A CLIMATOLOGY BASED FORECAST TOOL FOR COASTAL FLOODING IN THE LOWCOUNTRY

Abstract

AbstractCoastal nuisance flooding has increased by an order of magnitude over the past half century, but the National Weather Service has a limited suite of statistical tools to forecast them. Such a tool was developed using coastal flood events from 1996—2014 in Charleston, South Carolina, which were identified and classified by prevailing synoptic conditions based on composite mean sea level pressure anomalies. The synoptic climatology indicated low level northeasterly winds dominated the forcing in anticyclonic and cyclonic events, while a southeasterly surge was the main forcing component for frontal events. Tidal anomalies between flood events and previous low tides were used to create linear regression models for each composite classification studied for forecasting levels of coastal flood magnitude. Beta tests using data from 2018—2019 confirmed the effectiveness of the models with RMSE values less than 0.3 ft and MAE values less than 0.25 ft for each event type. The veracity of the methods was further verified by a multiple day case study from November 2018, where the model was tested against both statistically predicted heights and heights based on ETSS Model (v2.2). The RMSE and MAE for the statical model were 0.18 and 0.15 respectively, while the same values for the ETSS model were 0.28 and 0.23 respectively.