Deriving Forecast Inundation from SLOSH MEOWs: Initial Results from a Demonstration Project in Puerto Rico

Significant errors can exist in the track, forward speed, and intensity forecasts for a given tropical cyclone.   As a result, storm surge modeling and forecasting must rely on ensemble and/or probabilistic techniques that account for meteorological uncertainty. In CONUS, the National Hurricane Center (NHC) relies on such methods for operational forecasting using the Sea, Lake, and Overland Surges from Hurricanes (SLOSH) model via the Probabilistic Storm Surge (P-Surge) guidance. Outside of CONUS, there is a need to expand operational storm surge forecasts into areas dominated by the physics of breaking waves, and the huge computational costs of including the physics of breaking waves necessitates a new approach.

In this presentation, we introduce a new forecasting approach which derives inundation from pre-compiled Maximum Envelopes of Water, henceforth referred to as MEOWs.  MEOWs provide a basin composite storm surge snapshot for a particular storm category, forward speed and trajectory. Taking this methodology a step further, compositing multiple MEOWs allows a user to observe how maximum inundation values change for various perturbations of track, forward speed and intensity of a tropical cyclone. A strength of this approach is that, as opposed to real-time simulations using nearshore wave modeling, computational resources are conserved which allows for rapid processing and visualization of inundation values across a large number of storm types and tracks. This approach allows a user to better conceptualize total risk ahead of landfall enabling effective mitigation and messaging to end-users.  A MEOW blender was developed to allow testing this concept within a real-time environment.  This research presents preliminary results from the MEOW blender developed in the Storm Surge Unit to demonstrate how existing MEOWs can be leveraged to quickly derive storm surge plus wave forecasts rather than relying solely on computationally expensive modeling.