| 24 Jun 2026
Vertical Structure and Evolution of the Stratus-to-Fog-to-Stratus Transition: first high-resolution UAV-LCPIS Measurements over the Yellow Sea coast
Abstract. This study presents, for the first time, high‑resolution concurrent vertical profiles of temperature, humidity, water vapour, and fog/cloud droplet microphysics during a stratus‑fog‑stratus transition over the Yellow Sea (15–16 May 2025). The primary observational platform relied on a novel Lightweight Cloud Particle Imager sonde (LCPIS) deployed on an unmanned aerial vehicle (UAV) system. The event progressed through three distinct stages: pre‑fog stratus, dense fog (minimum visibility ~100 m), and lifting to stratus. Pre‑fog stratus featured a shallow mixed layer (Ri < 0.25) capped by a strong inversion (Ri > 0.25), trapping moisture below 300 m. During the fog transition, the inversion base stabilized at 270 m with a lapse rate of 5.0 °C·(100 m)⁻¹, while the saturated layer extended to the surface. Within the fog layer, the Richardson number (Ri) was negative, indicating vigorous turbulent mixing, while Ri > 0.25 near the fog top formed a dynamically stable cap. After dissipation, Ri turned positive and evaporated moisture re‑formed stratus. Peak droplet number concentration (NC) increased from 50–100 cm-3 (pre‑fog) to 100–120 cm-3 (fog), liquid water content (LWC) doubled from 0.1–0.2 to 0.2–0.45 cm-3, and mean diameter ranged 10–20 μm. Microphysical peaks occurred at ~200 m in both pre‑fog and fog stages. Key transition indicators include downward extension of the saturated layer to the surface, rise of the zero dew‑point depression height, and NC stabilization with LWC increase. This study provides the first synchronized vertical observations of thermodynamic, water vapour and microphysical parameters during a stratus-fog-stratus transition, offering a robust observational basis for improving sea fog forecasting.
This manuscript presents unique and valuable field observations of the bidirectional stratus-fog-stratus (SFS) transition along the Yellow Sea coast using a newly developed, UAV-mounted Lightweight Cloud Particle Imager Sonde (LCPIS). By bridging critical observational gaps in the lowest 100–300 m of the Marine Atmospheric Boundary Layer (MABL), this study delivers synchronized, high-resolution vertical profiles of thermodynamic, water vapor, and microphysical parameters. The authors integrate these UAV soundings with a comprehensive multi-platform suite—including ground microphysics, Ka-band cloud radar, FY-4B geostationary satellite imagery, and ERA5 reanalysis—providing a rich observational foundation for regional sea fog research and NWP model evaluation.
Despite the evident novelty of the observational system and dataset, several critical flaws weaken the reliability of core physical interpretations, reduce methodological reproducibility, and degrade manuscript readability. Specifically, the causal attribution of boundary-layer dynamics to radiative cooling lacks direct flux validation and prominent typographical and figure-caption errors exist. Full resolution of the major and minor comments detailed below are required before the manuscript can be considered suitable for publication.
Major Comments
The manuscript repeatedly invokes cloud-top radiative cooling as the driver of turbulent mixing (e.g., lines 405-410: “This instability is likely attributable to top-down convection induced by cloud-top radiative cooling”; line 510: “the core of stage 2 is the positive feedback triggered by radiative cooling at the fog top”). However, the LCPIS does not measure radiation fluxes, and no radiation data (e.g., from satellite or radiative transfer model) are presented to support this attribution. The negative Ri values could equally arise from strong vertical wind shear (dynamic instability) within the fog layer—a mechanism that does not require radiative cooling.
Recommendation: The authors should: (a) explicitly state that cloud-top radiative cooling is inferred rather than measured; (b) discuss whether the observed negative Ri is more likely due to shear or buoyancy based on the vertical wind shear magnitudes.
While the paper mentions a 1 Hz sensor sampling rate and 20 m vertical block-averaging (Lines 325–328), critical sampling metadata are missing. It is unclear whether profiling flights (SP01–SP09) were conducted during continuous vertical ascents/descents or step-and-stare hovering segments.
Recommendation: Add a detailed summary table in Section 2.1 or 3.1 specifying exact UAV vertical ascent/descent rates (m s⁻¹), hover durations (if any), maximum flight ceilings, and total profile durations for flights SP01–SP09.
The abstract states: “liquid water content (LWC) doubled from 0.1−0.2 to 0.2−0.45 cm−3”. The unit for LWC is clearly g⋅m−3 (as used throughout the manuscript), not cm−3. This is likely a simple typographical error, but it appears in a prominent location and could cause confusion.
Recommendation: Perform a comprehensive manuscript-wide audit to standardize units, mathematical symbols, and formatting across all text, equations (Eq. 1–8), figure axes, table headers, and appendix figures. Include a Nomenclature Table in the Appendix listing all key thermodynamic and microphysical variables, their symbols, and SI units.
Technical Revisions