As agriculture continues to expand across the tropics there is an urgent need to assess its effects on biodiversity and understand how to reconcile agricultural expansion with conservation and overall sustainable development. Protected areas are not large enough to sustain viable mammal populations, thus it is important to understand how to integrate agricultural regions into conservation strategies. The aims of this thesis were (1) to improve our understanding of jaguars and other medium-large terrestrial mammals across increasing agricultural landscapes; (2) assess the impact of human land uses such as oil palm cultivation on these species; and (3) inform strategies to reconcile biodiversity conservation with other sustainability aspects and regional development in rural areas in Colombia. The methods included field surveys using camera trapping, ecological analysis (e.g. capture-recapture and occupancy models), and scenario and network analysis combined with sustainability assessment. The findings conclude that there is an effect of agriculture on jaguar populations as densities were lower than in comparable natural areas, however there were resident individuals and breeding, highlighting that modified areas can be important for jaguar long-term survival and connectivity. Wetlands were the only variable explaining jaguar occurrence, while forests impacted puma's occupancy positively and were a predictor of mammal species richness. Conversely, both oil palm and pasture affected several mammal species negatively, and the remaining ones only displayed limited affinity to these land covers, showing that the expansion of oil palm plantations and pastures constitutes a threat for felids and mammals in general. These results suggest that maintaining natural areas such as forests and wetlands across agricultural regions is key to mammal survival, pointing at a land sparing strategy. Further oil palm expansion, when inevitable, should occur on pastures since they displayed limited to no conservation value for jaguars and other mammals. Overall, agriculture impacts mammal communities by decreasing their diversity and evenness, while increasing dominance, comparatively to pristine regions. The effect on species richness was not entirely evident, demonstrating that agricultural regions are not necessarily biological deserts. Data also show that jaguars did not affect the occupancy of other felid species and were a positive predictor of mammal species richness, hence conservation strategies focused on this declining keystone species can benefit the wider mammal community, even in modified regions. This thesis also highlights that rural areas can provide for both people and wildlife if the right conditions are in place. Under the current situation the main agricultural sectors (i.e. cattle ranching and oil palm cultivation) affect wildlife and other aspects of sustainability negatively. Both adopting a stronger regulatory framework with land use planning and applying incentive schemes are improvements, as they would enable to maintain natural habitats that are crucial for jaguar and other species, while improving overall sustainability. Relevant recommendations to reconcile biodiversity conservation with overall sustainable development include the design and adoption of strategic land use planning, making agricultural subsidies conditional to social and environmental standards, consolidating local institutions, designing incentives to foster the implementation of good agricultural practices, favouring small farmers, and creating a demand for certified agricultural commodities. Finally, this research proves that achieving conservation across agricultural regions is inherently complex. Interdisciplinary approaches are needed to study such landscapes and provide solutions that are effective and locally-relevant.

Invasive alien species threaten freshwater biodiversity, with uncertain impacts on ecosystem services. Invasive alien decapods are frequent components of invaded ecosystems, and have been found to impact biodiversity, ecosystem properties and processes. Studies however often do not consider comparisons to native analogues. Native analogues could mask impacts of invasive alien species through functional redundancy. The native White Clawed Crayfish (Austropotamobius pallipes) is declining at the advance of the invasive alien American Signal Crayfish (Pacifastacus leniusculus) and the Chinese Mitten Crab (Eriocheir sinensis). Consequences of this changeover of decapod species to the patterns and processes of river basins are not well understood. Laboratory microcosm experiments found invasive alien decapods processed native leaf litter at a greater rate than native crayfish, increasing smaller fragments, notably fine particulate organic matter and dissolved organic carbon (Chapter 2). However, these differences in transformation of leaf litter into other products did not increase the productivity of algal biofilms. Each decapod species excreted nutrients at different rates: A. pallipes and E. sinensis had similar nitrogen excretion, while P. leniusculus excreted much less nitrogen suggesting this species could be a nutrient sink. These differences in nutrients were apparently absorbed by biofilms growing on leaf and rock surfaces even though they did not grow as a result. Chapter 3 found the greater detritivory performance of invasive alien decapods was observed for other species of leaf litter, including increased breakdown of particularly troublesome invasive alien riparian plant Rhododendron ponticum. Derived products varied, with more fine particulate organic matter generated for Acer pseudoplatanus and R. ponticum, suggesting less assimilation of these litter types. This revealed a previously unknown feedback between invasive plants and detritivores, whereby the litter accumulations from riparian invasive plants is potentially consumed by invasive alien decapods. This in turn may subsidise the growth and reproduction of invasive alien aquatic decapods. Outdoor pond mesocosm experiments revealed that invasive alien decapods altered taxonomic richness, diversity, and structure, with snails particularly found to decline in ponds (Chapter 4). Primary productivity of periphyton was elevated in P. leniusculus ponds compared to natives, but not in the E. sinensis ponds, explained by a combination of trophic cascade and turbidity. Leaf litter decomposition rates, community respiration, gross and net primary productivity were otherwise unaffected. Outdoor flume mesocosms were utilised to replicate headwater streams to measure bioturbation of decapods, and associated changes to biodiversity, water quality and ecosystem processes (Chapter 5). Effects on sediment topology and transport were similar for controls, native and alien species, with the presence of refuges likely causing this lack of effect. Suspended organic matter was similar for all species, but sediment respiration was higher for P. leniusculus. Invertebrate community structure was significantly different for invasive alien decapods, with significantly higher densities of collector-gathering invertebrates such as Baetis sp. and Culicoides. The the lack of consensus between the laboratory (Chapters 2-3) and mesocosm studies (Chapters 4-5) lead to the conclusion that environmental practitioners should consider context and scaling when interpreting experimental results of biological invasions. The overall findings of this study however do show that P. leniusculus and E. sinensis are not functionally redundant for several ecosystem attributes compared to A. pallipes. These differences in species functional-traits are likely to be exacerbated by greater densities of invasives observed in situ. This supports the potential for invasive alien decapods to alter the ecosystem services of entire river basins, even when a native analogue was previously present.

Across the world, particularly in the tropics, the extent of forest clearance has been widespread. At present, few studies have been undertaken and little is known on the long-term effect of land use history following clearance, on forest recovery, a significant sink for atmospheric CO2. This study aimed at quantifying the capacity of regenerating forests in the Brazilian Legal Amazon (BLA) to recover carbon using a combination of Earth Observation (EO) data and the 3-PG forest growth model. Three sites were selected within the BLA, representative of diffrent clearance histories on which extensive deforestation has occurred. Land use history and forest age for these areas was obtained from time-series analysis of Landsat images (1974 - 2011). Long term trends of aboveground biomass (AGB) accumulation in secondary forests were studied using field inventory data from 52 secondary forest plots north of Manaus from 1993, 1995 and 2014. Plots were representative of different clearance histories; a combination of clearance frequency and period of active land use prior to abandonment. A variance based global sensitivity analysis (SA) was carried out on the 3-PG forest growth model to identify its most sensitive model inputs when applying it to a mixed tropical rainforest. A parameter set for mixed tropical forests was identified using a Monte-Carlo simulation and by comparing simulated outputs to field data. These parameters were used within 3-PG to provide estimates of total carbon sequestration and model sensitivity to future climate change. Results of this thesis showed that forest age derived from remote sensing time series was comparable to that derived from field observations and interviews. Sites with a higher land use intensity did not accumulate biomass at a significantly slower rate than those used less intensively. Accumulation rates predicted from the model closely matched those calculated from the forest inventory data gathered at each plot. SA results demonstrated scientifically credible behaviour of the model and allowed identification of the most responsive model inputs and interactions. Findings illustrated the suitability and potential of combining a process based model with EO data as a way to forecast the productivity of mixed secondary forest in Brazil. Comparisons with existing estimates highlight uncertainties in deriving secondary forest AGB from remote sensing using relationships fitted to primary forests. Development of these methodologies has applications to other tropical ecosystems that have experienced a similar history of disturbance and can provide invaluable information for future land-use planning and REDD+ monitoring.

The stock structure of edible crab (Cancer pagurus) and European lobster (Homarus gammarus) in the Irish Sea was investigated using eleven and twelve microsatellite markers, respectively. These shellfish species are of high economic value to the UK and Ireland and are being exploited with increasing intensity due to the decline of fin-fisheries. Population structure of H. gammarus was assessed by combining population genetic analysis of adults and biophysical modelling of larval dispersal, with particular emphasis on assessing recruitment patterns for the Lundy (South Wales) No-Take Zone (NTZ). Genetic structuring was statistically non-significant, compatible with recurrent spatial connectivity predicted by the larval dispersal modelling. The NTZ exhibited a statistically higher FIS and lower mean relatedness values which, in light of the predicted high proportion of allochthonous larval recruits, are attributed to increased variances in reproductive success linked to post-settlement processes. Likewise, microsatellite variation for C. pagurus supported a model of a single panmictic population within the Irish Sea. Integration of data collected throughout the Northeast Atlantic indicated high gene flow throughout the studied region but with some level of chaotic genetic patchiness likely due to sweepstakes recruitment episodes. As a comparison, the potential of Restriction-site Associated DNA sequencing (RADseq) as a future tool for population genetic analysis and fisheries management of C. pagurus was also investigated, with this being, to our knowledge, the first application of these methods to a crustacean. Seven RADseq libraries permitted genotyping of 566 polymorphic Single Nucleotide Polymorphisms (SNPs) among C. pagurus sampled throughout the Northeast Atlantic. Multiple marker-based neutrality tests revealed three consensus positive outlier SNPs. However, these were not significant in pairwise outlier tests and geographical patterns of allelic variation did not lend themselves to robust inference of environmental drivers. The neutral SNP data revealed a lack of wide scale geographic structure but more pronounced chaotic genetic patchiness than reported for microsatellites, indicative of greater sensitivity to neutral structuring. The implications of the findings for marine community ecology, fisheries management and NTZ design strategies are discussed.

Land use change and forest fragmentation threaten biodiversity in the tropics. One important mechanism by which species are affected by land use change is the change in microclimate that occurs with modifications to the vegetation cover. We quantified the link between vegetation and microclimate in the tropics for the first time, finding strong relationships between leaf area index and five key microclimate variables. The mean daily maximum air temperature 1.5 m above the ground was up to 6.5 °C hotter in oil palm plantations and up to 2.5 °C hotter in logged forest than in primary forest, which could have large impacts on the organisms living in these habitats. We also developed a process based microclimate model designed for use across this range of land use types. The model reproduced the direction of the observed trends, and performed well in predicting mean climate variables as well as inter- and intraday variability. Model results suggested that changes in canopy height associated with logging and agricultural conversion have a more important influence on the near-surface microclimate than changes in leaf area index. The hotter, drier and windier microclimate found near a forest edge could drive an increase in tree transpiration. For the first time, transpiration rates were measured in six trees before and immediately after the creation of a forest edge by monitoring sap flow. After edge creation, mean daily sap flow rates increased in all trees, with the increases ranging from 4-61%. The magnitude of the change in mean daily sap flow was found to be positively correlated with diameter at breast height and negatively correlated with the distance to the forest edge. These increases in tree water use near forest edges, particularly by large trees, could create water shortages and drive increased tree mortality.

Fisher behaviour remains a key source of uncertainty in fisheries management. Failing to account for the behavioural response of fishers can lead to unexpected or unintended consequences of management; our understanding of fisher behaviour, as well as our ability to translate this understanding into predictive management models, is underdeveloped. This thesis aimed to develop an individual-based model (IBM) that could be used by fishers and managers to evaluate the impacts of management scenarios in the Isle of Man scallop fishery. Questionnaire interview data and a conjoint analysis were used to understand fishing behaviour and to generate realistic parameters to input to an IBM of fishing activity. Vessel monitoring system (VMS) and logbook data were also analysed to inform the model development, and to provide the data against which the model could be validated. There is increasing interest in using automatic identification system (AIS) as an alternative to VMS when investigating fishing activity, so a comparison of AIS and VMS data was presented, highlighting substantial gaps in the coverage of AIS data. By using simple foraging decision rules, parameterised by questionnaire data, it was possible to build an IBM that could reproduce patterns seen in the Isle of Man scallop fishery with reasonable similarity. Comparing multiple submodels of fishing behaviour provided insights into predicting fishing activity, and identified the most structurally realistic models. It illustrated the importance of incorporating random behaviour in a model design, potentially to account for social aspects of fishing decisions that are more difficult to quantify. It also demonstrated that predicting responses to management by modelling fishers as optimal foragers that act in an economically rational manner may overestimate the capacity of the fleet to compensate for restrictions such as closed areas, and underestimate the fishing footprint. Fishery systems may be too complex to distil to a single simple and ‘accurate’ model, but having a suite of models that together give a reasonable representation of the fishery could allow the range of likely impacts of management to be better considered. This thesis demonstrates the value of individual-based modelling for both understanding fisher behaviour and predicting the outcomes of management. It has also provided strong evidence to support the use of questionnaire interview data in modelling fishing activity. Comprehensively documenting the stages of model development provided a transparent model validation which would enable managers to make informed decisions about how to apply such a model. Using an IBM to predict the response of fishers to management could facilitate more informed compromises between management objectives, and reduce uncertainty in fisheries management.

Modern fishing fleets have the capacity to over-exploit fish stocks. Inaccurate assessments could overestimate the stock size and as a result Total Allowable Catches (TAC) are set too high for sustainable stock conservation. Fisheries management need robust and reliable stock assessments to ensure that the species and environmental effect of fishing is sustainable. Since the demand for ecosystem based approaches to management has increased, the needs for improved estimates of un-assessed abundance have risen. Managers simply need to know how many fish left in the see and how much to limit the fishermen to fish to have sustainable fisheries. Therefore, accurate assessment of the market as well as by-catch stocks and records of true landings and discards are critical aspects of the scientific advice to the fisheries managers to accurately set TACs. Here, we consider the marine species that are left un-assessed. That is because they cannot be assessed by the existing methods. We therefore sought to fill the key gap with this matter. This thesis has five key elements. First we reviewed the stock assessment method with the emphasis on the length-structured models. Second, we produced a population model (so called survey-landings model) to make the use of survey frequency data extracted from International Bottom Trawl Survey and total annual landed biomass from commercial reports. Third, within a twin-experiment context and sensitivity analysis the model was assessed for accuracy and robustness in variability in initial parameter values and observational noise. Forth, applying the survey-landings model the population dynamics of the North Sea haddock was assessed and the results were compared with the International Council for Exploitation of the Sea assessment. Fifth, after the model proved to be reliable it is used as an alternative for age- or catch-at-length model, the population of the North Sea grey gurnards were modelled with confidence. This model enabled un-assessed species such as grey gurnards to be modelled and assessed for the first time.

For over a century there has been a push to preserve the areas of nature where the human impact is the smallest, often referred to as wilderness. In Europe the suitability of the concept is debated, as the entire continent has been heavily modified by humans, and the areas without visible human impacts are small and fragmented. At the same time there is a strong push for preserving these areas, including the areas at sea. At sea the wilderness concept faces unique challenges, as the environment is less understood than the terrestrial, and the potential wilderness areas are not necessarily accessible for recreational purposes. This thesis examines the use of the wilderness concept, especially as it relates to the marine environment, in both policy and common use. The aim is to contribute to the conceptual framework for marine wilderness, by studying how the concept of wilderness is understood and used by policymakers and stakeholders. The research is conducted using discourse analysis on legal texts and newspapers, and surveys, interviews, and social network analysis to examine the views of individual stakeholders. The results show that while there is political will in Scotland to conserve wild areas, which are more modest in size than wilderness but provide the wildness quality and its beneficial effects, the concept of wilderness has multiple interpretations, and can be rather political. To address the consequent issues, participation of stakeholders is considered vital for successful management. Marine wilderness remains a particularly ambiguous concept, and considering the ongoing tensions in marine resources management, it is suggested that marine wilderness is more useful if seen as an added benefit than the ultimate goal.

Biodiversity describes diversity at different levels of biological organisation, including: habitat diversity; species diversity; and genetic diversity. Understanding the processes that contribute to maintaining biodiversity is a primary concern for both ecology and evolutionary biology. To this end, research into the factors influencing the different levels of biodiversity independently are widespread. However, little is understood about the relationship between the different levels. This study investigates the patterns of habitat, species, and genetic diversity in fragmented internationally important calcareous grasslands, and analyses the spatial and temporal factors influencing them. Finally, the relationship between these levels of biodiversity is examined. Within the South Downs National Park study area, substantial change to habitat diversity and landscape structure was measured between the 1930s and 2012. The transition of semi-natural habitat to agricultural land was the predominate change. Loss of habitat between the 1930s and 2012 was found to influence both species richness and species evenness of vegetation in twelve calcareous grassland study sites. By contrast, none of the variables examined explained the variation in species composition between sites. Further analysis, at the genetic level, for two target species showed that the amount of habitat loss was important in explaining the genetic variation in Cirsium acaule, and soil nutrients were important in explaining the variation of Ranunculus bulbosus. In contrast to the predictions of the species genetic diversity correlation theory, no relationship was established between species and genetic diversity. Similarly, no relationships were found between habitat diversity and diversity at the species or genetic level. Although there were similarities in the factors influencing different levels of biodiversity, habitat diversity, species diversity, and genetic diversity appear to be responding independently to the processes acting on them. As such efforts to conserve biodiversity should consider the influence of conservation strategies on biodiversity holistically, and not focus on a single measure.

A better understanding of how biodiversity forms and is maintained is critical for informing our conservation efforts. Useful tools for achieving this are models of community assembly which help identify the dominant processes in different ecosystems. Neutral models in particular, have the benefit of being simple, tractable, mechanistic models that can produce good fits to empirical data for many different ecological communities.