Structure of Hurricanes and its Variation as observed by
Airborne Doppler Radar

Principal Investigator:John F. Gamache
Collaborating scientist(s):

Hugh E. Willoughby
Michael Black
Frank D. Marks, Jr.
Peter G. Black
Christopher Samsury/The Weather Channel
Objective : Describe the various inner core kinematic and reflectivity structures, and their relationships to environmental factors, such as wind shear and sea-surface temperature .  Also describe their short-term variation.
Rationale : Hurricanes exhibit a variety of structures. The structure of the storm winds and precipitation has an effect upon the distribution of damage, and may affect the track and intensity changes. Thus, there is a need to document the various types of structures seen in hurricanes and relate them to variations in environmental conditions such as shear and sea-surface temperature. 
Method : Fly coordinated missions of two NOAA WP-3D aircraft both carrying the now standard Doppler radars that scan in a plane perpendicular to aircraft track, or fore and aft a prescribed angle from the plane perpendicular to the aircraft axis.  Fly aircraft along simultaneous, but normal flight tracks through the storm center, thus obtaining winds out to 30-40 km radius from storm center within a span of 10 minutes. Repeat approximately every 40 minutes. Analyze Doppler data for each successive scan and note the core structure and its changes.


The analysis method used in this study is a three-dimensional variational scheme. Winds are matched as closely as possible to the observed Doppler radial velocities while satisfying the three-dimensional continuity equation. Filtering is also included within the cost functions used for the minimization. The wind grid and its filtering may be either cartesian or cylindrical (usually used for winds at the core of a hurricane).

A new variation has been added in the last two years to do variational analyses that are three-dimensional, but the representation is fourier in azimuth relative to storm center. This last method tends to smooth out small-scale spatial and temporal variations and is called MDHA (multiple-Doppler harmonic analysis). This version can be run in either an anelastic mode, or in a mode that constrains the vertical wind to be zero, and the horizontal divergence to be zero, and thus produces a non-divergent wind field. The divergent irrotational wind will then be the difference between the two analysis methods.

Results : Coordinated airborne Doppler missions have been flown into several hurricanes, notably Hurricane Gustav in 1990, Hurricanes Claudette and Jimena in 1991, Tropical Storm Tina in 1992, Hurricane Emily in 1993, Hurricane Olivia in 1994, Hurricanes Iris and Luis in 1995, Hurricane Edouard in 1996, and Hurricane Guillermo in 1997. The inner core Doppler analyses of Hurricane Guillermo will be placed in a larger 3-degree radius context, since an extensive pattern of GPS dropsondes launches was executed.

The most extensive analyses of the Doppler data have been performed for the following dates: Hurricane Jimena on 23 September 1991, and Hurricane Olivia on 24 and 25 September 1994. The hodographs obtained from the mean wind fields for all penetrations on a given date are shown in Figure 1 for Jimena on 23 September and in Figure 2 for Olivia on 24 and 25 September. The shear on 23 September 1991 in Hurricane Jimena was east-southeasterly in the low-to-mid troposphere, while on 24 September 1994 the shear was weaker in Hurricane Olivia. The mean shear was no stronger on 25 September in Olivia than on 23 September in Jimena, however it was westerly. Hurricane Jimena remained at a virtually constant intensity for 24 hours, while during the flights of 24 and 25 September in Olivia, the hurricane strengthened 1-2 mb/hour on the 24th and weakened by as much as 3 mb/hour on the 25th. Possible contributors to the various evolutions on these three days must include not only the shear as manifested in the inner core, but also the sea surface temperature. On 25 September Olivia was traveling north across a sharp gradient of sea surface temperature, and this probably contributed to its decreasing intensity, and to its evolution from a very symmetric (Figure 3) to a highly asymmetric structure (Figure 4).

Doppler observations have also been placed in a composite relative to moving cells. For simplicity, the cells were considered to move at the speed of the tangential wind. The composite showed the circulation of the first cell examined to be organized in both the vertical (Figure 5) and horizontal directions.

Beginning with the Hurricane Guillermo (1997) data set, scans are being edited and de-aliased using the SOLO software developed at NCAR/ATD/RSF.  The advantage of this method is a much less ambiguous processes than was true with the former methods used.  Each scan is dealiased separately, and thus all the data used to make dealiasing decisions are obtained within a six-second interval.  The editing processes is ongoing.

Low-wavenumber irrotational and non-divergent wind fields have been produced for the Olivia, Jimena, and some of the Guillermodata sets. In each case they show that the divergent flow sets up in such a manner as to oppose the environmental shear in the core (inside the eyewall) of the storm, thus protect it from the shear's effects.

Projects indexed by Scientist
Projects indexed by Subject
Return to Homepage

Last modified: 02/01/00