Doctor of Philosophy / Department of Biology / Brett K. Sandercock / Habitat heterogeneity is a key driver of biodiversity in many ecosystems. In native ecosystems, habitat heterogeneity can arise from multiple drivers including nutrients, topoedaphic conditions, and ecological disturbance. Historically, the prairies of North America existed as a heterogeneous mosaic of habitat conditions created by the interaction of fire and grazing by native ungulates. The focus of many grazing systems has been to minimize disturbance caused by grazing by promoting uniform distributions of grazing animals across management units. Patch-burn grazing is an alternative rangeland management practice that has been proposed to restore historical patch dynamics and biodiversity to rangelands by simulating historical disturbance processes. In my dissertation research, I tested the hypothesis that patch- burn grazing restores habitat heterogeneity to rangelands, and that the resulting habitat heterogeneity can promote biodiversity of native wildlife. I focus on responses of small mammals and coyotes to patch-burn grazing to gain a better understanding of wildlife responses to rangeland management, and because grassland mammals are an ecologically important group. My 3.5-year field study of habitat and small mammal responses to rangeland management showed that: 1) patch-burn grazing created greater heterogeneity in vegetative structure and composition of plant functional groups than in positive and negative controls; 2) habitat heterogeneity created by the interaction of fire and grazing increased small mammal richness and diversity compared to a negative control managed for uniform grazing distributions; 3) the interaction of fire and grazing structured small mammal communities in tallgrass prairie; and 4) population dynamic responses of small mammals to fire and grazing disturbance were species- specific. My 3-year study of coyote survival and resource selection revealed that: 1) rangeland management influences resource selection by coyotes in seasons when they depend on small mammal prey, but not during other seasons; and 2) anthropogenic sources of mortality are important for coyotes at a protected area, even in the absence of harvest. My field results show that restoring the drivers of historical patch dynamics to managed rangelands and publicly held grasslands that are not currently grazed could have profound effects on biodiversity conservation in North America, while continuing to provide ecosystem services to society.

Ocean acidification has the potential to profoundly affect marine ecosystems before the end of this century, but there are large uncertainties on its effects on temperate benthic communities. Volcanic CO2 seeps provide an opportunity to examine and improve our understanding of community responses to ocean acidification. In this thesis, two Mediterranean CO2 seeps (Methana in Greece and Vulcano in Italy) were used to investigate the responses of macroalgae and their epifaunal communities to increased CO2. Changes in plant-herbivore interactions at elevated CO2, as well as adaptation potential of dominant macroalgae and responses of macroalgae and epifauna to concurrent exposure to elevated CO2 and copper pollution, were also examined. Firstly, I determined that volcanic seeps off Methana (Greece) are suitable for ocean acidification studies as they do not have confounding gradients in temperature, salinity, total alkalinity, nutrients, hydrogen sulphide, heavy metals or wave exposure. Calcifying macroalgae abundance decreased as CO2 increased both at Methana and at Vulcano, while fucoid algae seemed to benefit from elevated pCO2 levels. Seasonality greatly affected macroalgal responses to increasing CO2, according to the annual cycles of dominant species. Epifaunal communities of dominant fucoid algae changed at elevated pCO2 as well, with calcifying invertebrates decreasing and polychaetes increasing near the seeps. Herbivore control of macroalgal biomass did not greatly change at elevated pCO2 levels, as limpets had a minor role in controlling macroalgal biomass off Vulcano (Italy) and sea urchins were replaced by herbivorous fish near seeps off Methana. The two macroalgal species examined for signs of long-term acclimatisation (Cystoseira corniculata (Turner) Zanardini and Jania rubens (Linnaeus) J.V.Lamouroux) to ocean acidification using reciprocal transplants did not appear to have permanently acclimatised to elevated pCO2 levels, but changed their physiology in four to nine months depending on the local environment. Furthermore, when exposed to a 36-hour copper pulse at elevated pCO2 levels both seaweed species accumulated more copper in their tissues compared to those exposed to copper in reference pCO2 conditions, and this resulted in altered epifaunal assemblages on C. corniculata. These observations suggest that benthic communities will significantly change as CO2 levels increase, and that long-term acclimatisation is not likely to play a significant role; this would have profound consequences for benthic ecosystems and the services they provide.

Doctor of Philosophy / Department of Biology / Brett K. Sandercock / Habitat heterogeneity is a key driver of biodiversity in many ecosystems. In native ecosystems, habitat heterogeneity can arise from multiple drivers including nutrients, topoedaphic conditions, and ecological disturbance. Historically, the prairies of North America existed as a heterogeneous mosaic of habitat conditions created by the interaction of fire and grazing by native ungulates. The focus of many grazing systems has been to minimize disturbance caused by grazing by promoting uniform distributions of grazing animals across management units. Patch-burn grazing is an alternative rangeland management practice that has been proposed to restore historical patch dynamics and biodiversity to rangelands by simulating historical disturbance processes. In my dissertation research, I tested the hypothesis that patch- burn grazing restores habitat heterogeneity to rangelands, and that the resulting habitat heterogeneity can promote biodiversity of native wildlife. I focus on responses of small mammals and coyotes to patch-burn grazing to gain a better understanding of wildlife responses to rangeland management, and because grassland mammals are an ecologically important group. My 3.5-year field study of habitat and small mammal responses to rangeland management showed that: 1) patch-burn grazing created greater heterogeneity in vegetative structure and composition of plant functional groups than in positive and negative controls; 2) habitat heterogeneity created by the interaction of fire and grazing increased small mammal richness and diversity compared to a negative control managed for uniform grazing distributions; 3) the interaction of fire and grazing structured small mammal communities in tallgrass prairie; and 4) population dynamic responses of small mammals to fire and grazing disturbance were species- specific. My 3-year study of coyote survival and resource selection revealed that: 1) rangeland management influences resource selection by coyotes in seasons when they depend on small mammal prey, but not during other seasons; and 2) anthropogenic sources of mortality are important for coyotes at a protected area, even in the absence of harvest. My field results show that restoring the drivers of historical patch dynamics to managed rangelands and publicly held grasslands that are not currently grazed could have profound effects on biodiversity conservation in North America, while continuing to provide ecosystem services to society.

O estudo teve por objetivo verificar se existe um gradiente de estresse nutricional entre fisionomias distintas na ARIE Pé-de-Gigante e verificar se às plantas de cerrado podem ser aplicadas (1) a hipótese da disponibilidade de recursos (Coley et al., 1985), segundo a qual plantas em ambientes com menos recursos investem mais em defesas e (2) a hipótese de que plantas sob estresse abiótico tornam-se mais susceptíveis a insetos herbívoros (Rhoades, 1979; White, 1969, 1984). Além disso, pretendeu-se verificar a composição da fauna de insetos herbívoros e de predadores e parasitóides associadas à planta. Foram comparados os padrões de herbivoria foliar, bem como a fauna de insetos herbívoros e de predadores e parasitóides associada a Didymopanax vinosum E. March. em fisionomias de campo cerrado, cerrado e cerradão, relacionando-os com possíveis variações nutritivas, da dureza foliar e dos teores de celulose, ligninas e taninos. Foi constatado que as folhas jovens de D. vinosum possuem maior conteúdo de água e nitrogênio, menor dureza e menores teores de celulose, ligninas e taninos. As análises de solo não apontaram diferenças entre as áreas em termos de composição química, indicando a ausência de um gradiente de estresse nutricional. No entanto, as plantas do cerrado e do cerradão apresentaram folhas mais ricas em água e nitrogênio e menor dureza que as plantas do campo cerrado. As plantas das duas primeiras áreas investem mais em taninos e menos em ligninas, os principais tipos de defesas quantitativas. Outros fatores abióticos podem exercer grande influência nas concentrações destes compostos, como a luminosidade, cuja deficiência parece colocar as plantas do sub-bosque do cerrado e do cerradão em situação de estresse. Desse modo, a hipótese 1 aplica-se a D. vinosum apenas quando se considera o investimento em taninos, mas não em ligninas. A herbivoria por insetos mastigadores foi muito baixa tanto em folhas jovens quanto em adultas (inferior a 2 e 5%, respectivamente). Em folhas adultas, foi maior apenas no cerradão, corroborando parcialmente a hipótese 2. Por outro lado, a herbivoria por insetos sugadores em folhas jovens foi intensa, principalmente no campo cerrado e no cerradão, sendo Lyothrips didymopanacis (Thysanoptera) a principal espécie responsável pelos danos foliares. As populações desta espécie demonstraram ser influenciadas primariamente pela disponibilidade de folhas jovens da planta. Populações de aranhas e coccinelídeos acompanharam parcialmente as flutuações das populações de L. didymopanacis. / This study aimed at verifying the existence of a nutritional stress gradient among distinct physiognomies at Pé-de-Gigante Relevant Ecological Interest Area (ARIE), and in such case, verifying whether the following hypotheses can be applied to cerrado plants: (1) the resource availability hypothesis (Coley et al., 1985), according to which plants in environments with fewer resources invest more in defence strategies, and (2) the hypothesis stating that plants under abiotic stress become more susceptible to herbivorous insects (Rhoades, 1979; White, 1969, 1984). In addition, the study also aimed at verifying the composition of the herbivorous insects fauna as well as those of predators and parasitoids associated with the plant. The leaf herbivory standards were compared and so were the faunas of herbivorous insects, predators and parasitoids associated with Didymopanax vinosum E. March in physiognomies of “campo cerrado", “cerrado" and “cerradão". Such standards were then related to possible variations in leaf nutrition, toughness and levels of cellulose, lignins and tannins. It was observed that young leaves of D. vinosum had more water and nitrogen content, less hardness and smaller levels of cellulose, lignins and tannins. Soil analyses did not show differences among the areas as to chemical composition, which indicates the absence of a nutritional stress gradient. However, cerrado and cerradão plants presented leaves with greater water and nitrogen contents as well as less toughness than campo cerrado plants. The plants in the two first areas invested more in tannins and less in lignins, which are the two major types of quantitative defences. Other abiotic factors can exert great influence on the concentration of these compounds, such as luminosity, whose deficiency seems to place cerrado and cerradão understory plants in a stressing situation. Therefore, hypothesis 1 applies to D. vinosum only when the investment in tannins is taken into account, but not that in lignins. Herbivory in chewer insects was low both in young and adult leaves (lower than 2 and 5%, respectively). As to adult leaves, it was higher only in cerradão, which partly supports hypothesis 2. On the other hand, herbivory by sucking insects on young leaves was intense, especially in campo cerrado and cerradão where Lyothrips didymopanacis (Thysanoptera) was the mainly responsible species for leaf damage. This species’ population showed to be primarily influenced by the availability of the plant’s young leaves. Populations of spiders and Coccinelideae partly followed L. didymonopacis population flows.

<p>In this study I investigate the sensitiveness of Bishop's goutweed (Aegopodium podagraria L.)to mechanical damages in an experiment with artificial clipping. Based on personalobservations of A. podagraria during several years I hypothesise that i) A. podagraria indifference with other species, for example different species of grass have difficult to recoverfrom mechanical damages ii) in a site with several other species present A. podagraria willhave harder to recover from mechanical damages than in a pure stand iii) as purer a stand withA. podagraria is, the harder other species will have to establish after the removal of aboveground parts of A. podagraria. The main conclusion of this work is that mechanical damagehas a significant effect on the vegetation cover of A. podagraria. The experiment also provethat a high grade of monoculture favours the ability of A. podagraria to recover frommechanical damage. It could however not be proven that high grade of monoculture of A.podagraria puts constraints on the ability of other species to establish on plots where A.podagraria is removed.</p>

<p>Studies in tropical regions have shown that trees and grasses respond differently to fire, grazing and browsing. In African savannas, the responses to fire, grazing and browsing are different, determined by negative or positive correlations. Browsing may have other consequences than grazing because instead of increasing woody biomass it reduces it, causing increase in grass growth, leading to increase in fuel that results in more intense fires and decrease in woody biomass. Fire and herbivory are an important interactive disturbance factors affecting vegetation succession and the tree-grass dynamics in savanna environment. Several of the fire-herbivory interactions are landscape level effects, which is shown in 2 models. My hypothesis is that the tree-grass balances are determined by interactions of both grazing and fire. There have come new scientific data about fire and herbivory and the interaction effects on tree-grass dynamic and succession in the savanna. By analyzing my hypothesis through a comparison between hypotheses, Intermediate disturbance hypothesis, Janzen-Connell hypothesis and the Huston hypothesis, I propose several scenarios of the savanna tree-grass dynamics in East Africa, as a result of this comparison.</p>

Nutrient availability and herbivory can regulate primary production in ecosystems, but little is known about how, or whether, they may interact with one another. Here I investigate how nitrogen availability and insect herbivory interact to alter above- and belowground plant community biomass in an old-field ecosystem. In 2004, 36 experimental plots were established in which soil nitrogen (N) availability (at three levels) was manipulated and insect abundance (at two levels) in a completely randomized plot design. In 2009, after six years of treatment, I measured aboveground biomass and assessed root production at peak growth. Overall, I found a significant effect of soil N availability on both above- and belowground plant biomass while insects affected only aboveground biomass of subdominant plant species and coarse root production; there were no statistical interactions between N availability and insect herbivory for any response variable. Specifically, responses of aboveground and belowground community biomass to nutrients were driven by reductions in soil N, but not additions, indicating that soil N may not be primarily limiting production in this ecosystem. Insect herbivory altered the aboveground biomass of the subdominant plant species and altered allocation patterns to coarse root production belowground. Overall, the results of six years of nutrient amendments and insect removals suggest strong bottom-up influences on total plant community productivity.

In this study I investigate the sensitiveness of Bishop's goutweed (Aegopodium podagraria L.)to mechanical damages in an experiment with artificial clipping. Based on personalobservations of A. podagraria during several years I hypothesise that i) A. podagraria indifference with other species, for example different species of grass have difficult to recoverfrom mechanical damages ii) in a site with several other species present A. podagraria willhave harder to recover from mechanical damages than in a pure stand iii) as purer a stand withA. podagraria is, the harder other species will have to establish after the removal of aboveground parts of A. podagraria. The main conclusion of this work is that mechanical damagehas a significant effect on the vegetation cover of A. podagraria. The experiment also provethat a high grade of monoculture favours the ability of A. podagraria to recover frommechanical damage. It could however not be proven that high grade of monoculture of A.podagraria puts constraints on the ability of other species to establish on plots where A.podagraria is removed.

Studies in tropical regions have shown that trees and grasses respond differently to fire, grazing and browsing. In African savannas, the responses to fire, grazing and browsing are different, determined by negative or positive correlations. Browsing may have other consequences than grazing because instead of increasing woody biomass it reduces it, causing increase in grass growth, leading to increase in fuel that results in more intense fires and decrease in woody biomass. Fire and herbivory are an important interactive disturbance factors affecting vegetation succession and the tree-grass dynamics in savanna environment. Several of the fire-herbivory interactions are landscape level effects, which is shown in 2 models. My hypothesis is that the tree-grass balances are determined by interactions of both grazing and fire. There have come new scientific data about fire and herbivory and the interaction effects on tree-grass dynamic and succession in the savanna. By analyzing my hypothesis through a comparison between hypotheses, Intermediate disturbance hypothesis, Janzen-Connell hypothesis and the Huston hypothesis, I propose several scenarios of the savanna tree-grass dynamics in East Africa, as a result of this comparison.

Determining the relative roles of top-down vs. bottom-up forces in controlling the structure of ecological communities is of primary importance because anthropogenic nutrient loading, overharvesting of consumers, and potential interactions of these forces are pervasively changing ecosystems throughout the world. Here I use both field experimentation and meta-analyses to investigate the role of predators in controlling community composition, the relative roles of herbivores vs. nutrient enrichment in controlling the abundance of benthic primary producers, and the influence of herbivore diversity on the community structure of coral reefs. On a coral reef in the Florida Keys, I showed that release from predation by large fishes and invertebrates via exclusion cages allowed population increases in the gorgonian-eating gastropod Cyphoma gibbosum which increased predation rates on gorgonian corals. To directly address the relative roles of top-down and bottom-up forces in controlling primary producers in benthic marine habitats, I used factorial meta-analysis of 54 field experiments that orthogonally manipulated herbivore pressure and nutrient loading to quantify the effects of consumers and nutrient enrichment on community structure. The relative effects of herbivores vs. nutrient enrichment were context dependant, varying with latitude, the type of primary producer, and the nutrient status of the system. To address the influence of herbivore diversity on the community structure of Caribbean coral reefs, I used manipulative field experiments over two years to show that a Caribbean reef changes dramatically as a function of herbivorous fish diversity. The effects of herbivore diversity on community structure were strong in both years of the experiment due to different diet preferences among herbivores. Higher herbivore diversity suppressed macroalgal abundance, increased abundance of crustose coralline algae, reduced coral mortality, and increased coral growth when compared to treatments with lower herbivore diversity. Complementary feeding by different fishes drove these patterns because macroalgae were unable to effectively deter feeding by fishes with different attack strategies. Thus, herbivore diversity appears to play an important role in the healthy function of coral reef ecosystems via complementary feeding of different herbivore species.