Polyphagy is a very common trait among insects. In this study, we focus on a generalist bark-beetle predator, Thanasimus formicarius (L.) (Coleoptera, Cleridae), which feeds on many scolytids in spruce, pine and broad-leaf stands. It is known to respond to the pheromones of many scolytids, among which the most harmful spruce bark beetle in Europe, Ips typographus (L.). The adults attack scolytid adults and oviposit on attacked trees where their larvae feed upon immature stages of the prey. However, a bottom-up process limits Thanasimus formicarius’ impact on spruce bark beetles, because in most cases the bark of spruce is too thin for sheltering pupal niches and mature larvae have to leave the trees. On pine however, pupation is quite successful and reproductive success is high. The present work estimates the advantages (complementary prey during gaps among the phenology of pine bark beetles or due to the population fluctuations of most scolytids) and constraints (landing on unsuitable host trees for the predator’s reproduction) for T. formicarius to have a wide range of prey. Passive barrier-trappings showed that the presence and abundance of scolytid species vary strongly from year to year. So, polyphagy in T. formicarius appears as a response to fluctuating prey supplies. This way of foraging may lead T. formicarius towards stands not always favourable for its development (for example, spruces). At the tree level, funnels and pitfall-traps caught high numbers of third-instar T. formicarius larvae walking on the bark surface of standing spruces infested by Ips typographus (respectively 365 and 70 L3s). After feeding into the whole infested part of the trunk, these larvae are obliged to migrate outside of the galleries to favourable pupation site (e.g. the base of the trees where the bark is thicker), or even to leave the trees and search for an acceptable pupation substrate in the litter. At the landscape level, different trapping experiments showed a correlation between catches of T. formicarius and the proportion of pines around each trap. Consequently, in a metapopulation landscape pattern, pines would act as “sources” of predators, whilst spruces are “sinks”. Indeed, Thanasimus formicarius are trapped in higher numbers in mixed stands comprising pines. This observation is also corroborated in a four-year trapping experiment in the North-East of France, following the storms of December 1999. The predator/prey ratios (T. formicarius/I. typographus) were higher in stands comprising pines than in stands without pines. The first step of a method to estimate Ips typographus infestation trends thanks to the predator/prey ratios was also developed.

Les paysages français sont de plus en plus soumis aux pressions anthropiques. Ces pressions engendrent des modifications de la qualité et de la configuration des habitats, y compris forestiers, et ont des effets sur les plantes. Cette thèse aborde le rôle de la composition et de la configuration de la mosaïque paysagère extra- et intra-forestière sur la diversité de la flore en forêt. Il s'appuie sur 1932 relevés de l'Inventaire Forestier National répartis dans 19 départements de la moitié nord de la France. Les indices paysagers ont été calculés à partir de fonds cartographiques et d'images aériennes photo-interprétées. L'étude de terrain a été consacrée à la portée de l'effet d'un élément linéaire particulier du paysage : la route forestière. Le paysage a un effet important au-delà de la qualité de l'habitat local. La flore répond à trois principaux gradients paysagers : le premier est lié à la nature forestière ou non forestière du paysage, le second à la présence de coupes et peuplements jeunes, et le dernier reflète la composition en essences (feuillus ou résineux). L'effet des ouvertures paysagères (extra-forestière cultivée, coupe, peuplement jeune) sur la flore passe par la quantité, et surtout le nombre de taches et la longueur d'interfaces. Cet effet est positif sur la plupart des espèces, ce qui confirme le rôle des ouvertures dans le maintien de la diversité floristique. Les traits des espèces discriminés sont l'épi-zoochorie et l'anémochorie, la masse x longueur des semences et certaines stratégies de Grime. Ce travail montre qu'il est primordial d'appréhender le paysage sur un rayon de plus d'1 km. En comparaison, la route a un faible impact sur la flore, et plus souvent un effet positif que négatif. Les mécanismes sous-jacents à l'influence du paysage sont discutés à la lumière d'un effet temporel sur les plantes. Les implications pour la gestion des habitats en faveur de la biodiversité sont abordées. / French landscapes are submitted to increasing anthropogenic pressures. These pressures can affect plant species diversity by modifying the quality and spatial configuration of habitats, including forest habitats. This project addressed the role of the composition and configuration of the forest and non forest landscape mosaic on forest plant diversity. It was based on 1932 plots from the National Forest Inventory located in 19 "départements" of Northern France. Landscape indices such as fragmentation were measured or calculated from aerial photographs and existing national forest maps. A specific field study was devoted to analyse the depth of the effect of a particular linear landscape feature: the forest road. Landscape patterns had a significant impact in addition to local habitat quality. Plant composition was influenced by three main landscape gradients: (1) forest or non-forest habitat, (2) presence of fellings and young stands and (3) species composition (broadleaves or coniferous). The openings in the surrounding landscape (agricultural lands, forest felling and young stands) had an effect on plants through their quantity, number of patches and edge length. This effect was positive for most species, which confirms the role of the openings in maintaining plant diversity. The species traits discriminated were epi-zoochory and anemochory, mass x length of seeds and some Grime adaptive strategies. This work also underlined the importance to analyse landscape pattern effect within a buffer zone of more than 1 km around the plots. In contrast, we found that forest roads had a short depth of edge influence on plants (< 5 m) and road effect was more often positive than negative. The mechanisms underlying the influence of landscape mosaic are discussed in light of a temporal effect on plants. The implications for habitat and forest management to protect biodiversity are presented.

Growth rate is an important life history trait, which impacts fitness indirectly through its effect on the age and size at maturity, as well as directly through costs associated with accelerated growth such as increased predation risk. Genetic variation and plasticity in growth are widespread in nature, and local adaptation of growth rate may evolve due to divergent selection in different environments, for example related to predation risk, temperature or time constraints. I studied local adaptation of larval life history in the moor frog Rana arvalis, in a local network of ponds close to Uppsala. Local adaptation of growth rate and survival was studied in a reciprocal transplant experiment between ponds with different habitat characteristics. Meanwhile, differences among the populations in intrinsic growth, activity and response to predation were studied in a common garden experiment in the laboratory, where tadpoles were raised in the presence or absence of a predator and tested in direct predation trials. In the field, differences in growth among populations were found, independent of which pond the tadpoles were raised in, indicating that the ponds were similar growth environments. Survival differences among the populations depended on the pond, but local populations did not do better than foreign ones. In the laboratory, similar patterns in growth rate were found. All populations were highly plastic in their response to predation, having lower growth and activity in the predator-induced treatment and decreased mortality in the predation trials. Tadpole size was an important factor in escaping predation. One population clearly grew faster than the others in the field and in the lab, which could be explained in terms of its habitat of origin but was most likely related to the relatively late hatching of this population. Future studies are necessary concerning the possible costs of this accelerated growth and the importance of breeding phenology. Apart from the one differential population, I did not find evidence of local adaptation in the field or in the laboratory. The influence of habitat characteristics on tadpole life history was difficult to study, due to the limited number of ponds and many environmental differences among them. However, this thesis was a valuable pilot study concerning the design of experiments to study factors promoting and constraining local adaptation in landscape mosaics. An understanding of local adaptation at the scale at which gene flow occurs is important for the conservation of populations in fragmented landscapes as well as for the study of ecological speciation.

Landscape modification resulting in habitat loss, fragmentation and intensification of land use is a serious threat to the earth’s biological diversity and the primary cause of the current extinction crisis. Recent research suggests the human-modified area of the landscape (that is not ‘traditional’ habitat for native species but potentially once was), or the ‘matrix’ (as it is hereafter called) has a major influence on wildlife persistence in modified landscapes. However, the matrix is a poorly studied and inadequately understood element of a modified landscape. There are at least two reasons for this. Firstly, the predominance of island biogeography and metapopulation theories in ecological thinking ensured that the matrix has historically been ignored in ecological research. Secondly, it is difficult to analyse matrix effects without confounding effects of habitat loss, habitat fragmentation and other aspects of landscape modification. As landscapes are modified attributes of the matrix co-vary with attributes of patches and landscapes, entangling their relative impacts on wildlife. Thus to investigate the independent influence of the matrix on mammals I selected 19 study landscapes by rigorous criteria to control for all potentially confounding patch and landscape attributes such as remnant forest patch size, shape, vegetation type, condition and position in the landscape, presence of any large (potential ‘source’) native forest patches within the landscape unit, as well as matrix land use composition and history. A study landscape was defined as the area within a 500 m radius of a remnant forest patch edge. Landscapes were selected along a gradient of rural-suburban residential development spanning the broadest achievable range of what I call ‘matrix development intensity’. This gradient was quantified by a novel weighted road-length metric that considered multiple road attributes to give significance to the ecological impact of different roads. Mammals and their habitat were sampled in three landscape elements within each landscape: remnant patch core, remnant patch edge and matrix to allow a landscape level inference. Mammals were sampled by a combination of Elliott traps, wire cage traps, hair funnels, scats and direct sightings. Thirty environmental variables were measured including habitat structure and disturbance and a full floristic survey was conducted. The matrix intensity gradient was characterised by increased anthropogenic disturbance such as increased housing density, closer proximity of sample sites to houses and higher human disturbance across the landscape mosaic, including in the core of remnant patches. However, matrix intensity was not the greatest source of overall variation in structural and floristic habitat attributes. Therefore the confounding of matrix effects with effects of remnant forest patch habitat attributes were successfully ruled out. Management actions of individual landholders can shape habitat attributes essential to mammals across the landscape mosaic. Mammal response to matrix intensity was species specific. Several native species declined in abundance, others were more resilient to moderate levels of matrix intensity, one species increased in abundance, and at least one species appeared unaffected by matrix intensity. Native species richness peaked at moderate levels, while exotic species richness and feral predators increased with matrix intensity and were negatively correlated with native species. Species response to matrix intensity appeared related to their use of edge or matrix habitat. However, an ability to use the matrix per se may not translate into an ability to persist in a landscape where development substantially reduces the habitat or movement value of the matrix. Seven a priori models of various remnant patch habitat, landscape and matrix influences on terrestrial mammal species richness were tested. Matrix attributes were the most important determinants of species richness. Matrix development intensity had a strong negative effect while matrix vegetation structural complexity had a strong positive effect on mammal species richness. Distance to the nearest remnant forest habitat was relatively unimportant. I hypothesised that thresholds of matrix intensity would exist where native species decline in abundance and exotic and native synanthropic species increase to dominate the mammal community. Thresholds were found for abundance of all native terrestrial species combined, macropod abundance and exotic rodent abundance. However, threshold models were only better than linear or cubic models for exotic rodent abundance. Matrix development intensity has a pervasive impact across the whole landscape mosaic that results in a complex range of environmental changes that individually and collectively impact the mammal community. Drawing on all results, I present a conceptual model of the overall impact of matrix development intensity on mammal community integrity. I conclude that a structurally complex matrix within a human-modified landscape can provide supplementary habitat resources and increase the probability of successful species movement across the landscape. Research needs to incorporate empirical data of specific matrix effects into models and theory of species distribution and abundance in human-modified landscapes. This can help guide application of management actions and landscape planning principles across different landscapes. Planning authorities and land managers need to explicitly acknowledge the importance of the matrix and the numerous factors that could be manipulated, such as retention or restoration of a structural vegetation layer that can assist mammal movement across the matrix, for greater landscape-wide conservation outcomes.

Landscape modification resulting in habitat loss, fragmentation and intensification of land use is a serious threat to the earth’s biological diversity and the primary cause of the current extinction crisis. Recent research suggests the human-modified area of the landscape (that is not ‘traditional’ habitat for native species but potentially once was), or the ‘matrix’ (as it is hereafter called) has a major influence on wildlife persistence in modified landscapes. However, the matrix is a poorly studied and inadequately understood element of a modified landscape. There are at least two reasons for this. Firstly, the predominance of island biogeography and metapopulation theories in ecological thinking ensured that the matrix has historically been ignored in ecological research. Secondly, it is difficult to analyse matrix effects without confounding effects of habitat loss, habitat fragmentation and other aspects of landscape modification. As landscapes are modified attributes of the matrix co-vary with attributes of patches and landscapes, entangling their relative impacts on wildlife. Thus to investigate the independent influence of the matrix on mammals I selected 19 study landscapes by rigorous criteria to control for all potentially confounding patch and landscape attributes such as remnant forest patch size, shape, vegetation type, condition and position in the landscape, presence of any large (potential ‘source’) native forest patches within the landscape unit, as well as matrix land use composition and history. A study landscape was defined as the area within a 500 m radius of a remnant forest patch edge. Landscapes were selected along a gradient of rural-suburban residential development spanning the broadest achievable range of what I call ‘matrix development intensity’. This gradient was quantified by a novel weighted road-length metric that considered multiple road attributes to give significance to the ecological impact of different roads. Mammals and their habitat were sampled in three landscape elements within each landscape: remnant patch core, remnant patch edge and matrix to allow a landscape level inference. Mammals were sampled by a combination of Elliott traps, wire cage traps, hair funnels, scats and direct sightings. Thirty environmental variables were measured including habitat structure and disturbance and a full floristic survey was conducted. The matrix intensity gradient was characterised by increased anthropogenic disturbance such as increased housing density, closer proximity of sample sites to houses and higher human disturbance across the landscape mosaic, including in the core of remnant patches. However, matrix intensity was not the greatest source of overall variation in structural and floristic habitat attributes. Therefore the confounding of matrix effects with effects of remnant forest patch habitat attributes were successfully ruled out. Management actions of individual landholders can shape habitat attributes essential to mammals across the landscape mosaic. Mammal response to matrix intensity was species specific. Several native species declined in abundance, others were more resilient to moderate levels of matrix intensity, one species increased in abundance, and at least one species appeared unaffected by matrix intensity. Native species richness peaked at moderate levels, while exotic species richness and feral predators increased with matrix intensity and were negatively correlated with native species. Species response to matrix intensity appeared related to their use of edge or matrix habitat. However, an ability to use the matrix per se may not translate into an ability to persist in a landscape where development substantially reduces the habitat or movement value of the matrix. Seven a priori models of various remnant patch habitat, landscape and matrix influences on terrestrial mammal species richness were tested. Matrix attributes were the most important determinants of species richness. Matrix development intensity had a strong negative effect while matrix vegetation structural complexity had a strong positive effect on mammal species richness. Distance to the nearest remnant forest habitat was relatively unimportant. I hypothesised that thresholds of matrix intensity would exist where native species decline in abundance and exotic and native synanthropic species increase to dominate the mammal community. Thresholds were found for abundance of all native terrestrial species combined, macropod abundance and exotic rodent abundance. However, threshold models were only better than linear or cubic models for exotic rodent abundance. Matrix development intensity has a pervasive impact across the whole landscape mosaic that results in a complex range of environmental changes that individually and collectively impact the mammal community. Drawing on all results, I present a conceptual model of the overall impact of matrix development intensity on mammal community integrity. I conclude that a structurally complex matrix within a human-modified landscape can provide supplementary habitat resources and increase the probability of successful species movement across the landscape. Research needs to incorporate empirical data of specific matrix effects into models and theory of species distribution and abundance in human-modified landscapes. This can help guide application of management actions and landscape planning principles across different landscapes. Planning authorities and land managers need to explicitly acknowledge the importance of the matrix and the numerous factors that could be manipulated, such as retention or restoration of a structural vegetation layer that can assist mammal movement across the matrix, for greater landscape-wide conservation outcomes.

The intertwined challenges of food insecurity, deforestation, and biodiversity loss remain perennial challenges in Ethiopia, despite increasing policy interventions. This thesis investigates smallholding farmers’ tree- and forest-based livelihoods and management practices, in the context of national development and conservation policies, and examines how these local management practices and policies transform the agriculture–forest mosaic landscapes of southwestern Ethiopia. The thesis is guided by a political ecology perspective, and focuses on an analytical framework of ecosystem services (ESs) and disservices (EDs). It uses a mixed research design with data from participatory field mapping, a tree ‘inventory’, interviews, focus group discussions, population censuses, and analysis of satellite images and aerial photos. The thesis presents four papers. Paper I investigates how smallholding farmers in an agriculture–forest mosaic landscape manage trees and forests in relation to a few selected ESs and EDs that they consider particularly beneficial or problematic. The farmers’ management practices were geared towards mitigating tree- and forest-related EDs such as wild mammal crop raiders, while at the same time augmenting ESs such as shaded coffee production, resulting in a restructuring of the agriculture–forest mosaic. Paper II builds further on the EDs introduced in paper I, to assess the effects of crop raids by forest-dwelling wild mammals on farmers’ livelihoods. The EDs of wild mammals and human–wildlife conflict are shown to constitute a problem that goes well beyond a narrow focus on yield loss. The paper illustrates the broader impacts of crop-raiding wild mammals on local agricultural and livelihood development (e.g. the effects on food security and children’s schooling), and how state forest and wildlife control and related conservation policy undermined farmers’ coping strategies. Paper III examines local forest-based livelihood sources and how smallholders’ access to forests is reduced by state transfer of forestland to private companies for coffee investment. This paper highlights how relatively small land areas appropriated for investment in relatively densely inhabited areas can harm the livelihoods of many farmers, and also negatively affect forest conservation. Paper IV investigates the patterns and drivers of forest cover change from 1958 to 2010. Between 1973 and 2010, 25% of the total forest was lost, and forest cover changes varied both spatially and temporally. State development and conservation policies spanning various political economies (feudal, socialist, and ‘free market-oriented’) directly or indirectly affected local ecosystem use, ecosystem management practices, and migration processes. These factors (policies, local practices, and migration) have thus together shaped the spatial patterns of forest cover change in the last 50 years. The thesis concludes that national development and conservation policies and the associated power relations and inequality have often undermined local livelihood security and forest conservation efforts. It also highlights how a conceptualization of a local ecosystem as a provider of both ESs and EDs can generate an understanding of local practices and decisions that shape development and conservation trajectories in mosaic landscapes. The thesis draws attention to the need to make development and conservation policies relevant and adaptable to local conditions as a means to promote local livelihood and food security, forest and biodiversity conservation, and ESs generated by agricultural mosaic landscapes. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 2: In press. Paper 3: Submitted. Paper 4: Manuscript.</p>

Les feuilles hébergent une grande diversité de micro-organismes. Parmi les facteurs responsables de la structuration des communautés microbiennes foliaires, l’effet du processus de dispersion reste peu étudié. Les structures paysagères telles que les lisières ou l’hétérogénéité du paysage influencent la migration et la dispersion de nombreuses espèces de macro-organismes, mais l’effet de tels facteurs sur la composition des communautés microbiennes foliaires n’a jamais été testé.Nous faisons l’hypothèse que les parcelles forestières sont une source de champignons foliaires pour les vignes adjacentes. Nous avons comparé les communautés fongique foliaires et aériennes de la vigne et des forêts adjacentes au cours d’une saison de végétation, testé l’effet d’une lisière forestière sur ces communautés et évalué l’effet de la composition du paysage sur ces communautés. Les communautés fongiques ont été caractérisées par métabarcoding.Les communautés fongiques foliaires viticoles et forestières divergent au cours de la saison. Ni la distance à la lisière, ni la proportion de forêt dans le paysage n’affectent les communautés foliaires de la vigne, mais les communautés aériennes diffèrent en fonction de la distance à la lisière forestière. Ces résultats suggèrent que la dispersion ne joue qu’un rôle mineur dans la structure des communautés fongiques foliaires. De nombreuses pressions de sélection telles que les pratiques agricoles semblent en revanche avoir un effet fort sur ces communautés. Des études supplémentaires sont nécessaires pour vérifier la contribution relative de ces pressions de sélection, ainsi que le potentiel service de régulation fourni aux cultures par la forêt. / Plant leaves host a large diversity of microorganisms. Among the factors shaping foliar microbial community structure, the effect of the dispersal process remain understudied. Landscape elements, such as edges or landscape heterogenity, influence migration and dispersal of many macro-organism species. However, the effect of such factors on foliar microbial communities has never been studied. We hypothesized that forests are a source of foliar fungi for adjacent vineyards. We compared foliar and airborne fungal communities in vineyard and adjacent forests along a vegetative season, we examined the effect of a forest edge on these communities in a vineyard, and weassessed the effect of landscape composition on these communities. Fungal communities were characterized with a metabarcoding method. Foliar fungal communities in vineyards and forests diverge over the course of the vegetative season. Neither the distance to the edge nor the proportion of forest in the landscape affect foliar fungal communities in vineyards, while airborne communities change with the distance to the forest edge. These results suggest that dispersal is not dominant in shaping foliar fungal communities. Instead, many selective pressures such as agricultural practices seem to shape strongly these communities. Further investigations are required in order to estimate the relative contribution of those processes, and the potential ecosystem service provided by the forest to crops.

Les bactéries et champignons des sols sont impliqués dans différentes fonctions des écosystèmes terrestres. Ils sont notamment investis dans la formation des sols, la stabilité des agrégats et les successions végétales. La Nouvelle- Calédonie est un archipel subtropical, classé comme hotspot de biodiversité et dont un tiers de la surface est recouvert par les substrats ultramafiques. Ces milieux sont caractérisés par de faibles concentrations en nutriments (N, K, P) et de fortes concentrations en métaux lourds (Ni, Co, Cr, Mn). Les écosystèmes présents sur ces substrats sont originaux et diversifiés. Ils sont aussi fortement menacés par l’activité minière. Cependant, pour pouvoir correctement restaurer ces milieux et relancer les dynamiques végétales, il est important de connaître les communautés microbiennes associées à ces écosystèmes ainsi que les facteurs les structurant. Ce travail de recherche a permis d’améliorer nos connaissances sur les communautés microbiennes issues de différents écosystèmes des sols ultramafiques néo-calédoniens, ainsi que sur les interactions existantes entre ces microorganismes et les facteurs biotiques et abiotiques. Les résultats obtenus ont montré que chaque formation végétale et chaque site possèdent une communauté microbienne qui lui est propre, d’où l’intérêt de conserver et protéger les écosystèmes calédoniens. De plus, ces travaux ont aussi montré la capacité des communautés bactériennes et fongiques de servir de bio-indicateurs, et plus particulièrement les communautés fongiques qui sont plus sensibles aux perturbations et variations de la couverture végétale. Il a aussi été démontré qu’en maquis ou forêts monospécifiques, les communautésectomycorhiziennes possèdent des fonctions similaires dans la production d’enzymes de dégradation de la matière organique. Ces travaux ont permis une meilleure connaissance des communautés microbiennes associées auxformations végétales des substrats ultramalfiques ainsi que des facteurs les structurant. Cela devrait améliorer la mise en place des futurs chantiers de restauration de ces écosystèmes. / Soil bacteria and fungi play different functions in terrestrial ecosystems. They are implicated in soil formations, aggregate stability, and plant succession. New Caledonia is a subtropical archipelago, classified as a biodiverse hotspot and a third of its surface is covered by ultramafic soils. These soils are characterised by low concentrations of nutrients (N, K, P) and high concentrations of heavy metals (Ni, Co, Cr, Mn). Ecosystems present in these soils are origina and diversified but strongly threatened by mining activity. It is a necessity to restore these ecosystems after ore exploitation. However, to correctly restorethese environments and relaunch plant dynamics, it is important to identify the microbial communities associated with these ecosystems as well as the structuring factors.This research enabled us to improve our knowledge of microbial communities from different ecosystems on New Caledonian ultramafic substrates, as well as the interactions which exist between these microorganisms and biotic and abiotic factors. Results obtained showed that each plant formation and each site possessed its own microbial community,hence the interest in conserving and protecting New Caledonian ecosystems. Moreover, these works also showed the capacity of bacterial and fungal communities to be used as bioindicators, and more particularly fungal communities which are more sensitive to disturbance and plant cover variations. It has also been demonstrated that in monospecific maquis and rainforests, ectomycorrhizal communities have similar functions in the production of degradative enzymes of organic matter. This research improved understanding of microbial communities associated with plant formations on ultramafic substrates as well as structuring factors. This should improve the implementation of future restoration projectson these ecosystems.

Polyphagy is a very common trait among insects. In this study, we focus on a generalist bark-beetle predator, Thanasimus formicarius (L.) (Coleoptera, Cleridae), which feeds on many scolytids in spruce, pine and broad-leaf stands. It is known to respond to the pheromones of many scolytids, among which the most harmful spruce bark beetle in Europe, Ips typographus (L.). The adults attack scolytid adults and oviposit on attacked trees where their larvae feed upon immature stages of the prey. <p>However, a bottom-up process limits Thanasimus formicarius’ impact on spruce bark beetles, because in most cases the bark of spruce is too thin for sheltering pupal niches and mature larvae have to leave the trees. On pine however, pupation is quite successful and reproductive success is high. <p><p>The present work estimates the advantages (complementary prey during gaps among the phenology of pine bark beetles or due to the population fluctuations of most scolytids) and constraints (landing on unsuitable host trees for the predator’s reproduction) for T. formicarius to have a wide range of prey. <p><p>Passive barrier-trappings showed that the presence and abundance of scolytid species vary strongly from year to year. So, polyphagy in T. formicarius appears as a response to fluctuating prey supplies. <p><p>This way of foraging may lead T. formicarius towards stands not always favourable for its development (for example, spruces). <p>At the tree level, funnels and pitfall-traps caught high numbers of third-instar T. formicarius larvae walking on the bark surface of standing spruces infested by Ips typographus (respectively 365 and 70 L3s). After feeding into the whole infested part of the trunk, these larvae are obliged to migrate outside of the galleries to favourable pupation site (e.g. the base of the trees where the bark is thicker), or even to leave the trees and search for an acceptable pupation substrate in the litter. <p><p>At the landscape level, different trapping experiments showed a correlation between catches of T. formicarius and the proportion of pines around each trap. Consequently, in a metapopulation landscape pattern, pines would act as “sources” of predators, whilst spruces are “sinks”. Indeed, Thanasimus formicarius are trapped in higher numbers in mixed stands comprising pines. This observation is also corroborated in a four-year trapping experiment in the North-East of France, following the storms of December 1999. The predator/prey ratios (T. formicarius/I. typographus) were higher in stands comprising pines than in stands without pines. The first step of a method to estimate Ips typographus infestation trends thanks to the predator/prey ratios was also developed. <p> / Doctorat en sciences, Spécialisation biologie animale / info:eu-repo/semantics/nonPublished

Les interactions systémiques entre l’Homme et son milieu sont étudiées dans cette thèse au travers de la dynamique de la végétation des domaines pastoraux de la Chaîne des Puys (Massif central, France). A partir d’analyses de terrain (phytosociologiques, agronomiques) et d’enquêtes agropastorales nous établissons un état des lieux des estives des Dômes. Une démarche diachronique vise ensuite la recherche des causes de l'extension ligneuse et ses conséquences sur la couverture végétale. Notre démarche d'analyse est fondée sur un triptyque végétation - milieu - pratiques établit sur plusieurséchelles spatio-temporelles. Les résultats montrent que la déprise agricole, symptomatique des zones de moyenne montagne, et la fermeture ligneusedes espaces qui lui est associée ne caractérisent pas les trajectoires observées de la dynamique végétale. Ainsi, les nombreuses perturbations affectant le système (crises, pratiques) remet en cause la linéarité des successions observées dans la bibliographie, tant sur les vitesses de progression que sur les trajectoires et les caractéristiques. Il apparaît alors que le concept de mosaïque phyto-morphologique anthropique, proposé ici soit plus adapté à la réalité du terrain et à l'intensité de l'usage de l'espace. Les composants de cette mosaïque et leurs évolutions suivent des trajectoires stochastiques voire chaotiques dont la caractérisation première tend vers une disparition des landes au profit de la strate herbacée et non à une fermeture généralisée des milieux. / Systemic interactions between Human and environment are studied in this thesis through the vegetation dynamics of the Chaîne des Puys pastoral area (French Massif central). From field analyses (phytosociological, agronomic) and agropastoral linguiries we establish a diagnosis oh the local pastures. The causes of the ligneous extension and its consequences on the vegetation cover explored through a diachronic survey and mapping of landscape. Our analytical approach is established on a triptyc vegetation - environment - practices based on three spatial and temporal scales. Results shows that agricultural abandonment symptomatic of mountain area correlated to space closure by woody strata do not characterize the observed trajectories of vegetation dynamics. Thus, many disturbances affecting the system (crises, practises) questions the linearity of successions observed in the literature, and affect the vegetation progression velocity and its species characteristics. It appears that the concept of landscape mosaic controlled by dynamically interacting ecological and anthropogenic drvers, is more suited to the field relaity and to intensity of the space use. The mosaic components and their evolutions follow even stochastic or chaotic trajectories which first characterization goes to a healthlang disappearance in favor oh herbaceous stratum and not in a generalized landscape closure.