MRMS Precipitation Estimates Using Specific Attenuation

An algorithm based on the use of specific attenuation A to estimate precipitation in pure rain has been developed using data from Dual Pol S-band radars. Specific attenuation along propagation path in rain is retrieved from the radial profile of radar reflectivity Z and the total span of differential phase along the path. Then rain rate is estimated from the power-law R(A) relation which is very robust with respect to the DSD variability. The technique works in rain only, hence the algorithm uses numerical model data to determine the likely location of the melting layer in order to estimate precipitation below it. In areas where rain is mixed with hail, rainfall rates are estimated using specific differential phase KDP.

The estimate of A is quite sensitive to the choice of the parameter α (the net ratio of A-to-KDP along the propagation path), which relates to the path-integrated attenuation PIA and the total span of differential phase ΔΦ_DP. It is determined automatically in real time using a special optimization procedure that uses Z and ZDR data. Using WSR-88D radar data collected from 37 radars on 56 calendar days over the Eastern U.S., QPE using specific attenuation were generated and compared to operational Dual Pol QPE estimates currently implemented on the WSR-88D fleet. The results indicated that the new R(A) QPE method performed better than the operational Dual Pol algorithm, particularly during moderate to high end precipitation events. Additionally, the new technique exhibited better performance in regions affected by partial blockage of the radar beam. The new technique has been integrated into the Multi-Radar-Multi-Sensor (MRMS) QPE algorithm according to which A is used to estimate precipitation in pure rain, KDP in areas where rain is mixed with hail, and Z, at longer distances from the radar where the radar resolution volume resides either within the melting layer or above. The new algorithm is currently undergoing testing and evaluation on the MRMS real-time research testbed and plans are for the product to be operational by the late fall of 2017.