Thesis (MSc)--University of Stellenbosch, 2011. / ENGLISH ABSTRACT: The current biodiversity crisis requires creative initiatives for mitigating further biodiversity loss. The use of ecological networks (ENs) is such an initiative. The South African forestry industry recognizes that there is loss of biodiversity at the smaller patch scale, while attempting to mitigate this loss at the landscape scale by implementing largescale ENs. The aim of this study was to determine how representative grassland biodiversity in ENs are of similar habitat in a nature reserves (NR). The study was conducted in the northeast of the KwaZulu-Natal Province, adjacent iSimangaliso Wetland Park, which is a natural World Heritage site. A systematic approach compared a wide range of taxa, namely plants, decomposition macrofungi, vertebrates (birds and large mammals) and faunal manifestations (e.g. molehills, dung and ant nests) between the EN and nature reserve, while controlling for differences in disturbance regime. Species richness was compared using Mann-Whitney U tests, while differences in species composition were determined using Correspondence Analyses, Multi-Dimensional Scaling and Analyses of Similarity. Grassland ENs had significantly less plant species. In addition, there were differences in plant and fungi species composition. Differences were probably caused by (1) degree of isolation i.e. proximity to source habitat patches in the surrounding matrix, and (2) habitat quality. Habitat quality was determined by local disturbance regimes (e.g. grazing and fire) and plantation-induced drought for plants, and size and amount of coarse woody debris for fungi. In addition, significant differences in abundances of mole hills (NR>EN) and small mammal burrows (EN>NR) might have had an effect on succession and regeneration of plant communities. There were differences in plant species composition between grassland EN and that at the adjacent nature reserve. However, differences between EN and NR were small when compared to differences between habitat types at the landscape spatial scale. It is concluded that grassland ENs among forestry plantations contribute to biodiversity conservation in the commercially-productive landscape. This approach to land use planning should be explored for other commercial land uses. / AFRIKAANSE OPSOMMING: Die huidige biodiversiteitskrisis vereis kreatiewe strategieë om ‘n verdere verlies in biodiversiteit te bekamp. Ekologiese Netwerke (EN’e) is een voorbeeld van sulke kreatiewe strategieë. Die Suid Afrikaanse bosbou-industrie erken die verlies in biodiversiteit wat kenmerkend in plantasies gevind word. Implementering van grootskaalse EN’e kan egter die verlies aan biodiversiteit, wat in plantasies ondervind word, temper. Die doel van hierdie studie was om vas te stel hoe goed biodiversiteit in grasveld EN’e soortgelyke habitat in ‘n nabygeleë natuurreservaat verteenwoordig. Die studie was uitgevoer in die noordooste van KwaZulu-Natal, langs iSimangaliso Wetland Park wat ‘n wêrelderfenisgebied is. Ons het ‘n stelselmatige benadering gevolg waartydens ‘n wye verskeidenheid taksa, naamlik plante, makro-fungi, vertebrate (groot soogdiere en voëls) en tekens van diere-aktiwiteit (bv. miersneste, dieremis en molshope), in die EN vergelyk is met die van ‘n natuurreservaat terwyl ons vir verskillende versteuringe gekontrolleer het. Spesiesrykheid is vergelyk met Mann-Whitney U toetse terwyl verskille in spesiessamestelling vasgestel is met Correspondence Analyses, Multi-Dimensional Scaling en Analyses of Similarity. Daar was ‘n statisties beduidende verskil in die hoeveelheid spesies tussen grasveld EN’e en die natuurreservaat. Grasveld EN’e het minder plant spesies gehad. Boonop was daar verskille in die samestelling van plant en fungi gemeenskappe. Verskille was waarskynlik veroorsaak deur (1) isolasie of die hoeveelheid nabygeleë habitatbronne in die omliggende omgewing, en (2) habitat kwaliteit. Habitat kwaliteit word bepaal deur versteuringe (bv. brand en beweiding) en die uitdrogingseffek van plantasies vir plante, en die hoeveelheid en grootte van growwe houtagtige puin vir fungi. Daar was ook beduidende verskille in die hoeveelheid molshope (NR>EN) en klein soogdier gate (EN>NR), wat moontlik ‘n effek kon hê op suksessie en herstel van plantgemeenskappe. Daar was verskille in plantspesiessamestelling tussen grasveld EN’e en die van die langsliggende natuurreservaat. Hierdie verskille was egter klein wanneer dit vergelyk word met die verskille tussen verskillende soorte habitatte in die landskap. Daarom kom ons tot die gevolgtrekking dat grasveld EN’e tussen bosbouplantasies bydra tot die bewaring van biodiversiteit in kommersiële landskappe. Hierdie benadering tot grondgebruik behoort verder verken te word vir ander kommersiële bosbou en boerderypraktyke.

This study focuses on the vegetation of the African city of Bujumbura, Burundi. Urban flora is detailed according to the origin of the species (native or introduced), their life form, the use of these species by the local population and their behaviour (invasive or not). <p>The correlation between different natural and anthropogenic environmental factors and the composition of plant communities has been demonstrated through the analysis of vegetation plots. Three types of green areas were studied: spontaneous vegetation on terra firma, wetlands <p>and domestic gardens.<p><p>The main ecological factor correlated to the composition of the spontaneous vegetation of terra firma is the degree of urbanization. From the suburb to the city centre, savannah species typical for the region are replaced by alien invasive species, mostly ruderals. The latter generally have a large distribution, often on several continents. The physiognomy of the vegetation also changes, with hemicryptophytes and phanerophytes in the outskirts and the dominance of therophytes and chamaephytes in the city centre.<p><p>In most of wetlands, a large proportion of the plant species are natives, but some wetlands are almost totally invaded by alien species. Soil type is the main environmental factor correlated with <p>the variability of the vegetation.<p><p>The flora of domestic gardens is mainly composed of introduced species used as ornamental, food or medicinal plants. The socio-economical status of the neighbourhood is the main factor correlated with the variability of the garden flora. In high standing neighbourhoods, most species are onamental, with many trees and shrubs. In popular neighbourhoods, the flora is mainly utilitarian (vegetables and fruits), with a significant proportion of native species, and composed mainly of annual species. <p>A small proportion of the alien species cultivated in gardens has become invasive in the semi-natural vegetation of the city and its surroundings. Many other species cultivated in the gardens could become invasive, based on their ecology and their invasive behaviour in other regions with a climate similar to that of Bujumbura.<p><p>Overall, the city of Bujumbura has a diverse flora, with many useful species (vegetables and fruits, traditional medicine, construction of fences and roofs, material for mats and art objects.). The vegetation also hosts wildlife: birds, fishes, hippos, crocodiles etc… <p><p>In conclusion, the vegetation of Bujumbura contributes to biodiversity conservation and provides services to the urban population, but the expansion of alien invasive species and the consequent homogenization of the vegetation is a challenge for the future.<p> / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

published_or_final_version / Real Estate and Construction / Doctoral / Doctor of Philosophy

Understanding the impacts of land-use change on biodiversity and the ecosystem services that it provides is of great importance given unprecedented growth of the human population. Past studies attempting to explore these impacts have described the overall structure of communities (i.e. species richness and trait diversity) across gradients of local scale degradation and fragmentation, and have sought to identify whether the loss of species following land-use change is non-random. Yet, despite a wealth of research we still lack a generalised understanding of how land-use change impacts on traits responsible for determining species sensitivity and their role within ecosystems, particularly for vertebrates. Moreover, despite the importance of niche-based processes in the assembly of communities, we have not yet elucidated whether these are important in mediating the collapse of communities in human-dominated landscapes. To fill these existing research gaps, I collated comprehensive avian species inventories from fragmented and degraded forests and compared their structure with communities existing in continuous forests. In Chapter 2, I tested whether sensitivity of species to forest fragmentation varies between the temperate zone and the tropics and whether there are key differences in the size of fragments required to maintain ecosystem processes in these regions. I found that sensitivity to fragmentation varies according to functional group and body mass, with the prevalence of insectivores and large frugivores declining in relation to fragment size, particularly in tropical fragments smaller than 100 ha. In Chapter 3, I tested whether functional diversity and the mean position of trait diversity of insectivores and frugivores, changed across a gradient of intensifying land-use change. I found a decline in the functional diversity of forest species and a shift in the mean community traits for both forest and non-forest species. In Chapter 4, I tested whether the structure of tropical bird communities are influenced by species interactions in a fragmented landscape. I found increasing over-dispersion in functional and phylogenetic trait relatedness among species with decreasing fragment size, suggesting that competitive interactions are important in the disassembly of avian communities. In Chapter 5, I modelled the impact of forest cover change on ecosystem function across the Brazilian Amazon, focusing on seed dispersal by birds. Furthermore, I tested whether ecosystem function declined linearly with decreased forest cover after accounting for differences in the underlying pools of species. I found the lowest levels of functional diversity along the southern arc of deforestation and that the dispersal of large seeds showed some resilience to declining forest cover. Taken together, my results suggest that the loss of species from communities in degraded and fragmented landscapes is strongly non-random. Insectivores and large frugivores are most sensitive to land-use change, with species located in the densest parts of trait space being most threatened by a decline in forest patch size, suggesting that species interactions regulate the collapse of avian diversity in human-modified forests. I conclude that land-use change has important implications for the provisioning of ecosystem services, including seed dispersal and the control of insect herbivores. The impact of future land-use change is likely to be mediated by the composition of the original pool of species and the amount of redundancy in the ecosystem services that they provide. I discuss the relevance of my findings to land-use management strategies and policy interventions, and in particular conclude that these should, where possible, maintain pristine forest patches above 1000 ha, improve connectivity among habitat patches, and ensure greater protection for logged and burnt forests. Future studies should focus on clarifying the link between shifts in vertebrate community structure and the functioning of forest ecosystems.

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. 24 March in Johannesburg, South Africa, 2014. / The Biosphere Reserve (BR) model of UNESCO’s Man and the Biosphere Programme (MAB) reflects a shift towards more socially accountable conservation, setting a basis for improving the relationships between people and their environment. The BR approach of dual ‘conservation’ and ‘sustainable development’ objectives is a long-term perspective that increases people’s ability to sustainably manage environmental resources into the future. By 2013 the MAB Programme had resulted in 621 sites designated globally since first inception in 1976, with 11 of these new listings since 2012. Yet regardless of this global spatial expansion, there are few examples of BRs successfully conforming to the MAB’s full criteria, as the reality of aligning conservation and development goals in this model has typically favoured one over the other, with examples of development objectives fuelling broad-scale landscape transformation and conservation objectives that have ignored socio-economic opportunities. However there are a number of case studies presented herein that have shown considerable success, and this thesis suggests that the potential opportunities inherent in the model may outweigh implementation challenges. This thesis critically reviews the history and evolution of the Programme internationally, and makes recommendations for the future of the MAB Programme. In the South African context, the BR model has been applied to a regional conservation landscape in the north-eastern part of the country. The Kruger to Canyons Biosphere Reserve (K2C) is a long-established land-use mosaic of socio-economic juxtapositions, where nationally important economic sectors and world-renowned protected areas (e.g. the Kruger National Park) neighbour extensive, mostly rural, human populations engaged in informal livelihood practices. These vast settlement expanses are artefacts of the former political system in South Africa, and have been prioritised by the South African National Government for urgent socio-economic upliftment. Given that understanding the underlying land-use mosaic is essential for successful implementation of the BR model in any landscape, this thesis assessed land-cover change across the BR (1993 – 2006 – 2012), through the creation of land-cover classification maps, analysing spatial changes in cover relative to the MAB theoretical aspirations. It focuses on three consolidated land-cover cover classes: Intact Vegetation, Impacted Vegetation and Settlement, that present a graduated scale of land-use intensities across the BR, acting as surrogates for the ‘environmental’ and ‘human utilisation’ aspects of the BR approach. Results indicate that the spatial distribution of land-cover change in K2C bears little regard for the theoretical BR zonation and the relative limitations on ‘use’ that typifies successful BR implementation. The spatial expansion of Settlement cover replaced adjacent Impacted Vegetation, which in turn encroached in original areas of Intact Vegetation cover, increasingly fragmenting Intact Vegetation across K2C. The expansion of Impacted Vegetation has resulted in the spatial consolidation of communal use areas, which has likely serious implications for the long-term sustainability of current harvesting levels. Anecdotal field observations have indicated community responses to these concerns with the formal fencing of rangelands to secure ownership and control access. Losses of Intact vegetation translate spatially into an area >400km2, with observed land-cover changes occurring beyond the transition zone, encroaching into the (fully-protected) core and (partially–protected) buffer zones. The increased rate of landscape change in the recent observation period (2.3%1993-2006 versus 5.7%2006-2012) poses challenges for landscape management, with future predictions of escalating transformation likely to undermine BR sustainability, as well as options for a regionally connected protected area network. Historical landscape trends predict a further 5.1% loss of intact vegetation in the next 6 years (2012 – 2018), yet this may over-estimate the potential for transformation in this landscape. However the possibility of a similar loss over a longer temporal period, i.e. 12 years, is not improbable, with the unprotected matrix landscape of the transition zone having experienced comparable losses since 1993, i.e. 4.17%. K2C has had substantial success with pilot socio-economic projects, and has received formal international recognition for the stakeholder engagement / alignment occurring in the subregion as a result of the BR’s listing here. Given these successes, and the considerable investment in its establishment, it is likely that the BR model here will endure. The results of this thesis contribute to improving implementation success long-term, informing policy and management priorities across the BR. These include contributions to i) a marketing approach for the BR, specifically the importance of a K2C ‘brand’ for enhancing co-operation and strengthening the BR’s regional identity, as well as ii) providing support for land-use planning and decisions related to land-use practices and BR sustainability.

A research report submitted to the Faculty of Science, University of the Witwatersrand, in partial fulfillment of the requirement for the degree of Master in Environmental Science Johannesburg, 2015 / Quantifying ecosystem services of urban forests has become an important subject for the national and international ecological economics agenda. This is in the wake of offsetting anthropogenic emissions of CO2, while promoting urban habitability and sustainability. This study estimates above-ground tree biomass, carbon stored and the associated economic value and net economic value of carbon sequestrated by the tree planting project in Soweto, Johannesburg, South Africa. Measurements of diameter at breast height (1.3 m) and tree height were done on all the individual trees that were recently planted (estimated to be about seven years) and other trees estimated to be over 25 years old in Petrus Molefe Park and Thokoza Park. A general allometric equation by Tietema (1993) was used to estimate above-ground biomass which was converted to carbon stocks. The economic value of carbon sequestrated was calculated at an equivalent price of R440.40 per tonne of carbon. The total above-ground biomass, carbon stored and economic value, and net economic value of the trees in Petrus Molefe Park was 7.45 tonnes, 3.35 tonnes, R1,475 and R-495,325, while the trees in Thokoza Park had 205.76 tonnes, 92.59 tonnes, R40,777 and R-312,023, respectively. The results indicated that the older trees in Thokoza Park had larger amounts of above-ground tree biomass, greater carbon storage and net economic value than the younger trees in Petrus Molefe Park. The economic values of carbon sequestrated were less than the cost of planting the trees, therefore the net economic value of carbon sequestrated were negative. The project is at an early, but promising stage, since the Greening Soweto Project provided a number of ecosystem services (i.e. beautifying the landscape, filtering air, recreation and amenity etc.), the performance of the project was evaluated by the extent to which it integrates the environmental and social benefits into the economic benefits and opportunities. Key words: Above-ground biomass, allometric equation, carbon stored, diameter at breast height, net economic value.

Em toda a superfície terrestre o uso humano está transformando grande parte das florestas, especialmente na região tropical. O desmatamento de florestas tropicais é uma das principais fontes de emissão de gases do efeito estufa e a principal ameaça à conservação de espécies e dos serviços ambientais prestados pela floresta. Dentro desse contexto que também envolve a fragmentação do habitat, o objetivo desse trabalho foi analisar como as relações entre a dinâmica do uso e cobertura da terra afetaram a estrutura e composição da comunidade de vertebrados, e também a qualidade da água em riachos no Sudoeste da Amazônia brasileira. Nós analisamos esse processo em uma área de aproximadamente 50000 km² por meio de mapas classificados a partir de imagens de satélite ao longo de 36 anos de monitoramento. Nós observamos uma perda de floresta da ordem de 24826 km² relacionada, dentre outros fatores, à expansão da malha viária de 3158 km para 12863 km de extensão, tornando a área de estudo como uma das mais desmatadas de toda a Amazônia brasileira. A dinâmica do uso e cobertura da terra ao longo da série temporal mostra que as paisagens com desmatamento mais recente apresentaram fragmentação mais intensa do que aquelas com desmatamento mais antigo e onde ocorreu pequena ou nenhuma derrubada de floresta. Os resultados da inferência com modelos generalizados de mínimos quadrados mostram que a proximidade das estradas é um dos fatores que contribuem para o aumento no desmatamento, principalmente em áreas com menor declividade e distantes dos rios. Os modelos lineares generalizados com os maiores valores de importância relativa para explicar a riqueza de espécies, biomassa agregada e o índice de vulnerabilidade à caça foram compostos pela área, número e distância entre os fragmentos como variáveis independentes. A área de floresta é importante também para a conservação da água, principalmente quanto à sua proporção na faixa de 100 m ao longo dos riachos amostrados. Levando-se em conta as características dos solos predominantes na microbacias e a sazonalidade, principalmente nos meses chuvosos, a redução de floresta próxima aos rios é responsável pelo aumento nas concentrações de alumínio, fosfato, sedimentos em suspensão, amônio e nitrato, de acordo com nossos modelos lineares generalizados. Ao final, considerações foram feitas a respeito do aumento da fiscalização e da regularização fundiária para impedir que a fronteira do desmatamento continue a avançar sobre áreas de floresta primária, e que o manejo das áreas fragmentadas inclua a restauração da floresta, principalmente ao longo dos rios para diminuir o isolamento dos fragmentos remanescentes. Nós ressaltamos também, a necessidade de se estudar como as perturbações recorrentes nos fragmentos como corte seletivo e incêndios continuam afetando a comunidade de vertebrados que ocorrem fora das áreas contínuas de floresta. / Extensive areas across the Earth surface have been modified by human land uses, particularly in the tropical region. Deforestation is one of main sources of emissions of greenhouse gases and the main threat to the conservation of biodiversity and ecosystem services provided by tropical forests. In the context of human-modified landscapes, including forest fragmentation, the main goal of this work is to understand the relationships between land use and land cover dynamics and the changes in the structure and composition of forest vertebrate community and also, the alterations in the quality of surface water in catchments. We assess land use dynamics about 50,000 km² of study area located in Southwestern Brazilian Amazon. By mean of maps of land cover classified from satellite images based in a 36-year monitoring, we observed a forest loss of 24,826 km² related to the expansion of road network from 3158 km to 12863 km in the same period. This result makes the study area as the most deforested of entire Brazilian Amazon. The land use dynamics show that fragmentation is higher in recent deforestation frontier than in both older deforestation and landscapes with few or no deforestation detected. Our results from inference with generalized least square models show that distance to the roads affects negatively the deforestation rates, mainly in flat areas and near to small and large rivers. The generalized linear models with higher relative importance to explain the species richness, aggregate biomass and the index of hunting vulnerability had as predictors the fragment area, number of patches and distance between fragments as major attributes of the landscape that affects the vertebrate community. The forest is important for maintenance of stream water quality, particularly if there is forest cover in the 100 m strips alongside the streams, considering the soil characteristics such as proportion of bases and seasonality. In our GLM multi-model analysis, the higher proportion of forest in surroundings the streams contributes to the increasing of concentrations in the rainy season of aluminum, phosphate, suspension sediments, ammonium and nitrate. At the end, we made consideration about law enforcement and landownership, to preclude the deforestation toward extensive areas of pristine forest. Moreover, we suggest that in fragmented landscapes the conservation efforts to be concentrated in forest restoration, mainly near to rivers, to reduce the isolation of the fragments. We highlighted the importance in monitoring the biodiversity responses to recurrent disturbing sources such as selective logging and accidental forest fires in fragmented landscapes.

Estima-se que atualmente 80% dos remanescentes florestais de Mata Atlântica estejam localizados em propriedades particulares, o que torna evidente a necessidade da conservação dessas áreas por entidades engajadas. A avaliação e o monitoramento das florestas restauradas são essenciais para o aperfeiçoamento das técnicas de restauração, contribuindo dessa forma para a manutenção e a permanência da biodiversidade nessas áreas. O presente estudo teve como objetivo avaliar a influência do tempo de restauração florestal em áreas plantadas com espécies arbóreas nativas e suas características estruturais sobre a riqueza e abundância da anurofauna de serapilheira. Os dados da presente pesquisa foram obtidos entre abril de 2011 e fevereiro de 2012, em parcelas de 5 × 5 m dispostas em 12 fragmentos localizados em três áreas de vegetação nativa e uma cronossequência (cinco meses e quatro e 11 anos de idade) de áreas restauradas, totalizando três fragmentos por idade. Em cada fragmento foram instaladas seis unidades amostrais, cada uma composta por quatro parcelas, perfazendo 288 parcelas de esforço amostral. Além da contagem de anuros, em cada parcela, com a intenção de caracterizar os micro-habitats disponíveis para os anuros dentro das diferentes florestas, foram medidas as seguintes variáveis abióticas: temperatura e umidade relativa do ar aprisionado na serapilheira, profundidade da serapilheira e cobertura do dossel. Observouse a presença de cinco espécies distribuídas em cinco famílias. A maior riqueza de espécies e abundância foram registradas nos fragmentos com 11 anos de restauração, seguida pela área de floresta nativa e pelo fragmento com cinco meses de idade. Na área com quatro anos de restauração nenhum indivíduo foi capturado. A análise de correlação de Spearman indicou que a riqueza foi significativamente correlacionada com a profundidade de serapilheira e a cobertura do dossel. Foi observada correlação positiva entre as variáveis cobertura do dossel e profundidade de serapilheira e correlação negativa entre umidade relativa do ar e temperatura relativa do ar. O teste de Kruskal-Wallis indicou diferença significativa entre os diferentes tipos de fragmentos avaliados. O teste de Wilcoxon revelou que os fragmentos de cinco meses e quatro anos não diferiram entre si, mas foram significativamente diferentes dos fragmentos de 11 anos e de floresta nativa. Ressalta-se, no entanto, a necessidade de estudos de médio e longo prazo para que se possa melhorar a compreensão dos processos envolvidos na sucessão ecológica e da ocupação desses habitats pelas espécies de anuros. / It is estimated that currently 80% of the remaining Atlantic forest are located on private properties, which makes evident the need for conservation of these areas by engaged entities. Evaluation and monitoring of restored forests are essential to improve restoration techniques, thereby contributing to the maintenance and persistence of biodiversity in these areas. The present study aimed to evaluate the influence of time of forest restoration in areas planted with native tree species and their structural characteristics on the richness and abundance of leaf litter frogs. The data set of this study was obtained between April 2011 and February 2012, on 5 × 5 m plots arranged in 12 fragments located in three areas of native vegetation and a chronosequence (five months, and four and 11 years old) of areas restored with native species, totaling three fragments by age. Four sampling stations with six sampling units were installed in each fragment, totaling 288 plots of sampling effort. In addition to counting the frogs in each plot the following abiotic variables were measured: temperature and relative humidity of the air trapped in the litter, litter depth and cover canopy. Five anuran species belonging to five families were recorded. The highest species richness and abundance were obtained in forested fragments with 11 years old, followed by the native forest area and the fragment with five months old. No individual was captured at the four years old area. The Spearman correlation analysis indicated that anuran richness was significantly correlated with the depth of leaf litter and canopy cover. Canopy cover and litter depth were positively correlated, and relative humidity and temperature of the air were negatively correlated. The Kruskal-Wallis test indicated a significant difference among different types of fragments evaluated. The Wilcoxon test revealed that five old months and four years old fragments did not differ, but they were significantly different from 11 years old and native forest fragments. It should be noted, however, the need for medium and long term studies to improve our understanding about processes involved in ecological succession and occupation of these habitats by anurans.

Por meio de uma amostragem padrozinada de longa duração, realizamos levantamentos de pequenos mamíferos com armadilhas de interceptação e queda em 68 sítios distribuídos em seis paisagens de 10.000 ha (três em mata contínua e três em paisagens fragmentadas) localizadas em três regiões do Planalto Atlântico Paulista. As paisagens fragmentadas compreendem diferentes quantidades de matas remanescentes, 50%, 30% e 10%, porcentagens acima e próximas dos limites superior e inferior do limiar teórico de fragmentação (10-30%). A presente tese de doutoramento foi dividida em quatro capítulos e duas abordagens principais. A primeira delas (Capítulo 2) teve por objetivo avaliar se os pequenos mamíferos listados como ameaçados de extinção são afetados pela fragmentação e pela qualidade dos remanescentes de Mata Atlântica do Planalto Atlântico Paulista. Para isso, utilizamos os dados coletados nos 68 sítios amostrados, os quais estavam distribuídos em oito categorias: nove em matas maduras contínuas, nove em matas secundárias contínuas, quatro em fragmentos grandes e 11 em fragmentos pequenos da paisagem com 50% de remanescentes, sete em fragmentos grandes e 13 em fragmentos pequenos da paisagem com 30% de remanescentes, e quatro em fragmentos grandes e 11 em fragmentos pequenos da paisagem com 10% de remanescentes. Avaliamos se 10 espécies de pequenos mamíferos listados como ameaçados de extinção e cinco espécies endêmicas comuns nas mata contínuas e ausentes das listas vermelhas são afetados igualmente pela fragmentação, na escala da paisagem e da mancha, e pela qualidade dos remanescentes de Mata Atlântica. Nenhuma das espécies analisadas, independentemente do grau de ameaça ou de raridade, respondeu a variação do estádio de regeneração das matas contínuas. Por outro lado, nossos dados mostraram que as espécies endêmicas comuns respondem de forma mais congruente e negativamente à perda e fragmentação da Mata Atlântica do que as ameaçadas, as quais tanto podem não ser afetadas quanto ser positivamente ou negativamente afetadas pela fragmentação. Assim, sugerimos a separação das espécies em dois grupos nas listas vermelhas (naturalmente raras versus afetadas pelas ações antrópicas) e a utilização de outros critérios para avaliar o status de ameaça das espécies mais comuns, como a resposta à perda e fragmentação do habitat e a especificidade ao habitat na forma de endemismos e de grau de tolerância a matriz. A segunda abordagem (Capítulo 3) teve como objetivo verificar a influência do contexto (paisagem) e do tamanho do fragmento em paisagens com quantidades diferentes de remanescentes sobre a riqueza e abundância de espécies endêmicas e não-endêmicas. Para isso, utilizamos dados de 50 dos 68 fragmentos amostrados, localizados nas três paisagens 114 fragmentadas com 50%, 30% e 10% de remanescentes. A partir da avaliação da plausibilidade de oito modelos de regressão, que expressam visões teóricas alternativas da importância do contexto e da área dos fragmentos, investigamos se a influência positiva da área das manchas de floresta é mais forte (1) para as espécies endêmicas e (2) na paisagem próxima ao limite superior do limiar de fragmentação (sensu Andrén, 1994), já que em contexto de muita mata remanescente, fragmentos pequenos e grandes poderiam abrigar populações viáveis, e em contexto de pouca mata remanescente, espécies sensíveis já teriam desaparecido. Com exceção da riqueza de espécies não-endêmicas, modelos que incluem o contexto foram as hipóteses mais plausíveis para descrever a variação da riqueza e abundância das espécies de pequenos mamíferos. Como esperado, a influência positiva da área do fragmento foi mais importante na paisagem com 30% de floresta para a maioria das espécies de pequenos mamíferos endêmicos, enquanto que os modelos que incluem a influência da área do fragmento não estiveram entre os mais plausíveis para as espécies não-endêmicas. Nossos resultados corroboram a existência de um limiar de fragmentação e indicam que, ainda que os limiares variem entre espécies, é possível identificar grupos com respostas semelhantes à perda e fragmentação do habitat, auxiliando as políticas de manejo e conservação. / A long-term standardized survey of the Atlantic Forest small mammals was conducted using pitfall traps in 68 sites distributed in six 10.000-ha landscapes (three in continuous forest and three in fragmented landscapes) located in three regions in the Atlantic Plateau of São Paulo. The fragmented landscapes harbored different amounts of remnants, 50%, 30% and 10%, percentages above or within the superior and inferior limits of the theoretical fragmentation threshold (10-30%). This thesis was divided in four chapters e two main approaches. The first approach (Chapter 2) aimed to evaluate if small mammals listed as threatened were affected by forest fragmentation and quality in the Atlantic Plateau of São Paulo state. For this, we used data from 68 sites distributed in eight categories: nine in mature continuous forests, nine in secondary continuous forests, four in large and 11 in small patches in the landscape with 50% of remnants, seven in large and 13 in small patches in the landscape with 30% of remnants and four in large and 11 in small patches in the landscape with 10% of remnants. We investigated if 10 threatened small mammals and five non-threatened endemic species commonly found in continuous forests were equally affected by fragmentation, at the landscape and patch scales, and by forest quality. Regardless of threat or rarity level, no analyzed species responded to differences in the regeneration stage in continuous forests. On the other hand, our data showed that common endemic species respond more strongly and negatively to the loss and fragmentation of the Atlantic Forest than threatened species, which either may not be affected by fragmentation, or be positively or negatively affected. we suggest separating species in two groups (naturally rare versus affected by human impact) in the Red Lists and including different criteria to evaluate common species such as response to habitat loss and fragmentation as well as habitat specificity in terms of endemism and level of matrix tolerance. The second approach (Chapter 3) aimed to evaluate the effects of context (landscape) and patch area in landscapes with different amounts of remnants on the richness and abundance of the endemic and non-endemic species. We used data from 50 sites located in the three fragmented landscapes. By analyzing the plausibility of eight regression models, which express alternative theoretical hypothesis about the importance of context and patch area, we investigate if the positive influence of patch area was stronger (1) for endemic species and (2) in the landscape within the superior limit of the fragmentation threshold (sensu Andrén, 1994), since in a context of high proportion of remnants, small and large patches could harbor viable populations and in a context of low proportion of remnants, sensitive species would have gone extinct. Except for the non116 endemic species richness, the models including context were the most plausible hypothesis to describe small mammal richness and abundance variations. As expected, the positive influence of patch area was more important in the landscape with 30% of remnants for the majority of the endemic small mammals, whereas the models including patch area were not among the most plausible ones for the non-endemic species. Our data corroborate the existence of a fragmentation threshold and point out that, although thresholds vary among species, it is possible to identify groups with similar response to habitat loss and fragmentation, directing management and conservation policies.

The challenge of balancing food production, ecosystem services (ESs) and biodiversity conservation evidences the disruption of the present agri-food system. This demands a social-ecological system transformation. For this, the replacement of conventional agricultural practices to agroecological ones has been suggested, since these practices could provide ESs provision, which agriculture and human life are dependent on, and a permeable matrix, which is indispensable for biodiversity conservation. However, the food production activity also needs to afford small farmers livelihood, otherwise, they will not be interested in taking part of this transition. In view of this, the present dissertation has the objective of analyzing agroecology as a proposal to provide farmers livelihood, generate ESs and contribute for biodiversity conservation, integrally. In order to reach the objective proposed, we selected the case study of Santa Rosa de Lima (SRL), located in Santa Catarina (SC) state, which is inserted in area of Atlantic Forest (AF), in Southern Brazil. This municipality has many dairy farmers associated with AF. AF is globally recognized as a priority biome for biodiversity conservation, due to its rich biodiversity and current level of degradation. Furthermore, the biome provides diverse ESs. Dairy activity has been extensively referred as a major driver of AF deforestation, on other hand, dairy activity is of relevance for the state economy and farmers livelihood. Due to the referred, it was brought to SC a potential agroecological and sustainable intensification (SI) dairy system, the management intensive grazing (MIG) system. The hypothesis was that the implementation of this system has potential to increase dairy profitability, reduce reliance on off-farm inputs, provide ESs and decrease pressure on remaining forest. Additionally, by increasing stocking rate and milk yields, farmers may be able to retire land from production to conservation without suffering an economic loss. To analyze the results of MIG implementation in SRL, as an agroeological system with potential to balance economic and ecological goals, we compared social, ecological and economic data of farmers applying MIG and farmers applying conventional systems. These data were collected through farmers interview and an accounting project conducted in the city during one year. As first research result, we found that MIG is still in process to become considered an agroecological system. Therefore, we analyzed this process of transformation through the lenses of social theories for social-ecological system transformation. In summary, our results has showed that MIG is more profitable than conventional systems, have increased land efficiency and showed signs of environmental improvements, however, still needs progresses to become an agroecological and sustainable intensification system. This process of transition has already started in SRL, since, among the mentioned and other evidences, MIG has gradually reduced the use of environmentally damaging inputs and farmers has showed to be more prone to meet environmental law. Our results also indicate social learning as indispensable, and financial capital as necessary to foment the process of transformation. To introduce both in the process, Participatory Processes and Payment for Ecosystem Services are advocated as suitable tools. / O desafio de equilibrar produção de alimentos, provisão de serviços ecossistêmicos (SEs) e conservação da biodiversidade evidencia a ruptura do sistema agroalimentar em vigor. Esta transição demanda uma transformação do sistema social-ecológico. Para isto, a substituição de práticas convencionais por práticas agroecológicas tem sido proposta, uma vez que estas geram SEs, do qual a agricultura e a vida humana são dependentes, e uma matriz permeável, que é indispensável para conservação da biodiversidade. Entretanto, a atividade de produção de alimentos também deve ser pensada no sentido de garantir a subsistência de pequenos produtores rurais, caso contrário, eles não apresentarão interesse em participar desta transição. Devido a isto, a presente tese tem por objetivo analisar a agroecologia como uma proposta para a subsistência de agricultores, a geração de SEs e contribuição na conservação da biodiversidade, integradamente. Para alcançar o objetivo proposto, foi selecionado um estudo de caso localizado em Santa Rosa de Lima (SRL), no estado de Santa Catarina (SC), em área de Mata Atlântica (MA), na região sul do Brasil. Este município possui vários produtores rurais associados à MA. A MA é globalmente reconhecida como um bioma prioritário para conservação da biodiversidade, devido a sua rica biodiversidade e estágio atual de degradação. Além do mais, o bioma oferece diversos SEs. A atividade de produção de leite tem sido recorrentemente referida como uma das principais causas de desmatamento de MA, por outro lado, esta atividade é de relevância para a economia do estado e subsistência de produtores rurais. Devido ao exposto, foi trazido para SC um sistema de produção de leite com potencial para ser agroecológico e de intensificação sustentável (IS), o sistema de Manejo Intensivo de Pastagem (MIP). A hipótese inicial consistia em que a implementação deste sistema tem o potencial de aumentar a lucratividade da atividade, reduzir a dependência de insumos externos à propriedade, gerar SEs e diminuir a pressão sobre os remanescentes florestais. Adicionalmente, através do aumento da taxa de estocagem e produção de leite, produtores podem estar aptos a separar terra, antes dedicada para a atividade, para conservação, sem sofrer perda econômica. Para analisar os resultados da implementação do MIP em SRL, como um sistema agroecológico com potencial para equilibrar metas econômicas e ecológicas, nós comparamos dados sociais, ecológicos e econômicos entre produtores que adotam MIP e produtores que adotam sistemas convencionais. Estes dados foram coletados através de entrevistas e um projeto contábil realizado na cidade durante um ano. Como primeiro resultado, nós diagnosticamos que o MIP está ainda em processo de tornar-se agroecológico. Entretanto, nós analisamos este processo de transformação através das lentes da teorias sociais. Em síntese, nossos resultados mostraram que o MIP é mais lucrativo, tem aumentado eficiência no uso da terra e tem mostrado sinais de melhoras ambientais, mas ainda necessita progressos para tornar-se um sistema agroecológico e de intensificação sustentável. Este processo de transição já tem se iniciado em SRL, uma vez que, além de outras evidencias e as já mencionadas, MIP tem reduzido gradativamente o uso de insumos danosos ao ambiente, e produtores tem se apresentado mais inclinados a atender a legislação ambiental. Nossos resultados também indicaram aprendizagem social como indispensável, e capital financeiro como necessário ao fomento do processo de transformação. Para a introdução de ambos no processo, Ações Participativas e Pagamentos por Serviços Ecossistêmicos são entendidos como ferramentas adequadas.