Analysis of trends and variability of water levels
Knowledge about the magnitude and frequency of extreme water levels is essential for assessing the risk of coastal flooding, which is the major coastal hazard worldwide. In the last years, significant effort has been directed to analyzing extreme water levels and their potential future changes under climate change. This thesis aims to contribute to this field by analyzing the historical variability of water levels and their components (i.e. tides and surges) at different temporal and spatial scales, and the potential effects of sea-level rise on the short-term variability of storm surges. Knowledge of both long and short term variability of water levels is required for reducing the uncertainties associated with the estimation of the likelihood of extreme water levels for flood impact assessments. Specifically, this thesis investigates long-term trends and inter-annual variability of water levels, tides and surges at two of the longest tide-gauge records of the southern hemisphere: Buenos Aires and Mar del Plata (Argentina). Over the last century, both water level series show an increasing trend caused by the rise of the mean sea-level, but also a decrease of the tidal amplitude. In the case of Buenos Aires, the changes in the tides are likely to be caused by the long-term changes in river discharge, as indicated by the high correlation observed between these two variables. In addition, river discharge is also highly correlated to the inter-annual variability of the extremes. Therefore, climate change induced changes in river discharge can in turn result in changes in the tides and storm surges at the Rio de la Plata estuary. Regarding the short-term variability of extreme water levels, two aspects are investigated in this thesis: namely the variability of the water level curve and the tide-surge interaction. The latter causes a dependency between the tide and the surge component, complicating their separation and posterior combination and thus the assessment of the water level curve evolution. To overcome this issue of tide-surge interaction, the skew surge parameter has been defined and shown to be independent on the high tidal level in semidiurnal regimes. Mixed semidiurnal regimes are characterized by a higher variability of high tidal levels, which can cause a dependency between the tide and the skew surge. This is statistically investigated for 15 sites worldwide, finding that half of these sites show a dependency between high tidal levels and extreme skew surges, and thus not any extreme skew surge can occur at any high tidal level. The skew surge does not contain information of the temporal evolution of the storm surge (water level curve), which is required as input of flood models. The uncertainties related to not accounting for the variability of the water curve when assessing coastal flooding is investigated for a coastal stretch of the German Bight. The high variability found at times around the water level peaks, when overflow/overtopping is more likely, can lead to a threefold increase of the total overflow volumes between events. Sea-level rise produces an increase of the water levels at times around the water level peaks, and this increase is relatively larger for low to moderate sea-level rise scenarios. Therefore, neglecting the variability of the water level curve can introduce large uncertainties in flood assessments and thus using the skew surge parameter might not be recommended when assessing coastal flooding. The results presented here provide important information about the long and short term variability of tides and storm surges that can be used for reducing uncertainties when estimating future extreme water levels and their associated coastal flood risk.