The manuscript by Mauder et al evaluates a new method to correct turbulent flux measurements for the widely observed energy balance non-closure. The manuscript addresses an important research question and tests a new approach to solve a long existing problem for measurements of turbulent fluxes. I have a few comments that hopefully can improve the quality of this manuscript.

• Additional statistical analyses could be used to test if the flux corrections result in statistically significant improvements of the energy balance closure. The results qualitatively indicate improvements, but further statistical support of the findings would strengthen this study (e.g., through additional regression uncertainty analysis).
• Unfortunately, the only site where the correction procedure can be applied directly has already a good energy balance closure, while the two other sites are characterised by a substantially worse closure. It would be helpful if sites with similar energy balance closures could be selected or at least if this issue would be discussed in more detail.
• Lastly, a new study by Sun et al. (https://doi.org/10.1029/2020JD034219) presenting a new hypothesis for the energy balance non-closure at flux tower sites related to non-hydrostatic energy transfer should be discussed in the manuscript. It would contribute to a comprehensive discussion of the universality of the correction procedures outlined in the manuscript.

The manuscript by Mauder et al., seeks to correct the local turbulent flux measurements based on a non-local parametric model from a set of LES. The authors used three different experimental sites. They showed that the accuracy of the turbulent flux estimates is improved after use the LES-based SEB closure correction method. These results are important for a better approximation of the closure of the SEB. The manuscript is well written. The text is clear and easy to read. Figures are very explanatory and aid understanding the manuscript. However, I have some minor concerns with the authors’ methodology, which I feel should be resolved before publication.

1) As the LES simulations performed by De Roo et al. (2018) were an important part of the methodology used in this manuscript, I suggest that the authors create a subsection in the "Methodology" to inform readers the main steps in the work of De Roo et al. (2018) were used by the authors in this manuscript . If the authors do this, all the equations in the introduction would go into this subsection;

2) The experimental measurements performed in this study were carried out with very similar instruments/methodologies, for the three different experimentais sites investigated. Except for one of them: the height of the boundary layer. Sometimes the authores used a ceilometer data and sometimes they used the slab model (Batchvarova and Gryning (1990)) to identify the BLH. What is the impact on the results presented by the authors if one of these devices/methodologies does not adequately represent the height of the convective boundary layer?

Technical corrections;

• L138: of 4x4 km;
• L200: ...(Emeis et al., 2011; Münkel et al., 2007). This method is...
• L211: 1.4 m
• L246-248: Isn't that logical? Hdis depends on H and the latter has higher values in summer.