Chapitre premier. The use of functional groups to predict the impact of conservation management on the vegetation of chalk grasslands
Utilisation de groupes fonctionnels pour prédire l'impact de différents régimes de gestion sur la végétation des pelouses calcicoles
p. 111-126
Résumés
L'impact de différents régimes de gestion (fauchage, labour, abandon) sur la diversité végétale de quatre pelouses calcicoles de la vallée de Seine (Haute-Normandie, France) a été analysé par la méthode des points quadrats durant une période de 3 ans. Le classement des espèces en groupes végétaux fonctionnels permet de dégager les effets de la gestion conservatoire sur les phytocénoses indépendamment de la composition taxonomique des différentes pelouses. Les résultats des analyses multivariées montrent une dominance de la contribution des espèces annuelles et bisannuelles dans les quadrats régulièrement fauchés au contraire des quadrats abandonnés et des témoins plutôt dominés par des espèces pérennes. La richesse et la diversité spécifique des quadrats fauchés pendant trois ans sont significativement supérieures aux témoins et aux quadrats abandonnés. L’utilisation de groupes fonctionnels basés sur quelques caractéristiques des végétaux est discutée dans une optique de généralisation et de modélisation prédictive des effets de la gestion pour d'autres sites.
The impact of different management regimes (mowing, ploughing, abandonment) on plant diversity of four chalk grasslands in the Seine valley (Upper-Normandy, France) have been analysed by the point quadrat method during a 3 year-period. Species classification by functional vegetation groups allows us to draw the effects of conservation management on phytocenosis independently of taxonomic composition of the different grasslands. Results of multivariate analysis show a dominance of annual and biennial species contribution in the regularly mown plots at contrary of controls and abandoned plots rather dominated by perennial species. Species richness and diversity of mown plots during three years are significatively higher than controls and abandoned plots. The use of functional groups based on several plant traits is discussed in a view of making predictive modélisation and generalisation of management effects to others sites.
Texte intégral
INTRODUCTION
1In Western Europe, chalk grasslands have been studied for a long time because of their ecological interest. Since World War II, these grasslands decreased considerably in area when they have lost their economic function as a food resource for livestock (Willems, 1990). After the abandonment of grazing, succession changes leading to an increase in phytomass of some tussock species (e.g. Brachypodium pinnatum, Bromus erectus) which is negatively correlated with species richness and diversity (Bobbink & Willems, 1987). Accordingly, the need of conservation management for these ecosystems has been discussed in general terms by Wells (1965) and is now applied in north-western Europe (Maubert & Dutoit, 1995).
2Despite an important body of literature on experimental management (Lloyd, 1968; Wells, 1969a; Ward & Jennings, 1990 a,b; Bobbink & Willems, 1991; Dolman & Sutherland, 1994), a global understanding of the mechanisms involved is still unknown, especially because predictive models couldn't be defined on floristic composition which differ between sites. These differences have been observed at the scale of the european continent (Willems, 1982b; Royer, 1985) but they also exist at the scale of a country (Géhu, 1984), and even at the scale of a region (Willems, 1978).
3If species oriented models cannot succeed in providing general prediction, several authors have proposed a new classification based on constructing groups with similar ecological traits for predictive ecology. This classification could be based on the works of several authors which have emphasized on function of species in communities (Grime, 1974, 1977, 1979, Van der Valk, 1981; Keddy, 1990). The classification of functional groups have been applied to the study of wetland plant communities and were useful to predict how a specified perturbation will convert one vegetation type to another. (Keddy & Shipley, 1989; Shipley et al., 1990).
4The impact of different management regimes (ploughing, mowing, abandonment) on chalk grassland vegetation have been analysed after three years of treatments with functional groups classification. Plant richness, diversity and evenness have been calculated to focus on treatments which provide a most diversified plant communities. Results are discussed in order to provide predictive ecological model at the regional level.
STUDY SITES
5The research was carried out in four sites of chalk grasslands sit uated in the Seine valley in Upper-Normandy, France (Fig. 1). These sites have been chosen because they form a set of samples of ecological (microclimate, soil characteristics) and human (former land use) factors influencing the chalk grassland vegetation in this region (Dutoit et al, 1994).
6The main difference between the four sites is the former utilisation in tillage (1820) for the south-facing slope, when the inclination is slight (1 & II). The soils of former tillage showed lower values for C/N, Total C, CaO, MgO and sand (Tab. 1).This difference could be explained by the former practice of ploughing which increased the mineralisation of the organic matter in comparison with the very old pastures (Wells et al., 1976; Dutoit & Alard, 1995a).
Sites | I | II | III | IV |
Altitude (m) | 40 | 70 | 95 | 80 |
Inclination (%) | 38 | 55 | 27 | 65 |
Orientation (°) | 180 | 270 | 120 | 210 |
Former land uses | ||||
1820 | Tillage | Heath | Tillage | Heath |
1914 | Pasture | Heath | Heath | Heath |
1994 | Heath | Heath | Heath | Heath |
Soils analysis | ||||
Total N (%) | 5.28 | 4.7 | 4.24 | 4.27 |
Total C (%) | 38 | 73 | 52 | 61 |
C/N | 7.2 | 15.5 | 12.3 | 14.3 |
CaCO3 (g/kg) | 281 | 593 | 769 | 593 |
CaO (g/kg) | 14 | 17 | 12 | 17 |
P2O5 (g/kg) | 0.072 | 0.092 | 0.024 | 0.1 |
MgO (g/kg) | 0.19 | 0.13 | 0.13 | 0.09 |
pH | 7.6 | 7.6 | 7.9 | 7.6 |
Clay (%) | 56 | 48 | 38 | 41 |
Silt (%) | 34 | 30 | 55 | 39 |
Sand (%) | 10 | 22 | 7 | 20 |
7From continental sites (III, IV) to the estuary site (II), a decrease of xerophilous species could be observed in relation with an increase of precipitation and humidity (Liger, 1952). In consequence, grassland vegetation of site I is now classified as a Festuco lemanii-Seslerietum albicantis Boullet 1986 while the other sites are considered as Pulsatillo vulgaris-Seslerietum albicantis Boullet 1986 (Alard, 1990).
MATERIALS AND METHODS
8Eight permanent plots (1 m2) were arranged in two rows laid out in both sites (Fig. 2). In January 1992, the four plots of each row were selected for one of the following treatments: (1) ploughing, (2) mowing. From June 1992 to June 1994, mowing is continued on b-plots using a clipper to cut annually the vegetation at 3 cm above the soil surface, a-plots were not removed after their first treatment in January 1992.
9For each of the 40 quadrats, point frequency records of 100 regularly spaced points (2 cm), consisting of 2 rows at regular distances of 33 cm were made with a needle which was long enough to contact even the tallest plants. These quadrats were used to assess the Specific Contribution (SC %) of dominant plant species. The SC % is the ratio between the frequency of a specie to the total of frequencies of all species recorded. For each plot, species richness (R), Shannon index of diversity (H) and plant evenness (E) were calculated based on SC% data. Nomenclature: Tutin et al. (1964-80).
10In this study, point quadrat method was selected because this quantitative method is the most suitable one for estimating changes in vegetation structure (Stampfli, 1991). Point quadrat method has also been preferred because other methods are less accurate: cover estimates (Beguinot, 1992), unsuitable: frequency determination using small quadrats (Van der Maarel & Sykes, 1993) or even more time-consuming: charting plant locations (Willems & Bobbink, 1990), biomass measurements of standing crop by species (Bobbink & Willems, 1991).
11Plants were recorded in the last week of June every year from 1992 to 1994. At this time, most species have passed over their flowering optimum and phenological changes in the structure of the grassland have slowed down. This fixed recording date avoided the major diversity changes which occur throughout the year (Hutchings 1983, Ward & Jennings 1990a), but could not avoid some seasonal variation. All the flowering plants in each grassland which occurred outside the permanent quadrats were noted, but the outer 0.5 m was excluded to avoid edge effects (Verkaar et al., 1983b).
12Functional vegetation groups were based on three traits easily accessible in literature: life-form, Raunkiaer's type and establishment strategy (Grime et al, 1988). Correspondence Analysis (CA) were performed on plant specific contribution using STAT-ITCF software. Plant richness, diversity and evenness were statistically analysed using ANOVA-TEST.
RESULTS
Multivariate analysis
13The first correspondence analysis was carried out on the 60 plots and the 75 taxa recorded by the point quadrat technique along the three years (Fig 3). The first two axes of the CA explained only 18.6% and 13.3% respectively, of the variation in the data set. The first axis can be interpreted as a climatic gradient, which opposes thermophilous species of site IV (Teucrium montanum, Geranium sanguineum) with mesophilous species of site I (Prunella vulgaris, Plantago lanceolata). The second axe could be interpreted as a trophic gradient which oppose oligotrophic plants of site IV (Hieracium pilosella. Stachys recta) with mesotrophic species of site III (Ononis spinosa, Hieracium murorum, Anthyllis vulneraria). On the ordination diagram, site plots are segregated independently of treatment effects.
14The second correspondence analysis was also carried out on the 60 plots but only on 34 taxa which have been recorded in more than one site. This CA revealed the effect of management treatments on plant distribution between the four sites and during the experimental period (Fig. 4). The first two axes explained 22,2 % and 16,4 % of the data matrix. From year to year, no major changes can be observed in the control plots. Whereas, most control plots and perennial species (e.g. Koeleria macrantha, Euphorbia amygdaloides, Sesleria albicans, Teucrium chamaedrys, Carex flacca) occur mainly to the left in the ordination diagram. At the opposite, occur species which are mostly annuals (e.g. Centaurium erythraea, Sonchus asper), biennials (e.g. Linum catharticum, Echium vulgare, Verbascum nigrum) or hemicryptophyte rosette species (Viola hirta, Centaurea jacea). Therefore, the variation along the first axis can be interpreted as a disturbance gradient.
15From 1992, a distinct separation between regularly mown (b) and abandoned plots (a) can be observed. Most abandoned plots are progressively returning to the initial situation. Nevertheless, this evolution is irregular for certain ones which show a progression to another direction in 1993.
16Most of the mown plots reveal also major changes which are different from year to year. No distinct separation between plots which have been mown or ploughed only in 1992 can be observed.
17CA realised individually for each site, with elimination of rare species, show the same plant distributions and plot dynamics (Fig. 5). Clustering can be made among the plots of June 1992 because of the importance of Bare ground surface (Bgr) after mowing or ploughing treatment in January 1992 (except for site IV). The control plots are mostly dominated by perennial species observed in figure 4. In site IV, Hieracium pilosella and ratio of bare ground surface are also important in control plots. This could explain why plots of 1992 couldn't be clustered and why no general direction could be observed in abandoned plots in this site.
18Regularly mown plots are generally characterised by the appearance and maintenance of some annuals, biennials or hemicryptophyte rosettes. Species present only in one or two sites can complete the main list given in figure 4 e.g. Veronica prostrata, Daucus carota and Hypericum perforatum in site I; Picris hieracioides, Medicago lupulina and Polygala calcarea for site II; Lotus corniculatus, Leontodon hispidus and Anthyllis vulneraria for site III; Digitalis lutea, Daucus carota and Galium pumilum for site IV.
19Table 2 and CA realised on plots and vegetation data means that inter-site variations are more important than inter-treatments variations, as far as the botanical compositions are concerned. This could be an important problem when drawing general models of management which have to be functional inside a given range of heterogeneity, even at a small regional scale
Plant richness, diversity and evenness
20Calculation of diversity index shows a general increase of species diversity and richness of regularly mown plots in site I to III (Fig. 6). Differences are generally significant with control plots from 1992 to 1994. (Tab. 3). Nevertheless, no significant differences appeared between initial treatments in 1992 and year to year evolution of specie evenness. Concerning, the abandoned plots, there is no significant difference between treatments, controls="true" and year to year changes for richness, diversity and evenness. Comparison of the mean richness, diversity and evenness (n=3) between control shows only significant difference between site I and the other ones.
Fig. 4: Ordination by correspondence analysis of vegetation data of the foursites after elimination of rare species. Time series (92→93→94), of control plots (*), ploughed (■) or mown (●) plots and abandoned since 1992; ploughed (□) or mown (○) plots in 1992 and regularly mown in June since this year. Key for species are given in table 2
DISCUSSION
21The first correspondence analysis confirm the results of Liger (1952) and De Foucault & Frileux (1988) who have shown the existence, in the Seine valley, of a climatic gradient responsible of the distribution of chalk grassland plant communities from upstream to downstream sites. In consequence, the results obtained after the application of conservation management treatments (mowing, ploughing) couldn't be applied to another sites because of the lack of correspondence between their plant composition. On the other hand, considering plant traits, our results show that a predictive model could be sketch based on four major functional vegetation groups and four regimes of conservation management (Fig. 7).
22The vegetation of control or the abandoned plots since 1992 is mostly composed of perennial species which can be split into two groups based upon traits associated with morphology and established strategy, (i) A group of large space holding species with clonal spread and competitor or stress-tolerant strategy mostly present in control plots, (ii) A group without vigorous clonal spread and with shallow rhizome or rosette, which can occurred in perturbed plots and persist after two years of abandonment. Control plots exhibit no year to year major changes in their plant composition. Without considering rare species, their vegetation are relatively homogeneous and clustering of the plots can be made independently of site and time recorded. The dominance of certain tussock species is responsible of a decrease in plant richness which is the slightest in these plots (Bobbink & Willems, 1987; Stampfli, 1992).
23In June 1992, several months after treatments, the vegetation recorded on perturbed plots is characterised by the dominance of annual and biennial species with ruderal or stress-ruderal strategies. Annuals lived only one growing season and emerge from buried seeds after the adult vegetation has been removed (Dolman & Sutherland, 1994; Dutoit & Alard, 1995b). Facultative annuals or biennials all flowered at the end of the first year
24These plots were also characterised by the importance of the percentage of bare ground, except for site IV. In this site, the depth slope (68%) seems to be responsible of a constant rejuvenation of the herbaceous layer, even in the control plots. No difference can be detected between the ploughed and the mown plots because rotovatation permits a rapid development of vegetation from pieces grass tussocks and fragment of perennial forbs in the soil surface (Dolman & Sutherland, 1994).
25Annuals and biennials could be maintained in plots perturbed in 1993 and 1994 in contrast to abandoned plots where new species appeared after one year of abandonment. These species are generally perennials with intermediate strategies (S/CSR, CSR/SC). Nevertheless, three years of abandonment after mowing or ploughing in 1992 cannot create a grassland similar to that of control plots which have been abandoned since the 1950's. Plant composition of the two types of plots are still quite different. The general increase of species richness and diversity of mown plots have been frequently described and is confirmed by our results (Bobbink & Willems, 1991). Despite of this, it could be noticed, that after three years of abandonment, these index decreases. No significant difference is found for species richness and diversity between the abandoned plots and the controls.
CONCLUSION
26Study of the impacts of conservation management on plant communities has yielded an overwhelming body of special cases but few general principles (Grubb, 1977; Van der Maarel & Sykes, 1993). This is due in part to the difference between study sites, methods and the phenomenological non-predictive approach (Keddy, 1990). Further progress requires a predictive approach (Gaudet & Keddy, 1988; Keddy, 1992) that will enable general principles to be deduced and that will apply beyond the species and conditions of a particular study site. The use of functional vegetation groups based on plant ecological traits have been already used with success for classifying European riverine wetland ecosystems (Hills et al, 1994) or for predicting changes in wetland plant communities along successional series (Keddy & Shipley, 1989; Boutin & Keddy, 1993).
27Our study shows that similar models could be also built for chalk grassland plant communities. Moreover, if we clearly need to extend our information about plant traits to define more effective functional vegetation groups, these models can be based initially stage on major plants traits easily available in literature (Tutin et al, 1964-80; Grime et al, 1988). This aspect is particularly important considering the need of biological management to conserve semi-natural areas (Sutherland & Hill, 1995) and the necessity to make accessible predictive models for biological conservationist.
Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
Monitoring the impacts of marine aggregate extraction
Knowledge Synthesis 2012 (GIS SIEGMA)
Robert Lafite et Michel Desprez (dir.)
2014
L’Impressionnisme, les arts, la fluidité
Philippe Fontaine, Frédéric Cousinié et Pierre-Albert Castanet (dir.)
2013
Suivi des impacts de l’extraction de granulats marins
Synthèse des connaissances 2012 (GIS SIEGMA)
Robert Lafite et Michel Desprez (dir.)
2012
Images sculptées au seuil des cathédrales
Les portails de Rouen, Lyon et Avignon (xiiie-xive siècles)
Franck Thénard-Duvivier
2012
Communistes au Havre
Histoire sociale, culturelle et politique (1930-1983)
Marie-Paule Dhaille-Hervieu
2009
Jean Lorrain, « produit d’extrême civilisation »
Éric Walbecq et Jean De Palacio (dir.) Marie-France David-de Palacio (éd.)
2009
Genre & Éducation
Former, se former, être formée au féminin
Paul Pasteur, Marie-Françoise Lemmonier-Delpy, Martine Gest et al. (dir.)
2009