Precipitation estimation from the ARM distributed radar network during the MC3E campaign

Abstract

This study presents radar-based precipitation estimates collected during the 2-month U.S. Department of Energy Atmospheric Radiation Measurement Program (ARM)-NASAMidlatitude Continental Convective Clouds Experiment (MC3E). Emphasis is on the usefulness of radar observations from the C-band and X-band scanningARMprecipitation radars (CSAPRandXSAPR, respectively) for rainfall estimation products to distances within 100km of the Lamont, Oklahoma, ARM facility. The study utilizes a dense collection of collocated ARM, NASA Global Precipitation Measurement, and nearby surface Oklahoma Mesonet gauge records to evaluate radar-based hourly rainfall products and campaign-optimized methods over individual gauges and for areal rainfall characterizations. Rainfall products are also evaluated against the performance of a regional NWS Weather Surveillance Radar-1988 Doppler (WSR-88D) S-band dual-polarization radar product. Results indicate that the CSAPR system may achieve similar point- and areal-gauge bias and rootmean-square (RMS) error performance to a WSR-88D reference for the variety of MC3E deep convective events sampled. The best campaign rainfall performance was achieved when using radar relations capitalizing on estimates of the specific attenuation from the CSAPR system. The XSAPRs demonstrate limited capabilities, having modest success in comparison with the WSR-88D reference for hourly rainfall accumulations that are under 10mm. All rainfall estimation methods exhibit a reduction by a factor of 1.5-2.5 in RMS errors for areal accumulations over a 15-km2 NASA dense gauge network, with the smallest errors typically associated with dual-polarization radar methods.