Principal Investigator: Frank Marks

Collaborating scientist(s): Peter Dodge

A major thrust in USWRP is to exploit new observing technology to obtain land-based observations in order to understand the ABL structure as the TC moves from open ocean to land. The WSR-88D radars provide an opportunity to learn much about the evolution of a TC as it makes landfall. This study explores the evolution of the vertical structure of the horizontal wind during the landfall of Hurricane Fran along the coast of North Carolina using data derived from the Wilmington, NC WSR-88D radar.
The U. S. Weather Research Program (USWRP) (Emanuel et al 1995) effort on tropical cyclone (TC) landfall (Marks and Shay 1998) is focused on what the meteorological research community can contribute to a reduction in the disastrous impacts on the nation. With the continuing shift of population and commerce to the U.S. coastal areas, the threat is continually rising. Although some of the losses from disruptions to industry, commerce, and transportation are not avoidable, better warnings would allow time for more damage prevention activities. Both science and technology opportunities are described in a 5-year (2000-2004) research plan (Elsberry and Marks 1998) that has as its ultimate goal to produce better guidance for hurricane forecasters.

During landfall, the key forecasts are of the inner wind structure and precipitation. The inner wind structure, which is modified in complex ways while passing from open ocean to coastal water and then to land is the primary determinant of localized wind damage, tornadoes, storm surge, ocean surface wave run-up, and even precipitation during landfall. Our knowledge of TC structure changes at landfall is in its infancy, as there are little hard data that survive the harsh conditions. The important scientific issues are:

i) Evolution and characteristics of the surface wind field before and after landfall need to be better understood. New data sets of the characteristics of the surface wind field at sea and at landfall must be acquired.

ii) Evolution of the vertical structure of the TC wind field after landfall. Results from the GFDL TC model suggest that the winds at the top of the atmospheric boundary layer (ABL) decay at a much slower rate than the surface winds. This result should be evaluated using sounding and radar observations, and may have important implications for the effect of topography and convection on the variability of the winds.

As storms make landfall the vertical structure of the horizontal wind in the ABL is investigated using the archived Level-ll WSR-88D Doppler radial velocity observations. The vertical profiles were constructed from the de-aliased radial velocity estimates using the velocity-azimuth display (VAD) technique as defined by Browning and Wexler (1968) for each 6-min volume scan. In precipitation-mode, the WSR-88D completes a volume scan every 6 min composed of 360° sweeps at 14 elevation angles between 0.5° and 19.5°.

To maximize the vertical resolution and minimize ground clutter contamination, horizontal wind speed and direction estimates were computed for the 20 250-m resolution radial bands between 5.0 and 9.75 km at each elevation angle. This approach yields horizontal wind estimates starting 62 m above the ground extending to 3.4 km with a mean vertical resolution of ~25 m (VAD wind profiles are computed operationally at 300 m resolution). One drawback to this approach is that the velocity estimates in each band fluctuate about a mean profile with an RMS of 2-3 m s-1. Smooth profiles were constructed (e.g., Figure 2) using a Gaussian-weighted filter with an e-folding distance of ~75 m. Tangential (Vq ) and radial (Vr) wind components were derived using the storm track and radar position shown in Figure1. Storm-relative, as well as, the total wind components were computed. Radius-height cross sections were constructed from the time-height cross sections using a linear fit to the storm motion between 1530 UTC, 5 September and 0730 UTC, 6 September (over this interval the mean storm motion was 330° at 8 m s-1).
Vertical profiles of the horizontal wind, analyzed using the VAD technique, reveal strong vertical wind shear in the lowest 1000 m above the ground. While the magnitude of the peak horizontal winds changed with radius from the storm center the magnitude of the vertical shear changed only slightly. The vertical profiles revealed a log- linear profile of wind speed from near the surface to the height of the wind maximum, ~1000 m (Figure 2). However, the wind direction changed little over the first 400 m, and then backed almost 50° to the height of the wind maximum.

The altitude of the maximum in Vq in Figure 3 was comparable to that for the wind direction, decreasing from 2-3 km altitude at 175 km radius to just above 1 km altitude in the eyewall (50-60 km radius). With slight variations in radius the vertical shear of Vq below the maximum is also nearly identical to that in wind speed in Figure 2, ~2X10-2 s-1, until the radius is <50 km.

However, the altitude of the maximum radial inflow is well below the Vq maximum, between 500 m at radii>175 km to 300-400 m just outside the eyewall (80-90 km radius), very the top of the layer of constant wind direction in Figure 2. Over the same period the altitude of the maximum radial outflow is always above the Vq maximum. Hence, both the tangential wind and radial inflow maxima decrease in altitude as radius decreases.

In radius the Vq maxima are nearly coincident with each reflectivity maxima, while the maximum inward Vr is always just radially outward of the reflectivity maxima, implying the maximum convergence of Vr in the vicinity of the reflectivity maximum. At the same radius that the inflow reaches a maximum, the depth of the inflow also increases, resulting in deeper convergence in the vicinity of the reflectivity maximum (rainband or eyewall). Conversely, the inflow layer, and subsequently the convergence, is reduced between the reflectivity maxima.

Wurman and Winslow (1998) showed that the lowest 200 m, characterized in the VAD profiles by little change in wind direction, was dominated by small (<10 km), narrow (<300 m) regions aligned along the vertical shear vector with large horizontal wind gradients (10-1 s-1) that resemble boundary layer rolls. The presence of these features is likely the cause for the small change in wind direction within this layer, and has possible implications in understanding ABL fluxes in high wind regimes.
Key references:
Browning, K.A., and R. Wexler, 1968: The determination of kinematic properties of a wind field using Doppler radar. J. Appl. Meteor., 7, 105-113.

Elsberry, R. L., and F.D. Marks, 1998: U. S. Weather Research Program Hurricane Landfall Workshop Report, NCAR Technical Note TN-442+PROC, 40 pp.

Emanuel, K., D. Raymond, A. Betts, L. Bosart, C. Bretherton, K. Droegemeir, B. Farrell, J.M. Fritsch, R. Houze, M. LeMone, D. Lilly, R. Rotunno, M. Shapiro, R. Smith, and A. Thorpe, 1995: Report of the first prospectus development team of the US Weather Research Program to NOAA and the NSF. Bull. Amer. Met. Soc., 76, 1194- 1208.

Crum, T.D., and R.L. Alberty, 1993: The WSR-88D and the WSR-88D operational support facility. Bull. Amer. Met. Soc., 74, 1669-1687.

Marks, F. and L.K. Shay, 1998: Landfalling tropical cyclones: Forecast problems and associated research opportunities: Report of the 5th prospectus development team to the U.S. Weather Research Program, Bull. Amer. Meteor. Soc., 79, 1-19.

Wurman, J., and J. Winslow, 1998: Intense Sub-Kilometer Boundary Layer Rolls in Hurricane Fran. Science, 280(5363), 555-557.

Frank Marks administrative information

Project page scientist index
Project page subject index

Main HRD home page

Last modified: 6/30/99