Performance of S-band Ground-Based Radar Precipitation Rate Retrieval Algorithms over a Dense Gauge Array

Significant progress has been made in the past two decades in developing quantitative precipitation estimation (QPE) algorithms and estimating rainfall using dual-polarization radars. Traditional approaches for QPE include using dual-polarimetric variables horizontal reflectivity Z_h, differential reflectivity Z_dr, and specific differential phase K_dp. Using a combination of dual-polarimetric variables for QPE such as R(Z_h), R(Z_h, Z_dr), R(K_dp), and R(K_dp, Z_dr) help mitigate uncertainties related to drop-size-distribution (DSD) variability. In the last five years, a new polarimetric method to estimate rainfall was proposed in literature using specific attenuation or R(A).

The Global Precipitation Measurement (GPM) mission Ground Validation (GV) supersite at NASA Wallops Flight Facility (WFF) has a multi-year record of precipitation observations dating back to 2014. The Precipitation Research Facility (PRF) at WFF serves as a semi-permanent laboratory with numerous ground-based instruments including the NASA Polarimetric (NPOL) research quality S-band radar. This study will quantify QPE algorithm performance from four radar-based algorithms by validating against the GPM GV 25km² quality-controlled gauge array located 30 km from NPOL. The gauge array dimensions are identical to the 25 km² footprint GPM Dual-frequency Precipitation Radar (DPR). Two traditional radar-based approaches that use a combination of dual-polarimetric variables for QPE will be compared to a snapshot DSD estimation method and a relatively new technique of rate estimation that uses specific attenuation.

The System for Integrating Multiplatform Data to Build the Atmospheric Column (SIMBA), developed in support of NASA’s GPM mission GV program will be used to fuse high temporal (3 minute) and spatial (250 m) resolution PPI radar data over the gauge array along with 20 tipping bucket gauges to construct a 3-dimensional grid to inter-compare QPE near the surface. The gauge 1 min rain-rates will be integrated to 5, 10, and 15 min rates and compared to instantaneous NPOL rates estimated from the four algorithms and averaged over similar time scales at 0.5 and 1.0 km above the ground over the 25 km² cartesian grid. The correlation and bias metric will be used to evaluate the performance of the radar rain rate algorithms from all events dating back to 2014. Specific attention will be focused on the attenuation-based algorithm and how well it performs against traditional methods.