Storm surge that results from tropical and extra-tropical cyclones is driven primarily by wind strength and inverse barometer effects which have global scale dynamics, but also local scale effects. In fact, storm surge is influenced by a multitude of local factors, including; shoreline geometry; the nearshore slope and bathymetry; cyclone wind obliquity to the coast and forward-moving speed; harmonic coupling with the astronomic tide; and cyclone intensity. The complexity of storm-surge dynamics means that both global-scale and local-scale processes need to both be resolved. In order to resolve the global scale processes, Deltares has developed the global numerical hydrodynamic Delft3D FM model “Global Tide and Storm surge Model” (GTSM). The Global Storm Surge Information System (GLOSSIS) runs the GTSM operationally, producing four forecast a day. To better understand the sensitivities of global to local scale storm surge modelling, the GTSM model has been applied in real-world operational forecasting systems (through GLOSSIS), as well as in a model sensitivity analysis study. Two operational systems are described in this paper: forecasting surface currents to gain an edge in the global sailing contest “Volvo Ocean Race” and storm surge and wave forecasting in TC prone Mozambique. The model sensitivity analysis described in this paper has been performed for a case study in North Queensland. In this study the global GTSM model provides forcing (boundary water level and atmospheric (velocity and pressure)) for a local Delft3D FM model operated by Risk Frontiers. This local model benefits from the availability of high-resolution bathymetry and coastal water level observations. The model sensitivity analysis looked at the impact of bathymetry and grid resolution. Separately, the GTSM v3 hydrodynamic model has been verified on a global scale through a validation study for a global observation dataset for both tide and surge, and compared to other state-of-the-art tidal models, as well as in a dedicated validation study for Europe. The validation shows good agreement of the model against both tide and surge observations, demonstrating the capability of the model to accurately represent water-levels at the coastal zone. The accuracy of the forecasting model is comparable to existing data-assimilated global tide models and outperforms modern non-assimilated global tide models. The surge results show very high correlation and low standard error (STDE) values for both datasets, and tropical-cyclone induced surges are also shown to be well captured. This paper describes the ongoing improvements of GTSM, as well as two forecasting applications of GLOSSIS in the real world as well as the use of GTSM in a model sensitivity analysis performed for a case study in North Queensland.

The physics of extreme high-water during coastal storms covers a wide range of scales, from tides and storms propagating over long distances to the shape of local bays, reefs and islands all contributing to what is potentially extreme high-water. Traditionally, modelling of storm-surges is mostly performed at a regional scale, sometimes extended with nested more detailed models to account for local effects. With the development of more and more accurate global tide and surge models this division may shift. In this study we show the recent advances of the Global Tide Surge Model, since our previous report [9] of version 1, including Self-Attraction and Loading (SAL) representation, dissipation of internal waves, grid refinements near the coast and bathymetry at the South Pole.

An open question is now, how to incorporate local effects with a global model. One option is to mimic the traditional approach and use the global model as a template for a regional model. Another option is to bypass the regional model and directly nest a detailed model in the global one. We will show an example of the first in this paper. An example of the second is planned as a collaborative initiative between Deltares and Risk Frontiers. An important question is what level is needed at the various scales needed to obtain accurate results. We will present both cases at the conference.