In this manuscript, Zabel et al. describe a new piece of software, CropSuite, that generates maps of crop suitability and related information based on climate, soils, and terrain. This builds on previous work by themselves and other authors to include, importantly, (a) a consideration of climate variability in addition to averages and (b) less-widespread but regionally important crop types. Noting that such crops are under-studied but are especially important in Africa, the authors focus their analyses there. The results look reasonable when compared to real-world crop distributions and sowing dates.

One of the goals of CropSuite was to make something that is easy-to-use and flexible enough to be used by a variety of stakeholders, not just scientists. As a scientist, I can’t really assess how accessible it is to less-technical users, but the inclusion of a graphical user interface (GUI) is a really important development. I do think, however, that this tool will be useful to scientists and model developers as well. Global gridded crop models and especially integrated assessment models need to be able to endogenously represent things like sowing date, the potential for multiple cropping, and shifts in what crops are planted where; tools like CropSuite can help.

That said, I do have some questions and concerns about the manuscript as currently written. Thus, I recommend it be considered for publication after minor revisions. See attachment for details.

This study improves the crop suitability assessment method by considering the climate variability. It compares results with SPAM2020, GAEZv4, and the crop calendar from GGCMI. The manuscript is clearly written. I have a few comments.

Major:

1. Make the dataset and results more accessible. Now, all the results in Zenodo are *.png, which contain no geographical information. I highly recommend authors to provide geographic file format, e.g., Geotif and netcdf.

 Minor:

1. line 77, Is the soil texture for >200cm also needed? if yes, where does it come from?

2. Why the soil layer in lines 79-80 is different with Table 2

3. Line 81, reference formate. And, why are the weights needed? How are these weights applied? It's a bit confusing here. Did the authors mean that weights were used to multiply with original value to generate the new value?

4. line 233. OK, MapSPAM2020 may introduce some uncertainties, then why not using MapSPAM2010?

5. Line 277, is it because that nutrient and soil fertility are not considered in this study?

6. In theory, I would expect a smaller area in this study because this study considers additional climate variability. However, Figure 8 shows a larger area by this study. Can the authors explain more about this?