Subject: A VERIFICATION OF NATIONAL HURRICANE CENTER FORECASTS OF SURFACE WIND SPEED RADII IN HURRICANES

Principal Investigators:
Mr. Samuel H. Houston (NOAA/AOML/HRD)
Mr. Miles Lawrence (NOAA/NWS/NHC)

Collaborating scientist(s):
Dr. Mark D. Powell (NOAA/AOML/HRD)
Ms. Shirley T. Murillo (NOAA/AOML/HRD)
Ms. Summer C. Spisak (NOAA/AOML/HRD)
Mr. Russell St.Fleur (MAST Academy)

Objective: A verification of the National Hurricane Center's (NHC) product containing the initial and forecast horizontal distribution of the surface wind fields in tropical cyclones.

Rationale: NHC's forecast/advisory product contains the initial and forecast horizontal distribution of surface winds (OFCM 1997). These radii are provided in four quadrants from the center of the tropical cyclone to three threshold wind speed values: 17.5, 25.7, and 32.9 m/s (i.e., 34, 50, and 64 kt, respectively). A set of opportunistic marine observations was collected to determine the wind speed radii from the center of each tropical cyclone to the observation site.

Method: The wind observations at National Data Buoy Center (NDBC) marine platforms collected by Murillo et al. (1997) were used to determine when each tropical cyclone's wind field affected a measurement platform with the required wind speed(s). These wind data were adjusted to maximum sustained 1-min values and 10 m heights following the methods of Powell et al. (1996). The wind speed radii based on the marine platform data in storm-relative coordinates were used to verify the official NHC radii forecasts. An example of the use of a single NDBC platform for this purpose was the Frying Pan Shoals (FSPN7) Coastal-Marine Automated Network (C-MAN) site in Hurricane Fran of 1996 (Figure 1).

Accomplishment: Figure 2 shows that the locations of the observed 17.5 m s-1 wind radii about the storm center are approximately evenly distributed. The preliminary results for 17.5 m s-1 wind radii are based on 53 cases. The mean absolute errors for the 0 and 12 h forecasts are somewhat larger than the errors at 24 and 36 h. The algebraic mean error (forecast radii minus observed radii) indicates that there is a mean positive bias at all forecast periods (i.e., the forecast wind radii were larger than observed wind radii). The bias decreases for longer forecast periods.

The small number of cases limits our analysis to simple averages. This is the first time that such a data set has been prepared and that forecasts of tropical cyclone intensity in terms of wind speed radii have been verified. Details on the method of verification will be provided by Houston et al. (1997), along with the final statistics of the study, including the 17.5 and 25.7 radii verification. In addition to the wind speed radii, NHC also provides radii of 3.65 m (12 ft) waves in the forecast/advisory product. Most of the data used in this study have hourly wave height measurements, which may be used to verify the radii of 3.65 m waves. This extends the study by Powell and Houston presented at the 1992 NOAA Hurricane Conference in Miami FL. This latest study examines the accuracy of wind radii forecasts in addition to verifying the wind radii valid for the time of the advisory.

Results will be presented at the Special Symposium on Tropical Cyclone Intensity Change at the American Meteorological Society (AMS) National Conference during 11-16 January 1998 in Phoenix, AZ and at the 52nd Interdepartmental Hurricane Conference during 27-30 January 1998 in Clearwater, FL.

Key reference:

Office of the Federal Coordinator of Meteorology (OFCM), 1997: National Hurricane Operations Plan. FCM-12-1997. OFCM, Silver Spring MD, 134 pp.

Houston, S. H., M. B. Lawrence, S. C. Spisak, S. T. Murillo, 1997: A verification of National Hurricane Center forecasts of surface wind speed radii. Preprints: Special Session on Tropical Cyclone Intensity Change, A.M.S., 11-16 Jan. 1998, Phoenix, AZ.

Murillo, S. T., S. H. Houston, and M. D. Powell, 1997: Composites of surface marine observations for hurricanes during 1975-1996. Proc.: 22nd Conf. on Hur. and Trop. Met., 19-23 May 1997, Ft. Collins, CO, 78-79.

Powell, M.D., S. H. Houston, and T. A. Reinhold, 1996: Hurricane Andrew's landfall in South Florida: Part I: Standardizing measurements for documentation of surface wind fields. Weather and Forecasting, 11, 304-328.

Return to project page