General comments

This paper presents a detailed observation of snow particle motion in order to understand the process of snow cornice formation. It also investigates the conditions under which snow particles adhere to snow cornices through particle-level force analysis. There are few cases where snow cornice formation has been observed, so even though this is a very small-scale experiment in a wind tunnel rather than a full-scale snow cornice, this study is very informative. In addition, it is expected that the detailed force analysis will lead to the construction of a model for the snow cornice formation process, making this work worthy of publication.

Specific comments

There is a big question about the force analysis, which is the main topic of this paper. The wind tunnel experiment in this paper uses dendritic snow particles. Although it is not clearly stated in the paper, my personal experimental experience and personal communications with researchers suggest that snow cornices can only form when dendritic snow particles are used, whereas they do not grow when spherical particles are used. The reason for this is not entirely clear, but it is thought that the large contact surface of dendritic particles makes it easier for snow particles to adhere to each other than for spherical particles. However, this paper discusses the balance of forces assuming that the particles are spherical, so it is possible that the contribution of the contact area of dendritic snow particles is sought in other forces when considering adhesion. To make this paper fruitful, I recommend that the author re-examine whether there are differences in snow cornice formation and adhesion forces between spherical and dendritic particles. Of course, it may not be easy to discuss adhesion forces between dendritic particles, but I expect that the contribution of dendritic shapes can be estimated from the parts that cannot be explained by considering spherical particles.

Technical corrections

Line 19: micr-mechanism -> micro-mechanism?

I appreciate very much for the efforts to observe the particle motion carefully in the wind tunnel and investigate the growing mechanism of the thin snow plate which extends to leeward from the edge. However, according to my observations in the fields and the wind tunnel experiments, it is extremely fragile and always breaks down after growing several centimeters long at the maximum. It never grows to the much larger one, such as, we find on the crest of a ridge along a mountain slope and occasionally causes the avalanche release. The authors need to make clear at the outset that the authors are looking at completely different phenomena. Even though it is allowed to say this miniature as “cornice” as well in a broad sense, the following numerous points should be taken into consideration before the publication.

Line 14: After “and snow cornice”, Seligman et al. (1936) should be put as the reference.

Line 29: “mechanical mechanism” sounds redundant and unnatural, although grammatically correct. I suppose only “mechanism” is fine.

Line 53 to 54: 4 m/s is the wind speed at the center of wind tunnel? If the authors would like to analyze the experimental output physically, the friction velocity u_* should be used in the manuscript instead. Furthermore, the threshold wind speed needs to be specified.

Figure 1: 0.125 m on the side view corresponds to the height of snow on the floor?

Line 118: “190 collision particles”: Are these particles with the snow surface? Such careful explanations are lacking overall. Please check the manuscript again by standing at the place of the readers.

Figure 4 (b): Although it is described that the size of points shows the particle diameter, no explanation are found in the figure caption. Further, the corresponding specific size should be added as well.   

Line 137: Although Stokes number: St is introduced, no explanations about the parameters are shown. Probably, ρp is particle density, d is particle diameter, U is wind speed, m is kinematic viscosity and L is the characteristic length scale. Specify each value you used, particularly L.

It looks the difference of St between the surface and the edge is caused by the particle diameter only. If it is the case, I am wondering it is needed to take the trouble to introduce St. As you see, St is a dimensionless number that characterizes the behavior of particles suspended in a fluid flow. It represents the ratio of the particle's inertial forces to the viscous forces exerted by the fluid. When St is much smaller than one, the particle closely follows the fluid flow, whereas St is much larger than one particle's inertia dominates, and it is less affected by the fluid flow, tending to maintain its original trajectory. On the other hand, at St ≈ 1 the particle behavior lies between these extremes, showing partial coupling with the fluid flow. In this case, both 1.8 and 3.1 are close to one and the difference is quite small; correspond to the third case together. Consequently, I have to say the discussions here are almost meaningless.    

Figures 5 and 7: Generally, the number of particles you analyzed is extremely small. If the authors would like to induce concrete and statistically significant conclusions, at least more than 300 particles data for edge and surface each should be corrected and examined.   

Figure 7: Authors are looking at the impact velocity and the angle separately. I suppose the combination of two factors for the particles on edge and surface reveals interesting findings, supposing the enough data exists.

Further, here, authors set focus on saltating particles only. How do you estimate the contribution of the creep particles?

Figure 8: Please make the figure clearer. Though many parameters are introduced, it is hard to recognize what they mean and indicate, such as, a. In addition, preferably, make the right and left sides reverse to align with Figures 4 and 6.   

Line 190-191: In fact, the wind speed near the surface is getting smaller. However, you cannot neglect the wind shear stress acting there.

Line 191: According to your data, particle speed is quite low (nearly 0.5 m/s), thus, the compression deformation due to the collision is unlikely.

Line 216: Authors say that the sintering time is much longer than the collision time in the manuscript at lines 26-27 and exclude the mechanism to create the wedged-shaped form. I do not understand why it appeared again. It looks contradictory statements.

Furthermore, the static electric forces have been observed in the specific cases so far, and not always.

Although authors tried to introduce all the conceivable forces which may act between particles on edge, time scales are not always consistent. Furthermore, all the discussions are qualitative from beginning to end and no quantitative estimates, which is enough to keep the thin plate growing, are shown.

To say the least, quantitative approval of their idea, based on the snow and environmental data obtained in the experiment, is essential to make the manuscript reasonable and worthwhile.