COMPARISON STUDY OF HRD AND SLOSH MODEL SURFACE WINDS IN RECENT HURRICANES
Samuel H. Houston
Dr. Wilson A. Shaffer (TDL)
Dr. Mark D. Powell
Dr. Jye Chen (TDL)
To compare the Hurricane Research Division's (HRD) surface wind fields in recent hurricanes with the input fields for the Sea, Lake, and Overland Surges from Hurricanes (SLOSH) model and develop methods for incorporating the HRD winds into the SLOSH mode
HRD has developed new methods to produce surface winds fields in tropical cyclones based on all available surface wind observations, including reconnaissance aircraft flight-level observations adjusted to the surface. HRD and the National Weather Service'
s (NWS) Techniques Development Laboratory (TDL) evaluated the parametric wind model used as input to the SLOSH model through comparisons with HRD's surface wind fields in recent hurricanes (Houston et al. 1995, 1996). Previously, the SLOSH model was teste
d for hurricane landfalls along the U.S. coastlines and calculated storm surge heights were found to have an accuracy of + 20% when the hurricane is adequately described.
The HRD surface wind analyses in hurricanes are based on all available surface wind observations from buoys, Coastal-Marine Automated Network (C-MAN) platforms, ships, and surface facilities. Because these types of data are often sparse near tropical cyc
lones, especially over the ocean, reconnaissance aircraft flight-level observations near the storm are used to supplement the in situ surface measurements after adjustment to the surface with a planetary boundary layer model. All of the observations in th
e data set are adjusted to conform to a common framework for exposure (oceanic or open-terrain over-land), height (10 m), and averaging time (10 min average for comparison with SLOSH model winds). The values input to the SLOSH parametric model are the tro
pical cyclone's position (either observed or forecast), size, and intensity. Using these input parameters, SLOSH computes a wind field which is the primary forcing mechanism for the oceanographic processes.
Comparisons of HRD and SLOSH wind fields for several recent hurricanes were conducted for this study (Figure 1). One example of these comparisons was for Hurricane Emily (1993) in which the peak HRD wind speeds (Figure 2) were nearly the same strength on the west and east sides of the northward moving storm during its closest approach to North Carolina. This resulted in higher than forecast storm surge near Hatteras, North Carolina caused by stro
ng winds across Pamlico Sound. The SLOSH model isotachs for Emily (Figure 3) indicate that the storm's peak winds were weaker and the strongest winds were mainly to the right of the storm unlike the HRD surface winds. More re
cently, Hurricane Opal (1995) struck the Florida Panhandle and caused considerable property damage (estimated property losses range from $2 to 3 billion). Most of the damage associated with Opal occurred along the immediate coast where the storm surge and
hurricane generated waves had the greatest impact. Opal's wind field at landfall (Figure 4), indicated that rapid weakening had occurred following the hurricanes peak intensity from approximately 10 hours earlier (not shown)
when the the storm was estimated to be a Category 4 hurricane on the Saffir-Simpson scale. The SLOSH parametric wind field at landfall (Figure 5) shows slightly stronger winds than the HRD wind fields. The peak storm surge v
alues computed by the SLOSH model were in close agreement with the observed storm surge. However, because the storm had been much stronger over the Gulf of Mexico a few hours prior to landfall, hurricane generated waves caused much higher wave runup along
the shoreline (the SLOSH model does not include waves or wave set up in its storm surge calculations). Hurricanes such as Andrew of 1992 (Powell and Houston, 1996), Bob of 1991, and Opal showed small differences between the SLOSH and HRD peak winds at la
ndfall in areas impacted by significant storm surge (the differences between the observed and SLOSH computed storm surge values were also small). These differences were within the expected range of errors associated with the wind observations and the assi
milation techniques used in the HRD wind analysis procedures. However, in the case of Emily, the SLOSH model could not capture the stronger winds to the left of the storm's track. This may have resulted in an underestimate of the forecast storm surge on
the Pamlico Sound side of the North Carolina Outer Banks near Hatteras. Methods are currently being developed to incorporate HRD winds into the SLOSH model and this may prove useful for improving storm surge forecasts.
Houston, S. H., and M. D. Powell, 1994: Observed and modeled water
level response from Tropical Storm Marco (1990). Wea. Forecasting,
Houston, S. H., W. A. Shaffer, M. D. Powell, and J. Chen, 1995:
Comparisons of SLOSH parametric and HRD analyzed surface wind fields
in recent hurricanes. A.M.S. 21st Conf. on Hur. and Trop. Met.,
24-28 Apr. 1995, Miami FL, 619-621.
Houston, S. H., W. A. Shaffer, M. D. Powell, and J. Chen, 1996:
Incorporating HRD surface wind fields into the SLOSH model. A.M.S.
Conf. on Coastal Ocean. and Atmos. Pred., 29 Jan. - 2 Feb. 1996,
Atlanta, GA, 265-267.
Powell, M. D., and S. H. Houston, 1996:
Hurricane Andrew's landfall in south Florida.
Part II: Surface wind fields and potential real-time applications.
Weather and Forecasting, 11, 329-349.
Last modified: 8/19/96