Reflectivity is a fundamental parameter for sensing the morphology and composition of clouds and precipitation.
However, attenuation due to varying amounts of precipitation, clouds, and water vapor along the propagation path
corrupts reflectivity estimates. In this paper, an algorithm to correct for these effects at 33 and 95 GHz is proposed.
This algorithm is then applied to corrupted reflectivity images collected with the University of Massachusetts Microwave
Remote Sensing Laboratory (MIRSL) Cloud Profiling Radar System (CPRS), which is a dual-frequency (33
and 95 GHz) , fully-polarimetric, pulse-Doppler, ground-based radar. The attenuation correction algorithm consists
of two steps. First, different sources of attenuation along the propagation path are identified by classifying each
image into regions of: air, ice particles, liquid droplets, rain, mixed-phase particles, and insects. This is accomplished
with a rule-based classifier that relies on collocated measurements of velocity, linear depolarization ratio, and height
to make classification decisions. The second step is correcting attenuation along the propagation path in a region
appropriate manner. By starting at the ground with the assumption that the reflectivity estimate is unattenuated,
and working away from the radar adding a region-appropriate amount to the reflectivity estimate at each range
gate, attenuation effects in the image can be largely removed. However, if a mixed-phase region where the rate of
attenuation is unknown is encountered along the propagation path, the correction is suspended and an alternative
approach that corrects attenuation from the top of the cloud down is used. The complete algorithm was applied to
the CPRS data and significantly improved reflectivity estimates.