This manuscript presents an at-launch algorithm for IWP retrievals constructed for the upcoming ICI instrument. The work is thoroughly performed and well described. I recommend publication following some minor revisions. Specific comments associated with line numbers below.

Line 24: I would leave out "now" in this sentence, as the reference is 30 years old

Line 44: the authors list the MWI frequencies later in the paper, but a range (19-183) would be useful here since it specifically mentions the frequency coverage.

Line 72: is the overpass time known? 1:00?

Line 89: reword "precipitation and snowfall" as both are precipitation. Suggest "liquid and frozen precipitation" or "rainfall and snowfall"

Line 170: this is done in section 3.2, but I would suggest adding information right at the beginning of this discussion about the period and coverage used, as I was asking this while reading this section. (two years, global CloudSat overpasses) and ancillary model data used

Figure 2: bottom panel - would prefer frequencies rather than channel#. Should the y axis be labeled Ta rather than Tb?

Line 451: suggest add "conically scanning"

Line 461-465: Just a note that this is almost certainly the case and emphasizes the need for better understanding of emissivity at these higher frequencies

Figure 5-6: would prefer labeling with center frequency rather than channel number (lots of flipping between pages)

Line is the same 614: Not sure if I missed it but should have a brief introduction of DARDAR somewhere

Line 618: There are a couple of features in the zonal mean plot that I think are worth discussing. Retrieved IWP is lower than the database for the northern midlatitudes. Perhaps a land/vegetation issue? Retrievals are conversely high for the ITCZ. Thoughts on this?

Line 784: Somewhere here it seems worth mentioning the importance of representativeness in the database for retrieving the highest and lowest values, as demonstrated by the author's plots 

The manuscript introduces the operational retrieval database for the upcoming Ice Cloud Imager (ICI) mission and conducts a comprehensive evaluation of its retrieval capabilities. ICI, an instrument aboard the EPS-SG, is a passive conically scanning radiometer operating within a frequency range of 183 GHz to 664 GHz. This innovative capability will provide unprecedented information on smaller ice crystals, effectively bridging the gap between microwave and infrared observations.

The work is meticulously executed and clearly articulated. To emphasize, the study’s simulations are among the most detailed to date, showing statistical consistency with real observations. And, the advancements presented in this paper provide significant added value to the scientific community.

To that end, I recommend it for publication after minor revisions.

General comments:

• There are inconsistencies in the formatting of the text preceding equations in the report. For example, a colon (:) is used before equations in some instances (e.g., line 374), while in others, no punctuation is used.
• The most precise way to refer to Metop-SG, is MetOp-SG.
• When a citation is given in parentheses, remove redundant brackets, e.g. (see Table 1 in Eriksson et. al (2020)) => (see Table 1 in Eriksson et. al, 2020).
• When a panel of a figure is described, please include the figure number as well for increasing clarity.
• When the discussion is centred around results that are not presented within the manuscript, please add the relevant info (in the beginning of the paragraph), e.g. not shown here.
• On some occasions, a space is missing between a number and its unit.
• Section 4.1, could be moved to the Appendix
• Please consider providing information on DARDAR and CALIPSO. Initially, the acronyms are not defined. Furthermore, enhancing clarity by including relevant details about the type and usage of any datasets referenced would greatly benefit readers.

Specific comments:

The following acronyms have not been introduced:

• EUMETSAT (mentioned in Affiliations, Abstract, and the first reference in the Introduction)
• DARDAR (line 15)
• ARTS (line 135)
• RTTOV (line 136)
• ECMWF (line 145)
• DISORT (line 304)
• CALIPSO (line 791)
• CloudSat (line 158)
• EarthCare (line 799)

On the other hand, the QRNN acronym (line 418) has already been defined.

Line 24: remove “now”; reference is more than 30 years old. The authors could consider updating the grammar as well.

Line 37: Adding some examples or references to passive optical and infrared missions that capture cloud top data could be beneficial.

Line 43: I have never seen the acronym to be written as Micro-Wave. Why not use the most common mention, i.e. Microwave Imager?

Line 44: The authors could consider including the MWI frequency range.

Line 71: Since the authors decided to add information on MetOp-SG A, including details on the differences between the two missions could be valuable. To elaborate, MetOp-SG A focuses on optical/infrared missions and atmospheric sounders, while MetOp-SG B focuses on microwave instruments. Otherwise, the mention does not provide any benefit. Please consider including any citation.

Line 104: There is no information on what an L1b product is. Plus, consider defining the acronym Level 1b (L1b).

Table 2: A full stop is missing at the end of the caption.

Line 89-90: precipitation and snowfall are both precipitation; maybe reward it to “liquid and solid precipitation”.

Line 94: "using different methods"; the word "retrieval" is redundant due to the aforementioned.

Line 235: No subscript should be in italic; please check this throughout the paper.

Line 251: There is no information on what the DARDAR product is. A brief introduction, since it is being used, could add value to the paper.

Line 260: “The scaling differs between V and H polarisation”, Could the authors elaborate on this a bit further? Please have a look at Barlakas, Geer, and Eriksson 2022 (Cloud particle orientation and polarisation for cross-track microwave sensors; NWP-SAF).

Line 304: To begin with, DISORT acronym has not been introduced. Why do not you add some brief information on what the DISORT solver is, including a reference.

Line 466: This can imply that the particle models could be further tailored. Could you please add some comments?

Figure 3: a space is missing from [-60, 60].

Line 519: Any comments on the differences in the polarisation signature between 243 and 664 GHz? Excluding the surface contamination, at which frequency do you expect to see larger polarisation differences and why (considering the ice amount)?

Line 623: A redundant bracket exists.

Line 646: the negative sign should be in $$.

Figure 13: The legend is incomplete; mid-latitude reference is missing

Line 799: EarthCare has been already launched; the authors could consider updating the information.

Line 801: remove “that”

Line 857: a redundant “-” exist

Line 858: a redundant “-” and a space exist.