Review of "Improving model-satellite comparisons of sea ice melt onset with a satellite simulator" by A. Smith at el.

The paper uses a brightness temperature satellite simulator to diagnose the timing of melt onset in model simulations, and compare this new metric to other metrics of melt onset. Overall, the paper is well written and worthy of publication subject to the minor corrections listed below:

Line 49: Write out the ACR3O acronym here as this is its first use.

Line 61: What method of regridding was used? 

Line 73: "Each .." Describe the method in slightly more detail - How does brightness temperature change as melt occurs?

Line 85: "Surface temperture" - clarify whether this is ice, snow or either surface temperature.

Line 123: What is the justification for the step function as opposed to another functional form?

Line 138: "So here" -> "Therefore" or another replacement.

Line 141: "CICE provides..." I believe CICE can provide the number of layers specificed so clarify this isn't the only choice possible. Maybe something like: "CICE was configured to provide"

Lines 140-150: Clarify what thermodynamic model CICE is used? Mushy layer? Bitz-Lipscomb 1999?

Line 152: Lagrangian tracking - I am unfamilar with this functionality in CICE - is it documented/referenced somewhere? There is a FY ice tracer, but it is not lagragian tracked, instead advected with the Eulerian transport scheme.

Line 154: Clarify what the correlation length is.

Line 159: "(greater than 30cm)(Fig. S2)" -> "(greater than 30cm; Fig. S2)"

Line 180-185: The two thresholds are presumably justified from a bimodal nature of the brightness temeperatures. It would be useful to see histograms demonstrating this bimodal nature.

Line 185: "with boundary" -> "with the boundary"

Figures 1, 5, 6, 8: Add the field name and units to the colorbars.

Figures 3, 4, 7: Add markers to the map of the Arctic ocean showing exactly where the a,b,c,d points are.

References: "SIMIP community ... 2020": This doesn't appear to be how to cite: https://agupubs.onlinelibrary.wiley.com/action/showCitFormats?doi=10.1029%2F2019GL086749

Review of “Improving model-satellite comparisons of sea ice melt onset with a satellite simulator”

This paper describes a novel use of model simulated brightness temperatures to compute a new metric to identify the timing of earliest snowmelt on Arctic sea ice. The paper compares the simulated brightness temperatures and earliest snowmelt dates with brightness temperatures and a melt onset data set from satellite data.

The paper is well written, and the study is clearly presented. I agree with comments from the prior reviewer and recommend that this paper is published pending a few additional clarifications as described in my comments below.

L61: Why was 2003 selected as the sample year? Is it representative of a normal (non-anomalous) melt onset year? Or something else?

L73: Why is the Steele et al. 2019 dataset used instead of the Markus melt onset product directly? Are they not the same data?

L206-207: To what extent do you think error in the observations or in the simulated brightness temperature contribute to the divergence between the simulated and observed brightness temperatures in the central Arctic (i.e, Figure 3d) seen after the SIC declines? What might the physical reason for this big difference be?

L288: Please add proper names for the geographic locations that are considered “inflow regions”.