The Cretaceous (~145–65 Ma) is widely regarded as a greenhouse period with warm, equable climates and elevated atmospheric CO2 relative to the modern. However, the earliest Cretaceous (Berriasian–Barremian; 145–125 Ma) is commonly characterised as a relatively colder “coolhouse” interval, typified by lower global temperatures than the mid-Cretaceous. Unfortunately, the lack of absolute sea surface temperature (SST) estimates prior to the Barremian has hampered efforts to definitively reconstruct Early Cretacous climate. Here, the TEX86 palaeotemperature proxy, for which a detailed review is provided, has been used to generate a 13 myr record of SST estimates for the Early Cretaceous, based on sediments from assorted deep-sea drilling sites. A consistent offset in the TEX86 ratio between transported mudstones and pelagic carbonates in the low-latitude marine sediments (DSDP Sites 603 and 534) has been identified, which may be linked to post-burial diagenesis or a difference in organic matter type between lithologies. Mindful of these apparent lithological effects on TEX86, only the pelagic sediments were used to subsequently reconstruct Early Cretaceous SSTs. These TEX86 records demonstrate both elevated SSTs (>27 ºC) at low and mid-latitudes relative to the modern, and the apparent stability of these high temperatures over long timescales. This lack of SST variation in the low-latitudes during the Valanginian positive carbon-isotope event (CIE; ~135–138 Ma), casts doubt on the warming-weathering feedback model put forward to explain this major perturbation. Additionally, new paired bulk organic (δ13Corg) and bulk carbonate (δ13Ccarb) carbon-isotope records from North Atlantic DSDP sites, have been used to reconstruct relative changes in pCO2 across the CIE. These observed fluctuations in Δ13C imply changes in carbon-cycling and a possible drawdown in CO2, due to excess organic matter burial associated with the CIE.

This thesis investigates climate variability during the Late Quaternary in northern Russia by analysis of subfossil Chironomid assemblages in lake sediment cores. The modern chironomid fauna was determined in surface-sediment samples from 94 lakes, located between 61° - 72°N and 52° - 131°E. The influence of chemical and physical environmental variables on the distribution and abundance of taxa was investigated using multivariate analysis and modelling of taxon responses. Chironomid distribution showed a statistically strong relationship to mean July air temperature and continentality, and inference models were developed to reconstruct these variables. Palaeoclimate reconstructions for the past 700 years are presented from lakes on the Putoran Plateau, western Siberia (68°N, 92°E). These reconstructions show that whilst July air temperatures have remained relatively stable over the last 50 years, continentality has declined, resulting in a longer growing season. The results also enabled the models to be validated against instrumental records. Reconstructions are then presented from lakes in north-east European Russia: Lake Kharinei, (67.36°N, 62.75°E) where sedimentation started approximately 12600 cal. yrs BP, and a mid-Holocene sequence (4500 – 6500 cal. yr BP) from VORK5 (67.86°N, 59.03°E). The chironomid-inferred reconstructions suggest the early Holocene was approximately 2°C warmer than present with a more continental climate. July air temperatures then declined but remained warmer than present until 6000 yrs BP. From approximately 6000 yrs BP, July temperatures declined and the climate became more continental, indicating a shift to short cool summers. Combined use of the July air temperature and continentality inference models enhances the explanatory power of the palaeoenvironmental reconstructions and helps to reconcile apparent discrepancies with other proxy records. The results improve our understanding of the nature and timing of climate change, such as the spatial extent of the Younger Dryas, in poorly-studied regions of northern Eurasia.

Bromate contamination over an area of more than 40 km2 in the Hertfordshire Chalk aquifer was first detected in 2000 and is the largest case of point-source groundwater contamination in the UK. Bromate is a possible human carcinogen, and a regulatory limit for drinking water of 10 μg l-1 had been implemented in the U.K. since 2003. Background concentrations of bromate in groundwater are believed to be effectively zero. In the affected area, bromate at concentrations of several 100 μg l-1 have forced the closure of a large public water supply source and restricted the use of seven other public supply boreholes up to 20 km from the contamination source. The source has been identified as a former industrial site which operated between 1955 and 1983. Residual contamination at the site provides a continuing source of bromate to groundwater. A range of conceptual scenarios for bromate mobilisation and release to groundwater have been developed and quantified based on interpretation of the available data, and constrained by interpolation of the observed concentrations. Analysis and interpretation of all available monitoring and investigation data throughout the catchment has revealed the influence of recharge, water level, and groundwater abstractions on bromate concentrations. These relationships, integrated with observations of the geology and hydrogeology of the area, support a conceptualisation of transport of bromate by dominantly double-porosity processes within the Vale of St. Albans area, which maintains a highly attenuated, stable contaminant distribution west of Hatfield. An extensive karst network related to the position of the Palaeogene overlap of the Chalk influences bromate transport to the east of Hatfield, dispersing bromate rapidly over large distances toward the Lea Valley. The revised conceptual understanding has enabled the development of a new interpretation of bromate transport within the catchment between 2000 and 2008. A new analytical network modelling approach has been developed to predict the long-term, largescale transport of bromate. The model simulates Fickian double-porosity diffusive exchange along interconnecting flow-lines, linked to rapid karst flow. The model is parameterised on the basis of single borehole dilution testing, catchment-scale natural gradient tracer testing, and literature derived values. The network model, combined with quantified bromate source terms, simulates bromate and bromide concentrations of the order of magnitude of those observed at locations within the Vale of St. Albans, and predicts bromate concentrations to remain above regulatory limits for around 200 years. This highlights the importance of double-porosity diffusion for the long-term evolution of contaminants at catchmentscale in the Chalk aquifer.

The integration of sedimentology, architectural element analysis and stratigraphy has been applied to characterise the complex depositional history of the Ainsa basin fill, and document the evolution of the proximal parts of eight, channelised deep-marine systems of the Hecho Group. The Eocene Ainsa basin provides an opportunity to research three-dimensional organisation through an entire deep-marine slope to proximal basin-floor fill, and records a range of depositional processes and sedimentary environments in a spatio-temporal framework. The Hecho Group can be divided into two tectono-sequences, TS-I and TS-II. TS-I accumulated during a period of strong flexural subsidence ahead of the main thrust front during a foredeep setting, whereas TS-II represents a more mature stage of basin development, characterised by anticlinal uplift as the basin became detached and evolved into a complex thrust-top basin. Four discrete systems and their constituent sandy sequences compose each tectono-sequence. The sequences comprising TS-I show very little lateral migration due to high basin-scale accommodation; however, westward lateral offset stacking is observed in the sandy sequences of TS-II due to the development of intrabasinal growth anticlines. These structurally controlled trends demonstrate that the timing of tectonic processes operated at frequencies consistent with the accumulation of the depositional systems. Depositional systems range between ~60–700 m thick, and were deposited in a number of deep-marine settings that include mid-slope canyons, lower-slope erosional channels and proximal basin-floor channel systems. Temporal variation in depositional style and architecture between systems reflects the tectonic regimes operating during the accumulation of the tectono-sequences. Alternatively, the 22 sandy sequences were controlled by the ~400 kyr Milankovitch frequency with higher-frequency orbital bands influencing the accumulation of channel complexes and channel fill elements. An important outcome of this study is the recognition of a complex hierarchical interaction between global climatic and tectonic drivers, operating at a variety of time scales to control the timing of coarse clastic sediment supply and the architectural styles of depositional systems.

The reemergence of tuberculosis (TB) throughout urban areas of the global North in the late 1970s and early 1980s prompted illuminating scholarship on the underpinnings of the disease’s modern resurgence. While valuable in their own right, these studies failed to elucidate fully the relationship between tuberculosis and public health infrastructure. In response, this thesis unpacks this complex relationship through a detailed case study of New York City, which experienced a profound resurgence of the disease between the late 1970s and its climax in the early 1990s. Principally, this work explores how changes in New York City’s public health infrastructure between 1970 and 2006 shaped the return of tuberculosis and influenced the City’s subsequent ability to control the disease. More specifically, the study examines if the strength of the public health infrastructure dictates the efficacy of TB control efforts; the pathways through which public health infrastructure and TB may be connected; and which public health infrastructure elements are most important to the City’s ability to keep the disease in check. An analytical framework developed by the Centers for Disease Control (CDC) is adopted and modified, thus organizing this research and atomizing public health infrastructure into four components: workforce capacity and competency, information and data systems, organizational capacity, and fiscal resources. This work charts and critically interrogates the role of these four elements in fluctuating New York City tuberculosis rates. An interdisciplinary approach is employed, which synthesizes insights gained from the medical sciences with analytic methods derived from medical geography, public health, history, and political science. Data comprise semi-structured interviews (with TB policy-makers, researchers, medical workers, and others), coupled with an analysis of public health legislative documents, newspaper articles, time-series tuberculosis rates, and newly uncovered staffing figures and fiscal resource data. Summarily, the thesis finds that public health infrastructure played a decisive role in New York City’s resurgence of tuberculosis and the City’s subsequent ability to control the disease. Each of the four facets of public health infrastructure proved integral to the success of the TB program. Additionally, interconnections between the four elements are revealed, highlighting the potentially far-reaching effects that changes to one element can cause. The findings of this thesis not only demystify the role of public health infrastructure as it relates to TB, but also bolster the work of past researchers who have proposed moving past bio-medically centric perspectives to include socio-geographic elements in tuberculosis initiatives.

This study investigates the relative importance of eutrophication and connectivity (dispersal) in structuring macrophyte and invertebrate lake assemblages across spatial and temporal scales in the Upper Lough Erne (ULE) system, Northern Ireland. Riverine systems and their associated flood-plains and lakes comprise dynamic diverse landscapes in which water flow plays a key role in affecting connectivity. However, as for many other freshwater systems, their ecological integrity is threatened by eutrophication and hydrological alteration. Eutrophication results in a shift from primarily benthic to primarily pelagic primary production and reductions in species diversity, while flow regulation often reduces water level fluctuation and hydrological connectivity in linked riverine systems. Low water levels promote isolation between areas and increases the importance of local driving forces (e.g. eutrophication). Conversely, enhanced water flow and flooding events promote connectivity in systems thus potentially increasing local diversity and homogenising habitats through the exchange of species. Therefore, connectivity may help to override the local effects of eutrophication. Attempts at testing the above ideas are rare and typically involve the examination of current community patterns using space for time substitution. However, biological community responses to eutrophication and changes in hydrological connectivity may involve lags, historical contingency, and may be manifested over intergenerational timescales (10s -100s of years), rendering modern studies less than satisfactory for building an understanding of processes that drive community structure and effect change. By combining contemporary and palaeolimnological data this study demonstrates that the ULE system is far from its pre-disturbance state as an oligotrophic-mesotrophic system. Furthermore, contemporary and palaeo-data suggest there has been a strong interaction between eutrophication and hydrological change, which influences the distributions and abundances of representative taxa in the ULE system. Thus, while eutrophication has promoted a decrease in compositional heterogeneity of organisms and has exerted a homogenising effect over time, connectivity has buffered the effects of eutrophication helping to maintain local diversity via re-introductions.

The natural physical and biological states of rivers have been altered increasingly by habitat modification and pollution. Successful river restoration is therefore critical for mitigating impacts on river ecosystems. However, restorations are typically constrained to small patches within individual rivers and often lack standardised or adequate monitoring and assessment. This has led to poor diagnosis of both the “problem” and the effectiveness of the “solution” as comparable data are rare. In a literature review and meta-analysis I identified three areas limiting the development of river restoration as a science: 1) prior monitoring to determine the chief ecological constraint was not common practice; 2) the primary driver for restoration was ecological (e.g. an increase in biodiversity) but monitoring was often inadequate to detect ecological response to restoration or not undertaken; 3) target species (e.g. brown trout) or assemblages (e.g. invertebrates) are often the focus of monitoring, but it is not yet clear how habitat restoration affects more complex organisational levels (e.g. food webs, community mass-abundance scaling, ecosystems) and how this regulates populations of target taxa. These three areas were the focus of the subsequent data chapters. Conserving naturally fallen large woody debris (LWD) in rivers and felling trees to mimic LWD is one current vogue in river restoration-management thought to increase habitat diversity and positively correlate with increasing invertebrate biodiversity. Invertebrates were sampled between reaches with and without LWD and across a nutrient gradient, which included 19 base-rich streams, to test whether the chief determinant of invertebrate density and composition was either habitat quality or water quality. Five chalk streams were sampled before and after at control, reference (i.e. natural LWD) and treatment (i.e. felled) sites to test for a general restoration effect. A combination of spatially nested scales (i.e. stream-reach-habitat), invertebrate structural measures (e.g. α-diversity, abundance and biomass) and food web metrics (e.g. trivariate analysis) have been tested so that the ecological impact of restoration could be quantified. Large-scale pressures, namely fine sediment and orthophosphate (and their potential covariates) were impacting invertebrate community structure. Quantitative estimates of abundance and biomass increased within LWD habitat and reference reaches respectively. Evenness, α- and β-diversity were not significantly higher in restored LWD. Community mass-abundance scaling was not affected by either experimental or natural LWD. This implies that LWD is having an effect, but this is restricted to the invertebrate community. Many of the patterns decribed here would not have been detected using standard protocol methods (e.g. invertebrate kick net sampling). This highlights the need for the development of standardised biomonitoring practices that are able to diagnose root causes of degradation and demonstrate ecological recovery following restoration. If the aim of restoration is to predictably increase fishes biomass or biodiversity, for example, I anticipate larger-scale restorations, which address multiple stressors, and longer-term monitoring will be needed.

Southern Africa is characterised by a high degree of rainfall variability affecting agriculture among other sectors. The focus of this study is to investigate such variability and to identify stable relationships with its potential drivers in the climate system. These relationships are used as the basis for the statistical downscaling of climate model (GCM) outputs. From the simulated rainfall, indices representative of growing season characteristics are computed with the final purpose of studying the implications on maize cropping under a future climate change scenario. The analysis uses generalized linear models (GLMs), which allow the investigation of the relationships between different components of the climate system (geographical and climatic drivers) simultaneously. Initially, the effects of various climate indicators upon monthly regional (for all southern Africa) precipitation occurrences and amounts are characterised. Six climate factors are found to drive part of the rainfall variability in the region and their modelled effect upon rainfall occurrences and amounts agrees broadly with previous studies. Among the retained indices, relative humidity and El Niño accounted for the highest degree of explained variability. The location and intensity of the jet stream is also found to have a statistically significant and physically meaningful effect upon rainfall variability. Although effective for the analysis of monthly regional precipitation, and used to investigate future regional projections, the models do not perform adequately at more local spatial scales such as station locations or few km grids. The same methodology is, therefore, applied to characterise daily precipitation variability at multiple locations within a smaller region. The small scale statistical models capture adequately the seasonal and annual rainfall structure in the area. Indeed, the observations can not be distinguished from the simulated time series. However, the simulated rainfall values tend to be slightly too high throughout the seasons, possibly due to the spatial correlation structure not completely appropriate for such a complex region. From the simulated rainfall sequences, seven growing season indices (including the onset and length of the growing season, proportion of rainy days and total precipitation during the growing season) are derived and their projected change investigated under a climate change scenario. There is little consensus between the 18 selected GCMs, regarding changes in growing season indices between two investigated periods in the 20th and 21st centuries. For the next couple of decades the dominant source of variation in the indices appears to be the natural rainfall variability. Such information should therefore be taken into account when planning adaptation and mitigation strategies. The research presented here emerges as the first comprehensive assessment of different climatic factors linked to southern Africa rainfall variability as well as the first attempt to evaluate the GLMs suitability for the generation of rainfall sequences for agricultural impact studies.

The sediment flux to a basin is controlled by a complex combination of tectonics, climatic variability and stochastic events, thereby creating a cryptic geological record. Deconvolving the effects of individual factors controlling sedimentation can often be challenging, due to a variety of reasons including a lack of outcrop, a poor understanding of the regional tectonic framework and insufficiently detailed geological maps. However, many of these deficiencies can be overcome within the deep-marine Ainsa basin, South Central Pyrenees where detailed research over the last decade has provided an extensive knowledge base. The Ainsa basin comprises ~4 km of middle Eocene deep-marine sediments. Basin stratigraphy consists of a succession of ~25 discrete sandy submarine fans and inter-fan deposits belonging to the Hecho Group. Recently it has been hypothesised that the supply of coarse-clastic sediment to the basin was paced by orbitally induced climate and/or sea level variability, whilst tectonics controlled the locus of deposition. This hypothesis is tested within the Upper Hecho Group using a refined basin age model and the creation of floating orbital time scales between submarine fans. Using calcareous nannofossil and large shallow benthic foraminifera, deposition of the Upper Hecho Group took place over a 6.0-8.3 Myr period between ~40.5-48.4 Ma, giving an average sediment accumulation rate (SAR) of 43.2±10.5 cm/kyr. Stratigraphic time series analyses of inter-fan fine-grained sediments indicate the presence of short eccentricity, obliquity and precession Milankovitch cycles. These floating time scales provide average SARs of 36, 28 and 25-33 cm/kyr for the Banaston, Ainsa and Guaso systems respectively. Applying these age models to the three systems suggest that submarine fan deposition potentially corresponds to specific eccentricity minima. As in the Pleistocene, such Milankovitch forcing could be linked with ephemeral glacio-eustatic low-stand conditions, associated with increased coarse sediment flux to the deep-marine Ainsa basin.

The Marine Isotope Stage (MIS) 11 interglacial has been considered to be analogous to the Holocene because of a similarity in the conditions of orbital forcing (low eccentricity and precession), but the role of climatic feedback processes linking orbital drivers to climate signatures is not well understood. This thesis examines palaeoceanographic sediment records from the South East Atlantic Ocean covering the last 50,000 years and the period encompassing MIS-11 (335-475 ka) to better understand how different facets of the climate systems were linked together in these two periods of low orbital eccentricity. It is shown that palaeoceanographic changes were closely linked to variations in Northern Hemisphere ice-volume via changes in the mode of the Atlantic Meridional Overturning Circulation (MOC). When ice-volume was high, a weaker MOC drove regional sea-surface warming and salinity increases; it also produced corresponding changes in the South African monsoon system by displacing the southern zone of Hadley Cell circulation over the study region. The interlinked nature of these connections makes it difficult to establish an ultimate driver; for example, a precession-pacing of monsoon and sea-surface temperature proxies during late MIS-11 could have been induced either by local (sub-tropical) insolation forcing or by precession-modulation of ice-sheet margins. A direct comparison between identical palaeoclimate datasets for the last 50,000 years and the MIS-11 interglacial suggests that the long duration of MIS-11 compared to subsequent interglacials may have arisen through a strengthening of oceanic heat transport during a period of high obliquity approximately 15,000 years into the interglaical optimum. Although several Atlantic Ocean records support this contention, refining this model will require the aquisition of more geographically diverse records with more precise chronological constraints than re currently available from the global ocean.