Developing Self-Similar Ice Scattering Regimes from Spaceborne Radar Observations

Passive microwave precipitation retrievals rely on the ice scattering signal as their primary source of information over land, though its indirect relation to precipitation also makes it a source of global and regional uncertainty. Quantifying this uncertainty is not straightforward owing to the nonunique impact of ice scattering on microwave brightness temperatures and the limited spatiotemporal availability of more direct observations of this signal. Ice scattering responds to the large-scale circulation, offering a way to both resolve the non-uniqueness issue and account for regional uncertainties. This study aims to identify self-similar behaviors in ice scattering properties of precipitating clouds and their relation to atmospheric conditions. Using information from the GPM Combined algorithm, ice scattering information from three tropical land regions is quantified using ice-rain ratio and grouped via K-Means clustering. This yields five clusters with coherent and interpretable spatiotemporal structures across the three regions. Analyses of both the cluster distributions of ice-rain ratio and coincident radar reflectivity and water content profiles indicate the presence of multiple distinct environmental characteristics like monsoonal circulations, deep convection and warm-phase processes. This information is related to airmass characteristics obtainable from reanalysis data, allowing for global assessment of the ice scattering signal and its associated uncertainties.