Geomorphological evidence has frequently been used to infer past environmental conditions, but in recent years the emergence of landscape evolution models (LEMs) has opened the possibility of us- ing numerical modelling as a tool in palaeo-environmental reconstruction. The application of LEMs for this purpose involves retrodictive modelling, each simulation scenario being congured with model variables (e.g. reflecting climate change) and parameters to reflect a specic hypothesis of environmental change. Plausible scenarios are then identied by matching contemporary observed and modelled landscapes. However, although considerable uncertainty is known to surround the specication of model driving conditions and parameters, previous studies have not considered this issue. This research applies a technique of accounting for the uncertainty surrounding the speci- cation of driving conditions and model parameters by using reduced complexity 'metamodels' to analyse the full model parameter space and thus constrain sources of uncertainty and plausible retrodicted scenarios more eectively. This study applies the developed techniques to a case study focused on a specic set of coastal gullies found on the Isle of Wight, UK. A key factor in the evolution of these gullies are the relative balance between rates of cliff retreat (which reduces gully extent) and headwards incision caused by knickpoint migration (which increases gully extent). To inform the choice and parameterisation of the numerical model used in this research an empirical- conceptual model of gully evolution was initially developed. To provide a long-term context for the evolution of the gullies and to identify the relative importance of the various driving factors, the Holocene erosional history of the Isle of Wight gullies was then simulated using a LEM. In a preliminary set of simulations a 'traditional' (i.e. with no consideration of parameter uncertainty) retrodictive modelling approach was applied, in which driving variables were arbitrarily altered and observed and simulated landscape topographies compared, under various scenarios of imposed environmental change. These initial results revealed that the coastal gullies have been ephemeral in nature for much of the Holocene, only becoming semi-permanent once cli retreat rates fall below a critical threshold at 2500 cal. years BP. Next, in an attempt to constrain more detailed erosional histories and to explore the extent to which retrodicted interpretations of landscape change were confounded by uncertainty, a Central Composite Design (CCD) sampling technique was employed to sample variations in the model driving variables, enabling the trajectories of gully response to dierent combinations of the driving conditions to be modelled explicitly. In some of these simulations, where the ranges of bedrock erodibility (0:03-0:04m0:2a).

The aim of this thesis is to develop an integrated methodology for investigating the potential for new local railway stations within a given area, with particular emphasis on the use of Geographical Information Systems (GIS). Existing methods for assessing the case for constructing new local railway stations have often been found wanting, with the forecasts produced proving to be inaccurate. A review of previous work in this field has been undertaken and methodologies with the potential to enhance local rail demand models have been identified. Trip rate and trip end models have been developed which are capable of forecasting usage at new station sites anywhere in England and Wales. Geographically Weighted Regression (GWR) has been used to enhance the performance of these models and to account for local variations in the effects of explanatory variables on rail demand. Flow level models have been produced for stations in South-East Wales, with a range of model formulations tested. A survey of ultimate passenger trip origins and destinations was carried out in the same area, enabling the accuracy of theoretical station catchment definition methods to be tested. A GIS-based procedure for locating potential sites for new railway stations within a given area has been developed. This was combined with the results from the demand models and estimates of associated costs and benefits to give a synthesised appraisal procedure capable of assessing the case for constructing particular stations. This procedure was applied to 14 sites in South-East Wales and, along with trip end forecasts for 421 sites across the country, this indicated that there is almost certainly a positive case for constructing a significant number of new railway stations in the UK.

Low-lying coastal zones are high-risk areas threatened by flooding due to extreme coastal events and rising sea-levels. The coastal floodplain system includes elements such as near-shore waves and water levels, inter-tidal beaches and coastal habitats, natural and artificial sea defences and multiple inland floodplain features. Flood risk studies generally achieve an integrated assessment of these elements using multiple numerical models for different floodplain elements. However fundamental choices of floodplain description and the appropriate data, methods and models can vary widely between different sites and flood risk studies. A comprehensive conceptual model is needed to describe the floodplain system and help inform these choices in each site. However a descriptive conceptual model for coastal floodplain systems does not exist at present. There is a bias in flood risk studies towards the direct use of numerical models with limited use of conceptual models – existing models are implicit and do not describe the coastal floodplain system. This thesis addresses this gap by developing, applying and testing a rapid appraisal tool that conceptually describes the coastal floodplain as a system of interacting elements. The tool is developed in two parts – i) a quasi-2D Source – Pathway – Receptor (SPR) model that provides a comprehensive qualitative description of the floodplain; and ii) a Bayesian network model that uses this description to quantify individual elements as sources, pathways and receptors of flood propagation. The quasi-2D SPR is applied in 8 diverse coastal zones across Europe 4 of which include nested case-studies. It is an effective way of gathering and describing information about the floodplain from stakeholders across multiple disciplines. The Bayesian network model is applied to two contrasting floodplain systems in England – Teignmouth and Portsmouth. The network model is effective in pinpointing critical flood pathways and identifying key knowledge gaps for further analyses. The two models together provide a comprehensive understanding of the coastal floodplain system that can be used to inform and target the use of more detailed numerical models. Hence this thesis provides a conceptual model and tool to improve flood risk assessment. It makes conceptual understanding of the floodplain explicit and stratifies quantitative analysis by application of a rapid assessment tool before the use of detailed numerical models.

Humans have interacted with their surroundings for over one million years, and researchers have only recently been able to assess the geomorphic impacts indigenous peoples had on their landscapes prior to the onset of European colonialism. The history of human occupation of Australia is noteworthy in that Aboriginal Australians arrived ~50 ka and remained relatively isolated from the rest of the world until the AD 1788 when Europeans established a permanent settlement in Sydney, New South Wales. The southeastern Australian Tablelands landscape, west of Sydney, has seemingly undergone drastic geomorphic change since European arrival. The introduction of European grazing practices reportedly led to the occurrence of deep erosional incisions, gullies, into valley bottoms and hillslopes, releasing sediment, which is subsequently deposited over downstream wetland environments – swampy meadows. This sediment is often called post-settlement alluvium (PSA); however, the age and genesis of PSA in Australia are debated. Questions regarding the geomorphic features and processes in the Tablelands remain unanswered because few studies quantify the timing of gully incision, PSA deposition, or the pre-human rate of landscape change. Erosion rates inferred from concentrations of in situ 10Be measured in fluvial sediment (n = 11) and bedrock outcrops (n = 6) range from 2.9–11.9 mm/kyr and 5.2–13.8 mm/kyr, respectively. The two sample populations are statistically indistinguishable, suggesting no relief has been generated since 600–110 ka. The overall erosion rate in the Tablelands is 7.5 mm/kyr, equal to long-term denudation rates integrated since ~20 Ma. Aboriginal Australians have been present in the Tablelands for at least 30 kyr, ~12–26% of the cosmogenic integration time, yet widespread Aboriginal fire use did not measurably affect landscape erosion until ~5.5 ka, in sync with increased charcoal in the sediment record. Portable optically stimulated luminescence (OSL) reader data from poly-mineral and poly-grain size samples collected from gully wall profiles of PSA and swampy meadow sediment show that swampy meadow environments were buried by PSA and that PSA is alluvium derived from upstream gully erosion. No relationships between bulk OSL and sample grain size or mineralogy exist, and inferences about bulk sediment mineralogy or grain size cannot be determined from portable OSL reader data. Large variability in adjacent PSA sample replicates, however, reveals incomplete sediment bleaching conditions during PSA deposition during floods. Greater bleaching efficiency is inferred from the small variability of bulk OSL data in the uppermost 10s of cm of PSA profiles. Measured concentrations of meteoric 10Be and bulk OSL in two PSA deposits in Birchams Creek show that initial gully incision eroded into weathered sandstone regolith and not swampy meadow environments, as previously believed. Initial gully incision was shallow (<15 cm) and PSA filled ponds in the lower reaches of the catchment. Continued erosion upstream led to a second depositional episode of PSA before headward gully incision from the mouth of Birchams Creek eroded through PSA deposits. Headward erosion of this gully created the continuous gully present at the site today. Initial gully incision was likely the result of livestock trampling in valley bottoms during droughts, creating localised slopes greater than the critical slope threshold required to erode the valley bottom. OSL burial ages of six PSA deposits collected throughout the Tablelands range from 195.1 ± 17.8 to 90.4 ± 8.9 a, corresponding to AD 1800–1932. The OSL burial ages are younger than European arrival in the Tablelands, and the term, PSA, is redefined as post-European settlement alluvium in Australia, recognising the earlier settlement of the region by Aboriginal Australians whose land use did not lead to PSA deposition. PSA burial ages agree with existing quantitative and anecdotal gully incision data. Contrary to previous assertions that gully incision began asynchronously in the Tablelands, three periods of synchronous gully erosion in localised areas within the Tablelands are recognized: 185 a, 158 a, and 94 a (AD 1828, 1855, and 1919, respectively) – in the southern, northern, and central Goulburn Plains, respectively. The AD 1828 and AD 1919 periods of gully incision correspond to the transition from drought-dominated climate regimes to flood-dominated regimes, and the AD 1855 period of gullying corresponds to a flood-dominated regime. Gully incision in the Tablelands is thus a result of European-introduced grazing practices, which primed the landscape for further erosion and degradation during climatic shifts. PSA deposits in the southeastern Australian Tablelands are some of the most recent examples of anthropogenic sedimentation in human history. The earliest preserved examples of PSA-type sediments are ~8,000 years old and found throughout the world. The establishment of an onset date for the Anthropocene is currently debated, and I believe the oldest PSA and PSA-type sediments around the world can define this modern epoch.

Ecosystems undergo permanent alteration and degradation as a result of the pressure exerted upon them by anthropogenic activities and natural causes. Amongst the natural causes, drought can have considerable impacts on vegetation productivity; these impacts include biomass reduction and changes in vegetation cycles, growth, and vitality. Such impacts have important implications for society, and understanding the impacts of drought in natural systems will help minimise vulnerability to it. The Yucatan Peninsula is a vulnerable area in climate change scenarios, where increased intensity and frequency of droughts is already occurring. This thesis aims to analyse how droughts impact upon vegetation on the Yucatan Peninsula, Mexico using both remotely sensed data (in particular the Advanced Very High Resolution Radiometer – AVHRR - data) and meteorological data (in situ precipitation data). Three major elements are addressed: (1) spatial and temporal precipitation variability, and the occurrence of droughts during the period 1980-2011; (2) vegetation productivity trends, focusing on the persistence and resilience of the different vegetation types present across the Yucatan Peninsula using remotely sensed data sets, specifically, the Normalised Difference Vegetation Index (NDVI) from the National Oceanic and Atmosphere Administration Advanced Very High Resolution Radiometer (NOAA-AVHRR) as a proxy for productivity in the period 1982 to 2011; and (3) the relationship between vegetation productivity and rainfall and the lack of it (i.e. droughts) at a more detailed temporal scale (monthly) with two vegetation indices (NDVI from AVHRR and MTCI from the MEdium Resolution Imaging Spectrometer (MERIS)). The main findings indicate that precipitation in the area follows a gradient from north to south which is linked to vegetation types in the area, from deciduous forest to evergreen forest. NDVI trends in the Yucatan Peninsula are generally stable during the period 1980-2011; however, negative trends are persistent in areas where human impacts and intensive agriculture are present. Analysis of resilience also shows that deciduous forest and grasslands are much less resilient to disturbance and have a longer recovery period compared to other vegetation types. Per-pixel regression analysis between precipitation and vegetation productivity shows high explained variances between NDVI and rainfall with a time-lag of 1 and 2 months and tropical dry forest is the biome most affected by droughts. These results are valuable for decreasing the vulnerability of the Yucatan Peninsula and they could be used for increasing the understanding of the area and therefore go further in the creation of warning and management programmes.

Flood risk management is increasingly seeking to involve local stakeholders in decision-making, both to harness and benefit from their tacit knowledge and to devolve responsibility for delivering local-scale, individual and community responses. Current techniques used in flood risk management centre on a techno-scientific approach, which is well-suited to appraising and modelling hazard, but often inaccessible to those without specialist, technical expertise. This leads to participation that is often limited to infrequent consultation periods, keeping local stakeholders at the periphery of the decision-making process. Their absence from the more technical elements of the process can leave local stakeholders struggling to understand how different options have been identified, appraised and/or prioritised. This can in turn lead to dissatisfaction with process outcomes, lack of support for selected options, and foster distrust in expert practitioners. This thesis explores how participatory modelling techniques could complement current approaches, facilitating the co-production of models with local stakeholders that explore social constructions of risk, and the vulnerability of different receptors. It identifies the qualities that are sought from participation, including the need to remain highly accessible, yet sufficiently robust to capture the complexities encountered when working at the interface of social and physical systems. Reporting on two UK case studies, it exemplifies the benefits that two popular techniques, Bayesian networks and system dynamics, can deliver at different stages in the flood risk decision-making process. In each case, the effectiveness of the participatory approach is assessed using an evaluative framework that advances current approaches by including an early assessment of context, as well as a detailed exploration of substantive (user-defined goals), and social change outcomes. The holistic nature of the evaluation framework, and its population with practical criteria bespoke to flood risk management, enhance its transferability between different contexts. The thesis finds that participatory modelling techniques support the collating of diffuse tacit knowledge, building of consensus, strengthening of social networks, and the empowerment of local citizens to become volunteer risk managers; provided that process managers are willing to simplify the techniques to maintain accessibility, and be open to different metrics of success.

Patagonia has one of the most extensive and well preserved glacial geomorphic records of anywhere else in the world. This study provides empirical constraints of Patagonian Ice Sheet (PIS) configuration and dynamics during the last two glacial cycles and the Holocene at 49°S. In particular a chronology of palaeo-ice surface elevations, thickness changes and ice-thinning rates is developed. Cosmogenic surface exposure dating across eight mountain transects at Hielo Patagonico Sur (HPS) outlet glaciers' Viedma and Chico, and from eastern facing mountain valley glaciers, is utilised to reconstruct PIS ice-surface elevation changes through time. The earliest dated glacier ice thickness marker falls into MIS6 (Marine Isotope Stage). The last glacial cycle is characterised by continuous ice-surface elevation lowering from a maximum at MIS5a to the Antarctic Cold Reversal (ACR). Large scale PIS cover occurred between ca. 40-47 kyrs, in both the HPS outlet glaciers and the eastern mountain valleys. This regional ice-cover considerably pre-dates the global Last Glacial Maximum (gLGM). A gLGM vertical ice surface expression is absent from the study area. A minor re-advance and/or stillstand at ca. 18 kyrs is recorded by the valley glaciers, but is not evident at the HPS outlet glaciers. Glacier Viedma records a prominent ice elevation during the ACR with rapid thinning at the end of the climate perturbation. Ice-surface lowering from the ACR limit continues with acceleration toward the late-Holocene. The PIS palaeo-ice elevation and thinning record presented here is sensitive to internal PIS dynamics, local and global climate forcing acting on different timescales, with glacier response to climate perturbation strongly linked to glacier catchment size and distance from the main source of precipitation at 49°S.

Freshwater systems are generally found to be sources of CO2 to the atmosphere and evasion of CO2 from fluvial systems is now recognised to be a significant component of the global carbon cycle. However detailed understanding of fluvial carbon dynamics and controls on the system is lacking and global coverage of published data is sparse, but thorough understanding across a broad range of locations is crucial if global carbon budgets are to be refined. This research addresses this lack of understanding by investigating the magnitude, controls and sources of CO2 efflux across five catchments with different catchment characteristics, global locations and climate-change sensitivities. In doing so new understanding is used to explore a novel method for large-scale upscaling of CO2 efflux, time series reconstruction of the source and magnitude of CO2 efflux and incorporation of an Arctic region into the global fluvial carbon budget. The magnitude of and controls on CO2 efflux are not well understood, although it has been suggested that increased flow velocity and turbulence can enhance CO2 efflux rates. This research uses direct and contemporaneous measurements of CO2 efflux (range: -3.53 to 107 μmol CO2 m−2 s−1), flow hydraulics (e.g. mean velocity range: 0.03 to 1.39 m s-1; shear Reynolds number range: 350 to 174000), and water chemistry (e.g. pCO2 range: 388 to 4660 ppm), at sites in three UK catchments to assess whether flow intensity (a term which is used to describe one or more measures of flow strength and turbulence) is a primary control on CO2 efflux. These field sites have been chosen as they have contrasting size and land use: Drumtee Water (DW), 5.7 km2 and rural, the River Kelvin (RK), 335 km2 and urban, and the River Etive (RE), remote and snow-melt influenced. At the more soil-dominated sites DW and RK, a strong positive logarithmic relationship exists between CO2 efflux and measures of flow intensity (e.g. shear Reynolds number, overall R2 = 0.69), but this relationship is strengthened by including pCO2 (overall R2 = 0.72). Flow intensity may have a key influence on CO2 influx, although data are limited. A method using visual classification of flow intensity shows promise for supporting large-scale upscaling of fluvial CO2 efflux, if classification of water surface state can be standardised. Movement of dissolved inorganic carbon (DIC) through the hydrological cycle is an important component of global carbon budgets, and how they may respond to changing climatic conditions. However uncertainty remains about the hydrological and biogeochemical controls on DIC transmission through a catchment. Using contemporaneous measurements of DIC concentration ([DIC]) and stable carbon isotope composition of the DIC pool (δ13CDIC), fluvial DIC at more soil dominated sites, DW and RK, is found to vary considerably in response to changes in catchment hydrology. At low flow groundwater dominates, and has similar composition in both systems ([DIC]: 1.5 mmol L-1 DW, 2.0 mmol L-1 RK; δ13CDIC: -9 ‰ DW and RK) indicating a common hydrogeological inheritance in DIC, that is comparable to that of other temperate and tropical locations. Differences in composition at high flow ([DIC]: 0.1 mmol L-1 DW, 1.0 mmol L-1 RK; δ13CDIC: -23 ‰ DW, -14 ‰ RK) reflect catchment land use, and a lower contribution of soil water to the DIC pool in the more urban catchment (RK). Measured diel cycles in DIC pool composition at DW indicate biological processes modify the pool, and time series reconstructions of pool composition and CO2 efflux at DW reveal seasonal- and flow-related patterns in this biological activity. Time series reconstructions also reveal that at DW terrestrial-aquatic-atmospheric carbon cycling is rapid during event flows, with large amounts of CO2, of soil-origin, effluxed to the atmosphere in relatively short periods of time. Conversely, at low flows, CO2 efflux is of smaller magnitude and primarily fuelled by groundwater, and terrestrial-aquatic-atmospheric carbon cycling is slower. The reconstructions allow for inter-year comparisons which are useful in assessing for behaviours in CO2 source and feedback that might be typical under climate change-induced changes in hydrology (e.g. wetter winters, drier summers, more frequent large flow events). Global ice melt and permafrost thaw are increasing due to climate change, effects of melting ice and thawing permafrost on the global carbon cycle, and carbon cycling dynamics of the melt/thaw waters are not well understood. Data from the River Etive has few similarities to that of DW and RK and indicates that snow- and ice- dominated systems may behave very differently to more soil-dominated systems in terms of magnitude and controls on efflux and sources and mixing of the DIC pool. This is confirmed by data collected from the melt/thaw waters of two cryospheric systems in Greenland: a Greenland Ice Sheet (GrIS) drainage river (Akuliarusiarsuup Kuua River, or AR) and the local permafrost-landscape surface-drainage systems (PLST). CO2 efflux appears independent of flow controls in both systems, and instead seems to be pCO2 limited (average pCO2: 115 ppm AR, 596 ppm PLST), with spatial variation in AR (efflux decreases downstream) and temporal variation in PLST (efflux decreases with melt season progression). The frequent occurrence of CO2 influx (measured in 64% and 14% of cases in AR and PLST respectively), which has rarely been reported from other rivers globally, reveals that Arctic fluvial systems can periodically act as net sinks of CO2 and this should be incorporated into global carbon budgets. The occurrence of CO2 influx, and dominance of air-water CO2 exchange in these low pCO2 systems, is reflected in the DIC pool composition which is 13C-enriched and approaches isotopic equilibrium with the atmosphere (~0 ‰), and indicates that soil and ground water contributes little to the DIC pool under frozen ground conditions. Radiocarbon analysis gives further insight into the source of carbon in these systems, revealing that the GrIS is releasing old DOC (~5200 to 6600 yrs BP) upon melting, which is considered to be highly biolabile and may prime bacterial activity and feedback to climate change, and meltwaters are returning old carbon (800 to 960 yrs BP) to the atmosphere via CO2 efflux. Thus it appears that climate change (via melting ice sheets) may be a driver of the age of atmospheric carbon composition. The effluxed CO2 being less old than the DOC indicates the source of CO2 efflux is a mixed pool of respired/UV-oxidised old DOC and modern atmospheric CO2 from drawdown. In contrast to GrIS meltwaters, and the permafrost of other global locations (e.g. the Siberian Yedoma deposits), the permafrost landscape of the Kangerlussuaq region of Greenland is cycling modern carbon and appears not to be degrading, as old carbon is not found in, or degassed from, the fluvial systems. In summary this research contributes to a greater understanding of fluvial carbon dynamics and the processes controlling the return of CO2 to the atmosphere via efflux, across an array of catchment types, sizes, land uses and global locations, and makes contributions of novel data to a number of areas of fluvial carbon cycling research where there are scarcities. Marked differences in the fluvial carbon cycling dynamics of cryospheric and snow-melt dominated systems compared to soil-dominated terrestrial systems are uncovered, novel upscaling attempts made using new findings of the research, and a number of exciting new research directions and opportunities that could enhance the findings of this work are identified. Overall, this research takes steps towards a greater understanding of fluvial carbon cycling dynamics on a global scale and improved projections of the likely response of fluvial systems to climate change, ultimately aiding the community to be more prepared for what our shifting climate will bring.

With over 200 pools, lakes and rivers supporting over 70 species of endemic flora and fauna, the Cuatro Cienegas Basin, Coahuila, NE Mexico is an extremely important and extensively studied area in terms of conservation. The palaeoenvironment, however, is relatively understudied with only two reconstructions published to date - Meyer [1973] and Minckley and Jackson [2008]. This project has analysed a 15 m carbonate sediment core for multi-proxy palaeoenvironmental information and combined this with stable isotope, modern hydrological and geoarchaeological information in the Cuatro Cienegas Basin.

The full extent of a landscape’s resilience to the environmental impact of siting wind-based renewables on peats is currently unknown. This research explores if windfarm construction activities have caused disturbance by investigating; time series of fluvial carbon (C) and nutrient concentrations; constructing aquatic organic C fluxes, before, during and after the windfarm construction period. Additionally, C sequestration rates of peat and nearby lake sediments (Loch Brora) were calculated to provide a historical context to, i) calculated aquatic C fluxes and ii) sediment export from surrounding catchments, considering both a catchment hosting the windfarm construction and one that does not. Furthermore, the effectiveness of a peatland restoration technique, drain-blocking, was assessed as a means of undertaking a whole system approach to assessing the potential impact of the windfarm development, considering how these management strategies can help mitigate potential C losses associated with construction. The research field site was located on the Gordonbush Estate, near Brora, where construction started in July 2010 (the same time this research began) on Scottish and Sothern Energy Renewables (SSER) 35 turbine windfarm. Construction work finished in May 2012 and data collection continued until September 2014. Throughout this period, fieldwork was focussed on storm event sampling (collecting samples for dissolved organic carbon (DOC), particulate organic carbon (POC), total phosphorous (TP), soluble reactive phosphorous (SRP) and total oxidised nitrogen (TON)), collecting peat and lake sediment cores, samples of modern day sediment export and monitoring water table depth in an area where old drainage channels were blocked as part of a peatland restoration initiative. Three river catchments were studied, two affected by windfarm construction activities (GB10 and GB11) and one control site (GB12), DOC concentrations ranged from 1.1 mg l-1 to 48.3 mg l-1, POC from <0.1 mg l-1 to 21.3 mg l-1, TP from <0.5 µg l-1 to 264 µg l-1, SRP from <0.5 µg l-1 to 39 µg l-1 and TON from <1 µg l-1 to 141 µg l-1. These were all within ranges of macronutrient concentrations measured at other northern temperate peatland sites. Comparing macronutrient concentrations between catchments, generally GB10 > GB11 > GB12 for all determinants. Seasonal patterns in fluvial macronutrient concentrations were observed at Gordonbush: summer maxima and winter minima in DOC and TP concentrations and the opposite trend in TON concentrations. SRP data collected indicates a legacy of forest felling in the Bull Burn Plantation has contributed to increased concentration in the Allt Mhuilin river (GB10) compared to the two other catchments, Allt Smeorail (GB11) and Old Town Burn (GB12) where no forest felling occurred during the data collection period. Differences in DOC and TP concentration in Allt Mhuilin compared to other catchments could also be related to forest felling activities but catchment characteristics such as peat coverage may have also influenced results. For all relevant measures of water quality, macronutrient concentrations from Gordonbush shows studied streams consistently achieved “Good” or “High” status throughout the data collection period. Apart from the legacy of forest felling, a discernible impact of windfarm construction was not observed from macronutrient concentration time series. Calculating annual aquatic C fluxes from studied catchments offered a means of assessing potential impact. Various techniques of estimating fluxes were explored but splitting storm event DOC and POC data based on time of year and whether samples were collected on the rising or falling limbs were concluded to give the best estimates. Calculated fluxes ranged from 3 – 38 g C m-2 yr-1 and DOC consistently accounted for ~90% of total aquatic C export. These values were within limits of other C flux based studies from peatlands but the time series constructed at Gordonbush suggested windfarm construction, between July 2010 and May 2012, may have contributed to an increase in aquatic C export from affected catchments during this time, relative to the control site. Long term C sequestration rates from within the Gordonbush estate were 20-25 g C m-2 yr-1, the same magnitude as aquatic organic C fluxes. However, peat C sequestration was shown to be variable over the last ~9000 years since Scottish peatlands became established, with rates ranging from 10-60 g C m-2 yr-1. Controls on this variation are likely climatic with delivery of moisture influenced by the North Atlantic Oscillation (NAO) a key factor. Calculated lake C sequestration also varied over time, 22-82 g C m-2 yr-1 but an inconclusive radiocarbon dating chronology meant historical comparison of C export dynamics between the, C ‘source’, peatland to the, C ‘sink’, lake was unfortunately not possible. Modern day sedimentary export data showed higher sediment yields from windfarm affected catchment than the control site. Physical characteristics varied considerably between the two catchments so although this observation could not definitively be attributed to a direct windfarm impact, it remains a possibility. Whilst studying and quantifying the impact of drain blocking, manual measurements of water table depth (WTD) ranged between -53 cm to +14 cm in dip-wells and -36 cm to +20 cm in automated logging pressure transducers. The response of WTD throughout both data sets indicates meteorological conditions were more influential as a factor controlling peat hydrology across the site compared to topography. Manual measurements from dip-wells shows the drainage channels investigated (~0.5-0.7 m deep and ~0.5 m wide) had the greatest influence on effect WTD 0-2 m from the main channel but no statistically significant difference was detected in mean WTDs measurements before or after blocking, in relation to distance from the drainage channels themselves or comparisons between drained and un-drained (control) areas. However, data from PTs indicate the net effect of multiple parallel drains can cause water table drawdown at a significant distance, ~ 25 m, from the drainage channel. This is an important finding as methodology used to calculate the C ‘payback time’ of windfarms utilises the lateral drainage extent of peat when turbines bases are excavated. Drain blocking had no obvious effect (either positive or negative) on WTDs however it is acknowledged positive effects can take up to five years, after blocking has taken place, to be observed. Maximum DOC concentrations increased the year after blocking however this result has been recorded at other sites and the exceptionally dry summer of 2013 could have contributed to the noticed increased by promoting more peat oxidation and subsequently DOC production. There was no statistically significant difference between [DOC] collected up and downstream of the drainage channel inputs for samples collected before and after blocking. This suggests drain-blocking has had little impact on the larger site [DOC] signature one year after drain-blocking. However, as discharge from drainage channels was not measured, a potential reduction in overall DOC export could not be fully assessed and this is a highlighted future research need. Combining averages of aquatic organic C fluxes and peat C sequestration rates calculated it is estimated net ecosystem exchange would have to be between -30 to -50 g C m-2 yr-1 for Gordonbush to be classed as a C ‘sink’. If the observed increases in sedimentary export could be attributed to windfarm construction, Loch Brora is unlikely to act as a strong C sink for any potential increased losses as it is estimated ~90% of POC exported is not sequestered on a long-term basis in the lake sediments. It has recently been recommended windfarms should not be developed on peatlands due to the marginal C savings achieved as our future energy mix changes (Smith et al., 2014). However, if similar projects are granted planning permission then findings from this research support the following recommendations: installation of buffer zones around areas of felled forestry to reduce nutrient export into surrounding streams; implementation of a water quality monitoring programme to assess impact of windfarm construction during construction and a period afterwards as it is still unclear from this research if there will be any lasting effects; installation of silt traps to reduce aquatic sediment export and disturbance; limit any high density excavation of drainage channels as the effects of water draw-down could be quite extensive; in addition, blocking all historical drainage channels and retaining as much moisture as possible within, and surrounding, areas of degraded peatland can increase long-term peat C sequestration rates and offset C losses experienced during construction. This research has been funded by SSER, Engineering and Physical Sciences Research Council (EPSRC) and Energy Technology Partnership (ETP). This research has been undertaken and supported at the University of Glasgow within the College of Science and Engineering, specifically aligned to the work of the Carbon Landscapes and Drainage (CLAD) research group headed by Prof. Susan Waldron in the School of Geographical and Earth Sciences. Finally, this research has also been supported in partnership with Stirling University.