Köcher et al. (2025) The spatial distribution of convective precipitation – an evaluation of cloud microphysics schemes with polarimetric radar observations

Identification

• Journal: Geoscientific model development
• Year: 2025
• Date: 2025-11-10
• Authors: Gregor Köcher, Tobias Zinner
• DOI: 10.5194/gmd-18-8363-2025

Research Groups

• Meteorologisches Institut, Fakultät für Physik, Ludwig-Maximilians-Universität München, Munich, Germany

Short Summary

This study statistically evaluates five cloud microphysics schemes in the WRF model for simulating convective precipitation events over a 30-day dataset, revealing that the choice of scheme significantly impacts the distribution of precipitation into convective and stratiform regions and their microphysical properties, mainly due to differences in simulated rain drop size distributions.

Objective

• Statistically evaluate the performance of five cloud microphysics schemes in simulating the distribution of precipitation into convective and stratiform regions.
• Assess the microphysical properties within these regions using observed and simulated polarimetric radar signals (reflectivity and differential reflectivity).

Study Configuration

• Spatial Scale: A 144 km x 144 km domain centered over Munich, Germany, with a horizontal grid spacing of 400 m and 40 vertical levels.
• Temporal Scale: 30 days of precipitation events collected during 2019 and 2020.

Methodology and Data

• Models used:
  • Weather Research and Forecasting (WRF) model, version 4.2.
  • Five microphysics schemes: Thompson 2-moment, Morrison 2-moment, Thompson aerosol-aware, Fast spectral bin (FSBM), and Predicted Particle Properties (P3).
  • Cloud-resolving model Radar SIMulator (CR-SIM), version 3.33, used as a forward operator.
• Data sources:
  • Observational: Polarimetric C-band radar data (radar reflectivity (Z) and differential reflectivity (Zdr)) from the German Meteorological Service (DWD) network (Isen radar, southern Germany).
  • Reanalysis: Global Forecast System (GFS) at 0.25° horizontal resolution for WRF initialization.
  • Cell-tracking algorithm: tobac (Sokolowsky et al., 2024) for identifying and segmenting convective and stratiform regions.

Main Results

• While total simulated precipitation amounts were comparable across schemes (295–321 mm, with FSBM at 390 mm), the number of identified convective cells varied by a factor of 2 (5261–11576).
• The choice of microphysics scheme significantly impacts the distribution of precipitation into convective and stratiform regions, with the median convective area fraction (CAF) varying by an order of magnitude between schemes.
• Thompson and P3 schemes exhibit unrealistically high CAFs (e.g., P3 5%, Thompson 3% vs. observed 0.8% at 1500 m altitude), whereas Morrison (0.5%) and FSBM (0.4%) are closer to observations.
• At 1500 m altitude in convective cores, FSBM and Morrison schemes frequently lack large rain drops (low Zdr bias for FSBM, less pronounced for Morrison), while Thompson and P3 schemes simulate too many large rain drops (high Zdr bias), leading to an overestimation of convective reflectivity.
• At 5500 m altitude in convective cores, Thompson schemes show consistent high reflectivity bias, suggesting issues originating in the ice phase (e.g., too large graupel particles). Morrison and FSBM underrepresent graupel, while P3 shows a deficit of convective reflectivities.
• At 1500 m altitude in stratiform regions, all schemes except P3 simulate too few large rain drops (low Zdr bias and reflectivity bias for Morrison and FSBM). Thompson schemes show a low Zdr bias, contrasting with their high bias in convective regions, possibly indicating inefficient transport of large particles to stratiform regions. P3 scheme's Zdr distribution closely resembles observations.
• At 5500 m altitude in stratiform regions, all schemes produce too large areas of reflectivity levels; Morrison shows a notable high bias (approximately 5 dB) in reflectivity, primarily associated with snow particles.
• Analysis of simulated rain drop size distributions (DSDs) at 1500 m confirms that Morrison and FSBM produce more small drops and fewer large drops, while P3 produces more large drops and fewer small drops, consistent with the Zdr interpretations.

Contributions

• Provides the first statistical evaluation of multiple cloud microphysics schemes for convective precipitation events using polarimetric radar observations over a 30-day dataset, offering more robust performance estimates than traditional case studies.
• Demonstrates the significant impact of microphysics scheme choice on the spatial distribution of precipitation (convective vs. stratiform regions) and the underlying microphysical properties.
• Identifies specific biases in rain drop size distributions (DSDs) and ice phase particles for different microphysics schemes using polarimetric radar signals (Z and Zdr).
• Establishes a robust framework for future statistical evaluations of weather models with polarimetric radar observations, extensible to other regions and weather scenarios.

Funding

• Project IcePolCKa (Investigation of convective evolution towards stratiform precipitation using simulations and polarimetric radar observations at C- and Ka-band)
• Special priority program on the Fusion of Radar Polarimetry and Atmospheric Modelling (DFG SPP-2115, PROM)
• Deutsche Forschungsgemeinschaft (grant-no. 408027579)

Citation

@article{Köcher2025spatial,
  author = {Köcher, Gregor and Zinner, Tobias},
  title = {The spatial distribution of convective precipitation – an evaluation of cloud microphysics schemes with polarimetric radar observations},
  journal = {Geoscientific model development},
  year = {2025},
  doi = {10.5194/gmd-18-8363-2025},
  url = {https://doi.org/10.5194/gmd-18-8363-2025}
}