Coastal dunes protect the hinterland against severe coastal storms. Rijkswaterstaat classifies 156 km of the 316 km Dutch coast as complex dune coast because of coastal curvature, multiple dune rows, or both. Current dune assessment methods rely on separate cross-sections. In complex dune systems, however, storm response may depend on interactions across the wider dune area. Improving the assessment of complex dunes is important because it guides decisions on flood-defence interventions.

To the author's knowledge, this thesis develops, applies, and evaluates the first fully probabilistic framework designed specifically for complex dune systems using XBeach 2DH area simulations. These simulations represent storm-driven erosion across the full dune area. The post-storm dune system was then assessed using a new failure criterion developed for complex dune systems.

Only the dune row forming the final barrier protecting the hinterland was assessed for failure. The other seaward dune rows were included through the XBeach simulations but were not assessed separately. The assessed row was divided into three-dimensional cells. For each cell, the remaining sand volume above the maximum local water-level threshold was compared with the minimum volume required for safety. This minimum was based on the Dutch critical-profile concept but expressed as sand volume instead of a fitted profile. A model factor of 1.5 was applied to account for the simplification. The assessed dune row was considered to fail when one or more cells did not meet the minimum-volume requirement.

Using this failure definition, the framework estimates the probability that the assessed dune row fails. This requires considering many possible storm conditions. Running XBeach for every possible storm would be too computationally expensive. Active Kriging–Monte Carlo Simulation (AK–MCS) was therefore used to iteratively select a limited number of informative XBeach simulations and build a surrogate model of the failure criterion. The surrogate model was then evaluated for many possible storms to estimate the failure probability.

The framework was applied to two complex dune systems on Schiermonnikoog: an exposed Island Head with coastal curvature and several dune rows, and a more sheltered Central Section with multiple dune rows of different spacing and orientation.

Stable failure-probability estimates were generally obtained after approximately 45 XBeach evaluations. The most stable configuration achieved a relative confidence-interval width of approximately 26%, whereas some configurations showed false convergence and widths exceeding 3500%. Uncertainty was higher for the exposed Island Head, where stronger wave action was hypothesised to increase numerical variability in the local water-level threshold. This variability may have contributed to unstable and uncertain Kriging surrogate models. Nevertheless, the stable configurations show that the developed AK–MCS framework can efficiently produce stable failure-probability estimates while retaining a surrogate response consistent with the expected physical behaviour of complex dune systems.

Surrogate-model analyses identified maximum total water level as the main driver of predicted dune failure in both case studies. Wave height also influenced the exposed Island Head, whereas storm-surge duration, surge–tide peak timing, and offshore wave direction showed no consistent effect. Offshore wave direction also did not alter the physical failure locations. Although these findings depend on the surrogate fits and the two XBeach models, they suggest that future analyses may be able to represent uncertainty using fewer input variables.

The framework demonstrates that failure probabilities of complex dune systems can be estimated with XBeach 2DH area simulations, a tailored failure criterion, and an active-learning reliability method. Further development should first establish a more stable residual-volume threshold and calibrate the associated dune-cell dimensions and model factor. The numerical variability of the limit-state function should then be quantified to improve the noise treatment within AK–MCS. Finally, the resulting failure-probability estimates should be verified against direct Monte Carlo simulations to assess the performance of the framework. ... Read more

Sea turtle nesting beaches are under increasing pressure from climate change, rising sea levels, and human activity, making the protection of critical habitats like Ostional Beach, Costa Rica, an urgent priority. This study integrates the analysis of wave runup dynamics, groundwater behavior, stakeholder collaboration, and coastal squeeze mitigation to address the unique challenges at this vital olive ridley turtle nesting site.

Wave runup was examined using timestack imagery analysis. Two automated extraction models— entropy-only and entropy-saturation—were compared against manually digitized data. Results show that the entropy-only model is more reliable in capturing peak wave run-up values, a critical measure for understanding inundation risks. Challenges in accuracy, particularly for the entropy-saturation model, were linked to the site’s unique environmental conditions, such as dark volcanic sand. The findings fill a gap in understanding how specific extraction methods perform under unique site conditions, with implications for improving future modeling efforts.

Groundwater dynamics were studied using pressure sensors installed in custom-built wells, revealing significant interactions with tidal forces. These measurements highlighted the role of tidal cycles in influencing groundwater levels, providing crucial insights into the potential for nest inundation. These findings extend existing knowledge by combining tidal and hydrodynamic factors specific to turtle nesting sites.

Collaboration within the TURTLE project was analyzed through semi-structured interviews and structural evaluation of partner interactions. A tailored framework was developed to enhance communication and coordination among stakeholders, addressing identified gaps and leveraging existing strengths. This framework contributes to more effective project management and the application of scientific insights in conservation strategies.

To mitigate coastal squeeze—a phenomenon where natural habitats are compressed by rising sea levels and human development—the study evaluated strategies such as foreland restoration and managed retreat. Findings suggest that integrating habitat restoration with community involvement is critical to preserving the ecological and social balance at Ostional Beach.

This study takes an interdisciplinary approach to explore how wave runup, groundwater behavior, and collaboration strategies can be effectively combined to support conservation efforts. The results stress the need for tailored, site-specific solutions that blend engineering, ecological, and social perspectives to protect endangered species and their habitats. By bridging knowledge gaps and offering practical recommendations, the research bolsters both local and global initiatives aimed at preserving vulnerable coastal ecosystems.  ... Read more