Preprints
https://doi.org/10.5194/egusphere-2026-6107
https://doi.org/10.5194/egusphere-2026-6107
22 Nov 2026
 | 22 Nov 2026
Status: this preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).

Insights on Convective–Stratiform Designation and Its Ambiguity from Radar Wind Profilers at Darwin, Australia

Scott Edward Giangrande, Christopher R. Williams, Israel Silber, and Alain Protat

Abstract. Convective–stratiform (C-S) classifications are widely used in deep convective cloud remote sensing studies and model evaluation using observations. Most C-S classification (CSC) relies on a limited set of canonical signatures. Dual-frequency wind profiler observations collected at Darwin, Australia, provide a unique opportunity to constrain C-S designation by directly retrieving hydrometeor fall speed Vt, which may better identify the presence of faster falling, convective-processed media. Here, Random Forest (RF) methods are used to compare how well simple conventional radar features reproduce these Vt-informed classifications.

Reflectivity factor (Z) properties associated with the vertical structure of precipitation outperform simpler measures of precipitation intensity. Combinations of Z-based characteristics often exceed 80 % skill across multiple C-S predictor sets. Mean Doppler velocity (MDV) provides additional discriminatory value when strong vertical motions (>2–3 ms-1) are identified above the melting level, with MDV pairings often exceeding 90 % skill in recovering C-S regimes. Analogous tests using Vt provide higher performance, though genuine Vt insight aloft provides negligible improvement over other Doppler fields once additional predictors are considered. These results suggest that direct Vt insight is not a requirement for streamlining CSC model-observational evaluation. Further, most RF predictions strongly favor one C-S class, with such decisive predictions agreeing with our reference >95 % of the time. Disagreement exceeds 30 % among less-decisive predictions, where characteristic C-S structures also become less distinct. This study considers RF prediction ambiguity as one means for reconciling binary C-S designations with the fluid nature of these precipitation states.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Share
Scott Edward Giangrande, Christopher R. Williams, Israel Silber, and Alain Protat

Status: open (until 03 Jan 2027)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
Scott Edward Giangrande, Christopher R. Williams, Israel Silber, and Alain Protat
Scott Edward Giangrande, Christopher R. Williams, Israel Silber, and Alain Protat
Metrics will be available soon.
Latest update: 22 Nov 2026
Download
Short summary
Deep convective precipitation spans a continuum of dynamical and microphysical processes, yet many applications rely on simple convective–stratiform classifications. Using unique Darwin observations, we show that a few measurements distinguish these regimes well for most precipitation while identifying where the distinction becomes less clear. Quantifying this ambiguity helps retain useful binary classifications while recognizing their physical limitations.
Share