Intraseasonal predictions of tropical cyclone activity in the North Atlantic Ocean

Outside of seasonal predictions for tropical cyclone (TC) activity, most TC forecasts concentrate on the synoptic time scale (i.e. up to five to seven days) for track movement and intensity change. However, as has been suggested by Vitart (2009), forecast skill may exist on intraseasonal time scales using dynamically based ensemble predictions such as from the ECMWF Monthly Forecasting System. This modeling framework features an ensemble of 51 coupled ocean—atmosphere integrations out to 32 days. Using the monthly ensemble forecasts from ECMWF that are produced weekly, we compare the predicted TC activity in the North Atlantic Ocean to the observed variability during the months of June through October for 2007–2009. Results indicate that the modeling system can broadly capture large-scale regions that have a higher/lower probability of TC activity, featuring skill above climatology for the Caribbean Sea and Main Development Region on intraseasonal time scales. The modeling system is shown to handle easterly wave induced TC genesis events better than other pathways for TC formation that require mid-latitude interactions. Finally, the monthly forecast results for TC activity are shown to vary in accordance with the phase and intensity of the Madden-Julian Oscillation at the time of model initialization. Discussion of how subseasonal TC forecasts can be used to assist decision makers in the energy, retail, and reinsurance sectors is also provided.