</p> I investigated the effects of habitat fragmentation in terms of spatial scale (enclosures of different sizes) and the degree of isolation (different mesh sizes) on benthic macroinvertebrates in the littoral zone of a shallow lake. I hypothesized that decreasing diversity and increasing temporal variation will occur with decreasing size and increasing isolation of enclosures. This hypothesis was tested by examining spatial and temporal variation (coefficient of variation) among replicate enclosures. Variation in benthic diversity was further examined within and among enclosures. The degree of isolation and habitat size interacted to determine variation of species abundance in the benthic community. Specifically, at higher levels of isolation (plastic enclosures), variance among enclosures of the same and of different sizes was the greatest compared to less isolated enclosures. We also investigated the persistence of benthic macroinvertebrates where persistence is defined as constancy in the number of organisms through time. We hypothesized that persistence is lower in small and more isolated enclosures as opposed to large less isolated ones. We simultaneously tested the hierarchical nature of community persistence as outlined by Rahel (1990) to determine if there is a difference in the assessment of persistence of the assemblage depending on the analytical scale used. There is a possible scale effect at the level of abundance rank since small (1 and 4 m2 ) enclosures had lower persistence than large (9 and 16m2 ) enclosures. Low persistence in both abundance and abundance rank over time prevented a conclusive test of the hierarchical nature of community persistence. I conclude that it is important for benthic enclosure experiments to be conducted at various spatial and analytical scales and that, where replication is possible, spatial and temporal variation allows a thorough examination of different community responses. </p> / Thesis / Master of Science (MSc)

Our ability to predict community responses to environmental stress remains limited. To address this issue, I investigated how species abundance, community composition and food-web structure varied across abiotic gradients (principally disturbance and habitat size) in New Zealand streams. In surveys, community composition, biomass and richness were all strongly influenced by flood-related habitat disturbance, although disturbance influenced each trophic level via different mechanisms. Experiments indicated that macroinvertebrate prey communities were primarily structured by physical disturbance effects, whereas predatory fish communities were structured by physical disturbance effects and disturbance-mediated changes to prey communities. Prey community biomass and composition affected fish species identity and abundance and an in situ stream channel experiment suggested that prey communities were structured by trade-offs between resisting biotic interactions in physically stable environments and successfully exploiting highly disturbed habitats. The prey community traits associated with different disturbance regimes then directly influenced the composition and predatory impact of the resultant fish communities. In addition to disturbance-mediated biotic interactions, abiotic gradients also provided strong selection pressures on predatory fish communities. In particular, disturbance and habitat size strongly influenced predator community responses (e.g., biomass and maximum body size) in surveys and experiments. However, a habitat’s capacity to support predator community biomass was largely determined by its size. Food-web structure changed with habitat size; small streams supported more prey than predator biomass, whereas large streams had inverted biomass pyramids (i.e., more predator than prey biomass). Similar relationships between food-web structure and habitat size were found in grassland and forested streams, but terrestrial invertebrate subsidies meant that forested streams supported more predator biomass per unit area than grassland stream food webs. My results indicate that human actions resulting in habitat loss (e.g., water abstraction or river impoundment) and increases in flood-related disturbance events (e.g., climate change) are likely to have significant impacts on stream food webs, ultimately leading to habitats that support smaller fish communities (i.e., less biomass, smaller body size). This means that ecologists and managers will need to consider the separate, interactive and indirect effects of disturbance and habitat size on ecological communities if we are to accurately predict and manage food-web responses to global environmental change.

Studien syftar till att testa teorier om effekter av habitatstorlek, konnektivitet, områdesskydd, habitatmängd och fragmentering på enskilda arters förekomstmönster, samt undersöka existensen av fragmenteringströskel och utdöendetröskel. Förekomstdata för revlummer Lycopodium annotinum och mattlummer Lycopodium clavatum har insamlats i 61 skogsfragment utspridda i 10 kartrutor à 2,5×2,5 km i landskapet Södermanland. I vardera skogsfragment undersöktes en yta om 500m2. Revlummer förekom i 13 och mattlummer i 10 av de undersökta skogsfragmenten. I 7 kartrutor förekom revlummer i minst 1 skogsfragment, för mattlummer var motsvarande siffra 8 kartrutor. Resultaten visar att habitatstorlek har en signifikant positiv effekt på förekomst av revlummer och mattlummer samt att konnektivitet har en signifikant positiv effekt på förekomst av mattlummer. Områdesskydd, habitatmängd och fragmentering i landskapet har ingen påvisad effekt på förekomst av dessa arter och ingen fragmenteringströskel eller utdöendetröskel har hittats. Dessa resultat indikerar att ett bevarande av skogsfragment med stor storlek och hög konnektivitet är viktigt för de undersökta arternas fortsatta livskraft. / The aim of this study is to test theories of the effects of habitat size, connectivity, area protection, habitat amount and fragmentation on the occurrence of single species, plus examining the existence of a fragmentation threshold and an extinction threshold. Occurrence of stiff clubmoss Lycopodium annotinum and stag’s-horn clubmoss Lycopodium clavatum was examined in 61 forest fragments scattered over 10 large 2,5×2,5 km plots, located in the county of Stockholm in Sweden. In each forest an area covering 500m2 was searched for the two species. Lycopodium annotinum occured in 13 and L. clavatum in 10 of the investigated fragments. Lycopodium annotinum occurred in at least 1 fragment in 7 large plots, 8 large plots was the corresponding number for L. clavatum. The results showed a significant positive effect of habitat size on the occurrence of both species and a significant positive effect of connectivity on the occurrence of Lycopodium clavatum. Area protection, habitat amount and fragmentation had no observed effect on the occurrence of these species, and no fragmentation threshold or extinction threshold was found. These results indicate the importance of preserving large forest fragments with high connectivity, to assure continued viability of the examined species.

Alors que la superficie des forêts de plantation continue d'augmenter dans le monde, leurcontribution à la conservation de la biodiversité reste controversée. L’objectif de cette thèse estd'identifier les facteurs clés, à la fois au niveau de l'habitat local et à celui du paysage, qui influent surla diversité des papillons de jour dans les paysages en mosaïque dominés par des plantations de pins.Les communautés de papillons ont été échantillonnées en lisière et à l’intérieur de plantations de pinmaritime, pare-feux, ripisylves et fragments de forêts de feuillus variant par la taille et le degréd’isolement spatial. Les traits biologiques et écologiques des papillons ont été liés auxcaractéristiques de l’habitat et aux variables paysagères.Les éléments les plus importants pour la conservation des papillons dans les paysages dominés parles plantations de pins sont les habitats semi-naturels: forêts de feuillus, pare-feux et lisières. Lesripisylves se révèlent être les plus riches en papillons forestiers, abritant des espèces spécialisées. Lespare-feux hébergent deux fois plus d'espèces que les autres types d'habitats et sont importants pourla conservation de plusieurs espèces menacées. Toutefois les plantations de pin ne sont pas vide depapillons. La qualité de l'habitat, notamment la présence de plantes hôtes, est le facteur le plusdéterminant de la composition des communautés de rhopalocères. La composition et laconfiguration du paysage ont également une influence importante sur la diversité des papillons. Denombreuses espèces de papillons ont été observées dans plusieurs types d'habitat suggérant que lacomplémentation et supplémentation des ressources soient des processus clés pour maintenir ladiversité des papillons dans les paysages forestiers hétérogènes. / While the area of plantation forests continues to increase worldwide, their contribution to theconservation of biodiversity is still controversial. The aim of this thesis is to identify key habitat andlandscape factors that drive butterfly diversity in mosaic landscapes dominated by pine plantations.Butterfly communities were sampled at edges and interiors of five successional stages of pine stands,in firebreaks, riparian forests and in deciduous woodlands varying in fragment size and isolation.Biological and ecological traits of butterflies were related to habitat patch attributes and tolandscape composition and configuration.The results highlighted the critical importance of semi-natural habitats for butterfly conservation inpine plantation mosaics, i.e. deciduous woodlands, firebreaks and edges. Riparian forests wereespecially rich in forest butterfly species, harboring specialized species with both narrow habitat andthermal ranges. Firebreaks had twice as many species as other habitat types and were ofconservation value for several threatened butterfly species. Our results also showed that pine standswere not ‘free of butterflies'. Habitat quality, particularly the presence of host plants, was the mostimportant driver of butterfly community composition. Landscape composition and configuration alsoinfluenced butterfly diversity. Many species used more than one distinct habitat type, suggestingthat resource complementation and supplementation are important mechanisms of butterflydiversity persistence in pine plantation mosaics.