency is not enough to modify the understory community; plots inundated frequently but with a low level of sediment deposition and erosion are more similar floristically to infrequently-disturbed ones than plots with both high frequency and intensity of disturbance (fig. 5). The impact of the combination of the two factors has been also shown as important by M. Gerard et al. (2008). In that study, flood frequency and mowing were important correlates with plant species composition in temperate lowland floodplain meadows. In the range of observed disturbances along the Ain River, flood inundation and sediment reworking are both necessary for variation in the understory vegetation to be expressed (tab. 2), and major changes in composition occur with high rates of both sedimentation and erosion (fig. 6A). As shown in experimental studies, the mechanical action of sediment deposit or removal is more destructive (Ewing, 1996) than increases in water level. Erosion uproots vegetation and sedimentation can partially or totally cover aboveground stems (Kent et al., 2001). When deposition is high enough to inhibit seedling emergence some biological characteristics, such as vegetative reproduction, can be favoured (Sluis and Tandarich, 2004). In both cases (erosion and deposition), growing space is increased and competition is decreased during subsequent vegetative recovery and recolonisation.

18Between-site differences in disturbance regime are not strong enough to result in highly different communities. The communities are fairly similar among the four disturbance conditions, with slight differences in cover and species importance rankings (Wolfert et al., 2002). The timing of floods on the Ain River is likely a limiting factor of disturbance effects on the vegetation. Over the period 1963-2003, 80% of floods with a discharge > 750 m3/s occurred between October and March, during the dormant season when interactions with plants are at their lowest level. In spite of predominantly winter flooding, the level of disturbance is great enough to increase local composition and variability at a scale relevant to understory vegetation (one to several metres); higher disturbance levels resulted in higher occurrence of empty sub-quadrats, a lower similarity index between quadrats, and a higher coefficient of variation of species richness within quadrat. Locally-variable disturbances generate small patches of bare substrate (i.e., subquadrats empty of vegetation) which favor disturbance-adapted species typically classified as ruderal and short-lived (Menges and Waller, 1983; Hughes and Fahey, 1991; McAlister et al., 2000; Palmer et al., 2000).

19Multivariate factorial analysis showed significant differences among disturbance conditions in term of species composition; however, the primary difference is expressed as the balance of species and not by a distinct species mix in the more disturbed plots. In the most frequently and intensely disturbed areas, the mesic species Hedera helix and Ligustrum vulgare do not dominate the community to the same degree as in less-disturbed environments. Instead, the most disturbed plots are characterised by higher levels of species evenness and a lack of any strongly dominant species. Thus a certain level of disturbance maintains some open space, which provides resources to maintain species diversity (constant richness; Wintle and Kirkpatrick, 2007). Sampling bracketing the flood of 2002 indicate that single disturbance events temporarily increase the number of species (mainly annuals), but this increase is not maintained over the long term.

20Concerning the effect of canopy closure, numerous authors have shown that in riverine environments, interactions occur between disturbance and light conditions (Hall and Harcombe, 1998). But disturbance-mediated changes in canopy composition and cover require important changes in overstory cover (Hughes and Fahey, 1991). In our case, differences between groups are slight because floods are not powerful enough to destroy overstory plants and therefore, it is still unclear how particular floods affect some overstory species’ distributions.

Spatial variability of disturbance regime at fine scale in alluvial corridor

21Flooding and fine sediment transport processes have significant effects on understory vegetation within the riparian forests. As expected, theses processes occur in a spatially patchy way throughout a river corridor and depend on both local elevation and lateral distance from active channels (Kupfer and Malanson, 1993). Furthermore, physical/ecological processes are linked to channel planform which is not stable in time. Distance to channel changes with channel migration (e.g., 10-20 m/a in more active reaches of the Ain River; Dufour, 2005). Channel bed degradation increases relative elevation and reduces flood-borne sedimentation of forested plots outside the degraded area (Steiger et al., 2001). High sedimentation rates (high disturbance regime) can also increase relative elevation (Wyzga, 1999) and thus after several years lead to a reduction in inundation frequency. Aggradation within the channel reduces elevation differences between the channel and adjacent terrain, and leads to more extensive out-channel sedimentation. However, high levels of sedimentation in turn increase the relative elevation of a floodplain surface, which works against the net reduction in elevation produced by channel aggradation. In areas prone to high-velocity floodwaters, increased sediment transport capacity may result in lower rates of sedimentation or even net erosion.

22Reach morphological context also has a direct influence on disturbance regime and plant communities. Reaches with high topographic heterogeneity will often support a greater diversity of microenvironments of resource availability and disturbance intensity, with local influences on plant community composition and diversity (Bratton, 1976; Vivian-Smith, 1997). Another influence on plant diversity is the structure of the vegetation itself, specifically the presence and abundance of woody debris (Piégay, 1997). At a local scale (from centimetre to metre) dead wood creates flow diversions which both concentrates flows in some areas (resulting in higher velocities and sediment transport/erosion) and slow it in others (resulting in back-water areas of greater deposition). The result is a topographic and vegetative mosaic that is characteristic of fluvially-disturbed environments (Townsend, 1989; Barrat-Ségretain and Amoros, 1996).

Perspectives

23Our study documents different understory communities occurring under the same overstory vegetation. This disconnection between strata, also observed by other authors in non-alluvial forested ecosystems (Reed et al., 1993), is in our case a result of hydrogeomorphic drivers (i.e., main channel vertical mobility and related disturbance regime). From a long-term perspective, lasting impacts are not likely on overall species diversity per se, as we do not observe important loss or gain of species. However, one lasting impact is probably the flood-mediated control over woody tree recruitment and the understory seedling pool. Our results show that one flood event negatively affects tree germinants (mainly ash) abundance in subsequent years. But, conversely, the presence of more open space in the community due to flooding has a positive effect on these tree germinants. Floods generally have the greatest effect on reducing cover within the shrub and understory strata, and thus create microsites with increased growing space and light availability (Lorimer et al., 1994; Küßner, 2003). Indeed, it is well known that understory vegetation can have a controlling effect on seedling recruitment through light limitation (George and Bazzaz, 1999 a and b) even more than the overstory (Beckage et al., 2000). For example, J. Emborg (1998) found in that deciduous forests dominated by a shade-tolerant species (beech, 60%) mixed with oak, ash and elm, effects of floods on plant competition hinged on whether they subsequently reduced the vegetation cover. However, the long and mid-term dynamics of communities, whatever ecosystem type, is difficult to predict from short-term observation (Collins, 2000) and observed differences have to be validated by long-term monitoring. To evaluate the long-term effects of fluvial forces on overstory vegetation, the next point to analyse could be the disturbance effect at all morphological stages: germinant and seedling (Wall et al., 2005) but also seeds (Deiller et al., 2003; Peterson and Baldwin, 2004), sapling and adults.

Haut de page

Bibliographie

Baron C., Vang E. (1996) StatView, references manual. Abacus-concepts, Berkeley, Californie, 477 p.

Barrat-Ségretain M.H., Amoros C. (1996)– Recovery of riverine vegetation after experimental disturbance: a field test of the Patch Dynamics Concept. Hydrobiologia 321, 53-68.

Beckage B., Clark J.S., Clinton B.D., Haines B.L. (2000)– A long-term study of tree seedling recruitment in the southern Appalachian forests: the effects of canopy gaps and shrub understories. Canadian Journal of Forest Research 30, 1617-1631.

Bornette G., Amoros C. (1991) – Aquatic vegetation and hydrology of a braided river floodplain. Journal of Vegetation Science 2, 497-512.

Bornette G., Amoros C. (1996)– Disturbance regimes and vegetation dynamics: role of floods in riverine wetlands. Journal of Vegetation Science 7, 615-622.

Bornette G., Heiler G. (1994)– Environmental and biological responses of the former channels to river incision: a diachronic study on the Upper Rhône River. Regulated Rivers: Research and Management 9, 79-92.

Bornette G., Amoros C., Rostan J.C. (1996) – River incision and decennial vegetation dynamics in cut-off channels. Aquatic Sciences 58, 31-51.

Bornette G., Amoros C., Lamouroux N. (1998)– Aquatic plant diversity in riverine wetlands: the role of connectivity. Freshwater Biology 39, 267-283.

Bornette G., Tabacchi E., Hupp C., Puijalon S., Rostan J.C. (2008)– A model of plant strategies in fluvial hydrosystems. Freshwater Biology 53, 1692-1705.

Bossuyt B., Hermy M. (2001) – Influence of land use history on seed banks in European temperate forest ecosystems: a review. Ecography 24, 225-238.

Bossuyt B., Hermy M. (2000)– Restoration of the understorey layer of recent forest bordering ancient forest. Applied Vegetation Science 3, 43-50.

Bratton S.P. (1976) – Ressources division in an understory herb community: responses to temporal and microtopographic gradients. The American Naturalist 110, 679-693.

Bravard J.-P., Amoros C., Pautou G., Bornette G., Bournaud M., Creuzé des Châtelliers M., Gibert J., Peiry J.-L., Perrin J.-F., Tachet H. (1997) – Stream incision in Southeast France: morphological phenomena and impacts upon biocenoses. Regulated Rivers: Research and Management 13, 1-16.

Brosofske K.D., Chen J., Crow T.R. (2001)– Understory vegetation and site factors: implications for a managed Wisconsin landscape. Forest Ecology and Management 146, 75-87.

Chaneton E.J., Facelli J.M. (1991)– Disturbance effects on plant community diversity: spatial scales and dominance hierarchies. Vegetatio 93, 143-155.

Collins S.L. (2000) – Disturbance frequency and community stability in native tall grass prairie. AmericanNaturalist 155, 311-325.

Connell J.H. (1978)– Diversity in tropical forests and coral reefs. Science 199, 1302-1310.

Deiller A.F., Walter J.M.N., Trémolières M. (2003) – Regeneration strategies in a temperate hardwood floodplain forest of the Upper Rhine: sexual versus vegetative reproduction of woody species. Forest Ecology and Management 180, 215-225.

Dufour S. (2001)Dynamique sédimentaire des bouchons alluviaux de la basse vallée de l’Ain : recherche de bio-indicateurs. Master thesis, university of Lyons, 50 p. (unpublished).

Dufour S. (2005)Contrôles naturels et anthropiques de la structure et de la dynamique des forêts riveraines des cours d’eau du bassin rhodanien (Ain, Arve, Drôme et Rhône). PhD thesis, university of Lyons, 244 p., unpublished (available from tel.archives-ouvertes.fr).

Dufour S., Piégay H. (2008) – Geomorphic controls of Fraxinus excelsior growth and regeneration in floodplain forests. Ecology 89, 205-215.

Dufour S., Barsoum N., Muller E., Piégay H. (2007) – Effects of channel confinement on pioneer woody plant community structure, composition and diversity along the River Drôme (SE, France). Earth Surfaces Processes and Landforms 32, 1244-1256.

Emborg J. (1998)– Understorey light conditions and regeneration with respect to the structural dynamics of a near natural temperate deciduous forest in Denmark. Forest Ecology and Management 106, 83-95.

Ewing K. (1996) – Tolerance of four wetland plant species to flooding and sediment deposition. Environmental and Experimental Botany 36, 131-146.

Fonda R.W. (1974) – Forest succession in relation to river terrace development in Olympic National Park, Washington. Ecology 55, 927-942.

Fox J.F. (1979) – Intermediate-disturbance hypothesis. Science 204, 1344-1345.

George L.O., Bazzaz F.A. (1999a)– The fen understory as an ecological filter: emergence and establishment of canopy-tree seedlings. Ecology 80, 833-845.

George L.O., Bazzaz F.A. (1999b) – The fen understory as an ecological filter: growth and survival of canopy-tree seedlings. Ecology 80, 846-856.

Gerard M., El Kahloun M., Rymen J., Beauchard O., Meire P. (2008)– Importance of mowing and flood frequency in promoting species richness in restored floodplains. Journal of Applied Ecology 45, 1780-1789.

Gilvear D.J., Cecil J., Parsons H. (2000)– Channel change and vegetation diversity on a low-angle alluvial fan, River Feshie, Scotland. Aquatic Conservation: Marine and Freshwater Ecosystems 10, 53-71.

Grime J.P. (1977) – Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist 111, 1169-1194.

Grime J.P. (2001)Plant strategies, vegetation processes and ecosystem properties. John Wiley and Sons, Chichester, 2nd edition, 417 p.

Hall R.B., Harcombe P.A. (1998) – Flooding alters apparent position of floodplain saplings on a light gradient. Ecology 79, 847-855.

Hammer Ø., Harper D.A.T., Ryan P.D. (2004)PAST-PAlaeontological STatistics, ver. 1.32, 65 p. (http://folk.uio.no/ohammer/past).

Hughes J.W., Fahey T.J. (1991) – Colonization dynamics of herbs and shrubs in a disturbed northern hardwood forest. Journal of Ecology 79, 605-616.

Hughes F.M.R. (1997) – Floodplain biogeomorphology. Progress in Physical Geography 21, 501-529.

Hupp C.R., Osterkamp W.R. (1985) – Bottomland vegetation distribution along Passage Creek, Virginia, in relation to fluvial landforms. Ecology 66, 670-681.

Hupp C.R., Osterkamp W.R. (1996) – Riparian vegetation and fluvial geomorphic processes. Geomorphology, 14, 277-295.

Huston M. (1979)– A general hypothesis of species diversity. The American Naturalist 113, 81-101.

Jacquemyn H., Butaye J., Hermy M. (2003)– Impacts of restored patch density and distance from natural forests on colonization success. Restoration Ecology 11, 417-423.

Kamisako M., Sannoh K., Kamitani T. (2007) – Does understory vegetation reflect the history of fluvial disturbance in a riparian forest? Ecological Research 22, 67-74.

Kent M., Owen N.M., Dale P., Newham R.M., Giles T.G. (2001)– Studies of vegetation burial: a focus for biogeography and biogeomorphology ? Progress in Physical Geography 25, 455-482.

Kupfer J.A., Malanson G.P. (1993) – Observed and modelled directional change in riparian forest composition at cutbank edge. Landscape Ecology 8, 185-199.

Küßner R. (2003)– Mortality patterns of Quercus, Tilia, and Fraxinus germinants in a floodplain forest on the river Elbe, Germany. Forest Ecology and Management 173, 37-48.

Lorimer C.G., Chapman J.W., Lambert W.D. (1994)– Tall understory vegetation as a factor in the poor development of oak seedlings beneath mature stands. Journal of Ecology 82, 227-237.

Mackey R.L., Currie D.J. (2001)– The diversity-disturbance relationship: is it generally strong and peaked ? Ecology 82, 3479-3492.

Marston R.A., Girel J., Pautou G., Piégay H., Bravard J.-P., Arneson C. (1995)– Channel metamorphosis, floodplain disturbance, and vegetation developpement: Ain river, France. Geomorphology 13, 121-131.

McAlister S.D., Palmer M.W., Arévalo J.R., DeCoster J.K. (2000)Species composition on tip up mounds and pits created by catastrophic windthrow in Minnesota forest. Proceedings IAVS Symposium, Opulus Press, Uppsala, 104-107.

Menges E.S., Waller D.M. (1983) – Plant strategies in relation to elevation and light in floodplain herbs. TheAmerican Naturalist 122, 454-473.

Nilsson C. (1987) – Distribution of stream-edge vegetation along a gradient of current velocity. Journal of Ecology 75, 513-522.

Palmer M.W., Dixon P.M. (1990) – Small-scale environmental heterogeneity and the analysis of species distributions along gradients. Journal of Vegetation Science 1, 57-65.

Palmer M.W., McAlister S.D., Arévalo J.R., DeCoster J.K. (2000) – Changes in the understory during 14 years following catastrophic windthrow in two Minnesota forests. Journal of Vegetation Science 11, 841-854.

Pautou G., Décamps H. (1985)– Ecological interactions between the alluvial forests and hydrology of the upper Rhône. Archiv für Hydrobiologie 104, 13-37.

Perumal V.J., Maun M.A. (2006) – Ecophysiological response of dune species to experimental burial under field and controlled conditions. Plant Ecology 184, 89-104.

Peterson J.E., Baldwin A.H. (2004) – Seedling emergence from seed banks of tidal freshwater wetlands: response to inundation and sedimentation. Aquatic Botany 78, 243-254.

Pianka E.R. (1970) – On r- and K- selection. American Naturalist 104, 592-597.

Pickett S.T.A., White P.S. (Eds.) (1985) The ecology of natural disturbances and patch dynamic. Academic Press, 456 p.

Piégay H. (1997) – Interactions between floodplain forests and overbank flows: data from three piedmont rivers of south-eastern France. Global Ecology and Biogeography Letters 6, 187-196.

Piégay H., Landon N., Saulnier D. (2000)Etude géomorphologique de la basse rivière d'Ain, propositions pour une gestion physique équilibrée des lits mineurs. SAGE “Basse vallée de l’Ain”, unpublished report, 86 p.

Piégay H., Hupp C.R., Citterio A., Dufour S., Moulin B., Walling D.E. (2008)– Spatial and temporal variability in sedimentation rates associated with cut-off channel infill deposits: Ain River, France. Water Resources Research 44, W05420, 18 p.

Pollock M.M., Naiman R.J., Hanley T.A. (1998)– Plant species richness in riparian Wetlands - A test of biodiversity theory. Ecology 79, 94-105.

Raunkiaër C. (1934)The life forms of plants and statistical plant geography. Clarendon Press, Oxford, 632 p.

Reed R.A., Peet R.K., Palmer M.W., White P.S. (1993)– Scale dependence of vegetation-environment correlations: a case study of a North Carolina piedmont woodland. Journal of Vegetation Science 4, 329-340.

Roxburgh S.H., Shea K., Wilson J.B. (2004)– The intermediate disturbance hypothesis: patch dynamics and mechanisms of species coexistence. Ecology 85, 359-371.

Sale P.P. (1977) – Maintenance of high diversity in coral reef communities. The American Naturalist 111, 337-359.

Salo J., Kalliola R., Hakkinen I., Makinen Y., Niemela P., Puhakka M., Coley P.D.(1986)– River dynamics and the diversity of Amazon lowland forest. Nature 332, 254-258.

Sluis W., Tandarich J. (2004) – Siltation and Hydrologic Regime Determine Species Composition in Herbaceous Floodplain Communities. Plant Ecology 173, 115-124.

Steiger J., Gurnell A.M., Ergenzinger P., Snelder D. (2001)– Sedimentation in the riparian zone of an incising river. Earth Surface Processes and Landforms 26, 91-108.

Steiger J., Tabacchi E., Dufour S., Corenblit D., Peiry J.-L. (2005)– Hydrogeomorphic processes affecting riparian habitat within alluvial channel-floodplain river systems: a review for the temperate zone.River Research and Applications 21, 719-737.

Strahler A.N. (1957)– Quantitative analysis of watershed geomorphology. American Geophysical Union Transactions 38, 913-920.

Suding K.N., Goldberg D. (2001)– Do disturbances alter competitive hierarchies ? Mechanisms of change following gap creation. Ecology 82, 2133-2149.

Suzuki W., Osumi K., Masaki T., Takahashi K., Daimaru H., Hoshizaki K. (2002)– Disturbance regimes and community structures of a riparian and an adjacent terrace stand in the Kanumazawa Riparian Research Forest, northern Japan. Forest Ecology and Management 157, 285-301.

Syms C., Jones G.P. (2000)– Disturbance, habitat structure and the dynamics of a coral-reef fish community. Ecology 81, 2714-2729.

Tabacchi E., Planty-Tabacchi A.-M., Salinas J., Décamps H. (1996) – Landscape structure and diversity in riparian plant communities: a longitudinal comparative study. Regulated River: Research and Management 12, 367-390.

Townsend C.R. (1989) – The patch dynamics concept of stream community ecology. Journal of North American Benthological Society 8, 36-50.

Van Coller A.L., Rogers K.H., Heritage J.L. (1997)– Linking riparian vegetation types and fluvial geomorphology along the Sabie River within the Kruger National Park, South Africa.African Journal of Ecology 35, 194-212.

Van der Maarel E. (1993)– Some remarks on disturbance and its relations to diversity and stability. Journal of Vegetation Science 4, 733-736.

Verheyen K., Bossuyt B., Honnay O., Hermy M. (2003)– Herbaceous plant community structure of ancient and recent forests in two contrasting forest types. Basic and Applied Ecology 4, 537-546.

Vervuren P.J.A., Blom C.W.P.M., De Kroon H. (2003)– Extreme flooding events on the Rhine and the survival and distribution of riparian plant species. Journal of Ecology 91, 135-146.

Vivian-Smith G. (1997)– Microtopography heterogeneity and floristic diversity in experimental wetland community. Journal of Ecology 85, 71-82.

Walkley A., Black I.A. (1934)– An examination of the Degjareff method for determining SOM and a proposed modification of the chromic acid titration method. Soil Science 37, 29-38.

Wall R.L., Heller Wardrop D., Brooks R.P. (2005)– The impact of experimental sedimentation and flooding on the growth and germination of floodplain trees. Plant Ecology 176, 203-213.

Ward J.V. (1998)– Riverine landscapes: biodiversity patterns, disturbance regimes and aquatic conservation. Biological Conservation 83, 269-278.

White P.S., Jentsch A. (2001)– The search of generality in studies of disturbance and ecosystem dynamics. Progress in Botany 62, 399-450.

Whittaker R.H., Goodman D. (1979) – Classifying species according to their demographic strategy. I. Population fluctuations and environmental heterogeneity. The American Naturalist 113, 185-200.

Wilkinson D.M. (1999)– The disturbing history of intermediate disturbance. Oikos 84, 145-147.

Wintle B.C., Kirkpatrick J.B. (2007)– The response of riparian vegetation to flood-maintained habitat heterogeneity. Austral Ecology 32, 592-599.

Wolfert H.P., Hommel P.W.F.M., Prins A.H., Stam M.H. (2002)– The formation of natural levees as a disturbance process significant to the conservation of riverine pastures. Landscape Ecology 17, 47-57.

Wootton J.T. (1998) – Effects of disturbance on species diversity: a multitrophic perspective. TheAmericanNaturalist 152, 803-825.

Wyzga B. (1999)– Estimating mean flow velocity in channel and floodplain areas and its use for explaining the pattern of overbank deposition and floodplain retention. Geomorphology 28, 281-297.

Haut de page

Annexe

Version abrégée en français

Dans les hydrosystèmes fluviaux, la complexité spatiale et la distribution des communautés biologiques sont en grande partie sous le contrôle de la mobilité du chenal (incision, érosion latérale) et des processus hydrologiques (inondation) et morphologiques (érosion et sédimentation). Ces processus interagissent et créent un large panel de niveaux de perturbations entre les différentes formes fluviales et au sein de chaque forme. Dans cette contribution, notre objectif est d’analyser quantitativement les liens qui existent entre la mobilité du chenal, le régime de perturbations généré par les crues et la végétation riveraine. L’étude porte sur la strate basse d’unités forestières localisées dans la plaine alluviale de l’Ain et similaires en termes de composition de la strate haute (milieux post-pionniers dominés par Fraxinus excelsior et Populus nigra). L’objectif n’est pas d’évaluer le rétablissement d’une communauté après une forte perturbation rare et très destructrice, mais plutôt la capacité d’évènements plus fréquents (Q1 à Q5) à modifier la strate basse et, donc, les conditions de régénération de la forêt. Au-delà de l’influence des crues, c’est également l’effet de la mobilité du chenal au sein du corridor qui est analysé. Pour cela, nous avons suivi deux hypothèses. Premièrement, le régime de perturbations enregistré par une placette est variable en fonction de sa position spatiale dans le corridor à un moment donné, mais aussi en fonction de la dynamique morphologique du chenal au droit de celle-ci. Deuxièmement, dans des unités végétales apparemment homogènes en termes de végétation de la strate arborescente, les différences en termes de perturbations influencent la composition, la structure ou encore la diversité de la strate basse. Afin de tester à la fois les gradients de fréquence et d’intensité des crues, l’échantillonnage a été stratifié sur la base d’une double entrée : 1) forte et faible fréquence. Les sites les plus fréquemment perturbés sont localisés dans les concavités de méandre, dans des tronçons stables ou exhaussés. Les autres sites sont situés dans les tronçons incisés, c’est-à-dire qu’ils sont positionnés à une altitude relative plus élevée par rapport au chenal et donc que des débits plus forts sont nécessaires pour les inonder ; 2) forte et faible intensité. La forte intensité de perturbation lors des inondations de la plaine alluviale est liée à des écoulements rapides ; elle est enregistrée au sein d’une zone « tampon » d’environ 40 m à 50 m de large située en bord de chenal au sein de laquelle des processus d’érosion/sédimentation sont plus intenses (sédimentation plus grossière et plus sableuse).

Dans les hydrosystèmes fluviaux, les apports d’eau et de sédiments pendant les crues sont généralement reconnus comme l’une des principales causes de perturbation à l’échelle du corridor fluvial. Nos résultats démontrent que ces apports influencent également les communautés végétales riveraines à une échelle plus fine. Ainsi, dans les milieux post-pionniers de la forêt riveraine de l’Ain, la fréquence de perturbation des sites varie théoriquement entre Q1 et Q1000, et des différences significatives en termes de végétation sont observées, dans la strate basse, pour un seuil compris entre Q2 et Q4. Mais nos résultats montrent également qu’une fréquence relativement importante de perturbations n’est pas suffisante pour modifier la communauté végétale occupant la strate basse. En effet, les sites fréquemment inondés, mais pour lesquels les processus hydrosédimentaires sont de moindre intensité, c’est-à-dire ceux situés en retrait du chenal, ressemblent plus aux sites moins fréquemment perturbés qu’à ceux connaissant la même fréquence de crue mais situés en bordure de chenal. Ceci semble pouvoir être mis en relation avec l’effet des crues : en effet, l’existence de processus de rajeunissement ou de remaniement du substrat semble indispensable pour affecter la végétation. De fait, l’action mécanique des phénomènes d’érosion et de sédimentation est plus destructrice, ou du moins plus contraignante, que celle d’une simple inondation. La gamme de perturbations analysée ici ne nous permet pas d’observer une flore originale et profondément différente d’un contexte à l’autre, comme cela peut être le cas dans des conditions plus tranchées. Cependant, certaines différences concernant la structure de la communauté végétale sont significatives. Localement, les perturbations augmentent la variabilité : présence de sous-quadrats sans végétation, similarité inter-quadrats plus faible, coefficient de variation de la richesse spécifique plus fort. Elles favorisent l’apparition de micro-tâches de substrat remanié qui, en retour, favorisent les espèces rudérales et à courte durée de vie. L’analyse factorielle des relevés de végétation montre des différences en termes de composition, même si elles apparaissent faibles du fait de l’absence d’espèces abondantes qui soient spécifiques des sites les plus perturbés ; les espèces qui dominent très largement la communauté végétale dans les sites moins perturbés (Hedera helix et Ligustrum vulgare) sont seulement moins fréquentes dans les sites plus perturbés. Les perturbations maintiennent donc des espaces ouverts en réduisant l’abondance des essences normalement dominantes, sans réduire la richesse. Comme il n’y a pas de différence en termes de richesse spécifique et que les sites perturbés ne sont pas dominés par un « pool » restreint d’espèces, la régularité y est supérieure. Ponctuellement, un événement de crue augmente le nombre d’espèces (surtout des annuelles), mais cette modification ne semble pas durable dans le temps. Enfin, contrairement à d’autres études réalisées en environnement alluvial, nous n’observons aucune interaction entre le régime des perturbations et les conditions de luminosité à l’échelle du quadrat.

Si les crues et le remaniement du substrat ont un effet significatif sur la strate basse des forêts riveraines, nous montrons également que la connexion hydrologique d’une parcelle et sa dynamique sédimentaire dépendent de sa position dans le corridor fluvial : position altitudinale et position en plan par rapport aux axes d’écoulement. De plus, nos résultats soulignent que la position spatiale d’une placette par rapport à l’axe d’énergie n’est pas stable dans le temps ; en effet, elle est modifiée par la mobilité du chenal. Dans les berges concaves, la migration latérale du chenal (jusqu’à 10-15 m/a sur l’Ain dans les secteurs les plus dynamiques) réduit la distance séparant la placette du chenal et augmente donc l’impact des crues. L’incision augmente l’altitude relative et réduit la fréquence de connexion de la placette ; à l’inverse, l’exhaussement du chenal permet d’accroître la fréquence d’inondation. La stabilité du chenal ne permet pas de conserver une fréquence identique car la sédimentation provoque un exhaussement progressif de la plaine. La sédimentation - sur l’Ain, nous avons enregistré en une crue entre 0 kg/m² et 45 kg/m² -, accroît l’altitude relative de la plaine, abaisse progressivement la fréquence des inondations et donc le régime de perturbations. Seul l’exhaussement du chenal est donc susceptible de compenser l’exhaussement de la plaine alluviale. Mais nos résultats indiquent également que l’élévation progressive de la plaine alluviale par sédimentation peut être ralentie, puisque dans les sites les plus perturbés, les fortes vitesses ne permettent pas les dépôts, mais favorise plutôt les phénomènes d’érosion du substrat : jusqu’à 30 cm d’épaisseur sur nos placettes.

Enfin, s’il est bien connu que l’incision du chenal peut se traduire par une modification des peuplements riverains du fait de la mise en place de conditions plus xériques, nos résultats indiquent également que la mobilité du chenal est susceptible d’avoir un effet sur la composition floristique, par le biais indirect du recrutement. Du fait des perturbations hydrosédimentaires, il existe des différences dans la strate basse d’unités forestières apparemment similaires en termes de strate haute. Ce hiatus entre les deux strates est, dans notre cas, sous le contrôle d’un facteur externe à l’écosystème : la mobilité du tronçon. Dans une perspective à moyen et long termes, cela pourrait entraîner une différenciation des trajectoires écologiques des différents sites étudiés, par modification de la composition du « pool » de recrutement en fonction du contexte morphologique du tronçon. Nos résultats indiquent un impact négatif direct de la crue sur les plantules des essences ligneuses (essentiellement du frêne) ; mais, en moyenne, une fréquence de perturbation plus importante se traduit par une présence plus élevée de plantules. La libération d’espaces lors des crues semble favorable à la germination des graines de frêne, probablement en raison d’une plus faible concurrence avec la strate basse, entre autres pour la lumière. En résumé, les crues semblent avoir un effet direct préjudiciable (sédimentation, érosion du substrat, destruction des organes aériens) mais un effet indirect bénéfique (libération de ressource) sur les plantules de frêne. Cependant, l’effet à moyen terme reste difficile à prévoir à partir de mesures réalisées à court terme et d’autres changements pourraient encore être observés par un suivi à plus long terme.

Au terme de cette étude, la dynamique sédimentaire sous forêt alluviale apparaît comme primordiale dans la structuration des communautés végétales. Évidemment, certains points restent à approfondir. L’analyse in situ des influences complexes et réciproques entre végétation et dynamique fluviale et des émergences suscitées par leurs interactions reste encore largement à développer. Ainsi, d’un point de vue géomorphologique, l’évaluation quantitative et spatio-temporelle des processus d’érosion et de sédimentation doit être encore affinée : par exemple, à l’échelle d’une boucle de méandre, quelles sont la gamme et la répartition spatiale des conditions de sédimentation ? Quelle est la permanence temporelle de ces processus ? Quel est l’impact de la végétation aux échelles spatiales fines ?

Haut de page

Table des illustrations

Titre Fig. 1 – Location of the Ain River (A) and the study sites (B), and sampling strategy (C): studied plots are located in geomorphically-stable or aggraded reaches or along river reaches with bed degradation over the 1976-1999 period (D).Fig. 1 – Carte de localisation de l’Ain (A) et des sites d’étude (B) et stratégie d’échantillonnage (C) : les sites d'études sont localisés soit dans un contexte d'incision soit dans un contexte d'exhaussement/stabilité du chenal principal sur la période 1976-1999 (D).
URL http://geomorphologie.revues.org/docannexe/image/8101/img-1.jpg
Fichier image/jpeg, 204k
Titre Fig. 2 – Daily discharge at Chazey gauging station (1998-2003) and floristic survey timing (A), and flow duration curve of daily discharge data (between 1963 and 2003; B).Fig. 2 – Débits moyens journaliers à la station hydrologique de Chazey (période 1998-2003) et dates des relevés de végétation (A) et courbe des débits moyens journaliers classés (période 1963-2003 ; B).
URL http://geomorphologie.revues.org/docannexe/image/8101/img-2.jpg
Fichier image/jpeg, 180k
Titre Fig. 3 – Comparison of relative elevation between morphological contexts.Fig. 3 – Comparaison entre l’altitude relative et le contexte géomorphologique.
Légende (mean values and 95% confidence interval, t test; A), and relation between elevation and overbank discharge (B).(valeurs moyennes avec un intervalle de confiance à 95 %, comparaison par test t ; A) et relation entre l’altitude et le débit de débordement (B).
URL http://geomorphologie.revues.org/docannexe/image/8101/img-3.jpg
Fichier image/jpeg, 76k
Titre Tab. 1 – Characteristics of the sampling plots, mean values with standard deviation within the parentheses (letters give homogeneous groups from ANOVA and ANOSIM).Tab. 1 – Caractéristiques des placettes étudiées, valeurs moyennes avec écart type indiqués entre parenthèses (les lettres indiquent la groupes homogènes issus des analyses ANOVA et ANOSIM).
Légende 1 measured at quadrat scale; 2 measured at plot scale; § evaluated over the 1963-2003 period at the Chazey-sur-Ain gauging station; # species< 10 % of the stand value are not indicated; Pn: Populus nigra, Fe: Fraxinus excelsior, Ap: Acer platanoïdes, Cm Crataegus monogyna; * when p is < 0.05 and *** < 0.0001 ; NS = not significant.
URL http://geomorphologie.revues.org/docannexe/image/8101/img-4.jpg
Fichier image/jpeg, 132k
Titre Fig. 4 – Plots of samples (A) and species (B) with respect to the first two axes of a Detrended Correspondence Analysis (DCA). Fig. 4 – Cartes factorielles des quadrats (A) et des espèces (B) selon les deux premiers axes de l’analyse des correspondances décentrée (DCA).
URL http://geomorphologie.revues.org/docannexe/image/8101/img-5.jpg
Fichier image/jpeg, 228k
Titre Fig. 5 – Abundance of main species for each disturbance regime.Fig. 5 – Abondance absolue des principales espèces de la strate basse en fonction du régime de perturbations.
URL http://geomorphologie.revues.org/docannexe/image/8101/img-6.jpg
Fichier image/jpeg, 184k
Titre Fig. 6 – Degree of community changes after a flood (i.e., similarity index before/after flood). Fig. 6 – Ampleur de la modification de la strate basse en termes de composition avant/après une crue.
Légende A. Relation with sediment erosion and sedimentation processes (mean values and 95% confidence interval, letters give homogeneous groups from ANOVA). B. Comparison of similarity (before/after flood) observed and predicted by a multiple regression based on disturbance attributes (connection frequency and sediment grain size). A. Relation avec les processus d’érosion et de sédimentation (valeurs moyennes avec un intervalle de confiance à 95 %, les lettres indiquant les groupes homogènes issus de l’ANOVA). B. Comparaison entre les valeurs observées et prédites des valeurs de la similarité avant/après crue, les valeurs prédites sont obtenues par une régression multiple basée sur le régime de perturbations (fréquence de connexion et granularité des sédiments).
URL http://geomorphologie.revues.org/docannexe/image/8101/img-7.jpg
Fichier image/jpeg, 79k
Haut de page

Pour citer cet article

Référence papier

Simon Dufour et Hervé Piégay, « Channel vertical mobility, hydro-geomorphic disturbances and understory vegetation in floodplain forests of the Ain River (France) », Géomorphologie : relief, processus, environnement, vol. 16 - n° 4 | 2010, 371-386.

Référence électronique

Simon Dufour et Hervé Piégay, « Channel vertical mobility, hydro-geomorphic disturbances and understory vegetation in floodplain forests of the Ain River (France) », Géomorphologie : relief, processus, environnement [En ligne], vol. 16 - n° 4 | 2010, mis en ligne le 01 décembre 2012, consulté le 25 novembre 2017. URL : http://geomorphologie.revues.org/8101 ; DOI : 10.4000/geomorphologie.8101

Haut de page

Auteurs

Simon Dufour

Université Rennes 2 - COSTEL/LETG CNRS UMR 6554 - Place recteur Henri le Moal - 35000 Rennes (simon.dufour@univ-rennes2.fr)

Articles du même auteur

Hervé Piégay

EVS/ENS LSH site CNRS UMR 5600 - Parvis Descartes - 69007 Lyon (Herve.Piegay@ens-lsh.fr)

Articles du même auteur

Haut de page

Droits d’auteur

© Groupe français de géomorphologie

Haut de page
  • Logo CNRS - Institut des sciences humaines et sociales
  • Logo Groupe français de géomorphologie
  • Revues.org