Three-dimensional hurricane structure change as revealed by automated airborne Doppler analyses

John F. Gamache, Peter G. Black, and Frank D. Marks, Jr.

NOAA/AOML/Hurricane Research Division

Wind fields from automatically quality-controlled and analyzed airborne Doppler-radar observations were produced in "real time" (aboard the NOAA WP-3D aircraft during the flight) for Hurricanes Katrina, Rita, Ophelia, and Wilma during the 2005 Hurricane Season.  Several steps are involved in the quality control and analysis:

1.    Remove observations with high spectral width in the velocity observations.  The spectral-width value used this season was 6.25 m/s. 

2.    Remove the reflection of the main and side lobes by the ocean surface. 

3.    Remove "speckles" of data from the observations. 

4.    De-alias the observations using the Bargen-Brown method, and an HRD-developed two-dimensional sweep dealiasing method. 

5.    Produce a wavenumber-0 and 1 analysis from these quality-controlled observations. 

6.    Use the low-wavenumber analysis to assist the Bargen-Brown and two-dimensional sweep dealias processes, and then produce a fully three-dimensional Doppler wind analysis

7.    Use a specialized interpolation method to produce higher-resolution (1.5 km radial resolution; 150 m vertical resolution) radial-vertical cross sections of wind speed, tangential wind, radial wind and vertical wind along the flight tracks.

In this presentation we describe comparisons of automatic airborne Doppler analyses with other observations, including flight-level observations, stepped frequency microwave radiometer (SFMR) data, and GPS dropwindsonde data.  These comparisons show the value of the airborne Doppler analyses in providing a context for the other observations.  In particular, near landfall in Hurricanes Katrina and Rita the Doppler analyses showed a jet near the 700 mb level where the reconnaissance aircraft were flying, and a vertical motion field that suggested a relatively stratified hurricane rather than a convectively active one.  Such stratification would suggest a stronger shear in the planetary boundary layer, and possibly weaker surface winds than suggested by the flight-level data.  Several Doppler analyses during the season suggested that the low-level Doppler wind maximum observed during a radial flight leg was displaced to the left or right of the flight track, and thus was not detected by flight level or SFMR surface measurements. This suggests a method for quantifying the level of uncertainty in the maximum surface wind speed determined from the flight-level, sonde,  and SFMR data.

The automatic process appeared to be fairly robust in its first full year of real-time testing, indicating promise as a future operational tool.  It also showed its use as a quick-look, higher-resolution tool in diagnosing hurricane structure and intensity in the days following hurricane landfall.  The automatic method can also be applied successfully to the large archive of airborne Doppler data, opening up a new data source for hurricane researchers.