Uncertainties in Microwave Properties of Frozen Precipitation: Implications for Remote Sensing and Data Assimilation

Abstract

Abstract A combined active/passive modeling system that converts CloudSat observations to simulated microwave brightness temperatures (TB) is used to assess different ice particle models under precipitating conditions. Simulation results indicate that certain ice models (e.g., low-density spheres) produce excessive scattering and implausibly low simulated TBs for stratiform precipitation events owing to excessive derived ice water paths (IWPs), while other ice models produce unphysical TB depressions due to the combined effects of elevated derived IWP and excessive particle size distribution–averaged extinction. An ensemble of nonspherical ice particle models, however, consistently produces realistic results under most circumstances and adequately captures the radiative properties of frozen hydrometeors associated with precipitation—with the possible exception of very high IWP events. Large derived IWP uncertainties exceeding 60% are also noted and may indicate IWP retrieval accuracy deficiencies using high-frequency passive microwave observations. Simulated TB uncertainties due to the ice particle model ensemble members approach 9 (5) K at 89 (157) GHz for high ice water path conditions associated with snowfall and ∼2–3 (∼1–2) K under typical stratiform rain conditions. These uncertainties, however, display considerable variability owing to ice water path, precipitation type, satellite zenith angle, and frequency. Comparisons between 157-GHz simulations and observations under precipitating conditions produce low biases (<1.5 K) and high correlations, but lower-frequency channels display consistent negative biases of 3–4 K in precipitating regions. Sample error correlations and covariance matrices for select microwave frequencies also show strong functional relationships with ice water path and variability depending on precipitation type.