Piston cores from the continental slope of Spitsbergen were analysed using micropalaeontological, stable isotope and sedimentological techniques. The attempted to determine the palaeoceanographic history of the Norwegian-Greenland Sea, and the glacial history of the Spitsbergen ice cap over the last 200ka. The chronostratigraphy of the cores was determined using stable oxygen isotope analysis of planktic foraminifera. These records were correlated to a composite benthic foraminiferal δ<SUP>18 </SUP>O stratigraphy from the Norwegian Sea (Duplessy <I>et al</I>., 1988), as well as to other radiometrically dated records from the Norwegian-Greenland Sea to produce an age model. X-ray photographs were used to assess the ice rafted detrital (IRD) content of the cores, dropstones were counted from these x-rays. IRD content was used as a measure of calving rate from the outlet glaciers of the Spitsbergen ice cap. From these counts and the δ<SUP>18</SUP>O record a model of relative advance and retreat of the Spitsbergen ice cap was produced. It was found that moisture supply was the over riding factor controlling dropstone input to the Spitsbergen margin. Periods of high moisture supply led to rapid ice advance and hence high calving rate producing peaks in IRD. Peak IRD events were found during interstadial periods, such as substage 5c, as well as during glacial periods when ice was at the shelf edge, and during deglacial periods. Absolute abundance counts of foraminifera suggest glacial periods had a much higher productivity than interglacial periods. This strongly suggests the Spitsbergen margin was at least seasonally ice free during glacial periods. Melting events were recognised from the δ<SUP>18</SUP>O record of planktic and benthic foraminifera. The influx of North Atlantic surface waters during deglaciation has been identified from the δ<SUP>18</SUP>O record and the faunal assemblages. The initial influx of North Atlantic water during the last deglaciation took place at about 12.3ka, disappeared at 12ka, was present again from 10.5 to 10.3ka, and then from 10ka onwards.

In this thesis I propose a framework to characterise the size and spacing of refugia in steep streams, based on a geomorphological classification of the streambed. Focusing primarily on four aspects, the physical attributes, particle movement and invertebrate assemblage associated with each morphological patch, and the hydraulic influence on invertebrate drift and settlement, I test the response of invertebrates to floods of different magnitudes on each patch type, Loose, Degraded and Structured. Patch type differed in sedimentological character. A strong association of invertebrate assemblage with patch type was detected. Distribution of some species was restricted to Structured patches. Contribution of Structured patches to invertebrate abundance is much greater than their areal extent suggests. I hypothesised that the hydraulic environment of a Structured patch, step, promotes drift settlement. <i>Baetis rhodani </i>and <i>Simulium </i>sp show no change in drift and settlement over a step, despite a large increase in velocity and turbulence. <i>Hydropsyche pellicudula </i>and <i>Ecdyonurus torrentis </i>drift further. The hydraulic environment of the step constrained the settlement and distribution of <i>E. torrentis. </i>The difference in drift response between the mayfly species is due to a complex interaction between behaviour and hydraulic environment. I contend this is driven by body morphology. A net loss occurred on all patch types during after flooding, although the assemblage on Structured patches was more resistant. Species-specific response was complex, and not related to particle movement. Loose patches, despite complete mobilisation, accumulated individuals, but this varied between flood events. Particle movement was only weakly linked with the processes responsible for the post-flood pattern of invertebrates. The patch classification does not discriminate on the basis of these processes, and may not provide a useable framework to characterise size and spacing of refugia in steep streams.

Instrumentation to monitor soil water chemistry and soil hydrology was installed within a sub-catchment of Loch Bradan water supply reservoir in SW Scotland. Following the dry summer of 1995, raw water samples from this reservoir contained Mn concentrations far greater than the EC standard. This increase in reservoir water Mn concentration is believed to have resulted from processes associated with the rewetting of exceptionally dry catchment soils. For this study, six field sites were chosen to allow comparison of the effect that major soil types (peat, peaty gley, peaty podzol) and land uses (Sitka spruce plantation, moorland) within the catchment had on Mn mobilisation upon rewetting following a dry period. Soil water and stream water samples were collected over an 18-month period from April 1999 to October 2000. Additional data were also collected concerning Mn concentration in soil, stream sediment, precipitation and vegetation samples in order to characterise potential Mn sources within the catchment. Results from the field monitoring programme indicate that land use did not influence soil water Mn concentrations and that there was no seasonal peak in soil water Mn concentration from any of the instrumented soil horizons. However, the absence of extended drying conditions during the monitoring period may explain this lack of anticipated seasonal Mn mobilisation, as it is possible that soils did not dry out to the extent where Mn mobilisation would occur. Laboratory experiments were undertaken to examine the potential of catchment soils to release Mn following prolonged drying conditions not encountered during the field monitoring programme. Soil cores were collected from the same major soil horizons and land uses that were instrumented during the field monitoring programme and allowed to dry under controlled conditions before rewetting. Soil water chemistry data from these experiments indicates that Mn mobilisation occurred within specific horizons, containing both mineral and organic matter, rather than from all soil horizons.

Examination of Landsat imagery of the bed of the last North American (Laurentide) Ice Sheet has revealed a previously undetected pattern of glacial streamlining. Superimposed, cross-cutting, sub-parallel sets of glacial lineations of the order of 2-50 km in length are found to be widespread. They are assumed to reflect successive phases of sustained ice flow, revealing major changes in the geometry of ice sheet flow. Using six scales of remotely sensed imagery to map the lineations, and establish their cross-cutting relationships, permits relative ages of continent-wide ice flow phases to be determined. Comparison of the principal ice flow sets with established stratigraphies suggests that the inferred shifts in ice flow occurred during the last (Wis-consinan) glacial cycle. The configuration of ice flow indicates that ice divides migrated by the order of a thousand kilometres. The evolution of the Laurentide Ice Sheet through the Wisconsinan is reconstructed. There is evidence that during the Early Wisconsinan, ice sheet formation in Keewatin may have been independent of that in Labrador-Quebec, and that these two ice masses coalesced to form a major Early Wisconsinan ice sheet. The palaeo-ice flow evidence indicates that this ice sheet configuration consisted of a trans-Laurentide divide aligned NW-SE across Hudson Bay. Subsequently, the western sector decayed whilst the eastern dome remained stable (Middle Wisconsinan). An ice dispersal centre in the west reformed and fused with the eastern ice mass to form the Late Wisconsinan Ice Sheet. Decay of this ice sheet is well known from existing analyses. The probability of strong coupling between ice sheet topography and atmospheric circulation suggests that the major changes in Laurenlidc Ice Sheet geometry must have been associated with large-scale atmospheric circulation changes. The corollary is that the high mobility of mid-latitude ice sheets may help explain the non-linear response of glacial climates to the insolation changes produced by external forcing.

A new detailed gravity and topographic dataset covering the Qattara Depression in Western Egypt is studied, as this topographic feature is thought to have formed by erosion and, as such, should not be associated with internal loads. A comparison of the results from this data is made with those from other point datasets for Kenya, southern UK, and Greece and the Aegean, as well as with admittance and coherence estimates determined by other workers for areas in North America, Siberia, India and Australia. Two different model variations are tested. The first involves loading at the surface and at the base of the crust, while the second allows internal loads to be positioned at much shallower levels in the crust. When the model explicitly places internal loads at the base of the crust, only the data from Egypt and Kenya can be fitted at an acceptable level and are associated with elastic thickness estimates of 19 and 23 km respectively. If loading is allowed to occur at shallower levels, then the fit between the models and the data is improved in all instances, but still only gives an acceptable fit to the Egyptian and Kenyan data. Models which include near surface loading are also correlated with elastic thickness estimates (18 km for Egypt and 14 km for Kenya). Possible explanations of the failure of the model to fit other datasets are given. The best-fit models to the Qattara data invoke a non-zero ratio of internal to external loading, indicating that contrary to the initial hypothsis, this region is affected by subsurface loads but that these may be upper crustal density variations and not necessarily related to underplating at the Moho. This analysis shows that the joint use of the admittance and coherence, when correctly applied, yields the small values of the elastic thickness of the continental lithosphere expected from thermal and seismogenic studies. Furthermore, it shows how the inclusion of near surface loading successfully models the short wavelength gravity variations, previously thought to invalidate the coherence method.

Riffle-pool sequences in gravel-bed rivers provide the template for a number ol fish and invertebrate habitats and their morphological complexities create spatial hydraulic variation over the flow regime that is logistically impossible to measure completely in the field. The mechanisms responsible for riffle-pool maintenance are unclear and despite contributions from many researchers, no conclusive explanation has been developed. There is much debate over the popular velocity reversal hypothesis, where it is generally acknowledged that at low flows the maximum bed shear stress in the channel occurs across the riffle, but at higher less frequent flows areas of maximum bed shear stress can switch to the pool. Sediment interactions control fluvial forms and processes, as such specific analysis of spatial hydraulic patterns in boundary shear stress are required to aid understanding of riffle-pool system behaviour and gain the insight in to the transport capacity that can scour sediment deposited in the predominandy low energy, lower competence pools and thereby provide a mechanism for maintenance ol riffle-pool morphology. Terrestrial LiDAR captures reach scale topographical data to yield a 0.02 m digital elevation model and provides input to three-dimensional computation fluid dynamics software (CFD), where the spatial distribution and area! extent of bed shear stress, surface flow velocities and potential sediment entrainment over the discharge range are investigated for a 188 m gravel-bed reach of Kingsdale Beck, UK. With increasing discharge, spatial distributions of shear stress are revealed, which, until now, have been largely overlooked using previous cross-sectional analysis, highlighting the influence of pool tails as discharge rises. Results show some agreement with past literature on reversal with higher shear stress zones associated with riffles. As discharge increases pools are shown to become more dominant, however the existence of multiple reversals shows that while reversals do occur, they may be short-lived and marginal in magnitude, providing inherent problems with current theory. The overall temporal maintenance of the riffle-pool sequence is more explicitly linked with the long term flow regime, with the flows responsible for moving pool sediment closely confined to near bankfull discharge. Results indicate that potential to entrain sediment from pools occurs only from 85% bankfull. Discharges less than 85% bankfull are suggested to be responsible for maintaining the riffle-pool morphology through flow routing and bed shear stress reversal, however, they are not capable of potentially entraining sediment from the deepest pools. Combined CFD simulations with temporal discharge dominance derived from triangulated rainfall and the Revitalised FSR/FEH rainfallrunoff methodology, highlight the role of rarer more extreme flows in terms of explaining sediment entrainment from pools. At high flow Kingsdale Beck creates a competent sub channel, where a continuous sinuous channel with higher levels of bed shear stress develop with embryonic lateral bars, increasing potential sediment entrainment following the line of thalweg. As flow velocity decreases conventional rifflepool hydromorphology re-establishes and provides a new proposal for the maintenance of a riffle-pool sequence.

No description available.

No description available.

No description available.

No description available.