Statistical Method for Estimating Wind Arrival Times from Tropical Cyclones

Planning for tropical cyclone hazards is dependent on the time of arrival of tropical-storm-force winds.  Currently the National Hurricane Center (NHC) displays a probabilistic image on its website that illustrates the chance that an area might receive tropical-storm-force winds over fixed time intervals. The wind arrival graphics discussed in this abstract were created to communicate both (1) probability that an area might receive damaging winds and (2) the timing of when to reasonably expect these winds. The wind arrival graphic aims to communicate concerns quickly and enable citizens and businesses to make an informed decision to act. While forecast tracks and intensity profiles may change with each advisory, arrival time images can be adjusted for changes in track and intensity. This discussion will focus on how the arrival time graphics were created using a statistical wind speed probability model and Geographic Information System (GIS) tools.

NHC has provided operational wind speed probability products since 2006, which include uncertainty information from the track, intensity and wind structure forecasts. These replaced Strike Probabilities, which only provided information about track forecast uncertainties and did not address specific information about the wind hazard. This paper will describe the modifications to NHC’s wind speed probability model needed for the time of arrival probabilities.   The output from the updated wind speed probability model is used as input to a GIS system to develop prototype products.  The probabilities are determined at equal interval points, which have geographic coordinates.  Current prototypes enable a GIS user or forecaster can choose to display a 90% (best case), 50% (most likely), or 10% (worst-case) chance of receiving winds by a specific date and time. Wind speeds can be specified as greater than 34 knots, greater than 50 knots, or greater than 64 knots. Nine total combinations are possible. Timing results are overlaid on top of wind speed probability graphics. While the timing results communicate when winds have a chance to arrive, the probability graphic portrays the probability that tropical-storm-force winds will occur at all. Currently the prototype graphics are undergoing a social science testing to incorporate feedback from end users.