This study provides a well-structured and comprehensive analysis of global aridity projections based on CMIP6 scenarios. The results are presented clearly, and the methodological approach appears to be sound and well explained. The study is relevant for understanding long-term trends in desertification and future climate impacts.

Comments:

• Aridity classification - The manuscript primarily focuses on desertification, but only includes 1–2 humid categories. Would it be possible to shift the focus slightly toward transitions between different aridity index (AI) classification states rather than focusing exclusively on desertification? If the authors prefer to maintain the current classification, a justification for this choice would be helpful.
• The AI classification used in this study appears to be slightly different from the classification used by the IPCC Sixth Assessment Report and UNCCD (2024), also cited in this study. See: [Dry sub-humid (0.5 ≤ AI < 0.65), Semi-arid (0.2 ≤ AI < 0.5), Arid (0.05 ≤ AI < 0.2), Hyper-arid (AI < 0.05)]. It is only a minor change to the classification but it would make it easier to compare your assessment to more recent publications.
  
  
  
  See for reference: e.g. Figure CCP3.1 in Mirzabaev, A., L.C. Stringer, T.A. Benjaminsen, P. Gonzalez, R. Harris, M. Jafari, N. Stevens, C.M. Tirado, and S. Zakieldeen, 2022: Cross-Chapter Paper 3: Deserts, Semiarid Areas and Desertification. In: Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [H.-O. Pörtner, D.C. Roberts, M. Tignor, E.S. Poloczanska, K. Mintenbeck, A. Alegría, M. Craig, S. Langsdorf, S. Löschke, V. Möller, A. Okem, B. Rama (eds.)]. Cambridge University Press, Cambridge, UK and New York, NY, USA, pp. 2195–2231, doi:10.1017/9781009325844.020.
• Will the dataset produced in this study be made publicly available? A dataset of time-series AI classifications would enable further studies on system-state transitions, which could be valuable for assessing long-term desertification and land degradation trends. Making such data accessible would enhance the impact and usability of this research.
• Formatting comments:
  • There are some inconsistencies in citation formatting. For example, "et" appears to be used as “and” in some cases (e.g., lines 56, 59, and 73). Standardizing the citation format to English would avoid confusion.
  • Lines 99–100: Please check the units—there appears to be a discrepancy of three orders of magnitude between mJ and MJ.
  • Lines 545 and 548: The formatting of "CO₂" should be corrected.
  • Table 2 should be formatted for easier readability.

Review comments on “Global projections of aridity index for mid and long-term future based on CMIP6 scenarios by Crapart et al.

The authors evaluated AI and dryland distribution projected by 13 CMIP6 models for three different socio-economic scenarios and for two timeframes of 2030-2060 and 2070-2100. The evaluation was done against WorldClim and ERA5 datasets for 1970-2000. Their projections indicate significant (SSP2-4.5 and SSP5-8.5) or limited (SSP3-7.0(drying) for mid-term but more consistent and continuing drying in the long-term.

Systematic analysis of aridity in the mid-term and long-term future periods for a selection of representative emission scenarios is well within the interests of HESS readership and the use of 13-member ensemble of CMIP6 outputs add values to the future aridity and drought studies. However, the manuscript needs clarifications in some of the methods and results presented, justification or change of the ensemble CMIP6 sampling, and overall improvement in writing before it is considered for acceptance to HESS. I recommend a major revision. More detailed comments are provided below.  

The manuscript contains analysis of internal variability in CMIP6 members (Section 3.1) and evaluation of CMIP6 against ERA5 and WorldClim (Section 3.2) over a historical period of 1970-2000. An important utility of CMIP6 evaluation over historical period is that it can provide insights on the behavioral features of projections produced by each ensemble member in terms of bias, trends, dynamic ranges (seasonal or interannual), mean difference, etc. Analysis of projected AIs needs elaboration based on the results of Sections 3.1 and 3.2.

In addition, for CMIP6 based projections, the candidate models are often 'conditioned' by comparing their results with historical observation or reanalysis data. The conditioning can be based on different evaluation metrics depending on the information of interests (e.g., trends vs. absolute values) for the projected period. The authors are recommended to try conditioning of the ensemble to improve the reliability of the ensemble of projected AIs. Assessment of the Section 3.2. appears to be an excellent source of information for this purpose.

Results include evolving aridity, for some regions, from wetting or moderate drying for 2030-2060 to more consistent or stronger drying for 2070-2100, particularly for SSP3-7.0.  Given that the projections are derived from models, the manuscript may include more process-based explanation for the inconsistent changes between the mid-term and long-term projections.

It appears that there exists similarity between the manuscript and the UNCCD report (Vincente-Serrao et al., 2024) cited in the manuscript. Provide the difference between the manuscript and the UNCCD report (Vincente-Serrano et al., 2024) in methods and datasets. The UNCCD report appears to include comprehensive assessment of similar projections with minor differences in the projected time windows.

Some paragraphs in Introduction are loosely related to the main focus of the manuscript (e.g., history of climate zones). Tighten words and improve focus in the Introduction section. There are mentions of 30-year or 30 year, but the three periods of analysis are all 31 years (1970-2000, 2030-2060, and 2070-2100), right? Also, description of methods needs improvement for clarity. For example, more detailed description is needed for how results in Figure 1 and Table 4 are produced. Were all 31-year (not 30) data over all grid cells put together for Figure 1?  Lines 245-246 indicated a use of 30-year (31?) average, but was it used for all results or just Figure 2?

Specific comments

- Lines 20-21: The subject for the second part the sentence is different from the earlier part. Revise the sentence.

- Line 21: ‘population’ would be a more suitable word since ‘inhabitants’ can refer to the entire animal, unless that is the intended meaning.

- Line 24: Change ‘contrarily’ to ‘in contrast’.

- Line 26: “widespread and persistent”.

- Lines 66-67: What does “changes” here refer to? Changes from one category to another over time?

- Line 69: Revise the following to a correct expression: “the future downscaled CMIP6 models”.

- Eq 1 and throughout the manuscript: ET0 has been used to indicate the potential ET in the manuscript but it is commonly used for the reference ET. ‘PET’ would be a better choice to mean the potential ET.

- ‘2-m’ instead of ‘2m’ when it is used as adjective.

- Table 1 and throughout the manuscript: AI = 0.75 is used as threshold between ‘Dry subhumid’ and ‘Humid’. UNESCO (1979) uses AI = 0.65 https://www.ipcc.ch/report/ar6/wg2/figures/chapter-ccp3/figure-ccp3-001

- Lines 120-121: Rewrite this sentence for clarity.

- Line 120: Use ‘lower’ or ‘smaller’ in place of “inferior to”.

- Line 129: What does the “3 datasets” refer to?

- Section 2.2.1: Use plain paragraphs or summary of lists in tables instead of bullet points.

- Table 1: Highlight the quantities that exhibit large deviations from the central tendency (mean or median).

- Table 2 and throughout the manuscript: Units appears to be inconsistent or not in compliant with HESS style.

- Quality of Table 3 needs improvement (e.g., margins, line thickness).

- Figure 2 and relevant section: The reference data, ERA5 and WorldClim, feature noticeable discrepancy in AI. What are the sources of them when they both incorporate ground data?

Reference

Vincente-Serrano, S.M., N.G. Pricope, A. Toreti, E. Moran-Tejeda, J. Spinoni, A. Ocampo-Melgar, E. Archer, et al. 2024. “The Global Threat of Drying Lands: Regional and Global Aridity Trends and Future Projections. A Report of the Science-Policy Interface. United Nations Convention to Combat Desertification”. Bonn, Germany: UNCCD.