Latest Developments in the NWS Probabilistic Extratropical Storm Surge Model

The National Weather Service’s (NWS) Meteorological Development Laboratory (MDL) implemented the Extra-Tropical Storm Surge (ETSS) model in 1995 and Probabilistic Extra Tropical Storm Surge (P-ETSS) model in 2017. In 2000, MDL also implemented a station bias adjustment post-processing methodology for ETSS to statistically account for components such as sea level rise, waves, river flooding and model error. More recently, MDL enhanced P-ETSS by using the 42-member North American Ensemble Forecast System (NAEFS) instead of the 21‑member Global Ensemble Forecast System. Additionally, a station bias post-processing methodology, similar to the one in ETSS, was applied to P-ETSS. These enhancements are scheduled to be implemented in 2020.

However a gap in skill exists between the NWS probabilistic guidance for tropical and extra tropical storms. This is because Probabilistic tropical cyclone storm Surge (P-Surge) uses approximately 630 ensemble wind members, which is fifteen times more than the 42 ensemble members within the NAEFS based P-ETSS model. To increase the skill of the P‑ETSS results, MDL has worked to incorporate the European Centre for Medium-Range Weather Forecasts 51‑member Ensemble Prediction System into the P-ETSS model. This paper will discuss the resulting improvements in estimates of storm surge based inundation risk from P-ETSS.

For storm surge based inundation guidance, the main source of uncertainty is the atmospheric forcing; however, the initial water level is also an important source. To improve initial water level estimate, MDL has developed a methodology to create an initial water condition from all observations within a given model domain. This paper will describe the details of these efforts and provide validation using historic events.