Synopsis:

The authors Xu Ning  and Mostafa Bakhoday-Paskyabi report in their manuscript entitled “Implementation of a sigma coordinate system in PALM-Sigma v1.0 (based on PALM v21.10) for LES study of the marine atmospheric boundary layer” on the development of a large‑eddy‑simulation (LES) code based on the PALM LES framework, incorporating a modified vertical coordinate that accounts for the actual, instantaneous position of the atmospheric lower boundary. The authors use their code to investigate the interaction of the wave‑induced effect between surface waves and the marine atmospheric boundary layer. Their results clearly demonstrate that the modelling approach commonly employed in atmospheric flow models—representing the ocean’s influence on the atmosphere solely through a roughness length parameterized as a function of wave characteristics such as significant wave height and wave period—constitutes a strong simplification, especially in situations where the wave field is not in equilibrium with the wind field. For example, the authors show that the mismatch between wave direction and wind direction also influences properties of the marine atmospheric boundary layer above the wave‑affected layer. When wind and waves are opposed, turbulence throughout the entire marine atmospheric boundary layer is enhanced compared to the case where wind and waves are aligned.

Evaluation:

I would like to thank the authors for what I consider to be an excellent piece of work. The manuscript is both very well structured and very well written. I have only minor comments on the manuscript, and therefore I recommend its acceptance for publication in EGUsphere after minor revisions. I’ll ask the authors to take the comments below into account when revising the manuscript.  

Comments:

1. Page 1, line 24/25: “The latest high-performance computing (HPC) equipment enables numerical simulations with a magnitude of grid points up to 1010 (Kröniger et al., 2018)” I suggest to find a more recently published paper as a reference for the number of grid points that can be handled by the latest generation of HPC clusters. I think it is slightly contradictory if you speak of latest HPC equipment (which should be from 2025), but refer to a paper from 2018. If no more recent publication is available I suggest to modify the sentence, e.g. as follows: “Already in 2018 high-performance computing (HPC) equipment enabled numerical simulations with a magnitude of up to 10°10 grid points (Kröniger et al., 2018)”.
2. Page 2, line 27: Please correct “In numerical modeling of the marine atmospheric boundary layer flows” to “In numerical modeling of marine atmospheric boundary layer flows”.
3. Page 3, line 68: Please change “The current work further develop PALM …” to “The current work further develops PALM …”.
4. Page 8, equation 23: The values of the weights applied in the Runge-Kutta scheme should be specified.
5. Page 12, line 246: To apply the grid stretching already inside the atmospheric boundary layer is rather uncommon at least in work that has been done with the LES code PALM. Typically, the grid stretching would be applied above the atmospheric boundary-layer where turbulent processes do not play a major role any longer. It would be an extremely interesting result if the authors could show that applying the grid stretching already inside the boundary layer does not change the results significantly as it would be of importance for the computing resources required to carry out LES runs. Therefore, I ask the authors to extend their study by one extra simulation that applies a uniform distance in z-direction inside the boundary layer.
6. Page 12, line 247: The authors report that their model domain has a size of 1200 m x 1200 x 850 m. The initial height of the bottom boundary of the inversion layer chosen by the authors is 600 m. I’m wondering whether the size of the model domain is actually large enough. I ask the authors to add a statement whether they checked the sensitivity of their results on the size of the model domain.
7. Page 12, line 249: The authors report that they have applied a total simulation time of 20 h. Is that length of the simulation run sufficient to get rid of the inertial oscillations that occur in simulations with PALM if the Coriolis force is switched on? g. Maas (2023) used a physical simulation time of 48 h to obtain a steady-state mean flow in his LES of an offshore flow at 55°N (Maas, 2023: From gigawatt to multi-gigawatt wind farms: wake effects, energy budgets and inertial gravity waves investigated by large-eddy simulations).In order to allow others to repeat their simulations the authors should provide also an information on which geographical latitude was assumed in their simulations. Moreover, information on the lateral boundary conditions chosen should be provided.  Page 14, figure 3: I doubt slightly the value of showing a comparison of instantaneous velocity fields. The flow is anyhow different. We do not see the impact of background turbulence and turbulence created by the different surface waves.-  I think to show statistical measures like the variance of velocity components makes more sense
8. Page 15, figure 4: The caption of figure 4 lacks an information which cases are actually shown in the figure.
9. General comment: I encourage the authors to mention also other possible applications of their modified PALM code such as flow in complex terrain in the outlook section of their manuscript.