The Use of Bulk Zdr to Mitigate Biases in MRMS Specific Attenuation Based QPE

The MRMS Dual Pol radar based quantitative precipitation estimates (QPE) utilizes specific attenuation to estimate precipitation below the melting layer, specific differential phase wherever convective cores have reflectivities (Z) ≥ 50 dBZ’s, and a selection of reflectivity-to-rain rate (Z-R) relations within and above the melting layer. This product has been operational within the National Weather Service (NWS) since October 2020. While the QPE has done a good job estimating precipitation for the majority of precipitation, especially widespread heavy rains, a number of cases have shown significant biases. For example, there is a wet bias for a few convective cases where the parameter alpha, used in the path integrated attenuation (PIA), did not match the environment as verified by quality-controlled rain gauges. In these cases, the Zdr versus Z slope, used to estimate alpha when convective and stratiform rain are present, is quite flat. This causes alpha to be high which, in-turn, leads to high rain rates. However, the bulk Zdr was generally larger than would be expected for very efficient rainfall cases.

A wet bias was also found in cases where weak convection was mixed with large areas of stratiform rain across the radar field of view (FOV). In these cases, there were not sufficient Zdr-Z samples in the high reflectivity range to obtain a robust Zdr-Z slope, thus a default alpha (0.035) typically associated with more efficient stratiform rain was applied. Originally, this value was selected in PIA calculations to minimize underestimates in stratiform rain along warm maritime coastlines as well as precipitation associated with tropical cyclones or their remnants. However, it is found to cause overestimation when weak convection was present.

This study shows how the use of bulk Zdr properties in the low (10 – 30 dBZ’s) and high (40 – 50 dBZ’s) reflectivity bands can be used to mitigate the aforementioned challenges. For the first challenge, bulk Zdr for low and high Z bands were used to ensure the slope derived alpha would better match the environment leading to more accurate hourly rain totals. For stratiform rainfall, bulk Zdr was also used to determine values of alpha used to calculate rain rates across the radar FOV. The initial results of the experimental QPE for 25 cases indicates significant improvement in reducing errors and biases previously observed in the MRMS operational QPE.