Storm surges pose a critical threat to the Dutch North Sea coast, where low-lying areas are highly vulnerable to extreme water levels. While short-term storm surge forecasting is well established, extending predictions beyond 10 days remains a challenge due to the complexity of atmospheric dynamics and limitations in numerical models, which require extensive computational resources and lack flexibility for alternative forecasting techniques. This study explores the potential of weather pattern-based classification as a method for improving mid-term storm surge prediction. Phase I evaluates a set of predefined weather patterns (Neal, et al., 2018), originally developed by the Met Office for probabilistic forecasting in the UK, to assess their applicability for surge forecasting along the Dutch coast. The results indicate that while certain patterns show associations with high-surge events, they systematically underestimate surge magnitudes and, at longer lead times, the surge distributions associated with different patterns become increasingly similar, reducing their ability to distinguish between high- and low-surge conditions. Phase II explores an alternative self-clustered classification approach using k-means clustering with Principal Component Analysis (PCA) for dimensionality reduction. Several data selection methods, including surge thresholding, Maximum Dissimilarity Algorithm (MDA), and stratified sampling, are tested to optimise clustering. While the self-clustered patterns show slight improvements over Neal’s predefined patterns, they still underestimate surge magnitudes and lack the accuracy needed for operational forecasting. A proof-of-concept evaluation using storm Pia (December 2023) and a representative SEAS5 storm reveals that the self-clustered weather patterns struggle to capture extreme surge events. Although these methods are not yet suitable for operational forecasting, this study suggest several possible refinements, such as sequential clustering and expanding the spatial domain, to be promising avenues for enhancing predictive skill.

Anthropogenic activities induced global warming have caused visible consequences in the increase of atmosphere and ocean temperatures. During the 20th century, the global sea level rise behaves a non-uniformed rising rate. In most of the 20th century, the global mean sea level (GMSL) has risen to around 1.8 mm/yr. However, since the early 90s, this rate has increased beyond 3 mm/yr, and from 2006 – 2018 the rate of rise is 3.7 mm/yr. Many researchers have suggested that relatively moderate changes in sea level can cause significant shoreline changes. One of the most significant impacts of sea level rise can be attributed to the retreat of the shoreline, owing to the permanent passive submersion of lands and coastal erosion. To comprehensively understand the effect of sea level rise on different coastal areas, multiple site-specific data-driven models are developed based on the collected global sandy beach datasets. Based on the long-term evolution of available beach indicators, this research are dedicated to studying spatial and temporal topographic morphological variation as a response to relative regional rising sea level.

Under tidal currents and wave action, sediment particles show complex transport patterns for coastal systems like tidal inlets. To model the sediment transport in such coastal systems, models have been derived that can predict that transport using Eulerian (grid-based) or Lagrangian (particle-tracking) methods. The sediment transport model SedTRAILS uses a Lagrangian method to predict the pathways that sediment particles will follow. As a process-based model SedTRAILS is limited in how fast it can compute results. Here we show a methodology to make a faster probabilistic metamodel from the process-based model SedTRAILS by creating an Auto-regressive Hidden Markov Model (ARHMM). Creating a probabilistic model of the sediment transport in a coastal system using Markov chains is
a novel approach in that it builds further on the connectivity approach of SedTRAILS, which is still quite a new development in the field of sediment transport at present. For different versions of the model, correspondence with physical properties of the system were found, that the development to estimate sediment pathways accurately look promising. The method of describing sediment transport pathways of a coastal system as an ARHMM is a form of probabilistic Machine Learning allowing for faster computation times by describing the patterns the system can make. This is in contrast to a process-based model that has to compute all of the underlying processes. The autoregressive component in this approach is expected to be essential in describing any type of dynamical system that consists of trackable pathways. Taking into account the position of a particle at the current time step highly limits the position a particle can have at the next time step in favor of accurately estimating sediment pathways. The modelling approach discussed can find uses in engineering applications where particle pathways are of importance like the dispersal of nourishments, but also other dispersion events like dispersed goods from when a vessel loses a container, as long as there is a process-based model available to create the necessary Lagrangian data.

Wave overtopping of coastal structures is generally expressed in terms of average discharge and maximum overtopping volume. While substantial research can be found on the relationship of such variables with incident spectral characteristics and other geometrical dimensions, limited research has been done to identify the conditions for which severe individual overtopping events occur. The objective of the present work is to further investigate the statistical dependencies between individual overtopping volumes and incident wave characteristics and to identify the conditions which result in maximum overtopping events. To this aim, we statistically analyse timeseries of waves and overtopping events for the case of a vertical wall breakwater on a horizontal bed under non-breaking wave conditions. The timeseries are generated using a validated numerical model for 6 sets of boundary conditions. The analysis of the data shows that, among the examined variables, the incident wave crest level at the toe of the structure has the highest rank correlation with the individual overtopping volumes. Conversely, the incident wave heights show a lower rank correlation while the wave periods show no correlation at all. Additionally, by leveraging the tool of copulas we find that the incident wave crest levels have strong upper tail dependence with the individual overtopping volumes while the wave heights show no tail dependence. Subsequently, we examine possible dependencies with the wave group periods and wave group energy. We find that a moderate correlation exists between wave group periods and overtopping volumes but no upper tail dependence. Moreover, our results show that the fraction between the wave group energy and the group period has a high correlation and upper tail dependence with the overtopping volumes. Finally, we discuss possible practical applications of our findings as well as points for future research.

This thesis explores a suitable method and appropriate input to determine the failure probability of thermal shrinkage cracking in a steel fibre-reinforced underwater concrete floor by examining the factors that influence this failure probability.

An underwater concrete (UWC) floor is a common construction type in the Netherlands that is often applied to create building pits under the groundwater table. Usually, these (generally unreinforced) UWC floors only function as a watertight bottom layer of the building pit with only a temporary sealing function. Water tightness and crack prevention are very important aspects of this type of construction. A permanent reinforced structural top floor is frequently used on top of the temporary UWC floor to make a completely watertight construction, which is not the most sustainable and cost-efficient solution. Advances in concrete technology such as fibre-reinforced concrete have made it possible to integrate both the UWC and structural top floor or even use the UWC floor as the permanent structural floor. In these cases, the UWC floor should already function as a watertight barrier and controlling leakage by limiting the crack width becomes even more important. The addition of fibres to concrete may prevent through crack formation in the UWC slab and the possible consequent leakage. An important cause of cracking in UWC floors is thermal shrinkage during the cooling phase of the hardening reaction shortly after casting the UWC floor.

Currently, there are no guidelines for the construction of fibre-reinforced concrete (FRC) floors and the prevention or limitation of thermal shrinkage cracking. A CROW-committee "Steel fibre reinforced underwater concrete (SFRUWC) floor as permanent structural floor" has been formed. This committee aims to set up a design recommendation and part of that is addressing the thermal shrinkage cracking problem. There is a lot of uncertainty in geometry, material properties and boundary conditions associated with SFRUWC floor design and construction. Because of this, there is a need for a probabilistic approach to investigate the influence of these uncertainties and tolerances on crack formation in SFRUWC floors during the hardening phase. This research aims to determine a suitable probabilistic method to investigate the failure probability of SFRUWC floors and the parameters that influence this.

In order to achieve this, a finite element model was developed that determines the shrinkage cracking behaviour in a part of a UWC floor in a building pit. Before this model could be developed, all the necessary input parameters and design equations were collected from literature and existing guidelines. To set up a model, first the behaviour of UWC floors was studied, followed by the development of a finite element model that includes the hardening behaviour and strength development of young concrete. An important aspect of this finite element model is the use of random fields to introduce spatial variation in the strength properties of the SFRUWC floor, which is a first step to include stochastic variation in the model for the probabilistic analysis. A probabilistic sensitivity study was performed with the finite element model by calculating multiple samples for each set of input parameters. Separate input parameters were varied and the influence on the results was investigated. The parameters that were considered, include the random field properties, material properties and thermal properties. Finally, a full Monte Carlo analysis was performed to give a proof of concept on how to calculate the failure probability regarding thermal shrinkage cracking in SFRUWC floors.

In order to prevent leaking cracks and satisfy the crack width criterion, the tension-hardening behaviour of FRC has to be utilised. Tension hardening behaviour will lead to a distributed cracking pattern, consisting of multiple small cracks which can satisfy the crack width criterion as opposed to a single large separation crack which does not satisfy the maximum allowable crack width criterion. It was found that the tensile behaviour of FRC and the use of random fields to model this tensile behaviour were major parameters that influenced the crack width and the occurrence of a distributed crack pattern. An important parameter is the standard deviation of the random field, which influences the difference between the maximum and minimum tensile strength in different locations in the slab. Looking at material behaviour, the main conclusion was that the introduction of a small hardening branch in the tensile material model was the governing influence factor. Both the standard deviation of the random field and the introduction of a hardening branch in the tensile behaviour affect the ratio between the tensile strength of the ordinary concrete and the residual tensile strength of the fibre-reinforced concrete. These two factors in combination with the model that was used, turned out to be dominating the results in such a way that the other input parameters only had a relatively small influence on the crack widths.

To determine maximum crack widths in SFRUWC floor it is essential to consider multiple samples with different random fields to find out if only a distributed crack pattern is possible or whether also single localised cracks can occur. The possibility of both these results being able to occur, leads to a large possible error in the maximum crack width. The results of this thesis have shown that the random field parameters remain uncertain and experimental research is needed to determine these correctly. Until then, it is important to investigate how significantly these parameters influence the results. It is recommended to extend this research by improving the model by making it more realistic and finding out if these two factors are still the dominating input parameters.

Groynes in the Lower Elbe estuary in Germany are subject to attack from overflowing long-period primary ship-induced waves. Damage can occur on the crest and lee side of the structure due to the high turbulent overflowing flow velocities. An actual design tool for groynes exposed to overflowing primary ship waves is not yet present. Therefore, experimental research based on physical modelling is performed to establish a relationship between the damages and corresponding overflowing primary ship waves. This can be used as a design tool for groynes exposed to these types of wave attack.

The Port of Rotterdam (PoR) is the largest port in Europe and in order to maintain its status, it would need to expand. For the expansion, an extensive survey of the subsurface is needed for the construction of new port areas and its geotechnical structures. As part of designing the geotechnical structures, the subsurface is often modelled as multiple homogeneous soil layers. However the soil properties are in reality heterogeneous and spatially variable. The spatial variability of a soil property is characterized by a mean, a standard deviation and the scale of fluctuation. The scale of fluctuation is the distance over which a soil property is significantly correlated and it is limited to the soil layer of that soil property. Therefore the change in the geological layering of the soil because of external depositional factors such as river, sea and wind can influence the scales of fluctuations and the resulting geotechnical design.

The objective of this thesis is to look at the spatial variability in the vertical direction of the Pleistocene sand from the Kreftenheye and Boxtel Formation in the Port of Rotterdam and see if the river Meuse has any influence on the spatial variability in the vertical direction. An additional question is asked if the spatial variability has any influence on the computation of the pile base capacity for a single foundation pile and what are the implications of the answer to that question. To answer these questions, four sites in the Maasvlakte, Botlek and Pernis were selected and the cone penetration tests (CPTs) taken at the twelve sites were used for this thesis. An empirical method which uses CPT data to identify soil layers was used to identify the Pleistocene sand layer in the CPT data. The first part of the thesis uses the cone resistance data of the CPTs to estimate the spatial variability of the sites. The second part of the thesis focuses on the additional research question by using random field theory. Per site, the mean, standard deviation and vertical scale of fluctuation θ_v were used to generate simulations of cone resistance data. For each combination of standard deviation and θ_v 500 simulations were carried out. For each simulation the pile base capacity was computed with two CPT-based averaging methods, Koppejan method and LCPC method. The coefficient of variation of pile base capacity is used to measure the uncertainty of the computed pile base capacity.

The results show that the range θ_v values are: 0.26 – 2 m in the Maasvlakte, 0.24 – 1.76 m in the Botlek and 0.14 – 1.18 m in Pernis. In terms of the mean θ_v, you see a gradual increase from the upstream area (Pernis) to the downstream area (Maasvlakte). The increase is from 0.27 – 0.63 m in Pernis to 0.64 – 0.80 m in Botlek to 0.84 – 1.86 m in Maasvlakte. However, it is not clear if this is due to the Meuse or due to the existence of sublayers in the geological formation or due to some other factor. Further investigation is needed before a conclusive answer can be given. The answer for the second part is that as long θ_v is significantly larger than the pile diameter D (θ_v ≥ 4D), it does not influence the uncertainty of the computed pile base capacity. However, the mean and standard deviation of cone resistance does influence the uncertainty of the computed pile base capacities. Finally, it is observed that spatial variability does not play a role in the uncertainty of the computed pile base capacity if the coefficient of variation of the cone resistance c_v (q_c) is small (c_v (q_c) ≤ 0.15). The implication for the uncertainty of the computed pile base capacity and therefore the pile design is that one can afford to have a less accurate description of the spatial variability from using fewer CPTs if θ_v ≥ 4D. The same holds true if c_v(q_c) ≤ 0.15.

The authorities of the Port of Genoa have requested the construction of a new vertical breakwater for the Sampierdarena canal, as well as the subsequent demolition of part of the existing breakwater. The reason for this project is the need to expand the size of the canal, to allow larger vessels to safely use it. A choice is made for a vertical-type breakwater, i.e., a concrete caisson placed on top of a rubble mound foundation. The main goal of this thesis is to provide a preliminary probabilistic design of a cross-section of the new breakwater at the port of Genoa by applying a vine-copula model for the wave climate in a Monte-Carlo simulation.

All possible regular vines were obtained by permuting the six equivalence classes for 5 nodes. 13 different copula types and all 480 possible 5-node regular vines were fitted on extreme wave data. The best vine-copula was selected based on the lowest AIC value. The performance of the vine-copula model was compared to that of an independent model.

Offshore waves were transformed to onshore waves by applying SWAN. A MATLAB loop-function was made to automate this process. Prof. Goda's method was used for determining the wave-induced loads on the vertical breakwater. 10 failure modes were considered in total. A Monte-Carlo simulation was built in MATLAB. Using the Monte-Carlo simulation, several dozen designs were tested in an iterative matter to find an optimized design. The proposed preliminary cross-sectional design fulfilled all design criteria.

The future hydrological response of a river system is often predicted using hydrological models that are calibrated on past observations. By using static model parameters it is assumed that hydrological systems do not change in time. For long-term predictions, this is in clear contrast with the notion that vegetation actively adapts its root-system. Moreover, it neglects the possible impact that climate change might have on vegetation species and land-use. The root-zone storage capacity (S_r) is a key component in hydrological models as it has a critical function in the partitioning of water fluxes. This research (1) proposes a regression relationship using climate analogy to estimate the time-dynamic root-zone storage capacity for the Meuse basin, and (2) quantifies the impact of this time-dynamic root-zone storage capacity on the change in hydrological response under future climate change.

A selection of catchments from large sample data sets have been used in the analysis. For each of these catchments we calculated 27 catchment descriptors and the S_r using the water balance method. The regression relationship that is developed based on climate analogy and Multi-Linear Regression analysis showed that the S_r value can be estimated based on 4 catchment descriptors, namely Holdridge Aridity Index, phase shift of precipitation, seasonal amplitude for the potential evaporation, and sand fraction. This regression relationship has an adjusted R² value of approximately 0.70.

We considered three model scenarios. Each scenario consists of a part that models the historical hydrological response which is forced using the simulated historical climate data, and a part that models the future hydrological response which is forced using the simulated 2K climate data. The differences between these model parts indicate how the hydrological response of a system will change in the future. The benchmark scenarios use a static S_r value which is estimated for the simulated historical climate data using different methods. We considered the benchmark scenario for the water balance method and for the regression relationship. The third model scenario is called the dynamic regression scenario. For this scenario the regression relationship is used to estimate the S_r of the simulated historical and simulated 2K climate data. These values are then used for the historical and future part of the model, respectively.

Comparing the benchmark scenarios shows that the regression relationship only has a minimal impact on the change in hydrological response. The impact of the time-dynamic S_r on the change in hydrological response of the system is quantified by comparing the benchmark scenario for the regression relationship with the dynamic regression scenario. The comparison indicates that a time-dynamic S_r results in an increase of absolute evaporation during the summer months (+6.6%), while at the same time resulting in lower values for the streamflow (-8.6%) and groundwater storage (-4.8%) during the winter months. The root-zone storage capacity shows an increase throughout the whole year (maximum +23.6%). In other words, the time-dynamic root-zone storage capacity has a significant impact on the seasonality of the change in the hydrological response. The results also show that the regression relationship is promising and suggests that it might be a good way to estimate the root-zone storage capacity for climate projections in temperate climates.

Floods and droughts, also known as hydro-hazards, are phenomena that generally involve detrimental consequences to society and environment. Traditional practices for risk assessment consider flood and drought independently. However, they are two opposite extremes of the same hydrological cycle. Omitting their interaction might lead to an under- or overestimation of the current and future risks associated with such natural hazards. In history, a number of drought-flood interactions have been observed in various parts of the world. Research to these drought to flood interactions is still in its infancy. Therefore, this research explores the concept of consecutive dry and wet (CDW) events in the Netherlands. The aim of this research is two sided. First, the Consecutive Events Graph (CEG) is introduced. This is a radar chart type of graph used to quantify spatial and temporal changes in dry and wet indicators in consecutive seasons. These can be used to identify hot-spots prone to opposite extremes. Second, a fully probabilistic framework based on Non-parametric Bayesian Network is developed to model the dependence between dry and wet indicators. Such model can be used to infer expected wet conditions in a given region when dry conditions are known.
For the CEG and probabilistic model a number of settings were introduced to quantify meteorological dry and wet extremes and to couple them spatial-temporally. First, a number of indicators were selected to quantify both type of extremes. Second, the dry period is defined in summer (June, July and August) and consecutive wet period in fall (September, October and November). Third, the Netherlands was subdivided in five homogeneous regions such that both wet and dry indicators were characterized on a regional scale. Maximum values of the indicators in its corresponding period were calculated for each single region and for every single year between 1965 to 2020.
This resulted in a dataset consisting of quantities for dry extremes in summer and wet extremes in fall for 5 unique regions over 56 years. Application to the CEG shows potential to identify and quantify CDW extremes. Region-to-region and year-to-year comparison is possible to quantify changes between years or regions. Application to the NPBNs disclosed limited interdependencies across the dry and wet indicators. Using the NPBNs for precise forecasting of expected wet conditions is deemed unsuitable as of low precision. Making inference of wet indicators based on hypothetical mild to extremely dry indicators revealed multiple trends of those wet indicators. These trends are increasing for short term precipitation indicators (R1D, R3D, and R5D) and simple precipitation intensity index (SDII) and are mildly decreasing for the total precipitation (Ptot).
Extreme dry events, extreme wet events and consecutive occurrences of these events are inevitable. It is expected that these phenomena will occur more frequently and become more severe due to a changing climate. A number of recommendations for future research is proposed. Findings from this thesis will help to smooth the path towards better understanding of the identification, quantification and interaction of CDW events or multi-hazard events in general.

For the transportation of the produced energy at offshore windfarms to the onshore grid, export cables have to be installed in the sea bed. Usually, at first trenches are dredged, in which approximately one month later the cables are installed. During this period the trench should remain sufficiently open to install the cables, fulfilling the depth requirements. This thesis focusses on the influence of sand wave fields on the sedimentation rate. It is investigated how trench sedimentation can best be assessed: with a simple model in combination with an extensive probabilistic assessment, or with a more complex model in which a less detailed probabilistic assessment is possible. The research is based on a case study at the Borssele Wind Farm Zone, for which the sedimentation rate is determined during the month April.
The sedimentation is primarily caused by a reduction in (the perpendicular-directed) flow velocity. Shifting of shoals or banks does not occur during the short timescale. The same holds for the sand waves, which do not migrate significantly during this time. Under influence of bedload transport however, the largest sedimentation rates are found for trenches around the sand wave crest. Closer to the troughs, less sedimentation can be expected.
During the probabilistic assessment of the trench sedimentation in flat bathymetries, it was found that the uncertainty obtained by Latin Hypercube Sampling (LHS) with an 2DV-model is reduced compared to the Monte Carlo-analysis (MC-analysis) in SedPit: the median is for both cases 0.06m, the 5-95 percentiles are 0.03m-0.09m and 0.04m-0.13m respectively. Including (idealized) sand wave perturbations in the 2DV-model, increases the range of the P05-P95 values for in trench sedimentation to 0.01 and 0.13m.
Based on the results, it is concluded that SedPit can be used as first assessment tool for the trench sedimentation, since it covers the range obtained with the more complex 2DV-model. The main advantage is the limited computation time of around 1.5 hours, compared to around 48 hours with the sand wave model. On the other hand, the 2DV-model is able to produce more accurate results.
The siltation rate in trenches located at sand wave crests is underestimated by SedPit compared to the results obtained from the 2DV-model. The opposite holds for trenches located at sand wave troughs, in which more sedimentation is predicted by SedPit. Adjusting the calibration parameter of sand transport in SedPit within the ranges provided in Van Rijn (2013) (0.5 for a sand wave trough and 1.0 for a sand wave crest), can at most decrease the difference between the models. For trenches located near a sand wave crest or trough, it is therefore advised to use at least the 2DV-model to assess trench sedimentation.
Furthermore, the 2DV-model can only be used in the case of perpendicular flow with respect to the trench-axis. Inclusion of the sand wave processes, which act in the 2DV-direction, for non-perpendicular flows therefore requires 3D modelling.

Nowadays, one of the main requirements from civil engineering structures is the level of reliability and safety that they must provide to the users. In structural analysis, the exponential application of numerical methods, as the nonlinear finite element analysis (NLFEA), is due to the increasing growth of computational power. These powerful tools provide the opportunity to study the global behavior of complex systems. The latter is especially relevant for cases where nonlinearities highly impact the performance of the structure. However, to ensure a certain level of safety, the stochastic nature of all the influencing parameters must be accounted for during the analysis and design stages. The latter is precisely the role of Safety Formats: a mathematical procedure to ensure an imposed level of reliability while computing the design resistance of structures. Several existing safety formats have been implemented along with the NLFEA in the structural reliability assessment of reinforced concrete structures. Nonetheless, these formats face some criticism in their usability due to some of the assumptions implied in their derivation. Therefore, Monti et al. (2021) recently proposed a safety format: a new Global Factor Method (GFM) that aims to be applicable in the nonlinear finite element analysis of reinforced concrete structures with concurrent failure mechanisms influencing the global performance at a predefined limit state. In a general description, the method requires the output obtained from two nonlinear finite element analyses and the calculation of a Global Safety Factor related to the resistance of the whole structural system. This thesis aims to provide scenarios to validate the new safety format for its future implementation in codes and the industry since, currently, they are extremely limited. Therefore, the GFM is implemented in three overall case studies with increasing levels of complexity, divided into two reinforced concrete cross-sections, and three simply supported beams. During the case studies, special attention is given to the size of the perturbation (described by the so-called $c$ parameter) that proportionally decreases the basic variables for the second nonlinear analysis. Several Engineer Decision-based Scenarios are applied when choosing the Critical Local Failure Mechanisms and the related basic variables to be perturbed. Finally, it is found that the method provides reliable resistance for structures with a single failure mechanism. However, for structures modelled with continuum elements where two concurrent failure mechanisms lead to global failure, the GFM still needs to be revised to provide more accurate guidelines and reduce the space for decisions made by the analyst that highly impact the performance of the method. These decisions are related to identifying and locating the LFM and their basic variables, selecting the solution strategy implemented in the NLFEA, and choosing a suitable modelling uncertainty. At this point, the application of the method demands an experienced analyst and several parametric studies regarding all those decisions, which reduces its efficiency in terms of engineering and computational time.

The root zone storage capacity is a critical determinant in hydrology, playing a major role in the partitioning of precipitation into evaporation and runoff. Besides, it is an important parameter in climatological and hydrological models. Understanding of the root zone storage capacity and its major determining processes is therefore fundamental in environmental sciences. Several studies have investigated root zone storage capacity magnitude and its descriptor variables, but mainly in snow absent regions. Computation and analysis of root zone storage capacities in snow dominant regions is therefore underexposed. As such, additional understanding of the major descriptor variables of the root zone storage capacity in boreal regions and in particular the influence of snow on the root zone storage capacity is desired. The aim of this study is therefore to quantify catchment average root zone storage capacities, identify its main descriptor variables and their regional variability and determine the influence of snow on root zone storage capacities in Canada. Root zone storage capacities were computed for 230 Canadian catchments using a simple water balance approach with additional snow module and were found to be normally distributed with mean magnitude of 183 mm and a standard deviation of 70 mm. Individual correlation of climate, discharge and landscape variables showed most relevant relationship between root zone storage capacities and yearly potential evaporation, runoff coefficient and seasonality index, although with considerable variance. Subsequent investigation on the mutual effect of several variables showed that the aridity index, runoff coefficient and seasonality timing index are major descriptor variables of the root zone storage capacity, by how they indicate the allocation of water for transpiration in a catchment and describe the degree of synchronisation between liquid input and atmospheric water demand. Application of a multiple linear regression model using the aridity index, runoff coefficient and seasonality timing index showed these variables can be used to predict root zone storage capacities in Canada with an R2 of 0.72. Subsequent tests of the predictive capability of this model Finland resulted in an R2 of 0.62. The influence of snow on the root zone storage capacity in Canada was identified by comparing its magnitude computed with and without a snow module. Whenever significantly present, snow effects showed a decrease in root zone storage capacity magnitude, caused by increased overlap between liquid input and transpiration output in a catchment. These effects are encapsulated by the seasonality timing index. To determine the regional variability of root zone storage capacity descriptors in Canada, catchments were clustered based on similar functioning. The results indicated that different variables have an effect on the root zone storage capacity in different functionally comparable regions and that a large part of the functional behaviour of the clusters can be explained by the geographical location of their catchments. The influence of these regionally dependent variables on the root zone storage capacity is encapsulated in the earlier defined main descriptor variables aridity index, runoff coefficient and seasonality timing index.

The objective of this study is find out whether maximum daily discharge of the Geul and Rur catchments can be forecast using machine learning (ML) methods, and if so, to what extent. In addition, these ML models are compared to a conceptual model to see which performs better. A second objective is to test whether soil moisture content (SMC) and NDVI increase performance of the two ML models. The Geul and Rur catchments are both partly situated in the administrative area of Waterschap Limburg, a water authority in the Netherlands. They use discharge forecasts in order to prepare flood defenses and to monitor high water levels more closely. Currently, discharge is forecast using the conceptual HBV model for the Geul. Forecasting is done based on experience for the Rur and only in case of high water levels. Conceptual and physical models are based on physical laws, i.e. conservation of mass and energy. However, some relations are not yet fully understood, or are hard to translate to equations, and assumptions have to be made. This is why a data-driven approach is used, as no explicit relationship between variables has to be specified. In this study a gradient boosted decision tree framework (XGBoost) and a recurrent deep learning model (LSTM) are used to map input to output. XGBoost is a relatively new framework that has shown promising results in other water resources related studies. Long Short-Term Memory is a type of recurrent neural network and chosen for its ability to handle long-term dependencies and for its ability to model non-linearities. In order to see whether the machine learning methods outperform conceptual models, they are compared to the GR4J model. GR4J is a simple yet effective soil moisture accounting model. Beside SMC and NDVI, meteorological variables are used as input. Results show that the deep learning model performs best for simulating today’s discharge and when forecasting up to three days ahead. The GBDT model has a slightly higher Nash-Sutcliffe Efficiency (NSE) for the daily simulation of the Geul, but also a higher mean absolute error (MAE) compared to the deep learning model. The same holds for the three-day-ahead forecast for the Geul and the Rur. Peak timing is accurate for most models but peak discharge is often underestimated. When comparing the ML models to the conceptual model for the daily simulation, deep learning performs best in terms of MAE, but GBDT is better in terms of NSE. When looking at the one-day-ahead forecast, deep learning outperforms the GBDT and conceptual model in both NSE and MAE. In any case, when looking at the metric they outperform the conceptual model. However, the conceptual model has only a couple of parameters to calibrate, is transparent and has only two input variables. The ML models, on the other hand, have my parameters to train, are difficult to physically interpret and have four to five input variables. Besides comparing the two types of models, it is tested whether adding soil moisture content and NDVI as input improve performance of the machine learning models. The former undoubtedly improves performance, whereas NDVI at best improves performance as much as some other meteorological variables. Overall, this study finds that a conceptual model still outperforms the two ML models from a holistic point of view. However, machine learning is not yet fully exploited in water resources management. It already gives promising results and is likely to perform even better in the future.

The existing hydrodynamic models consider full physics approaches to calculate storm surge at coastal regions. However, due to the complexity of the equations that model these processes, the computational time and power required to run them can be large, compared to models that consider simplified equations. By contrast, simplified hydraulic models lack physical background, what leads to a minor accuracy in the surge estimations with respect to hydrodynamic models. In this project, a stochastic model has been developed with the objective to estimate surge at the coast of Mississippi (United States) at a reasonable accuracy and time, without solving equations that represent complex physical processes. The stochastic model needs to be trained by means of hurricane data, including surge levels. This information must be generated beforehand, by simulating a limited number of hurricanes with an hydrodynamic model. In this project, the hydrodynamic model Delft3D Flexible Mesh has been used for this purpose. The approach followed to build the stochastic model has been based on three main steps. The first step has been setting up and validating the hydrodynamic model in Delft3D FM. Hurricane Katrina (2005) has been simulated to calibrate the input parameters of the model, by comparing the maximum simulated water levels at 41 stations along the shoreline of Mississippi to the high water marks observed at the same locations during the event. Tide, surge and wave setup have been considered in the validation. The results of the validation show a line best fit slope from the origin of 0.912 and an R-squared of 0.996. At Gulfport, the absolute error of the surge estimation is 19 centimeters, equivalent to a relative error of 2.5%. The second step in the construction of the stochastic model has been the generation of a historical hurricane data base. The hurricane best tracks have been retrieved from the HURDAT2 data base. The variables considered have been the forward speed and the forward direction of the hurricane at landfall, the wind speed at landfall, the distance from landfall location to a reference point (Galveston Bay) and the maximum storm surge during the hurricane. In this case, only the hurricane forcing is considered as external action. The storm surge has been recorded at Gulfport Harbour (central coast of Mississippi). The values of the surge have been obtained by using the validated model to simulate the historical hurricanes making landfall in a rectangular domain of 600 kilometers, being Gulfport Harbour the center of the rectangle. Due to the scarce number of hurricanes making landfall in this region, the tracks of the hurricanes making landfall in the North of the Gulf of Mexico but outside the rectangular domain have been shifted inside the domain, in order to generate a sufficiently large data base to train the stochastic model. A data base with 140 hurricanes has been built, from which the 85% (119 hurricanes) have been used for the training of the stochastic model and the other 15% (21 hurricanes) have been used for the validation of the stochastic model. The third and last step has been the setup and validation of the stochastic model, by comparing the storm surge obtained from the stochastic model to the surge obtained from the hydrodynamic simulations. The stochastic model used to estimate storm surge has been a Bayesian Network that assumes normal copulas to represent the joint distribution between nodes of the network. The calculated slope of the best fit line for the mean surge values has been 0.861, with an R-squared of 0.885. Moreover, the average standard deviation of the estimations is 1.16 meters. These results indicate a reasonable estimation of the surge by means of the Bayesian Network. This estimation can be made in the order of seconds.

Maturity model has proven to be helpful for an organization to identify its current capability, strength, and weaknesses through maturity levels. These levels are used to rank the various dimension or process of an organization. However, maturity levels only describe a general description of organization capability. The effect of different maturity levels towards the organizational outcomes is speculated in the basis of the description at each maturity levels. The use of the maturity levels is limited in the pre-determined description. This research is an attempt to foresee the effect of different management maturity towards asset cost and performance by using Dynamic Bayesian Network. The network is developed by using a semi-hypothetical case on grass revetment management. The network models the maintenance process which covers the asset degradation and the influence of different extent of management towards grass condition in time. The degradation probability was determined by using Cooke’s classical model expert judgment. We designed different network involving four management dimension that was translated as model scenarios. The result provides an insight into the potential gain or loss of different management maturity which is beneficial for organizations in evaluating different alternatives of management improvement.

This study examines a method for a stochastic beach width prediction with a process-based morphodynamic model. This is carried out by a case study at the Hondsbossche Dunes, after a beach nourishment in March 2018. Stochastic forcing conditions are generated for the location of the Hondsbossche Dunes. These stochastic forcing conditions consist of time series with significant wave heights, peak wave periods, wave directions, and water levels. The basis of the method for generating these stochastic forcing conditions are Autoregressive-Moving Average (ARMA) models. The morphological change of the beach width due to these stochastic forcing conditions is modelled using an XBeach model. This results in a range of the possible development of the beach width, with the probability of occurrence. With the method examined in this study, a more realistic idea of the possible future beach width development can be given.

Surrogate models are used to approximate the expensive ‘true’ simulation codes and thus have the potential to speed up the wind farm layout optimisation problem (WFLOP). One technique to make surrogate models is Polynomial Chaos Expansion (PCE). PCE can approximate a (wind farm) model by using orthogonal polynomials which are constructed based on input variables. In case of a wind farm model, these are wind speed and wind direction. The technique sounds promising, but up till now, PCE has mainly been used as an uncertainty quantification method and not as much in order to help with optimisation problems. This thesis research project aims to implement the PCE method in WFLOP by implementing a multivariate polynomial basis based on the wind speed and wind direction. The usability of the method will be determined based on a comparison between the WFLOP results of the PCE surrogate model and the conventional approach.

A low-lying country as The Netherland is prone to coastal flooding, and its risk may be enhanced by global-warming induced climate change. Sea level rise has been historically considered as the key factor in coastal retreat, but waves also play an important erosive role along the coast, which can also be affected by the changing climate. During the last years, important advances have been achieved in climate modeling, with a very detailed characterization of the different components of the climate system, for the present and for different future scenarios. However, characterization of future ocean waves is still a matter of discussion and ongoing research. In this thesis, a statistical downscaling methodology based on weather types has been chosen to model the present wave climate and explore potential changes in future waves. These changes are quantified in terms of the impact of these variations in the longshore sediment transport. The methodology is applied to Noordwijk, selected as a representative location of the central Dutch coast. The statistical downscaling methodology is based on a classification procedure of the predictor into similar atmospheric patterns over the wave generation areas, namely the weather types. Then, the wave data is grouped according to the occurrence of the weather types. The predictor is built from the sea level pressure fields (SLP) and the squared SLP gradients, while the predictant wave climate is characterized by significant wave height, wave peak period and mean wave direction of wind sea and swell components, resulting in unimodal or bimodal sea states. The chronology of the weather types is modeled using an autoregressive logistic model, which incorporates the seasonality, the interannual variability and the persistence observed from the historical data. For each weather type, wave parameters are modeled using the categorical distribution for the sea-state type, non-parametric kernel density functions for the central-mass regime and two generalized Pareto distributions for the lower and upper tails of wave height and wave period data. The statistical dependence between wave parameters for each sea state is included using a vine-copula approach, where the bivariate dependence of H_s and T_p is modeled using the AC skew t copula and the remaining relations are considered to be Gaussian. The effect of climate change is studied using SLP predictors from the global circulation model ACCESS1.0, under the RCP8.5 scenario and period 2070-2099. The statistical model is applied to identify changes in the occurrence probability of the weather types in the future. The importance of these changes are quantified in terms of the wave-induced longshore sediment transport using the process based model Unibest TC. The longshore sediment transport distribution for each weather type is computed and afterwards the changes in gross and net transports are estimated using the present and future probabilities of the weather types. In terms of the weather types, the results of this work suggest changes in the occurrence probability of the weather types in the future, with variations of ±60% with respect to present climate, and no relevant changes in the sequence of weather types, neither in terms of the transition probability matrix nor in terms of the persistence of each weather type. In terms of longshore sediment transport, significant changes are detected in the transport associated with some of the weather types. For the remaining weather types, the changes are in the order of the model uncertainty. Taking into account the contributions from all weather types, a net increase of the southward directed net longshore sediment transport with respect to the historical period is detected. This increase is driven by a decrease in northwards transport and a lower decrease in southwards transport. This result is in line with a poleward shift of the North Atlantic storm track reported in other studies. Most importantly, weather-type based climate classification has in this thesis been succesfully proven to be a reliable tool to analyze the wave climate at the location of interest. Furthermore, the statistical downscaling also provides a climate emulator that captures the climate dynamics at different time scales, which can be used for stochastic simulations of the atmospheric and wave climate, either for the recent past or future projections.

Flooding is one of the most frequent and terrible natural disasters. For the coastal areas connecting inland with the open sea, the occurrence of flooding could be caused by high flows from inland rivers or high-water levels at open sea (e.g., from tsunamis or storm surge). There is also one extreme case that high sea level and high inland flow happen simultaneously, which is also known as the compound flood event. If a compound flood event happens, it has the potential to cause huge losses.

The Houston Ship Channel (HSC) is an important passage connecting “Greater Houston” area to the Gulf of Mexico which plays a critical role in the economic development of Texas and even the whole country. However, because of its special location, the HSC area is prone to flooding caused by both storm surge and heavy rainfall, which has been demonstrated by many historical flood events e.g. Tropical Storm Allison (2001), Hurricane Ike (2008) and Hurricane Harvey (2017). In the wake of Hurricane Ike, the Severe Storm Prediction, Education, and Evacuation from Disasters (SSPEED) Center at Rice University proposed to build a storm surge barrier near the downstream of Fred Hartman Bridge for the protection of the industrial facilities along the HSC from storm surges. One of the major questions related to the design of the storm surge barrier is ‘what are the probable boundary conditions associated with compound flood events (i.e., the combination of storm surge and upstream rainfall-runoff) in the Houston Ship Channel?’. Because once the barrier is closed, the upstream flow cannot flow out to the Galveston Bay, there is the potential to cause the flooding behind the barrier when the closure time is long and upstream discharge is large enough. Therefore, in this thesis, it is focused on the exploration of potential combinations of compound floods in the downstream reach of the HSC.