The Weddell Polynya, a large hole in the Antarctic sea ice, reappeared in 2017. The polynya forms due to deep convection, which is caused by static instability of the water column. Observations and model studies show periodic heat accumulation in the subsurface layer prior to a polynya. This heat accumulation could be caused by internal ocean dynamics: the Southern Ocean Mode. Periodic subsurface heat and salt accumulation could be the major driver in causing periodic deep convection, which is in contrast with earlier studies. These studies focus on surface processes, and see the polynya as an irregular event. In this study a simple convective model is used to look into this contrast. Model simulations excluding and including periodic subsurface heat and salt fluxes have been performed. Multiple polynya events were only simulated in the model set up including subsurface fluxes. The dominant frequency for polynya events in these simulations equals the frequency of the subsurface heat and salt accumulation. This frequency is still visible in runs with white noise added to the freshwater flux, showing the importance and dominance of the subsurface forcing. In combination with earlier studies, this study suggests that periodic subsurface processes are most dominant and govern the initial formation and periodicity of the Weddell Polynya.

To combat the emission of greenhouse gasses and the corresponding climate change, emission reduction goals have been established in the recent Paris Agreement. In order to meet these reduction goals and minimise the global average temperature increase, implementation of renewable energy sources is critical. Wind energy is one of the fastest growing renewable energy sources in the European Union, since Europe has the largest offshore wind energy capacity with its shallow shelf sea: the North Sea. Over the last twenty years, the wind energy capacity has increased significantly to become the current second highest power generation form of the European Union. The downside to offshore wind energy however, are the high costs that come with its generation. Cable losses of alternating currents between the wind turbines and the shore become increasingly high, as the distances towards the shore increment. For long distances, this inefficiency becomes unacceptable and conversion to a direct current is deemed necessary. A recently presented concept that aims to reduce the costs of offshore wind energy, is the concept of a large scale roll-out of interlinked offshore turbines, coordinated at an European level, including a so-called hub island. This artificial island facilitates converters, as well as a home base for engineers and large scale power storage. Technically, no big problems are expected with the construction of such artificial island. The major reason that could stop the hub island concept, is resistance from environmental organisations. To predict ecological repercussions of an artificial island, the impact on hydrodynamic parameters most important to the bottom of the marine food chain are explored. The primary producers represent this foundation of the food chain, and are most affected by the island through changes in water stratification and residual currents. These parameters highly influence the light and nutrient availability, thereby regulating the primary production dynamics of the ecosystem. By developing and applying the Three Dimensional Dutch Continental Shelf Flexible Mesh (3D DCSM-FM) model, the impact of an artificial island on the stratification and residual currents is explored for five case studies. Each case study comprises a 6km$^2$ cylindrical island, for different North Sea locations with distinct hydrodynamic properties. The implementation of an artificial island alters residual currents up to 10km from its position. The location determines the impact pattern, but influences remain local without any large scale North Sea impacts. Since changes in nutrient availability are only expected for large scale residual current impacts, no significant alterations in primary production dynamics are to be expected. The impact on stratification however, can have a significant influence on the dynamics of primary production. The originally well-mixed areas remain mostly unchanged, while islands in the more stratified regions can cause significant changes in absolute stratification up to 20km from the banks, and alterations in regimes or the seasonal onset timing up to 2.5km. The location and its hydrodynamic properties are paramount to the type of expected stratification impact, and proves to be an important design parameter for ecology around an offshore energy hub island in the North Sea.

To limit global warming to 2 °C or below, the IPCC emphasizes the need for large-scale carbon dioxide removal (CDR) alongside emissions reductions. There is a growing recognition of ocean-based CDR techniques and expanded research is needed. This thesis investigated the long-term (>100 years) carbon sequestration potential of enhancing the biological carbon pump by ocean nitrogen fertilization, a research domain which so far has received little attention. The objective was to make an initial rough quantification of the (cumulative) CDR potential of one-off ocean nitrogen fertilization achieved within 5 years after the fertilization, to evaluate its potential as a global CDR technique and for commercial implementation. An idealized one-dimensional vertical (1DV) model framework, coupling a marine ecosystem model with a vertical eddy diffusion physical model, was established to simulate ocean nitrogen fertilization, incorporating essential processes of ocean productivity and the biological carbon pump. Half of the ocean was excluded from the framework since these regions are incompatible with 1DV modelling. Additionally, regions where nitrogen is not the first limiting nutrient were excluded. Various fertilization simulations were performed in which the nitrogen limitation in the mixed layer of the remaining ocean areas was replenished by a one-off addition. Long-term carbon sequestration resulting from fertilization was assessed by comparing these simulations with baseline simulations of the biological carbon pump. An aggregated global outcome, forming the basis for feasibility assessment, was derived by averaging over simulations, while excluding hexagons with low sequestration efficiency as they contribute insignificantly to CDR. In principle, the one-off nitrogen fertilization shows an effective carbon sequestration efficiency (2.95 kg CO2 per kg N). However, the high and challenging-to-reduce carbon footprint of current urea production, coupled with challenges in collecting manure and the priority of utilising it in agriculture rather than directing it for carbon removal, leads to a low overall global carbon sequestration potential. Nevertheless, it is demonstrated that implementing one-off nitrogen fertilization at a location with high sequestration efficiency using manure can be commercially viable. The cost per net tonne of CO2 sequestered is lower than the current market price of a carbon credit, considering the transportation is conducted with vessels that would have navigated the route regardless. Based on the findings of this study, it is concluded that nitrogen fertilization has very limited potential to serve as a large-scale global carbon dioxide removal strategy, but can be effectively employed on company-level using surplus manure without agricultural purpose. The initial quantification, despite inherent limitations such as the use of a 1DV modelling approach, serves as a valuable addition to the scarcity of research in the domain of ocean nitrogen fertilization. Further more detailed research into strategic nitrogen fertilization for company-level CDR is justified, with this study providing valuable guidance on optimal research locations.

The Greenland Ice Sheet has a total volume of 2900000 km3. In recent decades, the ice-sheet has been losing mass rapidly and has nearly doubled its contribution to sea-level rise. One main contributing factor has been the recent widespread acceleration of the tidewater glaciers that terminate in deep and narrow glacial fjords. However, our understanding of the subsurface water properties causing this acceleration has been extremely limited owing to the lack of observations around these locations. We hypothesize that while the ice-sheet is not coupled to the ocean in the current General Circulation Models (GCMs), they can still be used to improve our understanding of the subsurface warming around these regions. In this study, we first evaluate two GCMs (HADGEM2-ES and GISS-E2-R) around three tidewater glacier locations, by interpolating the high resolution CTD data to the GCM grid. Our results substantiate that GISS-E2-R performs better than the HADGEM2-ES model along the Western Margin. However, along the Eastern Margin, we find that the HADGEM2-ES model is more consistent with the observations. The modelled spatio-temporal variability of water masses was investigated in a Pre-Industrial and Historical simulation, later used to quantify the evolution of subsurface warming under a future warming (RCP 8.5) scenario. Our results show that the Kangerlussuaq glacier (shelf) is more sensitive to greenhouse gas (GHG) forcings than Helheim and Jakobshavn. With respect to the Pre-Industrial climate, the mean warming of the Kangerlussuaq shelf under the RCP 8.5 scenario is 3.76°C, which is considerably greater than Helheim (2.38°C) and Jakobshavn (1.92°C). This is predominantly driven by the subsurface Return Waters of the Arctic Atlantic which are seen to warm the most under a future warming (RCP 8.5) scenario. For each of the three simulation, we decompose the Ocean Heat Content (OHC) time-series into intrinsic oscillatory modes so as to discern and quantify the high frequency variability from the multi-year/decadal variability and trends in the OHC. Where the contribution from the high-frequency modes to the total energy contained in the OHC signal is considerable in a Pre-Industrial scenario; the Historical and RCP 8.5 derived OHC are seen to contain a very dominant low-frequency response to GHG forcings, which contains (almost) all of the energy contained in the signal. For the RCP 8.5 scenario, we infer that at Kangerlussuaq, such a response is considerably greater than other locations. Furthermore, a consistent long-term increasing trend is seen in the upper ocean heat content for Kangerlussuaq, throughout the 21st century (1.01 × 10^8 J/m^2/month), unlike our inferences from other locations. This trend is found to be an order of magnitude higher than Helheim (2.36 × 10^7 J/m^2/month) and Jakobshavn (1.27 × 10^7 J/m^2/month). Our results also indicate that sea-ice free months at the Nares Strait (A.D. 2055 onwards) allows for a relatively greater mixing of the Lincoln Sea and Baffin Bay waters, which is consistent with the decline and stagnation of the rising OHC at Jakobshavn.

Subsea Rock Installation vessels cover pipelines, which transport oil from oil and gas fields. This is done by means of a flexible fallpipe until depths of 1000 meters. Around the Shetland Islands, the workability of the vessels at two specific project locations is compromised by the conditions in the ocean. In this research, data from a 3D hydrodynamic model and logged information from the vessels has been analysed and combined to find an explanation for these problems. At the second location, for tidal flow to the east or east-northeast in the bottom and middle part of the water column, it is highly likely that the working activities have to be interrupted. Moreover, if the value of the turbulent vertical eddy viscosity exceeds 0.028 m²/s, the chance on non-working conditions increases with 70%. For the other project location, these kind of correlations have not been observed.

Sandy beaches comprise large parts of the world's shorelines and act as a natural buffer for many exposed people and assets that are concentrated in the coastal zone. Many coastal communities are vulnerable to the impact of sea-level rise (SLR) that can amplify the episodic erosion from storms and drive structural erosion. The way communities adapt to SLR hinge critically on future SLR projections. One of the major uncertainties is the potential rapid disintegration of large fractions of the Antarctic ice sheet (AIS) that can accelerate sea-level rise, albeit neglected in the latest SLR estimates of the 'Intergovernmental Panel on Climate Change (IPCC)'. Accounting for rapid AIS mass loss in coastal impact assessments is essential for risk-averse coastal managers that disfavour events with large consequences. Although methods to predict future erosion estimates under SLR have been developed, hitherto no study has assessed the impact of different cases of AIS dynamics to erosion estimates. Here, a case-study to the island of Sint Maarten is considered to evaluate the implications for strategies to manage coastal erosion under SLR uncertainty. Regional SLR projections are made for a case consistent with the IPCC, a case with a skewed probability distribution function of the AIS dynamics and a high-end scenario of Antarctic mass loss. SLR projections are incorporated within a probabilistic erosion framework using synthetic storm time series for two beaches on the island. Future retreat distances from storms and long term coastal recession are calculated, and the different scenarios are compared and contrasted. For a future 1/100 year retreat distance of storm erosion, often used for zoning policies, estimates may be exceeded up to 1.11-2.22 times as frequent for inclusion of skewness, and 2.22-67 times as frequent for the high-end scenario compared to the IPCC case. These numbers further increase when additional climate model uncertainty is introduced. In terms of long-term recession, the 1% exceedance probability in 2100 for the IPCC case has a 2-4.5% exceedance probability for a skewed distribution function and a 37-88% exceedance probability under a high-end scenario of the AIS. Lower exceedance probabilities, essential for risk-averse coastal managers, are underestimated relatively more leading to potential disillusion about the safety level that is set. In conclusion, precluding AIS uncertainty from SLR projections that feed coastal impact assessments may lead to ill-informed decisions on SLR adaptation. Risk-averse coastal managers should thus be better informed on deep uncertainty in SLR projections to prevent maladaptation of vulnerable areas.

The Atlantic Meridional Overturning Circulation (AMOC) is a key component in the Earth System. Given its important role in the climate system, variability in the AMOC strength is expected to have great impact on the global climate. The current observational timeseries are not long enough to make climate projections for the end of the century or even longer. Therefore, coupled climate models play an important role in the making of end of century climate projections on the AMOC strength. A known issue with the current generation of global coupled climate models is that the grid resolution is generally too coarse to resolve smaller scale processes such as mesoscale eddies. Observations and modelling studies suggest that mesoscale eddies play an important role in the exchange of water between convection regions and downwelling regions. When such processes are absent or parametrized incorrectly, it can have an influence on the climate projections based on these model simulations. This study analysed the AMOC characteristics in two different simulations of the Community Earth System Model; a reference simulation (referred to as piControl) and a simulation in which the atmospheric CO2 concentrations have been increased to four times the initial concentration (referred to as 1pctCO2). First, the AMOC characteristics in the piControl simulation are analysed using both a Eulerian and a Lagrangian approach. The Eulerian analysis shows that deep mixed layers, an indicator for convection, are present in the subpolar North Atlantic. Compared to observations and higher-resolution ocean-only models are these located closer to the West-Greenland coast. Strong vertical velocities are found over the continental slopes, especially over the steep continental slopes around Greenland. Second, the Lagrangian analysis showed the consequences of the coarse grid in the model. Only a single pathway around the subpolar gyre was observed. This implies that particles will experience convection while crossing the interior of the Labrador Sea, but only will experience downwelling when their individual pathway comes close enough to the continental slopes around Greenland. Furthermore, there is only limited exchange of particles with the regions north and south of the subpolar gyre. In the export of deep waters to the subtropical gyre, does it seem that the particles are being blocked by the North Atlantic Current. Third, the changes in the AMOC characteristics in the 1pctCO2 simulation are compared to the piControl simulation. In the 1pctCO2 simulation have deep mixed layers disappeared from the subpolar North Atlantic and convection in this region has shut down. A new fresh(er) surface layer in the Labrador Sea has intensified the stratification and prohibits the formation of deepmixed layers. Instead of the deep mixed layers in the subpolar North Atlantic have new deep mixed layers emerged between 30N and 40N. In general are velocities reduced in magnitude in the 1pctCO2 simulation, but the stronger vertical velocities can still be found over the continental slopes. In conclusion, the results show that the CESM model can reproduce the two components of the AMOC reasonably well, but the connection in the formof mesoscale eddies is missing. This illustrates that a overturning streamfunction simplifies the complexity of the overturning process. The overturning streamfunction is a measure for the overturning strength, but it does not take into account how and if the convection process and the exchange between convection and downwelling are represented.

The Labrador Sea is one of the deep convection sites in the world's oceans and the water masses formed here are an important component of the Atlantic Meridional Overturning Circulation (AMOC). To study this linkage, one study in particular used an idealized model of the Labrador Sea where the density variations consists only of temperature variations. In this study, it is questioned whether the assumption of neglecting salinity is appropriate, by analysing the pathways of water mass and water mass transformation in the Labrador Sea. This is investigated by using that same idealized model (here called the reference run) and comparing this to a model where salinity variations are added whilst keeping density variations the same (Sconstant) to produce a similar circulation pattern. Furthermore, a model configuration is created which investigates if a seasonal cycle in salinity impacts the circulation pattern of the Labrador Sea (Sseasonal). The pathways of water masses in these model configurations are analyzed by Lagrangian particle tracking from A to B. It was found that with the same initial density variations the maximum surface eddy kinetic energy (EKE) increases by 41 % when salinity is incorporated in the model. An increase in EKE is often associated with more water mass leaving the boundary current (BC) due to an increase in instabilities. Surprisingly, the opposite was found: 7.02 and 8.22 Sv are transported through the BC for the reference run and Sconstant, respectively. Furthermore it was found that most of the water mass leaves and re-enters the BC near the maximum EKE for each model configuration. An increase was found in maximum overturning in density space from an Eulerian perspective: from 3.9 to 4.8 Sv for the reference run and Sconstant, respectively, where about 10 % so called density compensation occurred for Sconstant. No significant annual changes are found when adding a seasonal cycle to the model. For all model configurations a large discrepancy exists between Eulerian and Lagrangian calculations in downwelling. This discrepancy is due to Lagrangian particles that reside in the models at the end of their simulation duration Thus the overturning in the Labrador Sea is significantly influenced by particles that have a long residence time (longer than 4 years in these model simulations). Between 22 and 25 % of the Lagrangian volume transport does not reach the outflow of each model simulation. There are also properties that salinity did not influence: no significant changes were found between the model configurations for the overturning in depth space, the annual MLD and barotropic streamfunction. In conclusion adding salinity to the idealized model showed only minor changes in the pathways of water mass and water mass transformation: the order of magnitude of all analyzed properties stays the same. Density compensation however is neglected when no salinity variations are added in the model. This means that for a highly idealized model of the Labrador Sea, salinity variations can be neglected, when density variations due to salinity variations are represented by temperature variations.

The Rhine Region of Fresh Water Influence (ROFI) is a shallow frictional river plume in front of the Dutch coast. Each tidal cycle a new tidal plume front with fresh-water is released. Recently, internal gravity waves have been observed in this plume. Using a Froude number analysis, support for the internal wave generation mechanism by a tidal plume front is found. It is shown, that on averaged neap tides the density stratification is large. This results in a larger area where the internal waves can be released from the tidal plume fronts compared to spring tides. As the Rhine ROFI is located in shallow water, it is investigated what the effect of bathymetry variation is on internal waves. This is studied by extending the standard KdV model derivation to account for a variable bed. This resulted in a small correction on the propagation speed inside the KdV equation. Observations of internal waves have been used to validate the KdV model. By scaling analysis of the observed waves it is obtained that the relative wave height and relative depth balance for most of the observed events. For these events the wave period and velocity amplitude of the KdV model are well matched with the measurements. This showed that the KdV model may be used for a first estimate of internal waves in the Rhine ROFI. By developing a TGE fitting procedure, it was possible to obtain the pycnocline depth and the direction of wave propagation with only limited data available. Therefore, the parameters have been varied and the error between the TGE solution and the velocity potential obtained from the velocity measurements has been minimized.

Key points:- Pre-closure salinity intrusion into Haringvliet-Hollands Diep is known to have reached Biesbosch National Park at flood tide. Post-Delta21 salinity intrusion is projected to be less extensive than pre-closure. Maximum intrusion is estimated to reach the westernmost Moerdijk ports mainly due to diminished tidal flow at the estuary mouth.- Freshwater intake in the Haringvliet is projected to be compromised for the majority of the year whereas freshwater intake in Hollands Diep is compromised during prolonged drought.- Salinity outwash from the Haringvliet on ebb tide is projected to be poor under both drought and normal conditions due to widespread salinity diffusion in lateral and vertical directions on flood tide. The Delta21 framework, in which this study is positioned, aims at increased flood safety and ecological restoration of the Haringvliet, a former estuary in the Dutch southwestern delta that has been closed as part of the Delta works. To achieve this, Delta21 proposes to reopen the Haringvliet freshwater basin to tidal effects in an attempt to restore its estuarine character. Implementation of Delta21 introduces the risk of compromising agricultural and industrial activities around the Haringvliet-Hollands Diep by reintroducing salinity and tidal movement in the basin. The central problem lies in quantifying the extent and stability of the expected periodic salt intrusion post-Delta21. This has been done by projecting a geographical bandwidth of salinity intrusion patterns depending on the severity of SLR (2020-2100) and low-to-mean Rhine-Meuse discharges (in 2100 CE) using a numerical model. Mitigation strategies are then recommended based on these numerical results. The OSR-HV model (owner: Port of Rotterdam) is used to run predictive scenarios of salinity intrusion for the lower reaches of the Rhine-Meuse basin. OSR-HV runs in TRIWAQ (Rijkswaterstaat), which is 3D numerical modelling software that employs coupled hydrodynamics and constituent transport thereby resolving salinity transport. An upper-end critical scenario consists of a Rhine (Lobith) discharge averaging 1000 m3/s for 31 days combined with 85cm SLR in 2100. This resulted in an intrusion pattern reaching the westernmost port of Moerdijk in Hollands Diep. Results show that the basin geometry, possibly nudged by Coriolis deflection, initially causes a preferential path of salinity intrusion along the southern bank of the Haringvliet. Lateral and vertical mixing is extensive in the western part of Haringvliet which is thought to be a combined effect of weakened tidal flow at the estuary mouth and robust and erratic geometry of the basin. Further up-estuary, the historical flood-ebb tidal channel structures is the main transporter of salinity. The Haringvliet shows distinctly different estuarine behaviour compared to the neighbouring Rotterdam Waterways where stratification is more stable, causing less up-estuary diffusion. The relative robustness of the Haringvliet and mild freshwater forcing give rise to extensive 3D mixing which subsequently limits the maximum horizontal excursion of salinity. Salinity intrusion into the Old Meuse is observed to aggravate upon opening of the Haringvliet sluices due to flow reversal in Spui which connects Haringvliet with the Rotterdam Waterways. This effect can even result in salt intrusion from the Rotterdam Waterways via Old Meuse and Spui back into the Haringvliet, causing a secondary spike in salinity on ebb tide there. Finally, height-limitation of the Haringvliet sluice gates is somewhat effective in countering horizontal excursion of salinity but runaway diffusion in the Haringvliet results in similar salinity profiles compared to full opening of the sluice gates. A shipping channel that crosses the current Haringvliet front delta is included in Delta21. This deep feature promotes advection of salt through the Haringvliet sluices. Subsequent diffusion patterns cause for poor washout of salinity from the basin upon ebb tide. Limiting the depth of this channel is recommended if salt intrusion is to be reduced. Full opening of the Haringvliet sluices furthermore causes approx. 0.50m lowering of Mean Low Water (MLW) near Moerdijk which affects busy shipping routes between Rotterdam-Moerdijk-Scheldt. Partial reduction of the conveying area of the Haringvliet sluices may be used to suppress the tidal wave penetration into Haringvliet-Hollands Diep. Significant gain in ecological value is likely post-Delta21 due to addition of approx. 1900 ha of intertidal areas and a 40 km salinity gradient (excl. front delta). Ample recommendations on further research have been made in this exploratory study. It is recommended to further study the effects of Delta21 interventions on macro hydrodynamics of the Dutch coastal shelf. The interconnected nature of the region requires a larger modelling domain to prevent the occurrence of non-physical effects obtained from the current numerical schematization. Furthermore, hydrological relations and bathymetry were generated from 2020 data. It is therefore recommended to precede the assessment of salinity intrusion into the Haringvliet with numerical projections on change to these environmental factors. Lastly, significant gain in accuracy may be obtained from applying spatially varying temperature and wind to the domain to better replicate baroclinic flows and turbulent mixing.

Thermohaline staircases are characterised by stepped vertical temperature, salinity and density profiles, which are formed and maintained by the double diffusion of heat and salt. Because double diffusion is the primary mixing agent in regions with staircases, it is the topic of extensive studies. Previous studies, however, are mainly theoretical and modelling orientated and observational evidence is needed to verify the results. In our study we use an extensive dataset with 460 vertical profiles of temperature and salinity. We found that staircases in the Caribbean Sea are related through temperature and salinity, indicating that staircases in the Caribbean Sea are constant in time and space. Individual steps, however, differ and were characterised in four types: well-developed steps, transitional layers, inversions and absence of steps. A case study of a strong anticyclonic eddy gave the indication that steps are influenced by short term processes. The eddy induces lateral gradients and hereby positions the water masses in the interior and exterior of the eddy such that thermohaline intrusions are initiated. The apparent preconditioning by the eddy, leading to thermohaline intrusions, allows us to speculate that the eddy is a catalyst in double diffusive diapycnal buoyancy transport.

Climate models predict increased Arctic precipitation and subsequent Arctic freshening as a response to increased green house gas concentrations. Eulerian studies have shown that with increased Arctic precipitation AMOC (Atlantic Meridional Overturning Circulation) strength decreases. Decrease in AMOC strength comes with a decreased redistribution of heat from lower to higher latitudes which can have severe effects on our climate. Therefore, understanding the effects and mechanisms of Arctic precipitation change is a crucial building block for predicting and possibly preventing climate change. This study used a Lagrangian approach. The pathways of water at Fram Strait were investigated for present-day climate (control run) and two scenario runs with increased Arctic precipitation (+50% and +300% respectively). Importantly, it was found that Arctic water reaches the Labrador Sea through Denmark Strait for all three runs. Thus, the extra fresh water in the Arctic can possibly impact sinking and convection zones in the Labrador Sea. The total amount of Arctic water, passing Denmark Strait from Fram Strait, increases for the weak scenario and decreases for the strong scenario of this study. On the other side of Iceland, for the strong scenario of this study, Arctic water stops passing the Iceland-Faroe-Ridge through the Faroe Bank Channel. The amount of Arctic water going into and staying in the Nordic Seas remained almost unchanged with increased Arctic precipitation. The two routes passing from Fram Strait into the North Atlantic were analysed further with respect to depth changes and properties. On both routes particles were fresher compared to the control run when increasing Arctic precipitation. For the weak scenario particles were usually colder than the control run on both routes. For the strong scenario, particles were only colder at Fram Strait, but got warmer than the control run along the pathway.