Assessment of the nourishment efficiency using a Bayesian modellng approach : case study: North Holland

In this study, a Bayesian probabiblistic model has been implemented to assess the effects of nourishments on a number of coastl indicators using, as input, data defined at Jarkus transect level for the Norh Holland coast. As indicators for short- and medium term saftey, probablity of failure, MKL and MDL has been selected. The Bayesian network has proved to be a useful toll to point out the positive effects of different types of nourishmentsbuilt in the past, as well as interrelations between different indicators and the related uncertainties. Moreover, by training the network using information derived from the past, the network can be used as predictive tool i.e. to plan a nourishment in order to reacth a prefined objective (e.g. average seaward migration of MKL of a predefined distance). Nevertheless, the strength of the correlation betwee variables was found to be highly dependent on data availability. On top al all, the fact that several transects have never been nourished or nog been nourished for several years, strongly influence some of the results. As a matter of fact, these transects can not provide information on the effects of nourishments. However, they provide valuable inofrmation on the natural morphological evolution of not-nourished stretched coast. This problem could be partly overcom by using synthetic data derived by numerical models to partly fill up the missing information.