Seasonal coastal sea level prediction using a dynamical model

Sea level varies on a range of time scales from tidal to decadal and centennial change. To date, little attention has been focussed on the prediction of interannual sea level anomalies. Here we demonstrate that forecasts of coastal sea level anomalies from the dynamical Predictive Ocean Atmosphere Model for Australia (POAMA) have significant skill throughout the equatorial Pacific and along the eastern boundaries of the Pacific and Indian Oceans at lead times out to 8 months. POAMA forecasts for the western Pacific generally have greater skill than persistence, particularly at longer lead times. POAMA also has comparable or greater skill than previously published statistical forecasts from both a Markov model and canonical correlation analysis. Our results indicate the capability of physically based models to address the challenge of providing skillful forecasts of seasonal sea level fluctuations for coastal communities over a broad area and at a range of lead times. Key Points Seasonal sea level anomalies are affected by large‐scale climate phenomena Dynamical models have skill in predicting seasonal sea level anomalies The dynamical model skill is larger than published statistical techniques