Retrieval of Fine-Scale Microphysical Characterization of Clouds with Combination of Terahertz Radar Observation and Multi-Component Gaussian Decomposition

Cloud microphysical evolution occurs on short time scales and within shallow layers, where cloud radars provide vertically resolved obser-vations and Doppler spectra contain information beyond bulk moments through their linkage to hydrometeor fall speeds and vertical air motion. In warm clouds, however, Doppler spectra are often spectrally structured, exhibiting multi-peak signatures and pronounced skewness caused by closely spaced droplet sub-populations together with finite velocity resolution and spectral broadening. Terahertz cloud radars improve the observability of such fine-scale spectral morphology and its rapid variability through enhanced Doppler sensitivity, finer spatial resolution, and higher attainable temporal sampling for a given velocity resolution. We develop a retrieval framework that treats terahertz-resolved spectral morphology as the information carrier and converts multi-peaked spectra into component-wise microphysics. A Multi-component Gaussian Decomposition provides a label-free spectral parameterization as a weighted sum of Gaussian spectral components, which are then constrained by a physics-based forward model linking droplet size to Doppler velocity with scattering weights, Gaussian broadening, and reflectivity integral closure. After range-resolved two-way gas-attenuation correction, log-normal droplet size distributions are retrieved for each com-ponent and used to derive liquid water content, total number concentration, and mass-weighted mean diameter. Dominant-component statistics further summarize time-resolved spectral evolution and support stage-dependent interpretation of a warm-cloud drizzle lifecycle. The results highlight the potential of terahertz spectral-morphology analysis for constraining warm-cloud initiation pathways and informing sub-grid parameterizations