Answers are grounded in this paper and cite their sources.
Assesses indirect and direct radar assimilation techniques for better rainfall predictions, indicating indirect methods are more effective.
Radar data offer high-resolution spatiotemporal information on the three-dimensional structure of cloud and precipitation systems. Assimilating these data into numerical weather prediction has been shown to substantially increase the accuracy of precipitation forecasts. However, effectively optimizing the model's initial conditions to better represent the true atmospheric state remains a critical challenge. In this study, cycling assimilation and forecasting experiments were conducted via the Ensemble Square Root Filter (EnSRF) method with NCEP Global Ensemble Forecast System (GEFS) forecast data for two heavy precipitation events in the Jianghuai and North China regions. Both indirect and direct assimilation approaches for radar reflectivity were employed to conduct a comparative evaluation. The results indicate that although both methods improve forecast accuracy, the indirect assimilation technique yields more skillful predictions. Specifically, forecasts based on indirect assimilation exhibit reflectivity and precipitation patterns closer to observations, along with higher threat scores for precipitation. Further mechanistic analysis reveals that both assimilation methods effectively enhance the hydrometeor variables of the background field, which improves the analysis of hydrometeor variables toward observed values. However, the direct assimilation approach generates physically inconsistent increments in upper-level hydrometeors, whereas the indirect assimilation produces more reasonable analysis increments and maintains better thermodynamic and dynamic consistency among variable increments. In terms of forecast fields, the direct assimilation introduces unrealistic circulation features, which adversely affect the evolution and trajectory of meteorological systems, ultimately leading to less accurate forecasts. Overall, within the context of this study, the results suggest that the indirect assimilation approach may be a more robust and effective strategy for leveraging radar reflectivity data to improve convective-scale numerical weather prediction. Further investigation with more cases is needed to confirm its general applicability.
No takes yet. Share an insight, caveat, or question.
Zhu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: