Radar-derived Quantitative Precipitation Estimation Using a Hybrid Rate Estimator Based on Hydrometeor Type

We have developed a dual-polarization radar-based QPE method that makes use of the estimated hydrometeor type to determine which precipitation estimator is relevant at each radar gate location. Algorithms of this type are now quite widely reported in the literature. Initially we used the NCAR Particle ID (PID) algorithm for estimating the hydrometeor type. Based on this we developed two estimators: (a) a hybrid estimator that selects the relevant rate relationship based on the dominant hydrometeor type; and (b) a weighted estimator that uses the weight associated with each hydrometeor type (from the classification algorithm) to produce a weighted mean of the various precipitation rate estimators. The estimated rates in radar space are then projected to the surface using a logic sequence that takes into account data quality, hail contamination, beam blockage and beam height. The algorithms were tested during the Plains Elevated Convection At Night (PECAN) project in Kansas during the summer of 2015, and the 24-hour accumulation totals were compared with gauge measurements. The skill was promising, with biases of the order of 5 to 10% compared with the gauges. However, there was some evidence that the melting layer was biasing the results. Therefore we have added improvements to the algorithm to handle corrections to the melting layer. And we have also added an option to use the NEXRAD HCA algorithm instead of the NCAR PID algorithm. Once again, we ran the algorith for the PECAN data set, and scored the results against the gauges. The results are presented.