Improving the Gaussianity of Radar Reflectivity Departures between Observations and Simulations by Using the Symmetric Rain Rate

• DOI: 10.5194/amt-2024-15
• Publication Date: 2024-02-19T09:07:43Z
• Authors: Yudong Gao; Lidou Huyan; Zheng Wu; Bojun Liu
• Keywords: Rain rate

Abstract. Given that the Gaussianity of observation error distribution is fundamental principle some data assimilation and machine learning algorithms, structure radar reflectivity becomes increasingly important with development high resolution forecasts nowcasts convective systems. This study examines discusses what give rise to non-Gaussian by using 6 month observations minus backgrounds (OmBs) composites vertical maximum (CVMRs) in mountainous hilly areas. By following symmetric model all-sky satellite radiance assimilation, we unveil CVMRs as a function rain rates, which average observed simulated rates. Unlike radiance, shows sharper slope light precipitations than moderate precipitations. Thus, three-piecewise fitting more suitable for CVMRs. The probability density functions OmBs normalized rates become Gaussian comparison whole samples. Moreover, possibility third-party predictor construct are also discussed this study. can be further improved accurate precipitation observations. According Jensen-Shannon divergence, linear predictor, logarithmic transformation rate, provide most other predictors.