Graduation of Vassia Dagalaki

12 December 2018 15:00 till 16:30 - Location: Lecture Hall A, Faculty of Civil Engineering - By: Webmaster Hydraulic Engineering

"Quantifying coastline change uncertainty using a multi-model aggregation approach" | Professor of graduation: Prof. dr. ir. S.G.J. Aarninkhof, supervisors: Ir. W. P. de Boer (Deltares/TU Delft), Ir. F. Scheel (Deltares), Ir. A. Kroon (TU Delft/ Svasek).

Coastal recession is driven by various processes acting on different timescales ranging from decades to hours. The output of process-based morphodynamic models currently used to assess the effect of different processes on the coast is characterized by uncertainty originating among others by input variability and parameter imprecision. Increasing exposure of coastal societies to coastal recession risks creates the need for the aggregated coastal recession estimates with quantified uncertainty in the context of risk informed coastal zone management.


This study investigates the application of different statistical methods on process-based morphodynamic models for the quantification of the forcing and parameter uncertainties propagated to the model output (coastline position change). Subsequently a methodology for the aggregation of the probabilistic coastal recession estimates from multiple models was formulated to account for the combined effect of processes acting on different timescales.


The methods of this study were applied on Anmok beach, South Korea, a coastal stretch afflicted by erosion caused from long, intermediate and short timescale processes. UNIBEST-CL+ and Delft3D model schematizations capable to simulate long term processes were sourced from CoMIDAS research program. First, the most significant uncertainty input sources for each model schematization were identified and quantified through sensitivity analysis and expert judgment. Following a literature review, two methods were considered applicable for process-based morphodynamic models: Standard Monte Carlo and Latin Hypercube Sampling. Both were applied on the UNIBEST-CL+ model schematization, while only the later was applied on a Delft3D model schematization.The methods performance was evaluated based on the precision achieved for the different sample sizes. SMC gives quantified estimates of the precision, enabling the achievement specific target precisions, with the respective computational cost. For the smaller sample sizes used, LHS gave better precision results, proving more suitable for models with longer computational time. The quantification of the precision achieved is however not possible without extra iterations of the procedure.


The second part of the thesis involves the aggregation of multi-model coastline change probabilities into a single probability density function of coastline change under the cumulative effect of the different processes in a management horizon. A Monte Carlo approach was selected for the linear superposition of the contributing probability distribution functions. The advantages of this approach include speed, ease of implementation, comprehensibility and high resolution even at the tails of the aggregated distributions. Utilising this approach, the effect of alternative interventions (combinations of various breakwater designs with a small-scale nourishment) on the coastline change probabilities was quantified. The effect of the considered interventions consists of shifting the aggregated coastline change distribution towards more accretive or erosive values (in and at the edges of the shadow zone respectively), as well as reducing the range of possible coastline change realisations when the sandbar dynamics effect was included in the aggregation. From the designs considered, the emerged breakwater-nourishment yielded the best results in terms of the expected accretion and positively affected area stretch.