The accuracy of radar reflectivity measurements impacts quantitative precipitation estimates that rely on reflectivity-to-rainfall equations, especially with heavy rainfall rates. These precipitation estimates are used operationally by the National Weather Service in the U.S. for flash flood forecasting and warning, by the modeling community for verification of quantitative precipitation forecasts and more recently for assimilation into numerical weather prediction models. Recent findings indicate variability and consistent biases in the calibration of some Weather Surveillance Radars-1998 Doppler (WSR-88Ds) in the U.S. Each of these radars can be calibrated by the use of scattering objects that have well-known properties to radar. In this paper, we present a software solution to diagnosing relative calibration differences for a given radar network and show initial results for approximately 25 WSR-88D radars over a 6-month period of data collection.

Initially, grid points between adjacent WSR-88D radars that have the same locations in 3D space and similar beam volumes are identified. If the adjacent radars are calibrated well, then these collocated grid points should have similar reflectivity values when examining a long record with significant precipitation. Preliminary results show that the reflectivity differences can be consistently large suggesting that at least one of the adjacent radars is out of calibration. It is hoped that this software can be used in conjunction with absolute calibration methods to calibrate an entire network in a timely, cost-effective manner.