Review of: Representing effects of surface heterogeneity in a multi-plume eddy diffusivity mass flux boundary layer parameterization

by Nathan Arnold

The manuscript presents and analyzes a new approach to allow subgrid scale heterogeneity in land surface to influence updraft thermodynamic properties in an EDMF scheme. A series of single column model experiments were conducted with conditions from the ARM SGP site in the summer of 2017, and the results showed an increase in the near-surface subgrid scale variances, and a relatively small impact on the mean state. Sensitivities to parameters were also conducted, and a discussion of the further development of the scheme to increase the impact on the mean state was included in the discussion.

The modification to the EDMF scheme has the potential to allow for the impacts of surface heterogeneity to have an impact at higher levels in the atmosphere at relatively small computational cost, and the continuation of the development of DMF as suggested in the discussion is warranted. The contribution to modeling of turbulent processes in a GCM of the DMF scheme makes this manuscript of interest to atmospheric modelers. I recommend publication of the manuscript with the minor revisions suggested here.

Comments:

Line 35 - Perhaps a little too simple - suggest "...by modifying the updraft properties of individual plumes in the context of an EDMF scheme"

or something like that.

Sections 2.2, 2.3 - Although the descriptions of EDMF and SHOC are quite elegant and clear, I am going to suggest that the level of detail included in these sections is not needed as part of this manuscript. What is needed are the deviations in GEOS from the more standard implementations in other model, and enough detail to understand the additions to the scheme in the DMF section (2.4).

Section 3 - 

Line 168: Please provide the fractional coverage for the different tiles in the "Het" case.

Line 170: A sentence or two addressing the use of the "observed" Q1 and Q2 with the "Het" surface case is warranted. Are they still the appropriate forcing terms?

Section 4 -

Line 182: Please comment on why the Het case (with decidious trees and a lake) has larger daytime sensible heat and a smaller daytime latent heat flux than the grassland Hom case. The higher skin temperature for the Hom case and perhaps an intuitive expectation could suggest the opposite. For example, does the forest tile have a smaller vegetation cover area than the grassland tile? Mention is made one line 193 of greater surface roughness, but this would increase both sensible and latent heat fluxes.

Line 205: There are several fractions being discussed here and above, and the relationships among them warrants some clarification: Fractional coverage of each tile type; fractional area of updrafts within each tile; fractional time the (any?) updrafts are active. 

Line 256: Perhaps add a clarification of how the higher order moments impact cloud properties, ie., macro and micro-physics. Perhaps also a mention of what of these impacts are included in this model.

Line 260 and Figure 8: Are there differences in condensate (cloud water here I assume)? Panels a and b include this information but may mask the role of liquid water. Cloud fraction may not adequately describe the impact of the differences in cloud properties. I recommend adding cloud water, replacing cloud fraction with cloud water, or just reporting on differences.

Line 265: Have you tried an experiment without the relaxation? The need for the relaxation terms may no manifest at the start of the simulation, so an experiment without the relaxation may shed light on the impact of the relaxation.

Line 290: Can you speculate on the impact of the 0.75 Beta parameter on the conclusions of this study?Would you have seen a bigger impact on the mean flow?

Section 5 -

Line 340 - The experiment design (even the het case) may not include enough sub-grid scale heterogeneity to fully explore the impact of DMF. The lake tile is small and the grassland and forested tiles are relatively similar. Perhaps a coastal grid box or half a box of bare soil tiles would exhibit a larger impact of DMF. Perhaps a more heterogeneous Het case would be worth exploring.

Minutae:

Line 179 - "among" rather than "between" for more than two in the comparison