Contrasted influence of climate modes teleconnections to the interannual variability of coastal sea level components–implications for statistical forecasts

Sea level variations at the coast can have drastic environmental and socio-economic impacts in particular in the context of an ever-increasing coastal population and anthropogenic climate change. Regional to global climate variability influences all these factors and exerts a strong control on the coastal sea level over a wide range of time scales. Here, we focus on understanding interannual changes which is paramount to improve interannual forecasting systems as well as to constrain and reduce uncertainties on the secular trend in global mean sea level. We consider the coastal total water level (TWL) as the compound effect of three main components: the wave setup, mean regional sea level anomaly ( i.e., steric and ocean circulation influences) and atmospheric surge ( i.e., influence of local wind and surface atmospheric pressure). To understand their variability at a global scale, we focus on the effect of four climate modes that affect the major oceanic basins: the El Niño Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO), the North Atlantic Oscillation (NAO) and the Southern Annual Mode (SAM). The contrasted regional influence of these different climate modes on the interannual variations of TWL components are quantified. Results suggest that even if the regional mean sea level is overall the main contributor to the interannual variations of TWL variations at the coast and mostly related to ENSO, the contributions from wave setup and atmospheric surge are not negligible in particular at high latitudes and mostly related to the NAO in the Northern Atlantic and to the SAM in the Southern Hemisphere. Such influences from the NAO and SAM can be seen far away from their extratropical regions of action due to their atmospheric forcing of ocean waves that can significantly propagate their imprint towards tropical areas. Implications for interannual to decadal forecasts of the coastal TWL and related hazards are discussed in the light of regression statistical models and the climate modes own predictability.