Quantifying uncertainty in tide, surge and wave modelling during extreme storms

Interactions between meteorological and hydrodynamic processes are poorly understood, and may result in large uncertainties when assessing the performance of sea defences in extreme conditions. This study integrates numerical weather prediction models with models of wave generation and propagation, and surge and tide propagation. By using an ensemble methodology, the uncertainty at each stage of the model cascade may be quantified. Subsequently, this information, either as a proxy or appropriately transformed into predictive uncertainty, will be valuable in calculating the likelihood of hydraulic and structural failure in extreme storms. This paper describes results for a domain centred on one of the locations for the proposed Severn Barrage. This barrage will be the focus for the world's largest marine renewable energy scheme and will potentially have a significant impact on the coastal flooding response of this part of the Severn Estuary. Dynamically downscaled, high resolution wind and pressure fields of historic extreme storms are generated using the Weather Research and Forecasting (WRF) modelling system. The state of the art tide and surge model, POLCOMS, in conjunction with a third generation wave model (ProWAM), utilises the meteorological data, producing hydrodynamic parameters such as surge and wave heights at a proposed location for the Severn Barrage. European Centre for Medium range Forecasting (ECMWF) Ensemble Prediction System data are used for boundary conditions in WRF, producing a 50-member ensemble. The variation in storm track and intensity between members allows the uncertainty in the model system to be quantified in terms of wave and surge heights. This work is part of the NERC funded EPIRUS consortium research but is closely allied to the interests of the EPSRC FRMRC project and the HEPEX international network focused on ensemble prediction in the context of hydrological prediction systems.