The authors using aircraft observations investigate the spatial variability of riming and its contribution to the clustering of ice within clouds. They employ observations from IMPACTS and HALOAC3 where collocated aircraft observations are unique for the analysis. Although I feel this work has some new contributions to the community, the current presentation and organization need major revisions. More detailed comments may be given after addressing these key issues. Sorrying for the delayed post, I tried to understand this work by reading it for multiple times, but it is very frustrating in interpreting its core logics and idea.   

Major comments.

1. The research motivation is poorly structured. The introduction just lists all relevant topics, from spatial distribution of MPC, properties of MPC in mid-latitudes and Arctic, to the data you have, and then riming. I do not see a clear logic and strong motivation for this research. Frankly, it left me the impression that you did the analysis just because you have these data available.

Also, the introduction omits very key details on the unique datasets from IMPACTS and HALO-(AC)3. So far, we have so many aircraft observations, why just these two campaigns fit your study?

2. Causality issue. Although the analysis and results are no doubt interesting, I question the statements in many places such as riming enhances the probability of IWC clusters. What I can expect is some dynamical mechanisms such as the generating cells as discussed influence the IWC clustering, and such mechanisms influence riming and clustering. You may say that IWC clusters at certain scales are rimed, but I do not agree with the reasoning that the clustering of ice (macrophysics) is influenced by riming (microphysics).

Similar causal issues apply to many statements on the relationship between riming and IWC. In many places, the authors state that riming affects/influence IWC variability, however, these is no given evidence showing the IWC variability is due to riming. Riming is one of the characteristics of the ice clusters, not the factor leading to the variability.

3. Some key methods lack details or are arbitrary.

(1) Quantifying riming. After reading several times of the section 3.1, I have no idea what is the combined method for quantifying riming. I am frustrated in understanding the logic.

(2) Quantifying IWC variability. This key statement at L350 lacks physics background. The logic of interpreting signs of η is straightforward. However, it is arbitrary to interpret the sign of η1 – η2 in the same way, since a positive η1 – η2 can be the results of two negative η.

4. Lacks of in-depth analysis of the observations from the two campaigns. The authors tried to use data from two campaigns for the analysis. However, I do no see clear physics explaining the observed differences between the two campaigns, nor general conclusions given.

Minor comments

• L28&29 I understand that there are very few studies on the spatial distributions of ice and liquid mass, but I do not appreciate inappropriate citations. The listed ones have discussed the impacts of different cloud phases, they did not mention the spatial distribution.
• L43 Literature should be given. Also, I do not agree the statement. Smaller-scale bands are mostly linked to dynamics and associated microphysics.
• Section 3.3 Sensitivity study is poorly structured. It is difficult to capture the logic in the present format. It seems that this section was splitted into two parts. It is recommended to combine section 3.3 and section 4.2.
• L263 Awkward logic. MPC properties vary between IMPACTS and HALO-(AC)3 just because of different synoptic situations (Sect. 2.3) and measurement locations? This is very misleading.
• Figure 5. Why the boxplot of W-band retrieval in IMPACTS is different from others? You did not explain it in the caption.
• L394 this conclusion lacks evidence.
• Figure 11. Observations from two campaigns were analyzed, but you only show the conceptual diagram for HALOAC3. This is something of an anticlimax.