Hydrometeor Classification Algorithms (HCA) Primer

Most HCAs use a "fuzzy logic" scheme to classify hydrometeor type. Fuzzy logic algorithms are used to quantify uncertainty, much like a system of probablistic equations. In the case of hydrometeor classification, the exact boundaries between hydrometeor types using polarimetric variables are "fuzzy", so a zero to one probability must be assigned in the "fuzzy" region.

For example, consider the discrimination between rain and hail. Using differential reflectivity ZDR alone is not helpful. Values of ZDR near zero may suggest the presence of either hail or drizzle/cloud droplets, because both have nearly spherical shape.

Combining reflectivity factor Z with ZDR can help, because hail has a much larger Z than drizzle/cloud droplets. However, consider a case where Z= 53 dBZ and ZDR= 1.5 dB. Studies have shown this combination of Z and ZDR may result from heavy rain, small hail, or a mixture of both. In assigning hydrometeor type to this case , the fuzzy logic algorithm would assign a lower weight to Z vs. ZDR than to potentially more useful combinations, such as Z vs. KDP.

Once all of the potential combinations of polarimetric variables are tested, weighting functions are applied to make the final classification, such that the highest confidence combination is weighted the most, and the lowest confidence is weighted least.

Different hydrometeor classification algorithms are constantly being developed and tested. Early results are encouraging, but there is still insufficient in-situ verification of the algorithm output to be certain the output is always accurate. Therefore, forecaster feedback about the quality and usefulness of the tested HCA is very important!

The NSSL WSR-88D Polarimetric Radar (KOUN) HCA has a "warm" mode and a "cold" mode.

NSSL Hydrometeor Classification Algorithm - Warm Mode NSSL Hydrometeor Classification Algorithm - Cold Mode

Back to the main page.