A Simple Model for Coastal Sea Level Prediction

Abstract

Reliable forecasting of wind-forced coastal sea level on the synoptic scale is available for most of the coastal areas of the United States through the National Weather Service Extratropical Storm Surge Model (ESSM). However, in many coastal areas around the world, especially in underdeveloped countries, little if any sea level forecasting is available, despite the often acute need. Here a simple linear-regression model based on modest wind forecast capability and records of local coastal sea level, wind, and pressure is developed and tested. Despite its simplicity, the model is based on robust ocean dynamics, in particular coastal Ekman circulation principles. The empirical model is tested using sea level observations at Atlantic City, New Jersey. The performance of the model is comparable to that of ESSM. For the 2-yr time period 1997?98, ESSM explains 79% of the total observed subtidal frequency sea level while the model presented here explains 74%. The root-mean-square errors in sea level for ESSM and the current model are 0.136 and 0.115 m, respectively. This performance indicates that the empirical model is adequate for general use in regions where reliable sea level forecasts from a circulation model such as ESSM are not available.