Title: Multiscale Assessment of Indian monsoon rainfall using ICON and CMIP6 model simulations 

By: Pokhrel et al 

Reviewer: Sugata Narsey, Australian Bureau of Meteorology 

Recommendation: Accept with minor revisions 

In this manuscript the authors systematically evaluate Indian monsoon rainfall at multiple time scales in the ICON non-hydrostatic model and 5 CMIP6 models under a prescribed SST set-up. They explore whether a non-hydrostatic model, and/or higher spatial resolution, improves the simulation of Indian monsoon rainfall across scales. Their findings are mixed – while ICON shows some improvements over the CMIP6 models it also has some larger deficiencies at some time scales and in geographic locations. Likewise, higher resolution alone was not a panacea for Indian monsoon rainfall, with relatively coarse models such as IITM-ESM outperforming or matching other models in some areas. The manuscript is well-written and presented, the analysis is of a high quality, the work is novel, and the conclusions are well-supported by evidence. I recommend that the manuscript is accepted with minor revisions. 

Main comments: 

1. While the multiscale error assessment is useful, I would find this manuscript to be more useful overall if there were a summary figure or analysis specifically showing the cross-scale relationships of errors, e.g. a heatmap of metrics for each model. I would cite this paper for that specifically! The abstract asks how scale-dependent biases shape monsoon variability, but I feel that this was not addressed clearly in the paper (but could be with one more figure). 

1. A little more clarity on the model choices and methodological choices earlier in the manuscript would improve readability. Why ICON? Will it be used for some other purpose soon? Why these 5 CMIP models? Yes they vary across scales, but is this an ensemble of convenience (data availability) or is it a model subset choice? If the latter then justify it. Harmonic analysis is ubiquitous in math and physics, but less common in atmospheric science than perhaps it should be. Add a few references showing examples of how it has usefully been applied to gain insights. 

1. Well done on a nice manuscript! I enjoyed reading it and learned some interesting things about Indian monsoon rainfall. 

Specific comments: 

L27: what does block level mean? 

L31: N. 2005 is a strange way to reference. 

L45: There is some evidence that monsoon depressions are barotropic:  

Diaz, M., & Boos, W. R. (2021). The influence of surface heat fluxes on the growth of idealized monsoon. Journal of the Atmospheric Sciences. https://doi.org/10.1175/JAS-D-20-0359.1 

L56: no need for etc 

L75: This is a key point that I think could be made much earlier!! Forgot full stop. 

L102: why these 5 models? And what is the motivation for ICON specifically? 

L170: can you provide some justification for using harmonic analysis? Some citations or examples of previous use. Pros and cons of this vs other potential methods. 

L173: throughout the manuscript, please use the acronyms from Table 1. 

Fig 2: I wonder how these biases would look as a percentage... I suspect ICON would look better! 

L384: full stop. 

L577: i can see how this is suggested by results, but its not clearly shown. here, and throughout the manuscript, the link between two or more evaluation metrics could be displayed through a figure (scatter plots for example). 

L640: what specifically do you mean by higher-frequency variance? these two statements seem circular, since higher amplitude surely means higher variance. 

L645: this seems important. i would clarify this earlier, perhaps even in the methodology. 

L671: is this shown somewhere? I guess this indicates a larger-scale e.g. circulation (or finer scale e.g. microphysics) control on precipitation? 

L677-681: this sentence is too long. 

L694: though here you have prescribed SST so not really well-investigated? 

L718: credible for what purpose? No parameterised convection scheme will represent squall lines, but may still be useful for other applications.