The authors investigated the ice nucleation ability of large size dust particles. The results are interesting and contribute towards advancement of ice nucleation research. I have couple of comments and suggest the publication.

1. It is not clear why various techniques are used to probe the INP efficiency.
2. Section 2.1: What was the detection efficiency of Malvern instrument. Did this instrument measured all the particles that went through the laser beam. Did the PSD have any errors/uncertainties?
3. Abstract: We know that size plays an important role in ice nucleation and this study supports this premise. The abstract section could be improved, discussing why size plays an important role. It is also important to discuss why Moroccan sample showed not size dependency, but Icelandic showed subtle size dependence.
4. Large particles often are observed coated with secondary compounds like organics and sulfates because they provide larger surface area. Are these coated particles poor INP? To further improve the implications of this study what are the limitations to using larger size particles? Some discussion along this line in the Introduction section would be useful to understand the atmospheric implications of large size particles.

The paper investigates the immersion freezing ice‑nucleating activity (INA) of natural mineral dust from Morocco (low latitude) and Iceland (high latitude). The authors focused on coarse and super‑coarse particles and how INA relates to particle size and mineralogy. It combines a large number of experiments in multiple instruments and shows that Moroccan dust exhibits INA comparable to K‑feldspar with no clear size dependence, whereas Icelandic dust is less ice active, with a weaker, mineralogy‑linked size dependence. This work is important because it extends INA characterization to larger particle sizes and high‑latitude dust sources that are poorly constrained in current INP parameterizations. Overall, the manuscript could be made more concise to improve readability and easier to follow. Authors pointed out importance of how INAS density calculations can be revised for coarse and "super-coarse" particles. The study would benefit from clearer, quantitative discussion of size effects, a more explicit separation of instrument‑related differences from true dust‑property differences.

Here are the comments to consider:

I feel the restructuring of the paper and highlights the main points would improve the readability of the paper. Maybe move some of the more detailed derivations, intermediate equations, or sensitivity analyses to the Appendix and keep the main text focused on key steps and conclusions.

For Morocco, you state that INA shows no dependence on particle size, while for Iceland you find a “subtle” size dependence linked to pyroxene content. Given the central role of size in the paper, I suggest bit more quantitative discussion of this if possible. For example, include more explicit statistics for the size effect.

Moroccan samples showed no dependence on particle size, likely because the K-feldspar content remained constant across all sizes. How confident are you about your composition measurements for larger particles?

Emphasizing what is new for high‑latitude dust parameterizations (e.g., capturing pyroxene‑associated variability, broader spread in ns) and what that implies for Arctic/boreal mixed‑phase cloud simulations.

You showed an important comparison of INAS densities from AIDA and INSEKT and show systematic offsets. However, it is sometimes unclear whether observed differences in INA are driven more by sample properties (size, mineralogy) or by methodological differences between instruments.

In its current form, the manuscript focuses heavily on instrument intercomparisons and systematic biases, ns calculations, giving it the tone of a technical report rather than a science-driven study. It would strengthen the paper to expand the scientific discussion, particularly regarding the implications of these biases.  

The link to atmospheric relevance part could be improved. For example, how using your Morocco vs Iceland parameterizations would change predicted INP concentrations in a simple idealized case?