| 09 Feb 2026
Evaluation of ice hydrometeor retrieval using multi-band radar and millimeter-wave radiometer measurements from the IMPACTS campaign
Abstract. Understanding the microphysical properties of ice hydrometeors remains challenging. This study develops and evaluates a cloud-ice retrieval algorithm that synergistically uses W-band radar, Ku/Ka-band radar, and millimeter-wave radiometer observations. This strategy enables to observe deep inside precipitating clouds unreachable by conventional combined radar–lidar measurements due to severe attenuation.
The retrieved cloud microphysical parameters are compared with aircraft in situ measurements from the Investigation of Microphysics and Precipitation for Atlantic Coast-Threatening Snowstorms (IMPACTS) campaign. The bias between retrieved values and in situ measurements varied by roughly two orders of magnitude depending on the particle habit assumptions. A mixture of rosette and snowflake habits based on physical considerations yielded the best overall performance. Under this habit assumption, the mean ratios of the retrieved values to in situ probe measurements of ice water content (IWC), total number concentration (Nt), mass-weighted mean diameter (Dm) are 1.01, 0.97, and 1.05. The mean bias and RMSE for terminal fall velocity (Vt) are −0.02 m s⁻¹ and 0.30 m s⁻¹. Forward-simulated measurements from the retrieved profiles reproduce the actual radars and radiometer observations well, confirming the self-consistency of the current algorithm.
A theoretical sensitivity analysis demonstrates that the radar–radiometer synergy becomes particularly effective in deep cloud layers where particle sizes are large, ensuring that the multi-sensor retrieval outperform the W-band radar-only retrieval especially deep inside clouds. These results highlight the potential of combining multi-band radar and millimeter-wave radiometer observations to advance our understanding of ice-hydrometeor microphysics in deep precipitating clouds.
The manuscript presents an extended version of the authors’ combined radar–radiometer ice hydrometeor retrieval applied to observations from the IMPACTS campaign.
The retrieval algorithm, previously introduced in an earlier study, is extended with an updated method for generating the initial guess, a new mixed particle habit, and the capability to ingest Ku- and Ka-band radar observations.
The authors demonstrate convincingly that the algorithm yields good agreement with in-situ measurements. The manuscript is clearly written and the presentation meets the standards of the journal. I therefore consider the manuscript suitable for publication after minor revisions.
## General Comments:
- The manuscript would benefit from a discussion of the simplifying assumptions underlying the combined retrieval algorithm and how these assumptions may affect the results. For example, the algorithm currently does not retrieve temperature or water vapor, which will introduce forward-model errors in the passive microwave (PMW) observations. While some of these aspects are discussed in O25, it would be helpful to briefly summarize them here and refer to that work. Please also describe how surface emissivities are modeled.
- The current results do not show a clear benefit of the triple-radar combined retrieval compared to the retrieval using only W-band observations. This may be related to the fact that the IMPACTS campaign primarily sampled winter storms, where Ku- and Ka-band observations may provide limited additional sensitivity. However, this point should be discussed explicitly, as the current results suggest that W-band + PMW observations may already be sufficient to retrieve ice microphysics with good accuracy.
- While the study shows that the mixed particle habit yields the best agreement with the in-situ measurements, an important scientific question is whether the available observations alone are sufficient to constrain particle habit. I therefore suggest adding a plot similar to Figure 9 that shows the bias and mean-squared error of the simulated PMW observations for the different habit assumptions.
## Specific Comments
- l. 117: It may be helpful to note that these correspond to the high-frequency channels of GMI, since GMI includes additional lower-frequency channels.
- Figure 2:
- Please increase the line width/marker size in panel (a), (b), (e).
- Please also increase the font size and use vector graphics or higher-resolution images to improve readability.
- l. 374: In Fig. 2b, I only see one strong updraft that actually intersects the flight path. Could the observed biases simply result from the in-situ measurements being taken near the cloud base?
- l. 486: This statement appears to contradict the discussion in l.296. Please clarify.