Improvements of storm surge forecasting in the Gulf of Venice exploiting the potential of satellite data

eSurge-Venice (www.esurge-venice.eu), a project funded by the Data User Element (DUE) Programme of the European Space Agency (ESA) and completed recently, was aimed to demonstrate the improvement of the storm surge forecasting with the use of earth observation data. It was specifically focused on the Gulf of Venice in the northern Adriatic Sea. Small basins like the Adriatic Sea (~1000 km by 300 km) are essentially coastal seas, and the orography plays an important role in shaping the winds. Storm surge forecasting in such a basins is based on predicted fields of sea level pressure and sea surface wind, which force the dynamics of the surge. The surge level has to be added to the initial level of the sea surface at the onset of the simulation: this is the initial field of sea level, which also needs to be set. The accuracy of the storm surge simulations is therefore constrained by those of the atmospheric forcing and of the initial sea level fields: a better knowledge of both would increase the accuracy of the storm surge forecast itself. On the other hand, scatterometer winds are not yet suitable to be assimilated into the regional Limited Area Atmospheric models because of the poor re-visitation period (1.5 datum/day maximum). Nevertheless they are very important to understand and quantify the discrepancies of Numerical Weather Prediction (NWP) model fields. As part of the project, dedicated re-analyses of selected storm surge events have demonstrated that scatterometer data can be used to limit the bias between model forecasts and observations, thus reducing the uncertainties concerning the forcing NWP wind fields. Moreover, coastal altimetry has proven to bring a significant improvement of the knowledge of the initial sea level field across the basin. Indeed, measurements of Total Water Level Envelope (TWLE) have been assimilated into a Storm Surge Model (SSM) with a dual 4d-VAR (4d-PSAS) assimilation technique which allows the assimilation of model errors. Due to the lack of a common terrestrial reference frame, the TWLE observations and the surge level simulated by SSMs cannot be compared directly. However, their profiles should have comparable shapes, up to an additive constant, thus permitting the assimilation into SSMs of their differences. This contribution describes the methodology able to bring satellite observations of radar altimetry and scatterometry into storm surge modelling. Considering the re-analysed storm surge events, the rms error on the estimation of the maximum surge peak (observed - SSM) reduced by 35% using only scatterometer data, by 11 % using only altimetry data, and 40% using both. The technique of direct assimilation of altimetry data into the storm surge model, although the promising results achieved, needs further refinements, while the strategy of mitigation of the bias between model and scatterometer winds resulted reliable and easy to set up in the operational context.