General comments

The paper compares the extreme events between two simulations, a control and an Amazon deforestation, using the ICON global model at 5-km resolution. The simulations are based on the two three-year runs, using climatological sea surface temperature.  It is an interesting article, well-written and well-organized. However, several conclusions should be drawn more carefully. It is not clear that the CTL run is producing a realistic Amazon climate;  validation at the local scale with direct model output variables would give a clearer picture. The unchanged total precipitation with deforestation has implications on the precipitation recycling topic, which requires more attention.

Specific comments

Major concerns about the model setup

1. The result of no change to the total precipitation after deforestation should be treated with care.

As models increase horizontal resolution and switch off the convective parameterization, convective mixing is not treated within its timescale, and stronger updrafts are produced at the grid scale. This lifting of the air by the updraughts leads more easily to air saturation. The cloud microphysics produces a lot of rain due to saturation, but it does not treat column mixing. If some convective mixing is allowed, those extreme precipitation events should probably decrease.

2. Local validation of control run. Switching off convective parameterization completely may need additional verification at the local scale. I recommend including maps of precipitation over the Amazon region to validate precipitation, temperature, and winds at the local scale.

3. Forest parameters: Rooting depth is too shallow; Amazon forest roots are much deeper and should sustain evapotranspiration during dry periods.

4. Integration length: Three-year length is short for the runs to reach a stable climatic condition. The 15-day spin-up time to reach climatic conditions is also short.

5. Validations:The work requires validation at local scale of direct output variables. Differences and statistics are not enough to show the realism of the simulations (precipitation, temperature, evapotranspiration, winds, in different seasons) in the CTL run, There is not enough discussion on the precipitation recycling. This is a major topic 

6. Concerning Citations that deserve to be mentioned:

• Works on deforestation in the Amazon, that have carried out analysis on the impacts and extremes.

Bottino et al. 2024 (https://doi.org/10.1038/s41598-024-55176-5)

Brito et al. 2023 ( https://doi.org/ 10.1002/joc.8158),

Pilotto et al. 2023 – (https://doi.org/10.1007/s00382-023-06872-x)

• Works on precipitation recycling:

Rocha et al. 2017 (http://dx.doi.org/10.1590/0102-77863230006)

Salati et al 1979: https://doi.org/10.1029/WR015i005p01250 . Classic paper

Technical corrections

1. Missing the reference page 1: RAISG, 2022
2. Line 190: violent rains can be attributed to increased updraughts, not the other way round.
3. Line 63: Typo: not global but globe.
4. Line 86: Typo: not ourpur but output

Review of “Extreme events in the Amazon after deforestation”

This study employs a global storm-resolving climate model (ICON-Sapphire, 5 km resolution) to simulate the impacts of complete Amazon deforestation on short-duration extreme events. The model captures precipitation, temperature, and wind extremes more realistically than coarse-resolution models. The authors find that while annual mean precipitation remains largely unchanged, the tails of the distributions shift markedly: violent rainfall and no-rain events increase, heat stress intensifies, and extreme winds strengthen. The analysis attributes violent rainfall increases to enhanced moisture convergence, and stronger winds to both reduced surface roughness and storm downdrafts. Overall, the study concludes that deforestation exacerbates climatic extremes. This is an excellent and important paper that provides novel insights into how Amazon deforestation alters extreme events. I have only a few comments, detailed below.

Major Comments

• At the bottom of page 2, the authors mention that several biophysical changes following deforestation but do not explicitly discuss the role of surface albedo. Since Table 1 shows a notable increase in albedo after deforestation, please clarify how this factor interacts with evapotranspiration and surface energy fluxes in your interpretation.
• Page 4: Please clarify what vegetation or land cover is prescribed after deforestation. Relatedly, explain why the leaf area index is still set to 2.7 rather than 0, despite “complete” deforestation.
• Page 7: how about CAPE a few hours earlier? Why was 1 hour selected? Justify why CAPE was calculated only one hour before violent rainfall events. Would the results differ if CAPE were considered several hours earlier?
• On page 10, lines 198–200, the authors state: “The post-deforestation nighttime temperatures become comparable to pre-deforestation daytime values.” What does this mean in practice? Please clarify the significance. Do you mean that the nighttime minimum after deforestation is as large as daytime maximum before deforestation? A clearer formulation would help readers interpret the magnitude and implications of this result.
• Wet bulb temperature is a widely used indicator of heat stress. You could use both temperature and humidity changes to represent heat stress. If the sign change of wet bulb temperature differs from other indexes, it is possible that the heat stress change is not significant.
• Page 14: I do not understand: In summary, the relative contributions to the total wind speed anomaly are 60% R/C, 13% D, and 27%. Please re-explain the factor separation results more clearly.

Minor comments

On page 3, please add a map clearly showing the spatial extent of the deforested region in the simulations.