In addition, this method does not correct for very high Z values found in hail, leading to the potential for gross over estimation of rain totals where hail is present. False returns, such as found in side lobe contamination, ground clutter and anomolous propagation, may also contaminate the rainfall estimates. Attenuation of the radar signal due to heavy rain or partial beam blockage often leads to an underestimate of rainfall accumulation. Finally, Z-based algorithms are sensitive to radar calibration.
The standard (non-tropical) Z-R relation used by the WSR-88D is Z = 300R1.4. In addition, Z values above 53 dBZ are assumed to be hail and are not considered. This assumption is not often valid and is another possible source of error.
Z/ZDR Algorithms
In high reflectivity regions, differential reflectivity (ZDR) can provide an estimate
of the DSD, reducing errors from calculations using Z alone by a factor of two. However, the estimate is not precise enough
to yield an improvement for rates less than 20 mm h-1. In addition, Z/ZDR estimates can still suffer from
errors due to hail, non-precipitating echoes, attenuation, radar calibration, etc.
KDP Algorithms
Specific differential phase (KDP) is much more directly related to the DSD and rain
rate than either Z or ZDR, particularly in heavy rain. Analysis of experimental data suggests significant
improvement may be possible over Z-based algorithms, with errors as low as 10 to 15 percent for a well-calibrated radar.
KDP/ZDR Algorithms
KDP and ZDR can be used in unison for another rainfall estimate. A complete evaluation of this
procedure and comparison with other procedures has not yet been made.