Quality Control Method for S-band Dual-Polarization Radar Based on Linear Programming
Abstract：To improve the data quality of S-band dual-polarization radar parameters, a quality control method based on linear programming (LP) was designed to process the differential propagation phase shift (ΦDP) and calculate the differential propagation phase shift rate (KDP) in segments. First, by calculating the texture features of ΦDP and differential reflectivity (ZDR), it was determined that setting the constraints of sliding standard deviation SD(ZDR) < 1 dB and SD(ΦDP) < 5° could effectively distinguish between meteorological and non-meteorological clutter. Then, based on the physical property that radial ΦDP should accumulate in precipitation areas, LP processing with third-order difference filtering can suppress ΦDP fluctuations, thus maintaining the physical properties of ΦDP and the non-negativity of KDP. A segmented approach was used to recalculate KDP values: in light to moderate rain areas, KDP values were recalculated using LP-processed ΦDP data; in heavy rain areas, the original KDP values were retained. This segmented processing method can eliminate the sawtooth fluctuations of ΦDP in light to moderate rain areas, ensuring the non-negativity of KDP values, while avoiding the problem of KDP values being too small due to excessive smoothing in heavy rain areas, thereby optimizing the overall calculation effect of KDP. After testing with data from seven heavy rain events between 2016 and 2020, the normalized error (NE) of R(KDP) decreased from 59.54% to 50.52%, the root mean square error (RMSE) decreased from 4.575 to 4.274, and the correlation coefficient (CC) increased from 0.779 to 0.872, indicating that this quality control method helps improve the accuracy of quantitative precipitation estimation (QPE).
Keywords：Atmospheric detection and instrumentsRadar observation parametersLinear programmingQuantitative precipitation estimationSliding standard deviationS-band dual-polarization radar
Publication Date：2025-12-20
Online Publishing Date：2026-01-07(First online date of this platform, not the publication date of the document)
Pages：6( 79-84 )