REAL-TIME TSUNAMI FORECASTING FOR THE CARIBBEAN REGION

Due to the very short response times necessary for tsunami warnings, alerts provided by NOAA’s Tsunami Warning Centers (TWCs) are initially based on processed seismic information and pre-determined criteria. Supplementary information is based on forecasted impact along the coast. The forecasts are generated by numerical tsunami models constrained by earthquake parameters and real-time tsunami observations. Presently, three models are used by the TWCs to generate the forecasts in the Caribbean: Short-term Inundation Forecasting of Tsunamis (SIFT), the Alaska Tsunami Forecast Model (ATFM), and Real-time Forecasting of Tsunamis (RIFT). Each model uses a different approach to forecasting, though all are based on shallow water wave physics solved by finite difference methods. The SIFT approach utilizes pre-computed propagation forecasts which drive real-time coastal inundation forecast models at selected sites in Puerto Rico and the US Virgin Islands. ATFM is a fully pre-computed approach with forecast results pulled from a database and scaled with real-time observations. RIFT is a fully real-time approach which combines a propagation forecast which is based on the actual earthquake source with Green’s Law estimates for wave amplitudes along the coast. Comparing output from the three models provides TWC scientists with levels of confidence on the forecast at hand. This presentation will demonstrate use of the three models and output generated by each.