Scatterometer-retrieved hurricane wind direction ambiguity removal using spiral dealias

A notable shortcoming of scatterometry of high wind speeds is the apparent under estimation of the speed of the winds, but wind direction errors are also important. For the direction, a major source of error is associated with difficulty in selecting the proper wind direction from multiple possible solutions known as aliases. The wind retrieval process is based on a relation between wind-induced sea surface roughness and /spl sigma//sup 0/. This relation, denoted as geophysical model function (GMF), is semiempirically derived from collocated sets of radar /spl sigma//sup 0/ measurements and in-situ wind condition observations. Unfortunately, the bi-harmonic nature of the /spl sigma//sup 0/ anisotropy with wind direction results in multiple wind vectors solutions (speed and directions combinations) from a collocated set of scatterometer measurements. Both /spl sigma//sup 0/ measurement noise and a weak GMF wind direction signature contribute to ambiguity in the retrieved wind. In most wind retrieval algorithms, ambiguity in direction is a major component of the direction error. This paper investigates a simple method aimed at improving ambiguity removal during wind retrievals in tropical cyclones. The method uses the a priori knowledge of the general wind circulation about the center of rotation (eye) to produce a spiral "de-alias" of directions in multiple wind solutions. Technique is illustrated using observations from the SeaWinds instrument on board the QuikScat satellite.