TROPICAL CYCLONE WINDFIELDS AT LANDFALL FROM AIRBORNE AND LAND-BASED DOPPLER RADAR DATA

Principal Investigator:
Peter Dodge
Sam Houston

Collaborating scientist(s):
  • Dr. Frank Marks, Jr. 
    Objective: Analyze the three-dimensional wind structure of landfalling tropical cyclones by combining airborne Doppler radar data collected by NOAA research aircraft with Doppler data archived by National Weather Service WSR-88D coastal radars. 
    Rationale: NOAA has installed land-based Doppler (WSR-88D) radars along the coast of the continental US. The WSR-88D's measure only one component of the wind, towards or away from the radar. However, if a NOAA P-3 flies an appropriate track relative to the hurricane and WSR-88D, then data from the two platforms can be analyzed with dual-Doppler algorithms to yield a time-series of three-dimensional wind fields. HRD designed the Tropical Cyclone Windfields Near Landfall experiment, part of the Hurricane Field Program, to gather flight-level wind data and make surface wind estimates to improve real-time and post-storm surface wind analyses in tropical storms and hurricanes (see REAL-TIME HURRICANE DAMAGE MITIGATION). One of the objectives of this experiment is to collect airborne Doppler radar data to combine with WSR-88D radar data. 
    Method: If a tropical cyclone with sufficient radar scatterers to define the vortex moves within 230 km (Doppler range) of one or more WSR-88D radars, then a NOAA P-3 will fly on tracks defined by the WSR-88D and the storm center. Because the airborne Doppler radar scans in a vertical plane perpendicular to the aircraft track, the airborne and land-based Doppler rays will be nearly orthogonal, optimal for dual-Doppler analyses. During the storm, the WSR-88D will record base data. After the storm HRD will obtain the WSR-88D data from the National Climatic Data Center and process the Doppler data sets. 
    Accomplishment :
    1. Hurricane Fran: Hurricane Fran (1996) was a category three hurricane at landfall on 6 September 1996. The hurricane moved through a region covered by several WSR-88D Doppler radars, with overlapping Doppler coverage (FIGURE 1). To supplement the marine surface stations, the US Air Force deployed 3 National Data Buoy Center drifting buoys ahead of the storm. The University of Oklahoma sent a mobile Doppler radar (DOW1 ) to North Carolina to collect data, and NOAA flew two research missions into the storm. The second P3 flight collected Doppler radar, scatterometer, and flight-level data for the Tropical Cyclone Windfields at Landfall Experiment. A preliminary analysis of the overland leg was constucted by combining Airborne Doppler radar and WSR-88D data for a region 250 km long and 75 km wide. This comprised a region of stratiform rain ~ 70 km north of Fran's center (FIGURE 2). The winds at 1.0 km altitude were ~ 40 m/s in most of this region (FIGURE 3a). An east-west vertical slice through the horizontal wind maximum (FIGURE 3b) shows that the wind maximum appeared to be elevated at ~2 km in the eastern part of the storm, but was nearer the surface in the western half. Although some of this may be the result of the  geometry of the slice, it is also possible that this reflects boundary-layer processes modifying the  windfield as the air moves inland. A better understanding will require further careful analysis of the Doppler data sets.
    2. Hurricane Danny: Hurricane Danny (1997), a category one hurricane, also moved through a data-rich region (FIGURE 4). In addition to overlapping WSR-88D coverage, there are many surface marine stations in the Gulf of Mexico. GPS sondes were dropped near several of those to obtain boundary layer data in onshore and offshore flow. We were especially interested in the onshore flow because of the possibile storm surge in Lake Ponchatrain and Mobile Bay. The dBZ maximum in Danny was on the south side (FIGURE 5). The winds were calculated on an 80x80 km domain centered on the storm. This region was within the dual-88D coverage, although the 88D beams were ~ 1.5 km high at these ranges. The additional airborne Doppler radar data provide some additional details of the winds below that level. The estimate of windspeed at 1.5 km (FIGURE 6) shows a large swath of winds > 30 m/s in the southern eyewall, with embedded regions of winds > 35 m/s. The higher winds compare favorably with the wind peak found in the last sonde drop. We plan to produce several realizations of Danny's windfields with the multiple Doppler data. 
    Key reference:
    Click here to return to the AOML project overview page.

    Last modified: 24 November 1997
    Please report any problems to webmaster.