The action of sea storms is one of the most complex littoral processes with deep management implications. Along the Catalan shoreline which is about 700 km long, 190 km are subject to erosion and/or flooding. Around one million people live in areas potentially affected. Sea Level Rise could exacerbate this problem in the near future. Reactive interventions have been the norm in coastal engineering and management. This dissertation proposes a pre-storm strategy that foster costeffective eco-compatible measures, termed Quick Defence Measures (QDM). Prestorm intervention requires to forecast the future post-storm state. Hence, the main objective of this thesis is to assess present coastal risk through a Coastal Early Warning System (CEWS), termed LIM-COPAS, that forecasts the more relevant episodic coastal hazards at the area.LIM-COPAS consists of four modules: (i) meteorological model; (ii) wave generation/ propagation code; (iii) coupled morpho-hydrodynamic model and (iv) risk module via non-stationary multivariate probabilistic models. The performance of this suite of models has been tested with (i) a set of hindcast events and (ii) synthetic storm conditions. The hindcasted events have been: December 2008 (D-08); October-2015 (O-15); November 2015 (N-15); January 2016 (J-16); February 2016 (F-16); December 2016 (D-16) and January 2017 (J-17).
In D-08, errors in nearshore spectral wave parameters have been about twice than those in the offshore area. The error was around 20% in hydrodynamics and 50% in morphodynamics. The post-storm response has been acceptably reproduced, with a Brier Skill Score near 0.4.
LIM-COPAS has shown good accuracy with high return period events (i.e. Tr,waves _ 10 yrs, D-16 and J-17), but lower agreement was found for milder storms (i.e. O-15 and F-16).
The meteorological module provided wind fields that were systematically overestimated. The integrated Mean Bias (MB) was −1.52 ± 0.78 m/s. Tarragona (Coefficient of Efficiency, COE = 0.27±0.13) and Begur (COE = 0.29±0.17) had metrics above the average value (COE = 0.24 ± 0.14); but lower agreement was found at Mahón (COE = 0.13 ± 0.16) and Dragonera.
Wave metrics were more accurate than for the wind fields. The integrated Hs COE was 0.52±0.12 and Tm02 COE was 0.36±0.14. At the central coast, Hs has presented good metrics: low MB (−0.06 ± 0.08 m) and high COE (0.58 ± 0.11). The northern coast metrics were the most stable.
The newly developed risk module has been implemented at 79 beaches. Erosion has been estimated as a bounded cost, whereas flooding as a high upside cost. Dissipative beaches tend to exhibit higher costs than reflective beaches under high sea levels. Tr,waves _ 10 yrs events joint with storm-surges can lead to significant damage costs. The estimated losses for the N-15 event (2510·103euros) do not differ excessively from J-17 (3200 · 103 euros).
Two types of QDM have been numerically tested: (i) sand dunes and (ii) geotextile detached breakwaters. The benefits from maintaining the sand volumes outperform the flooding cost reduction. In general terms, the detached breakwater can be a suitable option for beaches in an intermediate morphodynamic state against low to moderate sea levels and high wave return periods. At dissipative beaches, dunes are the best option, but they require a minimum beach width (around 30 m) that ensures their lifetime.
QDM functionality can be enhanced with compatible long-term actions (nourishments, sand bypasses, submerged vegetation, etc.). A healthy beach state is essential for the QDM effectiveness. A higher sustainable management under present and future climate can be reached with the joint combination of (i) CEWS as a shortterm forecasting tool; (ii) QDM that mitigate storm impacts and (iii) long-term interventions that improves the beach health.
