Narrow Results

This paper analyses the influence on the measured mean overtopping discharge of the duration of physical model tests of wave overtopping, bearing in mind the practical purpose of the studies, and the required accuracy of the measurements. The case study of the South Breakwater of Póvoa de Varzim Harbor, Portugal, is used to investigate this subject. During the two-dimensional physical model tests, three main target test conditions were used. For each one, different wave trains were utilized, all conforming to the same target JONSWAP spectrum, and three different test durations were employed. The number of random waves ranged from about 300 to 2400. Then, one of the three main test conditions was again used, but for twelve different test durations. In this case, the number of waves ranged from about 150 to 1900.

The results suggest that the convergence of the mean overtopping discharge to a constant value with increasing test duration is not obvious. Regardless of the test duration, the information obtained with a single test gives limited information about the expected mean discharge, as the mean overtopping discharge varies even for the same wave and structure characteristics. Consequently, more information is obtained on the mean discharge if several tests of the same short duration (but with different time series) are undertaken rather than if one test of long duration is carried out.

Storms are responsible for large impacts on coastal areas leading to loss of lives, substantial damage to properties and consequent economic losses. Several mitigation and adaptation strategies have been developed to minimize risks associated with storms in coastal areas and early warning systems are one of those strategies. However, warning systems devoted to alerting about risks triggered by storms are not widely implemented yet. This work presents the development of an early warning system for risks induced by wave overtopping and dune/beach erosion, for three locations in Portugal (Costa da Caparica, Praia de Faro and Quarteira). The system uses a model chain (SWAN and XBeach) to generate a risk dataset with runs previously executed for all considered initial conditions. Data are incorporated in a Bayesian Network that, after development and training, is used to predict the risks quasi instantaneously. The system is running daily and disseminates alerts to decision-makers. These findings have implications for management of risk and property damage in densely-occupied coastal areas, especially in light of increasing populations and climate change.

A previously available software package that uses Level II probabilistic methods for safety assessment was extended to include the Hudson's and van der Meer's formulae for armour layer hydraulic instability of rubble-mound breakwaters. This new package is employed to assess the safety of Sines' west breakwater by using the Hudson's formula. A sensitivity analysis of the influence of the several random variables on the failure probability confirms the need to continuously collect and analyse wave data.

The effects of wave steepness and wave breaking on stability of newly developed blocks are examined by hydraulic experiments under conditions with and without wave breaking. It was confirmed that there is a clear dependence of the effect of wave steepness on the block stability through wave breaking type, such as plunging breaker and spilling breaker; the stability number for spilling breaker case is larger than that for plunging breaker case. On the other hand, under nonbreaking conditions, there is no effect of wave steepness; the stability number for non-breaking conditions is similar to that for plunging breaker conditions. Empirical stability formulae are proposed by considering the effects of wave steepness and wave breaker type. Finally, by using these stability formulae, failure probabilities of the wave dissipating blocks are estimated by the First Order Reliability Method (FORM).