On the correlation coefficient |ρhv| between horizontally and vertically polarized radar returns

The magnitude of the correlation coefficient between vertically and horizontally polarized returns is one of the very useful polarimetric variables for identifying the type of scatterers. For most hydrometeors such as rain and various types of dry snow the values are larger than about 0.97. Similar high values can be found in chaff and some type of ground clutter. For wet snow and giant hail at 10 cm wavelength values are lower and could reach about 0.8. At shorter wavelengths these can fall to about 0.4. Thus the span from 0.97 to 1 corresponding to several important categories of scatterers is rather small. It is argued that expansion of this interval could benefit data quality and possibly interpretation. The quadratic operation |ρhv|^2 accomplishes an expansion that has the following nice properties. The |ρhv|^2 is a correlation coefficient between reflectivity factors Zh and Zv at horizontal and vertical polarizations and therefore has physical meaning. The expansion occurs for values of |ρhv| between 0.5 and 1, i.e., the corresponding |ρhv|^2 is between 0.25 so that in case of linear quantization as on the WSR-88D the number of quantization intervals in the region corresponding to hydrometeors increases by 50%. Therefore the “signal” to quantization noise is enhanced. The quadratic function compresses the values between 0 and 0.5 to which mainly non meteorological scatterers contribute. That way the visual contrasts between color categories of |ρhv|^2 is enhanced compared to the categories of |ρhv|. The |ρhv|^2 has been used by early investigators, including the first author, but subsequently abandoned as these advantages were not apparent. Examples of both the |ρhv|^2 and the |ρhv| field will be used to illustrate these points. Included will be the signatures of the melting layer and some tornado debris observations in cases of violent tornadoes.