An efficient early warning system for typhoon storm surge based on time-varying advisories by coupled ADCIRC and SWAN

In order to mitigate storm surge impacts, precise surge guidance computations for forecasters must be finished within a short period of time to allow them to provide early warning to the public. For this purpose, a coupled ADCIRC and SWAN model was applied based on multiple scenario-based, deterministic model runs for each time-varying meteorological forecast advisory on a relatively lightweight mesh with 57 k nodes covering the North Western Pacific (NWP) ocean. The mesh was designed to achieve an optimal combination of speed and accuracy on a cost-effective parallel computer with 64 cores. These models were applied for two events in 2012: typhoon Bolaven (on the west coast of Korea) and typhoon Sanba (on the south coast of Korea). The surge results for a 72-h forecast yielded relative surge height error of 34.1 to 46.4 % in ADCIRC + SWAN. The surge results from a meteorological forecast 24 h from landfall improved to 21.7 to 26.8 %. Furthermore, surge elevation results progressively approached measured values (i.e., improved) with each successive typhoon advisory owing to diminishing uncertainties in the meteorological input. In conclusion, this new efficient early warning forecast guidance workflow successfully achieved its goals of real-time storm surge simulations for forecasters, early warning, and understanding of ocean dynamics.