Search

International audience; Questions: Evaluating reasons for the success (or failure) of restoration projects is one of the major goals for applied ecologists in the context of the dramatic worldwide decline in biodiversity. To that end, finding appropriate indicators and reference ecosystems is mandatory, especially for habitats in which restoration projects have been barely evaluated over the long term, such as coastal sand dunes. Do different indicators provide complementary information to evaluate restoration success? How successful is sand dune restoration after 30 years?Location: Brittany (France).Methods: We report changes in the plant community over a period of 30 years following restoration using trampling protection and marram grass (Ammophila arenaria) planting in four sites that were highly degraded by sand extraction and over-frequentation. We used several indicator types (geomorphological, taxonomic, functional) at different spatial scales (alpha, beta, gamma) to assess recovery since restoration in four sites of the northern coast of Brittany and undertook comparison with reference sites.Results: Our results indicate that over the 30 years after restoration the gamma richness of typical species and overall vegetation conservation status increased. We also found that restoration induces the recovery of zonation in plant communities, which become organized along the typical sea-inland gradient. We showed that use of functional traits and diversity indices is effective to compare restored communities with reference data from the literature. Based on this approach, we demonstrated that, according to most functional indicators, restored communities converge over time with the patterns found in reference data.Conclusions: We demonstrated that restoration of coastal dunes after 30 years induced the recovery of sand accumulation, typical species cover, sea-inland plant community zonation and community functional characteristics. Despite these positive results, which indicated overall success and confirmed interest in such restoration projects, we found an important discrepancy in restoration success among study sites (e.g., level of recovery of typical species cover, functional diversity). Finally, our study confirms the interest of functional composition and functional diversity as restoration success metrics.

La désertification des parcours pastoraux steppiques arides constitue une des problématiques fondamentales de développement dans le sud de la Méditerranée. Elle pose avec acuité les questions de restauration et de gestion de ces écosystèmes qui ont été, dans la plupart des cas, soumis au surpâturage depuis des décennies. Divers modèles ont été développés pour expliquer la dynamique des écosystèmes arides dont l’origine, les causes et les mécanismes des changements sont souvent peu ou pas connus. Le suivi à long terme et la connaissance de l’histoire d’un écosystème dégradé ont été souvent avancés comme le meilleur moyen de comprendre ces changements et surtout de définir les références pour le pilotage des actions de réhabilitation de cet écosystème. Dans le cas présent, le système préexistant est une steppe aride dans laquelle, l’Alfa (Stipa tenacissima L.) a été l’espèce dominante et dont a été montré le caractère «clef de voûte» vis-à-vis de la composition spécifique et des conditions édaphiques. Trente-cinq années de surveillance ont permis d’évaluer le fonctionnement de cet écosystème au niveau de la production primaire utilisable par les troupeaux et de faire ressortir l’importance des fluctuations dues au climat par rapport aux changements de végétation et de sol, induits par le surpâturage ovin. Les résultats de ce travail permettent de discuter les modèles dynamiques existants et d’avancer des propositions quant aux opérations de réhabilitation et les nouveaux modes de gestion préconisés dans les parcours steppiques d’Algérie., Aidoud Ahmed, Slimani Halima, Rozé Françoise. La surveillance à long terme des écosystèmes arides méditerranéens : quels enseignements pour la restauration ? Cas d’une steppe d’Alfa (Stipa tenacissima L.) en Algérie. In: Ecologia mediterranea, tome 37 n°2, 2011. Numéro spécial, 7e conférence européenne SER en écologie de la restauration, 23-27 août 2010, Avignon, France. pp. 17-32.

Question: Which biological traits (persistence, regeneration, dispersion traits and resource requirements) may explain the distribution of woodland flora along an urban–rural gradient? Location: The study was carried out in three medium-sized conurbations of north-western France: Angers, Nantes and Rennes. Methods: We sampled the vegetation of 36 small woodlands of about 1.5 ha composed of non-planted vegetation along an urban–rural gradient. We characterized the position of woodlands along the urban–rural gradient by examining adjacent land cover. By using an ordination analysis (RLQ), we analysed which traits out of –23 tested were related to the contrasted distribution of species along the urban–rural gradient. Results: Species that are more likely to be found in urban woodlands than rural woodlands have different persistence traits (higher specific leaf area, more often rosette or semi-rosette form, less underground vegetative multiplication), resource requirements (affinity for base-rich and fertile soils) and regeneration traits (short life-span). Dispersion traits were not related to the distribution of species along the urban–rural gradient. Conclusions: Our approach identifies traits that can help to determine the vulnerability of forest species as a result of the environmental changes that follow urbanization. Limiting the influence of the urban environment on habitat quality (particularly disturbance and soil enrichment) is likely to be of major importance in maintaining the plant biodiversity in woodlands.

International audience; To assess the possibility of restoring the original plant community of grey dune from a pine plantation, a synchronic study of the seed banks of three habitats, the grey dune, the pine plantation, and a clearing, was conducted at Quiberon (France). Five stations were selected: a Grey dune, Pine plantation 1, Pine plantation 2, Clearing 1, and Clearing 2. In contrast to the pine plantations, clearing 1, subject to heavy human disturbance, had a great similarity between its seed banks and its established vegetation. The larger clearing (C2) tend to resemble the grey dune, whereas the other was dominated by ruderal species. The analysis of the seed banks demonstrated differences between the clearings and the pine plantations from which they were derived. The potential vegetation of the clearings was closer to that of the grey dune and had a larger number of species with a short life cycle than the pine plantations. The seed banks of the pine plantations were not similar to that of the grey dune, but also differed from that of closed vegetation. This study suggested that it may not be possible to obtain a short term restoration of a grey dune after just clear felling only with the help of seed banks. On the other hand, seed dissemination could facilitate a restoration in a larger time scale as for clearings.

Dunes, which account for 13% of the Ille et Vilaine north Brittany coast, France, were degraded by high tourist pressure, and they were restored from 1988 onward. Ten years after commencing work an assessment of the restoration was made on three dunes: Les Chevrets, L'Anse Du Guesclin, and Le Verger. Annual monitoring of the vegetation and dune morphology provided an opportunity to study the restoration process. The dune front, which is similar to a reference pioneer dune, lies in front of the mobile dune. The highest part of the dune is an advanced stage of mobile dune. The restoration of the fixed rear dune was accomplished neither in terms of vegetation composition nor in terms of species richness. The vegetation parameter study allows differentiation between the dune front, the dune summit, and the rear dune. The variation in species richness and floristic composition from one zone to another can be explained by abiotic factors such as salinity and the accretion of sand. The restoration was satisfactory in terms of the geomorphology. Marram grass is a good tool for restoring the topography, but it will take a very long time to restore the conservation value of the dune.

The sand-dunes of Quiberon was chosen as a pilot site to investigate experimentation in conservatory management. Sand burial is necessary to conserve the semi-fixed dune which is a transitory dynamic stage. In the fixed dune, low disturbances benefit the vegetation diversity while heavy ones create serious injury. An opening of the milieu can restore very fast but a naked substrate is difficult to heal. The deposition of gorse branches is then efficient to facilitate the restoration. The fixed dune biodiversity is linked to human activities. Disturbances, natural or not, may be used as management tools. To cite this article: S. Lemauviel et al., C. R. Biologies 326 (2003).

In recent decades, human trampling has become an important factor in the degradation of Atlantic heathlands. In a previous paper we showed that short-term responses of heathland to trampling were dependant on environmental parameters such as season and weather conditions (Gallet, S., Roze, F., 2001b. Resistance of Atlantic Heathlands to trampling in Brittany (France): influence of vegetation type, season and weather conditions. Biological Conservation, 97, 189–198). Here, long term responses of experimental trampling, characterised by the resilience and the tolerance of communities and plants appear also dependant on these parameters. Dry and mesophilous heathlands are more tolerant to trampling in winter than in summer. In the case of mesophilous heathland, this is linked to high plant resilience, especially of Erica ciliaris . In summer, influence of meteorological conditions at the time of trampling depend on the species; indeed wet conditions lead to lower tolerance of Erica cinerea but to higher tolerance of E. ciliaris . Management of tourist pressure on natural or semi-natural sites must take visitors on the more tolerant communities, so it must take into account the variability of tolerance to trampling of the different types of vegetation present and needs to be adaptable to environmental conditions.

As a result of the large increase in tourist pressure, trampling has become an important factor in the degradation of natural areas and particularly Atlantic Heathlands. In this study, the resistance of dry and mesophilous heathlands was studied experimentally, under different environmental conditions. The trampling was applied in winter, and in summer in periods of wet and dry weather. The types of heathlands studied appeared to be moderately resistant to trampling. Mesophilous heathlands were on the whole more sensitive to trampling than dry heathlands. The resistance of these communities and of their component species varied greatly in relation to season and weather conditions. For example, Erica ciliaris was more resistant to trampling in summer than in winter, and in summer Erica cinerea was more sensitive to trampling in wet weather than in dry conditions. The analysis of the resistance of species showed that heather species were more sensitive than the rest of the plant cover. These heathers are characteristic species of the heathlands that were studied. The definition of acceptable visitor pressure thresholds must not be based on the index of resistance of the whole community, but on that of the characteristic species, in this case heathers. The aim of visitor pressure management must be to preserve the vegetation structure and not just simply plant cover.

The 18th century afforestation campaign carried out in western France to fix sand dunes has left only a small surface of non-forested dunes. As dune plantations have only a small conservation value, it would be of great interest to restore grey dunes. But, for the moment, there is no proof that pine felling would lead to grey dunes. So, a study has been carried out on pine forest clearings along the French Atlantic shoreline. Floral and ecological data were carried out and analysed with two statistical tools, CA (canonical analysis) and CCA (canonical correspondence analysis). The clearings appear very different from one bank of the Gironde river to the other. South of the river, the clearing vegetation develops on acid soils and possess some thermophilous species. North of the river, the clearings are composed mostly of species growing best on calcareous soils. Apart from geographical variations, clearings showed two patterns of response: development of closed thickets and tall heathlands; or an open vegetation similar to growing heathlands and grasslands. The cover, the height, the floral richness per strata as well as pedological characteristics are the parameters which explain the best floristic composition of pine clearings. The vigour of pine forests and the degree of exposure to coastal influences determine whether close or open vegetation develops. Where there is an open vegetation, restoration of grey dunes may be possible. In the other case, alternative solutions, such as the restoration of dune woodlands or dune heath, may provide best conservation values.

Trampling responses of an Atlantic heathland dominated by furzes were experimentally determined. They concerned two levels of biological systems: the vegetal community level (community scale) and the species level (intracommunity scale). Two trampling experiments were conducted: the first one was carried out in summer and the second one in winter. Heathland was trampled over 4 months, at differents intensities: 10, 30 and 90 persons a day. Whether trampling was conducted in the summer or in the winter, immediate responses of heathland (community scale) were similar: resistance of this vegetal community did not vary with the trampling season. Responses also showed that there is no linear relationship between vegetation degradation and trampling intensity. Deferred vegetation responses were similar, but within the community heathland (intracommunity scale) responses of a taxonomic group can depend on the trampling season. For example, resilience of furzes trampled in winter was greater than resilience of furzes trampled in summer. All these results should be considered in order to establish management orientations of trampled heathlands.

Cellulose represents an important part of the organic matter returned to soil with the litter of alfa. Thus cellulolytic activity is a good indicator of the soil microbial activity. Different experiments in vitro compared the influence of different conditions, particularly the gradient of aridity, corresponding to two sites : arid alfa steppe and semi-arid alfa steppe. Soil samples were collected underneath and between tufts of vegetation. Different temperature and humidity were chosen in function of climatic data. Two levels soil depth were studied. From our results, it can be concluded that cellulolytic activity in the arid alfa steppe soils is less intense than in the semi-arid ones, as determined in vitro., La cellulose constitue une part importante de la matière organique apportée par les litières d’alfa. L’activité cellulolytique étant un bon indicateur de l’activité microbiologique des sols, différentes expériences in vitro ont été effectuées afin de comparer l’influence des conditions stationnelles et en particulier du gradient d’aridité dans deux stations : une steppe d’alfa aride et une steppe d’alfa semi-aride. Des prélèvements ont été effectués sous les touffes d’alfa et entre les touffes. Différents taux d’humidité et températures ont été choisis en fonction des conditions climatiques. Deux niveaux de sol ont été étudiés. Il apparaît que l’activité cellulolytique in vitro est moins intense dans la station la plus aride. Les sols de la station aride sont proportionnellement plus actifs aux fortes températures et humidités faibles que ceux de la station semi-aride., Bessah Rahma, Rozé Françoise, Nedjraoui Dalila. Activité cellulolytique in vitro de sols de deux steppes à alfa (Stipa tenacissima L.) d’Algérie. In: Ecologia mediterranea, tome 25 n°2, 1999. pp. 185-192.

Summary Our study is concerning with plant succession in neglected fields and grasslands, using diachronic and synchronic analysis. It is aiming to build dynamic series and to characterize the ecological parameters during these successions. Two spatial scales and two temporal scales is used: on one hand, the bocage landscape scale for long term analysis and on the other hand the parcel scale for short term studies on permanent plots and on permanent lines. Dynamic processes occur very rapidly, using two strategies: by marginal growth from the hedges and by patch growth into the parcel. Various succession scenarios have been observed. They are depending on the parcel history (particularly its original state before abandonment) and also on the biotic environment of the parcel. Most of the plant succession theories can be applied to recolonization of neglected lands.

FR2116; International audience; The increasing rate of urban sprawl continues to fragment European landscapes threatening the persistence of native woodland plant communities. The dynamics of woodland edges depend on the characteristics of woodland patches and also on landscape context. Our aim was to assess the extent of edge influences on the understorey vegetation of small native woodlands in rural and urban landscapes. The study was carried out in two cities of north-western France. Ten comparable woodlands, each of about 1.5 ha, were surveyed; five were situated adjacent to crops and five adjacent to built-up land. Vascular plant species were recorded in 420 3 9 3 m plots placed at seven different distances from the edge (from 0 to about 45 m from the edge). Soil pH, light levels, level of disturbance and tree and shrub cover were also recorded. Plant species were first classified as nonindigenous or indigenous and then three groups of indigenous species were distinguished according to their affinity for forest habitat (forest specialists, forest generalists and nonforest species). We inferred certain ecological characteristics of understorey vegetation by using Ellenberg values. An inter-class correspondence analysis was carried out to detect patterns of variation in plant community composition. Linear mixed models were used to test the effects of adjacent land use, distance from the edge and their interactions on the species richness of the different groups and on the ecological characteristics of vegetation. Total species richness, richness of forest generalists and of non-forest species decreased from edge to interior in both urban and rural woodlands. The number of non-indigenous species depended mainly on urban-rural landscape context. Urban woodland edges were not as rich in forest specialists as rural edges. More surprisingly, the number of forest specialists was higher in rural edges than in rural interiors. Community composition was mainly affected by urban-rural context and to a lesser degree by the edge effect: the community composition of urban edges resembled that of urban interiors whereas in rural woodlands vegetation near edges (up to 10 m) strongly differed from interiors with a pool of species specific to edges. Urban woodland vegetation was more nitrophilous than rural vegetation in both edges and interiors. A major difference between urban and rural vegetation was the distribution of basiphilous species according to distance from the edge. Generally edge vegetation was more basiphilous than interior vegetation however the presence of basiphilous species fell off quickly with distance from the edge in rural woodlands (in the first 10-15 m) and more slowly (from 25 m onwards) in urban woodlands. This pattern was linked to variation in measured soil pH. As regards the conservation of flora in small native woodlands, it appeared that invasion of exotic and non-forest species was currently limited in both urban and rural landscape contexts but might pose problems in the future, especially in urban woodlands. Forest species were not negatively affected by the edge effect and indeed edges seemed to provide important habitats for this group. Hence conservationists should pay particular attention to the protection of edges in urban woodlands.

Questions: How does urbanisation influence soil mineral nitrogen stocks (nitrate and ammonium stocks) and what are the consequences of these modifications on the functional diversity of the herbaceous vegetation (vascular plants) and the seed bank?Location: Nine study sites were located on an Urbanisation gradient in the city of Rennes, France.Methods: Three Urbanisation levels were defined: urban areas (high grey/green ratio,), suburban (medium grey/green ratio) and periurban (low grey/green ratio). For each urbanisation level. nitrates and ammonium stocks were quantified; the herbaceous vegetation was surveyed as well as the soil seed bank (using the seedling emergence method).Results: Nitrate concentration increased with Urbanisation (the nitrate level in urban plots was twice the concentration in periurban ones) whereas the ammonium level was higher in periurban areas than in urban areas. In urban plots, the vegetation and the seed bank were more nitrophilous, whereas the nitrogen requirement was lower for periurban species. The relationship between the seed bank and the above-ground vegetation was not significant.Conclusions: The higher nitrate concentration in the urban area appeared to be related to higher concentrations of atmospheric pollutants found in this area and lower ammonium levels may be related to the higher temperature in urban areas (leading to higher nitrification rates). The shift in the composition of the seed bank and vegetation appeared to be a consequence of higher nitrate stocks. The dissimilarity between the seed bank and vegetation may be caused by enhanced emergence of nitrophilous species in urban areas.

FR2116; International audience; The effect of urbanization on species distribution has been extensively documented, but a main challenge in urban ecology is to better understand the factors causing different distributions among species in response to urbanization. Hence, this paper aims to compare the effects of urbanization on woodland plant assemblages in two cities and to describe species responses by using several indicators. The study was carried out in the cities of Angers and Rennes (North-Western France) where 11 isolated woodlands were surveyed along an urban-rural gradient in each city. Abundance data of spontaneous species were collected from 220 quadrats. The effect of land cover (within a 500 m buffer around each woodland) on species assemblages was investigated by Canonical Correspondence Analysis. Buildings and pavement areas were the most significant predictors of species composition, and the effect of location in Angers or Rennes appeared on the second axis. More than 60% of the most frequent plant specieswere indicator of urban or rural location and their preferences were similar in the two cities. These lists of urban and rural indicator species were compared with Ellenberg's indicator values and two other indicators specific to forest environment. The species which grow preferentially in urban woodlands are species which are already known to be associated with recent forests rather than ancient forests; with hedgerows rather than woodlands. The opposite pattern was observed concerning rural species. Moreover, urban indicator species have higher optima for soil pH and soil nitrogen content than rural indicator species. Different characteristics and history of forest habitat—continuity of the forest land cover, linearity of the habitat, change in adjacent land cover and land use—could select the same species, and the responses of the latter might involve different preferences concerning soil alkalinity and nutrient status.

In the last few decades, trampling has become a cause of major disturbance to natural areas. Trampling can occur regularly either seasonally or throughout the year, or as a single impact depending on a recreational activity or an event. This study, carried out on a coastal heathland in north Brittany (France), was based on the hypothesis that the vegetation response can differ depending on whether the trampling is repeated or due to a single event. The aim was to compare the resistance and the short-term recovery capacity of selected species in coastal heathland communities subjected to a trampling experiment, applied once or five times (with equal total number of passes from 0 to 800). The temporal distribution of the trampling did not affect the response of Erica ciliaris, whereas for Erica cinerea and Ulex gallii, trampling was slightly less damaging when applied once rather than five times. These differences could be linked to the lower resistance and the higher recovery capacity of E. ciliaris, compared with the two other species. This study also emphasized the influence of site location by comparing observations with previous studies. Plant resistance calculated here appeared to be lower than that calculated for a less stressful site. The relationship between plant traits and response to trampling was also investigated. The effects of the different times between disturbance application and observations of the short-term response of woody vegetation to experimental trampling is also discussed.

Dunes that are protected because of their very rich and diverse plant communities are often exposed to excessive visitor pressure. The effects of trampling on the habitat must be known from a conservation viewpoint but also are important for management. To determine the response of plant assemblages to trampling by people, an experimental study was conducted on the state-owned dunes at Quiberon (Brittany, France). Indices of resistance and resilience were used to compare three typical plant communities belonging to the various landscape units: mobile dune, semifixed dune, and fixed dune. The strong contrasts between communities belonging to different successional stages reflect their ecological functioning. The mobile dune and semifixed dune with their low resistance contrasted with the fixed dune. Only the vegetation cover of the semifixed dune benefited from long-term trampling and had a very high resilience (134%). This response could be explained by a good balance of two opposite factors: soil compaction increasing soil stability and moisture content, and vegetation destruction. Because of their low resilience, trampling seems to be harmful for fixed dunes in the long term. The tourist pressure seems easier to integrate in to the mobile dunes and the semifixed dunes if periods of recovery are included in the management.

The soil dynamics of old-fields, characterized by the early establishment of broom ( Cytisus scoparius L.), is analysed in the Breton bocage. The role of former land use practices is estimated by comparing two similar plant successions, post cultivation and post pasture. Different variables (organic matter, mineral N content, microbial biomass, total microbial activity and mineralization) are measured within the soil layer disturbed by farming (between –5 and –25 cm). One year after abandonment, the parameters taken into account show similar values for both series. After 8 years, in broom thickets, the two soil series are characterized by an increase in microbial activity, organic matter and mineral N contents. However, the post cultivation thicket differed by higher N content and mineralization rate than the post pasture thicket, but also by a smaller microbial biomass. It appears that, if the development of similar vegetational communities promotes the same kind of soil dynamics, the former land use practices significantly modify this evolution, for at least the first 8 years following abandonment.

Middle Brittany landscape have been studied at St Julien. Many alterations occur since 30 years, especially, following the break caused by the recession of agriculture. At the level of the study area, landscape diversity and complexity increased strongly after agriculture abandonment. This trend seems to be linked with the introduction of new landscape units and the abandonment of parcels. However, at the village's scale the large abandoned areas promote more shaded effects because of the uniformisation of the space. The importance of vegetation dynamics in former and abandoned agricultural lands is then emphasized. Concurrently, hedgerows' network has sustain deep damages that caused a loss of landscape's diversity in term of corridors., Le paysage en Bretagne intérieure, étudié au travers de la commune de St Julien (Côtes d’Armor) a subi d’importantes mutations depuis trente ans. En particulier, une cassure provoquée par la régression de l’agriculture s’est révélée avec force. Les abandons de parcelles sont une des principales causes de changement. La diversité et la complexité du paysage se sont alors accrues à l’échelle du groupe de villages. L’introduction de nouvelles unités écologiques et la dynamique propre des unités abandonnées sont apparues être le moteur de cette évolution. Par contre à l’échelle du village la forte vague de déprise agricole apparaît avoir des effets plus nuancés, provoquant l’homogénéisation de l’espace. La dynamique de la végétation des zones abandonnées est désormais le moteur de la dynamique paysagère. Parallèlement, le réseau bocager a connu en trente ans de profondes dégradations, sources d’une diminution de la diversité paysagère en terme de corridors., Diquélou Sylvain, Rozé Françoise. Effets de l'intensification de l'agriculture et de la déprise agricole sur la dynamique paysagère en Bretagne intérieure. In: Ecologia mediterranea, tome 23 n°1-2, 1997. pp. 91-106.

International audience

The increase of urbanization during the last decades (and the related losses of biodiversity) has resulted in a need to consider biodiversity conservation in urban areas. The present study focuses on the changes in plant communities of small woodlands (about 1.5ha) along an urban-rural gradient in three urban areas of the Massif armoricain in north-western France. Plant communities are strongly affected by the degree of urbanization. Urban woodlands are richer in exotic species but indigenous and forest species are dominant in all woodlands. However, the abundance of some indigenous species changes along the urban-rural gradient which may be related to their biological attributes. The local environment, in particular soil pH and soil fertility, seems to be the most important factor determining community differences along the gradient. Forest history and the dispersal ability of species are of minor importance in explaining these differences. At the scale of individual woodlands, the edge effect is an important factor influencing species richness: edges are richer in species in urban as well as in rural woodlands. The species richness of the litter seed bank changes with distance from the edge but is not much different between urban and rural woodlands. Species composition differs largely between litter seed bank and established vegetation. Non-forest species are more frequent in the litter seed bank than in the established vegetation indicating reduced seedling recruitment compared with forest species (that are dominant in the established vegetation). This work contributes to a better understanding of ecological mechanisms involved in the response of plant species to urbanization and could improve the integration of woodlands in urban planning and management.; L'urbanisation croissante de ces dernières décennies amène aujourd'hui à considérer la place de la biodiversité dans les territoires urbains. Ce travail est centré sur l'étude des modifications de communautés végétales de petits espaces boisés (1.5 ha en moyenne) le long d'un gradient d'urbanisation dans trois agglomérations du Massif armoricain : Angers, Nantes et Rennes. Les communautés végétales sont fortement modifiées par l'urbanisation. Les bois urbains sont plus riches en espèces exotiques mais restent dominés par des espèces indigènes et forestières. Cependant, la distribution des espèces indigènes est modifiée le long du gradient d'urbanisation et certains traits biologiques semblent associés à cette distribution. L'environnement local et en particulier le pH et la fertilité des sols semblent être les facteurs les plus importants en lien avec les modifications des communautés végétales. L'histoire des espaces boisés semble avoir une importance moindre. De plus, les caractéristiques de dispersion des espèces ne semblent pas être impliquées dans la distribution des espèces le long du gradient d'urbanisation. Au niveau de l'espace boisé, l'effet lisière est un déterminant important de la richesse spécifique, les lisières étant les plus riches en espèces aussi bien dans les bois urbains que ruraux. La banque de graines de la litière est affectée par la distance à la lisière mais elle est peu différente entre les bois urbains et ruraux. La composition de la banque de graines de la litière diffère largement de la composition de la végétation du bois. Les caractéristiques du milieu forestier empêchent probablement la germination d'un grand nombre d'espèces non forestières venant des habitats adjacents aux espaces boisés et sont donc importantes pour le maintien d'une végétation dominée par des espèces forestières. Ce travail permet d'ouvrir des perspectives pour une meilleure compréhension des mécanismes écologiques impliqués dans la réponse des espèces à l'urbanisation et pourrait contribuer à une meilleure intégration de ces espaces boisés dans l'aménagement des villes.

The increase of urbanization during the last decades (and the related losses of biodiversity) has resulted in a need to consider biodiversity conservation in urban areas. The present study focuses on the changes in plant communities of small woodlands (about 1.5ha) along an urban-rural gradient in three urban areas of the Massif armoricain in north-western France. Plant communities are strongly affected by the degree of urbanization. Urban woodlands are richer in exotic species but indigenous and forest species are dominant in all woodlands. However, the abundance of some indigenous species changes along the urban-rural gradient which may be related to their biological attributes. The local environment, in particular soil pH and soil fertility, seems to be the most important factor determining community differences along the gradient. Forest history and the dispersal ability of species are of minor importance in explaining these differences. At the scale of individual woodlands, the edge effect is an important factor influencing species richness: edges are richer in species in urban as well as in rural woodlands. The species richness of the litter seed bank changes with distance from the edge but is not much different between urban and rural woodlands. Species composition differs largely between litter seed bank and established vegetation. Non-forest species are more frequent in the litter seed bank than in the established vegetation indicating reduced seedling recruitment compared with forest species (that are dominant in the established vegetation). This work contributes to a better understanding of ecological mechanisms involved in the response of plant species to urbanization and could improve the integration of woodlands in urban planning and management.; L'urbanisation croissante de ces dernières décennies amène aujourd'hui à considérer la place de la biodiversité dans les territoires urbains. Ce travail est centré sur l'étude des modifications de communautés végétales de petits espaces boisés (1,5 ha en moyenne) le long d'un gradient d'urbanisation dans trois agglomérations du Massif armoricain : Angers, Nantes et Rennes. Les communautés végétales sont fortement modifiées par l'urbanisation. Les bois urbains sont plus riches en espèces exotiques mais restent dominés par des espèces indigènes et forestières. Cependant, la distribution des espèces indigènes est modifiée le long du gradient d'urbanisation et certains traits biologiques semblent associés à cette distribution. L'environnement local et en particulier le pH et la fertilité des sols semblent être les facteurs les plus importants en lien avec les modifications des communautés végétales. L'âge des espaces boisés semble avoir une importance moindre. De plus, les caractéristiques de dispersion des espèces ne semblent pas être impliquées dans la distribution des espèces le long du gradient d'urbanisation. Au niveau de l'espace boisé, l'effet lisière est un déterminant important de la richesse spécifique, les lisières étant les plus riches en espèces aussi bien dans les bois urbains que ruraux. La banque de graines de la litière est affectée par la distance à la lisière mais elle est peu différente entre les bois urbains et ruraux. La composition de la banque de graines de la litière diffère largement de la composition de la végétation du bois. Les caractéristiques du milieu forestier empêchent probablement la germination d'un grand nombre d'espèces non forestières venant des habitats adjacents aux espaces boisés et sont donc importantes pour le maintien d'une végétation dominée par des espèces forestières. Ce travail permet d'ouvrir des perspectives pour une meilleure compréhension des mécanismes écologiques impliqués dans la réponse des espèces à l'urbanisation et pourrait contribuer à une meilleure intégration de ces espaces boisés dans l'aménagement des villes.