The development of a technique to automatically prepare airborne Doppler observations of a hurricane eyewall for subsequent analysis is discussed. The goal of this development is to provide real-time data for operational forecast models, and real-time analyses for forecasters; however, there can be real gains for scientific understanding. Airborne Doppler radars can depict the three-dimensional structure of hurricane eyewalls, and thus are essential to probe the dynamics of hurricane structure and intensity change. Since kinematic and thermodynamic structure varies substantially in different hurricanes, it is necessary to study many cases to understand the relationships between the vertical wind shear, vorticity, convection, and thermodynamic structure in the hurricane eyewall.

Since 1982, wind fields in the cores of hurricanes have been observed by airborne Doppler radar. Scores of hurricane eyewalls have been probed and the data are stored at the Hurricane Research Division; however, the manpower necessary to examine many of these cases has been lacking, since the preparation of the data for analysis is a very labor-intensive process, requiring at least two full weeks of a person’s time for each snapshot of the eyewall structure. Preparation consists of the removal of artifacts and the de-aliasing of the radial (with respect to the radar) velocities.

Artifacts include reflection from the sea surface, “second-trip” echoes, radar-receiver noise, antenna side-lobe contamination, and extremely turbulent flows that disturb the automatic de-aliasing process. Automatic de-aliasing in a hurricane environment is particularly difficult. De-aliasing requires an estimate of wind velocity (“first guess”) at the location of the Doppler observation closest to the radar, essentially a boundary condition. To date that “first guess” has been the in-situ wind velocity measured by the aircraft; however, many of the Doppler observations of the eyewall are obtained while the aircraft is within the precipitation-free region of the eye, removed as much as 10-20 of km from any reflecting precipitation. Since the gradients in wind components within the eye and eyewall of a hurricane are large, many poor “first guesses” are made, resulting in badly de-aliased velocities that must be corrected manually by a human analyst. The technique described here provides a better automatic estimate of the “first guess”. It also applies automatic techniques to remove receiver noise, measurements of excessively turbulent winds, second-trip echoes, and sea-surface reflections.

A description of the data preparation techinque is given. Then examples of three-dimensional syntheses of the automatically prepared data are compared to those of data prepared by a human analyst.