Many parts of the Holocene climate system (e.g. impacts of freshwater input on polar oceans) are not fully understood due to a paucity of data. These data gaps limit our ability to examine the drivers of Holocene climate variations, restricting the accuracy of climate projection models. Freshwater input to polar oceans may have significant impacts on the behaviour of the thermohaline circulation (THC). In particular, freshwater input to the North Atlantic sector of the THC may cause reduced salinity slowing water mass sinking which drives the THC. Currently, our understanding of Holocene salinity change is temporally restricted due to limited instrumental runoff and salinity records. One approach to extend the records is to use runoff proxies; environmental recorders of historic runoff. The high temporal resolution, marine, red coralline algae (Lithothamnion glaciale), has been suggested as a potential solution to this problem. Coralline algae are long-lived with slow growth rates that display well developed annual growth bands at seasonal time-scales. In Greenland, they have been shown to record changes in water salinity and temperature at annual—5 year resolution. However, before L. glaciale, is used more widely to reconstruct freshwater runoff, we need to understand; 1) the water source that drives changes to salinity and 2) if the different water sources can be detected in the marine environment isotopically (δ18OVSMOW and δDVSMOW). To achieve that, this thesis aims to understand the oxygen (δ18OVSMOW) and hydrogen (δDVSMOW) isotope compositions in the routing of freshwater, from its source into the marine environment. This will enable determination of the environmental conditions which L. glaciale is exposed to and thus recording. Isotopic composition of source and marine waters were determined at locations that have a terrestrial-fjord/loch system and locations that are influenced by snow and ice melt (Glencoe, Scotland and Kangerlussuaq, Greenland respectively). At both locations, snowmelt, ice melt, riverine, lacustrine and marine waters were sampled for δ18OVSMOW and δDVSMOW analysis to determine; 1) the starting source isotope compositions in areas of freshwater influences into the marine environment, 2) any processes influencing the oxygen and hydrogen isotope composition change from the source to ocean, 3) the extent to which the isotopic composition changes throughout the runoff season and 4) the consequences of 1)-3) for using L. glaciale as an in-situ palaeo-runoff indicator. Similar results were found at both collection sites; 1) isotope compositions enriched from the main freshwater source, through a river system and into the marine environment, 2) δ18OVSMOW and δDVSMOW compositions could pinpoint the exact areas of freshwater input and mixing along the fjord/loch, 3) the fjord/loch have fast hydrological responses to changing runoff conditions, 4) additional freshwater inputs and changes in the source are the primary processes altered the isotopic composition of the freshwater surface layer, 5) established individual sources can be separated from the bulk water mass and can be detected 118 km (Søndre Strømfjord) and 20.4 km (Loch Etive) away from the main freshwater terminus and 6) a suitable collection site can be determined from the isotopic signal ensuring that the algae record the full runoff signal. Knowing the exact areas of freshwater input along with spatial mixing and tidal action within the surface layer of the fjord/loch has given insights into the environmental forcing acting on the algae at the time of growth. To relate key findings of this thesis with the algal signals detected in previous studies; salinity-temperature-δ18OPDB relationships (within Søndre Strømfjord) were used to determine any considerations for the use of L. glaciale as a climate recorder if salinity-driven δ18O patterns are not accounted for. However, it is important to note that as algal samples were from other studies, the temporal miss-match between water (this study) and algal collection may drive some of the observations made. It was found that; 1) Isotopic compositions extracted from L. glaciale growth bands in previous studies are enriched by 8.7-10.72 ‰ compared to calculated δ18OVSMOW at the algae collection sites, 2) the δ18OVSMOW environmental signal is thus not directly incorporated into the algae’s skeletal structure providing a consistent offset or enrichment of the freshwater signal, 3) calibrating algal δ18O with predicted sea surface temperatures (SST) at the algae collection site generates an algal-derived summer SST of ~4 °C which is comparable to instrumental records, 4) variations in the marine δ18OVSMOW composition due to a freshwater input can cause a δ18OPDB change within the algae and may influence the reconstructed temperature in areas of high freshwater input, 5) greater GrIS runoff corresponds to depleted δ18OPDB and cooler reconstructed temperatures and thus indicates that freshwater provides a temperature error in reconstructed temperatures. However, the runoff itself cools the water and this effect could be less than initially expected. This investigation has suggested the optimal collection locations for algae to increase the likelihood of them having recorded freshwater discharge signals. This had made a major advance as in the past algal collection locations were largely chosen using chart assessments. A dedicated study, which uses Lithothamnion glaciale collected from the suggested optimal recording locations, is now needed. Overall, knowledge generated in this study will enable advances in reconstructions of runoff from ice sheets, glaciers and calibrations using marine proxies, enabling detailed reconstruction of Holocene climate variability.

The Earth’s system has witnessed abrupt climate changes throughout its history. It is widely understood that the pacing of the large-scale glacial-interglacial cycles (tens to hundreds of thousands of years) of the Pleistocene Epoch (~2 million years) are not simply the result of changes in the incoming solar insolation (orbital forcing) alone and as such, this large-scale variability calls upon feedback mechanisms in order to explain the magnitude of such changes. This thesis investigates the shorter millennial scale events (thousands of years) and their role in determining the magnitude of glacial-interglacial cyclicity. In order to investigate this, in this thesis three sediment cores will be used, ODP Site 1063 and ODP Site 983 in the North Atlantic and ODP Site 1089 in the South Atlantic. Datasets from these core locations provide insights in to the magnitude and duration of millennial scale variability across terminations 1, 2 and 5. The two regions studied here (North and South Atlantic) are important regions for recording changes in ocean circulation, specifically the Atlantic Meridional Overturning Circulation. Upper water column reconstructions from the North and South Atlantic based on both planktonic foraminiferal faunal assemblage reconstructions and paired δ18O and Mg/Ca analysis provide insight into the abrupt changes in ocean dynamics across glacial cycles. Additionally with the use of neodymium isotope analysis, changes in the deep North Atlantic, likely the result of changes in the Meridional Overturning Circulation, have also been identified. Importantly, this data provides evidence that the timing of the resumption of the Atlantic Meridional Overturning Circulation following glacial conditions is integral in determining the extent of deglacial CO2 release. The work presented here suggests that rather than it being the magnitude of millennial scale events determining the magnitude of glacial-interglacial climate variability, instead, it is the relative timing of the millennial scale events that plays a role in controlling the magnitude of glacial-interglacial cyclicity.

In the present century, ecosystems across the globe will be subject to profound changes in climate. Forests are expected to be particularly sensitive to such change as the long life span of trees limits the potential for rapid adaptation. In order to preserve commercial viability and the essential ecosystem services provided by forests, there has been much interest in strategies for managing the adaptation of trees to their climatic environment. Climate Matching has emerged as one such strategy, whereby climate models are used to identify provenances – tree populations at a particular locality - with seed expected to be well adapted to the future conditions of a particular planting site. Debate continues about the feasibility and merit of this and other approaches, but it has yet to be demonstrated that the underlying assumptions of Climate Matching are valid for focal European tree species. Furthermore, a potentially major omission thus far has been consideration of how the Climate Matching strategy might influence associated organisms. Given the widely demonstrated bottom-up effects of foundation species genotype that have emerged from the field of community genetics, it is possible that planting seed of non-local provenance could effect forest organisms such as insect herbivores. In this thesis, I investigate the underlying assumptions of Climate Matching and its community level consequences using a model system of cynipid oak galls on Quercus petraea. Following a general introduction to Climate Matching and the study system, in Chapter 2 I use data from a provenance trial of Q. petraea in France to explore a central assumption of the Climate Matching strategy: that provenances of focal tree species show climate associated variation in adaptive phenotypic traits. In Chapter 3, I explore correlations between these phenotypic traits and the abundance, diversity, and community composition of an associated guild of specialist gall-inducing herbivores. Tree phenological traits in particular showed strong patterns of adaptation to climatic gradients, and influenced the abundance and community structure of galling species. However, as the response to non-local tree provenances was not strongly negative, it was considered unlikely that mixed planting of local and Climate Matched provenances would have sever impact on the gallwasp community. Having assessed the bottom-up effects of provenance phenotypic variation on the galling community, my ultimate aim is to extend analysis to include associated hymenopteran inquilines and parasitoids. However, interpretation of effects at this level is hindered by taxonomic uncertainty, with a growing appreciation that morpho-taxa may not represent independently evolving lineages (i.e. ‘true’ species). In Chapters 4 & 5 I therefore develop approaches for addressing taxonomic uncertainty with this ultimate aim in mind. In Chapter 4, I apply a DNA barcoding approach to parasitoid and inquiline specimens reared from the provenance trial, and compare taxa based on barcodes with those based on morphology to identify points of taxonomic uncertainty. I also investigate the extent to which networks based on morphological and molecular taxa support contrasting conclusions of network properties. In Chapter 5 I explore the potential for molecular based resolution of species level taxonomic error in a challenging group of parasitoids: the genus Cecidostiba. Beginning with a framework of single locus DNA barcoding, I use data from multiple nuclear loci to reveal the existence of cryptic species. Finally, in Chapter 6 I explore the practicalities of Climate Matching in light of my empirical results, and suggest fruitful avenues for further research.

Accelerated climate change and increasing climate variability caused by increasing anthropogenic greenhouse gas emissions is the single largest threat to the international goals of sustainable development, the Millennium Development Goals (MDGs) and disaster risk reduction. Global discourses recognise the need for effective and sustainable responses to produced climate risks. The risk types likely to occur are known, but only in broad terms. That they are produced by human action is accepted; but their scale, severity, longevity and frequency are not known. The challenge for policymakers is developing an effective framework within which sustainable responses can be formulated. Addressing the problems of produced risks requires a comprehensive approach to risk management to be effective. The mechanisms within the climate change, sustainable development and disaster risk reduction discourses are not sufficiently effective or integrated to respond to this challenge. Fundamental reform to current modes of risk reduction is needed, but this can only be achieved by a shift in the dominant perspective on formulating sustainable responses. This requires a shift to an enabling policy framework that encourages bottom-up resilient responses. Resilience is argued as a tool for policy development that can enhance adaptive capacity to current climate risks and shape energy policy to respond to mitigate future climate risks.

The mid- to late Miocene marks one of the last warm periods of the Neogene, before the descent in to the ice house climate of the late Pliocene and Pleistocene. The mid- to late Miocene climate was long overlooked, but receives increasingly more interest. This study used organic geochemical biomarker distributions and their isotopic signatures in sediment samples from the Japan Sea, recovered during IODP Expedition 346 and ODP Leg 127/128. It was the aim to reconstruct paleoceanographic and paleoclimatic changes over the last 18 Ma. Reconstructed sea surface temperatures revealed declining temperatures since the Mid-Miocene Climatic Optimum, possibly driven by a long-term decline in atmospheric CO2. Reconstructed pCO2 using alkenone paleobarometry, revealed a late Miocene decline from circa 520 μatm to 380 μatm. A decline in this range is required to trigger the onset of the late Miocene C4 plant expansion, which was also fully detected in the sedimentary record from the Japan Sea. Stable isotope analysis of n-alkanes (δD and δ13C) revealed an intensified summer monsoon from 5-3 Ma, which also led to a freshening of surface waters in the Japan Sea. Overall, the Japan Sea is a suitable study area for the late Miocene climate change.

This thesis presents new micro palaeontological and geochemical palaenvironmental proxy data through the late Eocene into the earliest Oligocene from a clay-rich succession from the US Gulf Coast. It is based on samples from the Yazoo Clay Formation, recovered in the Mossy Grove Core near Jackson, Mississippi. This represents an apparently continuous section of relatively uniform lithology, clay-rich deposits that host very well preserved assemblages of calcareous nannofossils, foraminifera and organic biomarkers. This thesis makes use of these to generate a detailed calcareous nannofossil taxonomy, high-resolution calcareous nannofossil assemblage data and coccolith-fraction bulk isotope data, as well as pilot planktonic foraminifera abundance and isotope records. It also develops pilot data for organic biomarkers that demonstrate the presence and utility of biomarker proxies for ancient sea surface temperatures within the Yazoo Clay. The results of this project characterize the late Eocene US Gulf Coast as a sub-tropical shelf sea environment that experiences several stages of sea surface cooling and increasing nutrient contents - potentially linked to sea level fall - in the late Eocene and earliest Oligocene. The most important finding of this project is evidence for cooling and major perturbations to the climate-carbon cycle significantly before the onset of the major phases of Antarctic glaciation.

There is major concern that river temperature changes driven by a changing climate and associated hydrological changes will have profound impacts on freshwater ecosystems. To identify the rivers most sensitive to change and implement effective strategies to mitigate high thermal extremes, this thesis aims to improve understanding of the influences of hydrometeorology and riparian landuse on river temperature dynamics, controls and processes within a UK context. Four studies are presented within a multi-scale research design which aimed to improve understanding of: (1) spatial patterns and inter-annual variability in the shape and magnitude of annual river temperature regimes across England and Wales, and regime sensitivity to air temperature and river basin properties, (2) the effects of riparian vegetation on water temperature under a range of hydrometeorological conditions, (2) the processes by which cool water refugia are produced beneath semi-natural, deciduous forest canopies, and (4) how minimal riparian planting can be used to produce thermal refugia in reaches of differing aspect and hydraulic characteristics. The aims are achieved by combining observational, statistical and deterministic modelling techniques. The outcomes of the research contribute significant new knowledge and tools for evidence based management of river and stream temperature under present and future climates.

- / -

Rapidly rising levels of atmospheric greenhouse gases including carbon dioxide and methane since the industrial revolution have drawn scientific attention to the importance of the global carbon cycle to the climate (Cubasch et al., 2013). Tropical peatlands, the majority of which are located in the Indonesian region, are a major source of uncertainty in the carbon cycle as the relationships between carbon accumulation and hydrological changes remain poorly understood (Hergoualc’h & Verchot, 2011, Page et al., 2011). An important driver of carbon emissions in tropical peatlands is fire, which in the Indonesian region is strongly influenced on interannual timescales by the El Niño/Southern Oscillation (ENSO). However, it is not clear how ENSO and fire have varied at decadal to centennial scales over the past two millennia. This thesis explores long term tropical hydrological variability and ENSO-like climate change from palaeorecords and their interactions with fire. Using a wide range of instrumental, proxy and model datasets and a novel reconstruction method, two separate reconstructions of long-term ENSO-like climate change are produced based on precipitation and temperature data. These show no evidence of a difference between the ENSO-like behaviour of precipitation and temperature. There is limited evidence for a difference in long-term ENSO-like state between the Medieval Climate Anomaly and the Little Ice Age. Reconstructions of hydrological variability and biomass burning in the Indonesian region suggest that precipitation and fire have been positively correlated over the past 2,000 years, which is contrary to the modern-day relationship on ENSO timescales. This throws up questions of long-term versus short-term interactions and feedbacks between fire, climate and vegetation. It is likely that anthropogenic activity in the Indonesian region has significantly altered the stability of the fire regime. Further research combining proxy data, climate and fire models, and using more robust statistical analysis is necessary to untangle the natural and anthropogenic driving factors at different time resolutions.

The potential impacts of climate change and environmental variability are already evident in most parts of the world, which is witnessing increasing temperature rates and prolonged flood or drought conditions that affect agriculture activities and nature-dependent livelihoods. This study was conducted in Mwanga District in the Kilimanjaro region of Tanzania to assess the nature and impacts of climate change and environmental variability on agriculture-dependent livelihoods and the adaptation strategies adopted by small-scale rural farmers. To attain its objective, the study employed a mixed methods approach in which both qualitative and quantitative techniques were used. The study shows that farmers are highly aware of their local environment and are conscious of the ways environmental changes affect their livelihoods. Farmers perceived that changes in climatic variables such as rainfall and temperature had occurred in their area over the period of three decades, and associated these changes with climate change and environmental variability. Farmers’ perceptions were confirmed by the evidence from rainfall and temperature data obtained from local and national weather stations, which showed that temperature and rainfall in the study area had become more variable over the past three decades. Farmers’ knowledge and perceptions of climate change vary depending on the location, age and gender of the respondents. The findings show that the farmers have limited understanding of the causes of climatic conditions and environmental variability, as some respondents associated climate change and environmental variability with social, cultural and religious factors. This study suggests that, despite the changing climatic conditions and environmental variability, farmers have developed and implemented a number of agriculture adaptation strategies that enable them to reduce their vulnerability to the changing conditions. The findings show that agriculture adaptation strategies employ both planned and autonomous adaptation strategies. However, the study shows that increasing drought conditions, rainfall variability, declining soil fertility and use of cheap farming technology are among the challenges that limit effective implementation of agriculture adaptation strategies. This study recommends further research on the varieties of drought-resilient crops, the development of small-scale irrigation schemes to reduce dependence on rain-fed agriculture, and the improvement of crop production in a given plot of land. In respect of the development of adaptation strategies, the study recommends the involvement of the local farmers and consideration of their knowledge and experience in the farming activities as well as the conditions of their local environment. Thus, the findings of this study may be helpful at various levels of decision making with regard to the development of climate change and environmental variability policies and strategies towards reducing farmers’ vulnerability to current and expected future changes.