The formation of sea ice in the presence of turbulence was studied using data from drifting buoy deployments and ice sampling in the Weddell Sea during April 2000. The study sought to improve understanding of pancake ice in terms of dynamics, heat fluxes, ice growth rates and mechanisms. Ice motion at high frequencies was examined using GPS buoy positions at a 20-minute sampling interval. Relative motions of the buoy array were characterised by a marked oscillation at the highest frequencies, with an RMS value two orders of magnitude higher than previously seen in the Weddell Sea. This motion ceased overnight as the pancakes consolidated. Wave forcing, either surface gravity or internal, was postulated as the cause. The oscillation was found to significantly influence the proportions of pancake and frazil ice, though the nature of the ice cover meant that ice production rates were unaffected, in contrast to the enhanced growth this would imply for congelation ice. Momentum transfer parameters were found to be similar to those found for the Greenland Sea Odden ice tongue. Pancakes were found to be dominantly thickened by over-topping of the surrounding frazil ice crystals, termed ‘scavenging’, and gave rise to distinct morphologies, which were classified. A physical model was developed to describe the evolution of the pancake ice cover to consolidation. Ice production in the pancake/frazil process was found to proceed at approximately double the rate of the equivalent congelation ice cover, or 0.58 times the limiting free-surface frazil production. It was suggested that the discrepancy will seriously impact largescale modelling attempts to simulate heat and momentum fluxes between the ocean and atmosphere, as well as salt rejection and subsequent water mass modification, though it is acknowledged that further field measurements are required to place some currently empirical parameters into a physical context.

Blooms of toxic, or otherwise harmful phytoplankton species are known to occur in eastern boundary upwelling systems, coincident with the relaxation of upwelling in late summer/autumn. Field studies were carried out in 3 consecutive summers (March/April 06-08) in the Benguela and in the autumn (Sept 06) and summer (June 07) in the Iberian upwelling system (Ría de Vigo), with the aim of identifying common nitrogen nutrition strategies of HAB species that may allow them to succeed in upwelling systems. Two summer field studies were also carried out in the Fal Estuary (UK) to identify possible differences between a UK estuary and these upwelling systems. In the Benguela, three toxic phytoplankton species were dominant under different nutrient conditions. Pseudo-nitzschia spp. were abundant during a period of strong upwelling and high NO3-, peaking during short periods of wind relaxation. During these periods, a switch from high nitrate uptake [(NO3 -)] to regenerated nitrogen uptake [(NH4 +) and (urea)] occurred, with ƒratios dropping from 0.79 to 0.12. Alexandrium catenella bloomed during a period of upwelling, displaying high (NO3-) and ƒ-ratios up to 0.87. Dinophysis acuminata dominated when NO3 - concentrations were <0.5 μmol l-1 and ƒ-ratios <0.1 in 2007, although in 2008 it formed a subsurface maximum, often associated with high NO3- concentrations. Nutrient uptake kinetics showed that Pseudo-nitzschia spp. displayed the highest maximum specific uptake rates (max). D. acuminata displayed the highest affinity for NH4+, as shown by its values (slope of the nutrient uptake vs. concentration curve). Thus, A. catenella was adapted to utilising high NO3- concentrations during upwelling pulses, whereas both Pseudo-nitzschia and D. acuminata were able to acclimate to both high and low NO3- concentrations during the upwelling/relaxation cycles. In the Ría de Vigo, warm water from the stratified shelf entered the ría and downwelled in September, resulting in a well-mixed water column. The phytoplankton assemblage, dominated by Ceratium spp., Dinophysis acuminata and Gymnodinium catenatum, appeared to be advected in to the ría. Nitrate concentrations were consistently low, whereas NH4+ concentrations increased towards the head of the ría and with depth. The phytoplankton community was dependent on regenerated nitrogen, with ƒ-ratios <0.2. In contrast, positive circulation in June resulted in strong vertical gradients in temperature, salinity and nutrients and a community dominated by diatoms. Nitrate and NH4+ were depleted in surface waters although uptake rates were higher than in September, as were the ƒ-ratios (0.1-0.3). In both systems, upwelling winds favoured diatoms, although they were able to utilise regenerated nitrogen when NO3- was depleted, whereas upwelling relaxation created favourable conditions for HAB development. Dinophysis spp. occurred in both systems and were able to grow on recycled nitrogen in the absence of NO3-. The Benguela showed high variability in the selection of particular HAB species, perhaps due to greater variability in upwelling-downwelling cycles. In the Ría de Vigo, the occurrence of downwelling and associated nutrient conditions leading to blooms of Dinophysis spp. and Gymnodinium catenatum seems more predictable. In the Fal Estuary, Alexandrium spp. was favoured by low irradiance and the combination of strong stratification and high nutrient concentrations, and its growth was sustained predominantly by NH4+. A. minutum strains isolated from both upwelling systems and from a UK Lagoon all displayed higher max for NH4+ relative to NO3- but higher growth rates on the latter. This was consistent with field results from all 3 regions, suggesting that the upwelling systems did not display a different order of nitrogen preference, although they did display a higher affinity for NO3-.

Subtropical gyres represent the single largest oceanic biome and may be very important for carbon cycling on account of their areal extent, yet our understanding of how these regions operate biologically is lacking. To address this issue measurements of NO3-, NH4+ and urea uptake were made using the 15N technique on 2 Atlantic Meridional Transect (AMT) cruises between 50oS and 50oN in May-June 2003 (AMT12) and April-June 2004 (AMT14). Distinct vertical profiles in the uptake of each nutrient were evident with urea uptake primarily confined to tropical and subtropical surface waters, NH4+ uptake to surface and intermediate waters and NO3- uptake peaking at depth leading to the observation of deep maxima in NO3- uptake within the subtropical gyres often coincident with the nutricline and with the deep chlorophyll maximum. Comparisons between the two cruises reveal interannual changes including an increase in the uptake rates of all three nutrients during AMT14, particularly NO3- which is driven by an increase in ambient NO3- concentrations. Several other parameters (e.g. chlorophyll, 14C carbon fixation, other nutrients) were also higher during AMT14 compared to AMT12 suggesting gyre scale variation. Estimates of NO3- based new production, obtained via a stoichiometric (Redfield) ratio, for the two cruises reveal relatively consistent rates within the tropics and subtropics within each cruise but higher rates of new production during AMT14. A greater proportion of new production was observed to occur in the lower 15% of the euphotic zone than in the surface layer reflecting the position and influence of the nutricline. A northward increase in new production in the lower regions of the euphotic zone was also observed which is hypothesized to represent a seasonal signal implying a degree of seasonality in the behaviour of the deep chlorophyll maximum and of deep NO3- uptake. The historical AMT database (AMT cruises 1-14) has been used to further examine the deep chlorophyll maximum. The observations of high NO3- uptake at depth, in excess of 14C based productivity nitrogen requirements and new production estimates exceeding total production rates suggests that nitrogen and carbon uptake may be decoupled at depth or that the measured NO3- uptake may not in fact represent new production. Integrated rates of new production were also occasionally found to locally exceed integrated community gross production and community respiration. Comparisons between new production and 234Th export production during AMT14 reveal that the two measurements are not equal, most likely on account of the different time scales of the two techniques, but comparisons with other parameters (carbon fixation, gross and net production and respiration) reveal that none of the parameters produce exact answers although similar patterns and magnitudes of carbon flux can be obtained. This would suggest that local decoupling of all rate processes is likely.

Obtaining information regarding the resistivity structure of the subsurface from marine CSEM data involves complex processes. 1D and 2D forward and inverse modelling are currently the standard approaches used to produce geoelectrical models, with 3D inversion fast becoming a realizable method. However, these methods are time consuming, require expert knowledge to produce reliable results, and suffer from the non-uniqueness of the EM problem. There is therefore considerable scope for developing imaging techniques for marine CSEM data that do not require lengthy, time consuming computations, but make use of entire datasets. These could provide a “first look” for possible structural information conveyed by the data, and may provide starting points or other constraints for inversion. In this thesis, a number of different imaging techniques for marine CSEM data are assessed, with particular reference to applications in hydrocarbon exploration. T-X and F-K imaging are widely used seismic reflection processing techniques that can be applied to CSEM data. Features produced in the T-X and F-K domains by 1D subsurface resistivity structures are investigated. The dip of an arrival corresponding to a subsurface resistive feature is found to depend on its resistivity, with reduction in resistivity producing steeper dipping events. The separation of arrivals according to their dips in the T-X domain is used as a basis for the attempted separation of the airwave, by filtering in the F-K domain. However, this does not prove to be useful. Secondly, in a adaptation of the F-K migration method used in seismic processing, EM migration is investigated, following the approach by (Tompkins, 2004b). The results of the migration method are compared and contrasted to a 1D smooth inversion algorithm. It is found that the migration is mostly dependent on the conductivity contrast across a geoelectrical boundary, whereas the inversion recovers the resistivity thickness product (transverse resistance). Hence, EM migration is a viable alternative to inversion and usefully complements it in regions of large conductivity contrasts. Normalized ElectroMagnetic Imaging (NEMI) extends the standard approach of normalizing the recorded electric field data by a 1D background model, to identify large lateral resistivity variations over a survey area. This is achieved by firstly sorting the data based on sensitivity to the target layer, and then distributing the normalized anomaly in the horizontal plane between the source and receiver using a simple quasitomographical approach. In some scenarios this provides a reasonable estimation of the lateral extent of a 3D resistive body buried in a conductive background. Lastly, Apparent Resistivity Imaging (ARI) is adapted for the use with the marine CSEM method. This generates pseudo-sections in which offsets are mapped into apparent depths. This study shows that whilst vertical resolution of resistive bodies is poor, lateral resolution is high and provides a good estimate of the true extent of a target body. Apparent resistivity pseudo-sections therefore provide a very effective means of “first look” imaging and assessment of marine CSEM data.

A theoretical and laboratory Investigation of rotating gravity currents and channel flows is presented. The study is applicable to buoyancy driven flows through straits, mid ocean ridge valleys and fracture zones, and intermittent gravity currents. In the theoretical study two extensions are achieved to the energy conserving theory ofHacker (1996). Hacker considered three flow geometries, case A - weak rotation, case B - intermediate rotation and case C - strong rotation. Firstly, the theory is extended to include dissipation. This is achieved in a similar manner to that used by Benjamin (1968) to include energy loss in the non-rotating gravity current theory. The governing equations and numerical solutions for the three flow geometries are presented. For shallow currents the energy loss theory predicts that the Froude number tends to 2* irrespective of the rotation rate. For deeper currents the Froude number increases with rotation. The second extension to the energy conserving theory is the inclusion of an upstream potential vorticity boundary condition in the current. The approach taken is based on a method used by van Heijst (1985). The governing equations and preliminary solutions for each case are derived. The potential vorticity theory provides an insight into the circulation that develops within the current. However, varying the pre-set potential vorticity in the source region does not appear to have a significant effect upon the front speed and the other principle variables. In the laboratory investigation the effects of fractional depth and rotation rate on the velocity and other parameters which characterise the flow are quantified. For weakly rotating currents, w/R < 0.7 (where w is the width of the channel and R the Rossby radius), the measured front speed is in fair agreement with the energy loss and potential vorticity theories. At higher rotation rates the front speed is lower than predicted. However, the theories assume that the fluid is inviscid, the no-slip condition is not applied at the boundary, potential vorticity is conserved and that energy loss is uniform across the channel. The theory does not include factors such as the enhanced vertical mixing and the development of a geostrophic eddy. These are associated with strong rotation rates and could account for the divergence of the experimental results from the theory.

Kuwait Bay presents a unique ecosystem and a significant nursery ground for many fish and shrimp species. In the last three decades, the bay has been under pressure from urbanization as well as from development from the entire region of the Arabian Gulf. Seawater temperature has an important impact on the marine environment. The aim of this study is to evaluate the drivers of seawater temperature in Kuwait Bay over the last two decades. Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat satellites images were used to describe the spatial and temporal distribution of sea surface temperature (SST) in the Arabian Gulf and Kuwait Bay. Hourly temperature measurements collected during the winter, 2007 and summer, 2008 were used to define the vertical temperature gradient in the water column, sea-bed and the intertidal flats of Kuwait Bay. Advanced Very High Resolution Radiometric (AVHRR) satellite data collected between 1985 and 2007 was also used to study the trends and drivers of increasing SST in Kuwait Bay over the last two decades. Satellite imaging showed that generally SST of Kuwait Bay is higher in the south than in the north. Highest SST was recorded near local human activities especially close to power and desalination plants due to thermal discharges. The field measurements showed that the water temperature is well-mixed in the bay. Seawater temperature reached 37 ºC in summer and dropped to 15.2 ºC in winter. Measurements beneath the sea-bed and within the intertidal flats showed that the temperature increased with depth in winter and decreased in summer. In winter, the sea-bed and intertidal flats are a source of heat to the water column, during summer the opposite is true. AVHRR data showed that the seawater temperature increased in Kuwait Bay by 0.62 (± 0.01)ºC/decade in the last two decades. This trend is three times greater than the global average. The defined trends were substantiated by routine in situ monthly measurements of SST made by the EPA in the bay, and were also similar in pattern and trend to air temperature recorded at Kuwait airport. Temperature trends have been affected by drivers, conveniently sub-divided into global (which contributes 37% of the change), regional (which contributes 50% of the change) and local (which contributes 13% of the change). SST measurements showed peaks in summer temperature coincident with El Niño events in 1998 and 2003. The measurements also showed a relatively-low summertime peak during 1991 in the aftermath of Iraqi invasion of Kuwait due to atmospheric dimming brought about by dense smoke that persisted in the region for most of that year. The long term trend also showed a drop in temperature after 2004 as a result of increasing dust storm frequency in the region. Air temperature was found to be the most dominant driver of seawater temperature in Kuwait Bay and operates at a regional scale. At the local scale the seawater temperature of Kuwait Bay is influenced by intertidal flat exposure time and the thermal discharge from power and desalination plants.

During a three year occupation of Station Mike (66°N 2°E), the Norwegian Ocean Weather Ship Polarfront was equipped with a range of meteorological and seastate measuring instruments, including the autonomous air-sea flux system “AutoFlux” (Yelland et al., 2009) and an underway ΔpCO2 system. An extensive set of direct, eddy covariance measurements of momentum, latent heat, sensible heat and CO2 flux was obtained over a wide range of open ocean conditions. The maximum recorded 20-minute mean wind speed was 25 m.s-1. The maximum significant wave height was 11 m. The initial CO2 flux results were subject to a large, commonly observed humidity cross-sensitivity error. A novel iterative correction procedure was developed, tested against an independent data set and proved to be robust (Prytherch et al., 2010a). Open-path sensors may now be used for air-sea CO2 flux measurement, greatly increasing the number of measurements available for analysis. There are large differences between existing gas transfer to wind speed relationships, particularly at high wind speeds, and there is significant uncertainty over the form (quadratic or cubic) of the relationship. From the 3938 direct CO2 flux measurements made onboard Polarfront, a new relationship between gas transfer velocity, k660 , and wind speed, U10n has been obtained: k660 = −0.51+ 0.095U10n 2.7 0 ≤U10n ≤ 20 m.s-1 The motion corrected fluxes were found to have a large signal at frequencies associated with platform motion. This signal is also apparent in results from previous air-sea experiments from both fixed and moving platforms. The cause of this signal, whether error or real wind-wave nteraction, remains unknown. The gas transfer relationship obtained after removal of this signal is: k660 = −0.09 + 0.02U10n 3.1 2 ≤U10n ≤ 20 m.s-1 demonstrating that the observed near cubic dependence on wind speed, also reported in some previous experiments over a more limited wind speed range (McGillis et al., 2001a), is a robust result. This suggests a significant role for wave breaking and bubble-mediated exchange in air-sea gas transfer.

The inter- and intra-annual variability of the western boundary North Atlantic 26.5 ◦N region has been central in the observations of the strength and structure of the Atlantic Meridional Overturning Circulation (AMOC). Interest in this work began when some recent work estimated the inter-annual fluctuations of the AMOC at 26.5 ◦N to be up to 3 Sv, and with a 25% reduction in strength over the last 50 years. There was increased need to understand both the short and long-term changes in the region and the responsible mechanisms for its variability. With the unique use of RAPID-MOC and MOCHA transatlantic mooring array in combination with satellite altimetry and transatlantic hydrographic observations, we find evidence that a significant amount of the variability can be accounted for by various mechanisms on different time-space scales including propagating features. Here we present simultaneous assimilation of surface and sub-surface observations that shows fresh insights into the contribution of the propagating features in the vertical structure of the temporal-spatial evolution in the western boundary 26.5 ◦N Atlantic. There is great prospect in using altimetry observations to reflect and infer the variability throughout the water-column - an effort vital in future interpretations of the AMOC fluctuations using altimetry and numerical models.

By means of numerical modelling the cycling of iron between its various physical (dissolved, colloidal, particulate) and chemical (redox state and organic complexation) forms in the upper mixed layer of the ocean is analysed. Using the model an initial quantitative assessment is made of how this cycling influences iron uptake by phytoplankton and its loss via particulate export. The model is forced with observed dust deposition rates, mixed layer depths, and solar radiation at the site of the Bermuda Atlantic Timeseries Study (BATS). It contains an optimised ecosystem model which yields results close to the observational data from BATS. Firstly, the results of a zero-dimensional model approach show that the mixed layer cycle strongly influences the cycling of iron between its various forms. This was mainly attributed to the light dependency of photoreductive processes and to the seasonality of primary production. The daily photochemical cycle is driven primarily by the production of superoxide and its amplitude depends on the concentration and speciation of dissolved copper. Model results are largely insensitive to the dominant form of dissolved iron introduced via dust deposition, and also to the form of iron that is taken up directly during algal growth. In the model solutions, the role of the colloidal pumping mechanism depends strongly on assumptions made regarding rates of colloid aggregation and photoreduction rate. Secondly, a one-dimensional approach of the model is coupled with the General Ocean Turbulence Model (GOTM). The combined model was able to simulate the temporal patterns and vertical profiles of dissolved iron in the upper ocean at the BATS site reasonably well. Subsurface model profiles depended strongly on the parameter values chosen for loss processes affecting iron, colloidal aggregation and scavenging onto particles. Current estimates for these parameters result in depletion of dFe. A high stability constant of iron-binding organic ligands is required to reproduce the observed degree of organic complexation below the mixed layer. Solubility of atmospherically deposited iron higher than 2% leads to dissolved iron concentrations higher than observations. Despite neglecting ultraviolet radiation, the model produces diurnal variations and mean vertical profiles of dFe which are in good agreement with observations.

A new methodology to estimate the turbulent air – sea heat and moisture fluxes and their uncertainty is developed and assessed using Voluntary Observing Ship (VOS) observations. Whilst important drivers of the global oceanic and atmospheric circulation these fluxes remain poorly quantified, both in terms of mean value and uncertainty. The new methodology addresses both of these issues and is extensible to other data sources. The individual observations are first bias and height adjusted to remove systematic errors and the impact of changing observing heights. They are then characterised in terms of random errors using a semi-variogram analysis and a range of variogram models. The data quality and sampling are then taken into account using optimal interpolation (OI) to grid the observations, producing daily mean fields and uncertainty estimates. These are then used to estimate the fluxes and flux uncertainty on both daily and monthly time scales. Comparisons of the mean fields and fluxes to the original input data and to independent buoy observations show the fields not to be significantly biased. The adjustments applied before gridding and flux calculation are also shown to improve the agreement with the buoy observations. The uncertainty estimates are assessed using a series of cross validation experiments and 3-way error analyses to make alternative estimates of the uncertainty. These alternative estimates are shown to be of the same order of magnitude as the OI uncertainty estimates and generally to be within 10 – 20% of the OI estimate. Whilst all three estimates are similar there are some systematic differences. The OI uncertainty estimates tend to be lower (higher) than the alternative estimates in high (low) variability regions. The representation of the variability in the new dataset is examined and shown to be improved compared to previous VOS based datasets. The adjustments are shown to have little impact on the temporal trends in temperature and humidity whilst reducing the wind speed and sensible and latent heat flux trends. These reduced trends are thought to be more realistic. The wind speed trend after adjustment is more similar to the trends reported in previous studies using reanalysis model output. However, there are still some differences in the trends, with the VOS based estimates larger, leading to uncertainty in trend estimates. The trends in the adjusted latent and sensible heat flux estimates are similar to those seen in other flux datasets but when compared to changes in the upper ocean heat content may still be too large. This may be due to the overestimate of the wind speed trend. Overall the uncertainty in the wind speed trend gives the largest uncertainty in the flux trends. Finally, the advances made in developing the new methodology are summarised and the potential uses of the new dataset identified. Future work and improvements are then suggested.