In this manuscript the authors introduce a scalable, data-driven framework for the analysis and the interpretation of CG lightning activity large datasets. This framework consists of Number Distribution Lightning Differential Space, and the Current Ratio Lightning Differential Space that combined can provide insights and information regarding the CG activity in storm regions. Overall, the manuscript is very interesting, well written and well presented. I suggest it should be published after some minor comments are addressed that could improve the readability of the presented work.

1) Figure 3 is not explained in the text, but it is discussed in combination with Figure 4. I suggest either a paragraph to be dedicated for the discussion of Figure 3, or Figures 3 and 4 to be combined into a multi panel figure and then the panels of this figure to be discussed in the manuscript.

2) Is the caption of Figure 4 correct? Does it really refer to Fig. 1 or does it refer to Fig. 3?

3) Wouldn't a projection of Fig. 5 onto dR and dT (similarly to Fig. 4) be useful in presenting and discussing the results?

4) The authors state that the presented framework is suitable for comparing storm regions, validating lightning models and enhancing early warning systems. The whole discussion is dedicated in presenting a comparison between different storm regions, and thus it clear the contribution of the presented framework. There is no discussion how can the presented framework be used for the validation of lightning models and how can it enhance early warning systems. How can someone use the presented framework into achieving these goals?

5) Can this analysis be also  used for the investigation of Intra Cloud (IC) lightning activity? After all, the IC lightning activity dominates over the CG lightning activity in terms of occurrence. Why is this analysis focused only to CG flashes?

• Authors should provide some additional information on existing differentiation methods in introduction
• I don’t agree with the argument L80-81 “(c) they are all characterized by flat terrain which is important for minimizing the effect of orographic convection on the analysis” since the region of interest in Eastern Mediterranean Sea is covered mostly by sea where the physical processes and magnitude of lightning activity differ compare two other two selected regions. Authors should provide more detailed information on the choice of this particular region.
• Two years (2020-2021) of observed stroke density are not enough to minimize the impact of the interannual variability. Authors should rephrase the lines 121-122.
• Authors should provide a short overview of the mean atmospheric conditions (e.g. mean height of iso 0, -2, -10, average liquid water and ice contend, cloud base height) during the evaluation period of their differentiation method and try to correlate their results with these conditions.

L63: Pls provide some references for Cyprus-Lows

L64: LDS abbreviation is missing (although it is mentioned in abstract)