Garzon, Juan L.Zozimo, CatarinaFerreira, Andreia M. MarquesFerreira, OscarFortes, Conceição JuanaReis, Maria Teresa2022-02-222022-02-222022http://hdl.handle.net/10400.1/17582Many coastal zones worldwide are heavily populated and host very important socio-economic sectors. Portugal is a good example of countries whose economy is highly dependent on tourism activities, especially those sea-related activities. The two sites selected in this project (Quarteira and Faro) receive thousands of national and international visitors annually, not only during the summertime but also in the rest of the seasons because of the favorable weather conditions. However, these sites have been acknowledged as coastal risk hotspots due to their exposure to wave-induced flooding and erosion. Under this threat, the implementation of effective disaster risk reduction (DRR) plans is vital for minimizing damages in occupied areas. In this regard, Early Warning Systems (EWSs) play an important role in allowing for preparedness, namely, timely site evacuation or effective intervention prior to the approaching storm. The successful implementation of EWSs is one of the most cost-effective and efficient measures for disaster risk reduction and the saving of lives. EWSs can rely on complex tools such as process-based models to simulate coastal hazards namely erosion and flooding. However, they are normally highly time-consuming and this aspect might represent a major limitation for operational systems. Conversely, Bayesian Networks (BNs) can provide risk probabilities instantly after being trained and they have been successfully used to make predictions of storm impacts in several coastal applications. The main disadvantage of Bayesian Networks is that they are data-intensive, requiring large input information in order to derive the probabilistic relationships used in their predictions. Under the lack of field observations, process-based models can be used to generate this required information. Once trained, the BN can be used as a surrogate for a process-based model in an EWS.engreport10.34623/bdcs-3z27