Radar-Driven Temperature Nudging for Nowcasting Winter Precipitation

Winter time hydrometeor classification at the surface is highly reliant on detection of temperature variability aloft. The presence of melting and refreezing layers can change the phase of the hydrometeors from solid to liquid or vice-versa at the surface. However, detection of these layers can be problematic during winter when melting layers are typically shallow. Weather radars often have insufficient resolution inside the melting layer for its accurate detection or tend to overshoot these layers altogether. The presence of a possible refreezing layer presents an additional difficulty, as these layers are not easily identified with traditional plane position indicator (PPI) scans. Refreezing layers are relatively rare, but important for distinguishing freezing rain from ice pellets. While multiple melting layer detection algorithms exist, refreezing layer detection methods have yet to be developed for operational use. The new hybrid phase transition detection algorithm exploits Quasi-vertical profiles (QVPs) of reflectivity, differential reflectivity, and co-polar correlation at different elevation angles. QVPs offer the opportunity to observe small enhancements of polarimetric properties often overlooked in PPI scans, where short dwell times produce errors of polarimetric variables, and to reveal to temporal trends. The suggested method was tested on several cases in Germany, including documented freezing rain, ice pellets and wet snow events. The quality of the QVP melting layer detection was then compared to the traditional radar methods for ML designation and radio soundings. The procedure was further developed in order to detect the presence of the refreezing layer, characterised by small bumps in Z_DR, and dips in Z_H and 𝜌_hv. The locations of the melting and refreezing layers are then used to nudge the temperature profiles obtained from the regional weather prediction model, COSMO-DE. The corrected temperature profile is further implemented into a spectral bin model to predict surface precipitation type. The spectral bin model predicts surface precipitation type based from profiles of atmospheric profiles, and accurate profiles of temperature are essential for classification. The hybrid phase transition algorithm presented here identifies both melting and refreezing layers, which when used in tandem with other methods, further improves the accuracy and reliability of the phase change detection aloft, and allows a radar driven temperature nudge to model profiles.