Latest Development in the NWS’ Extratropical Storm Surge Model and Probabilistic Extratropical Storm Surge Model

The National Weather Service’s (NWS) Meteorological Development Laboratory (MDL) developed the Extra-Tropical Storm Surge (ETSS) model in 1995 and Probabilistic Extra-Tropical Storm Surge (P-ETSS) model in 2017. In 2000, MDL also developed 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. When the first version of P-ETSS was successfully implemented in 2017, MDL had established the capability to operationally provide both deterministic and probabilistic inundation guidance based on tide and storm surge from extra-tropical storms four times a day along the United States’ Eastern and Western Seaboards, the Gulf of Mexico and Alaska.

However, the deterministic guidance can be improved as higher resolution (13 km vs 0.5 degree) wind and pressure input are now available from the GFS. Additionally, the East Coast basin hasn't been updated in 9 years, so updating it could improve its accuracy and enable it to cover Puerto Rico and Virgin Islands in anticipation of wave coupling. Also, the Gulf of Mexico basin hasn't been updated in 7 years, so updating it could improve its accuracy and allow it to cover the entire Gulf of Mexico further improving its accuracy and potentially assisting Mexico. These improvements are planned for implementation in November 2018.

For the probabilistic guidance, one improvement is to add the same bias correction at stations that is done for ETSS, which is planned for implementation in November 2018. Another is to expand the ensemble spread by increasing the number of ensemble members. For reference, P-ETSS uses the 21 Global Ensemble Forecast System (GEFS) members, whereas probabilistic tropical storm surge guidance uses approximately 630 members. MDL attempted to do so by using the 42 North American Ensemble Forecast System (NAEFS) members, which is only available at 0 and 12Z. However, MDL's internal NWS customers advised against implementing a system that alternated between NAEFS based products and GEFS based products due to the potential for the alternating characteristics of the spread to cause confusion. So while the validation results for the NAEFS based products are presented here, they will not be implemented until a method is developed to smooth out the cycle variations.

This paper will describe the details of these efforts and provide validation using historic events.