This papers investigated the dependence of cloud and updraft properties on mesoscale organization using doppler-wind lidar observations from the ATOMIC/EUREC4A field campaign. They split the days into more organized (MO) and less organized (LO) days based on satellite images. They show that MO have stronger and more variable updrafts, and this holds when constraining by size of updraft.  They show that LO has distinct diurnal cycle in updraft properties whereas MO does not, suggesting MO is reinforced by dynamical processes associated with mesoscale organization, consistent with LES studies. They also show that the environmental differences between MO/LO (e.g. wind speed, lower tropospheric stability) are consistent with the previous studies using sugar/gravel/fish/flowers categories.



This is a well-written and thorough paper, with interesting results that should be read by everyone working on cloud organization of trade-wind cumulus. I have a few (very) minor comments below, but I am otherwise happy for this to be published without any further review.

•  L250. The discussion from this sentence to me doesn't quite match up with what is shown in Fig. 4. The statistical tests show that there are significant differences between MO and LO, but to me the differences among different times within MO or LO don't always look significant. Particularly 4c, seems to indicate there is little to no significant diurnal cycle in W_cb,core. The statement that the largest mean in MO is 6-12UTC, could be 12-18 depending on where you look. Similary the "gradual ramp up in mean and variance" for LO does not look so clear and depends on where you look
• L265. Could you just show the fraction successful. There still does appear to be a small right shift in L_plume,clear, even though the shift on L_plume,cloud is more apparent
• L306. The formatting of W_CB,Core and L_CB,Core is different. Lower case "B" and missing a space
• L328. You say likely from Feb 9th. Would it be possible to check. Replot the same curve with data from Feb 9th excluded

General comments:

This manuscript uses ship-borne stabilized Doppler wind lidar measurements to study vertical motion underneath trade cumulus clouds over the tropical Atlantic during the ATOMIC/EUREC4A field campaign. It delivers an important, previously unrecognized finding: Their observations indicate that trade cumuli can achieve appreciable cloud-base mass flux variability at a constant thermal size, by varying the vertical velocity (wcbcore) per unit thermal witdth, especially in the upper sub-cloud layer. This “variability” is quantified by studying differences between scenes where the clouds are “less organized” (LO) and “more organized” (MO).

Additional conclusions are:

• The raised wcbcore in MO scenes relative to LO scenes increases as the cloud-base core sizes get bigger, because MO clouds continue to attain larger w as they grow, while LO clouds do not. The authors hypothesize that this can be explained by dynamics preventing the LO clouds to continue raising wcbore, and tie it to raised stability in the profile aloft.
• By quantifying the “environment” throughout the diurnal cycle, it is suggested that LO thermals and clouds are closely tied to the diurnal cycles in surface fluxes and winds, but MO wcbcore is kept high for other reasons. It is suggested that LO clouds are inhibited by stability aloft, and that MO clouds’ heightened wcbcore create moist layers aloft at the expense of drying the subcloud layer.

This work is unique, novel and interesting to a large community studying the role of trade cumuli in climate, and I would recommend it be published, though I would request the authors consider a few modifications and thoughts.

First, I would ask the authors to explain more precisely how they distinguished “more/less organized” scenes (see comment line 187). Their interpretation throughout the paper leans entirely on this distinction, but while they do quantify it using measures of organization, they never actually explain how they distinguished one from the other. Presenting at least some typical examples of the different scenes would help readers relate to what they are actually quantifying the differences between.

Second, where I find the parts of the manuscript that quantify and present the results clear, I find the interpretations often to be less precise. Most of my specific comments address parts that I struggled to parse, so addressing them will help. Generally, I would suggest the authors present something like a conceptual picture, or framework, that ties their observations to the mechanisms they propose to connect them, or that they consider developing the ideas they hypothesize to explain their result more explicitly in the text. Here are a few concrete questions I was hoping to find answers to:

• What (exactly) are moisture-convection feedback, “cloud-layer circulation” and how do the authors suggest they drive or feed back on the dynamically different MO/LO thermals?
• How can stability at 3000m explain the inhibition of LO (but not of MO) plumes at cloud base?
• If, as the authors hypothesize, stability controls LO’s wcbcore, then can they be sure that the fact that the clouds are differently organized actually matters? After all, could LO’s inhibition then be entirely controlled by stability, while it is merely a coincidence that the cloud scenes visually look different?

I am not saying the authors need to answer these questions conclusively in this manuscript, I would simply ask them to consider placing the results more specifically in the context they have chosen.

Specific comments

Introduction: Would the authors consider sharpening their broad introduction of several issues relating to shallow cumulus organisation, to knowledge gaps that their study fills? How does their “focus on observationally examining the influence that mesoscale organization has on the updraft velocities of wintertime trade Cu” relate to the motivating questions they introduce: i) cloud-radiative effects, ii) hydrological cycle changes, iii) cloud feedbacks?

Line 99: I believe those authors find that vertical velocity variability only becomes important at large values of mesoscale vertical ascent, while cloud amount variability retains leading-order importance?

Line 136: The two cloud fraction composites broadly agree, though there is a low bias in the doppler lidar-derived cloud fraction. Are these results sensitive to the “range-corrected intensity” threshold they mention? Is there additional salient information to be shared about this procedure?

Line 140: So a “cloud scene” includes all points between -1.5 to 1.5 x/Lchord, with Lchord defined by the RCI thresholding, and 0-2 z/CBH, with CBH defined how? From the ceilometer? It might be worthwhile defining “cloud scene” explicitly, for a reader’s convenience

Line 150: What is the motivation for, and is there sensitivity to, choosing an hourly-averaged wind speed at cloud base to define Lchord? And what is the reason to choose a different wind, the surface wind, (also hourly averaged?) to scale the w-plumes?

Line 187: Could the authors elaborate a little more on how they distinguish between LO and MO scenes from Terra/Aqua imagery? Purely by eye? They separate very nicely in their Iorg/S space, why then not just use a more objective distinction? And is the subjective distinction the classification that is used the rest of the paper? I would invest this effort, because it is presently not entirely clear what “more organized” and “less organized”, the fundamental distinction on which the authors base their conclusions, are defined. I wonder if the results’ interpretability would be enhanced if it is made clear what physical hypothesis underlies this distinction.

Line 215: Perhaps I missed this: Is “variance” horizontal variance at a height level, computed on single identified plumes, and then composite-averaged? I.e. is it really a composite measure of intra-plume turbulence?

Line 235: Do the authors mean figure 2 ac?

Line 241-245: This is an intriguing finding, and I wonder if the authors may wish to suggest more specifically how cloud-layer organisation influences the thermals. Are the authors suggesting that there may be pressure gradients extending down from the cloud layer, thus accelerating the thermals vertically? How would they be driven? Also pertains to discussion around line 270: How specifically would the cloud layer increase the success rate of the plumes in MO conditions? And 285: What specifically could a mesoscale ascending cloud layer do to thermals underneath? How is it related to “moisture-convection feedback”? And how does the test the authors perform (correlating Lplumecloud to Lcbcore and wcbcore) actually test these mechanisms? Finally, a “classical” reading of subcloud-layer thermals might involve a w*, defined solely on the basis of a surface buoyancy flux and relevant scales. My understanding is that most theories suppose this cannot change, so long as the surface forcing doesn’t change. Are the authors suggesting that this velocity scale is no longer appropriate in MO situations, because there are cloud-layer forces controlling the velocity scale? It may be worthwhile for the authors to juxtapose their findings against such classical frameworks, to place their novel findings in context.

Line 413: How do the authors suggest that a stronger capping inversion could flatten the wcbcore-Lcbcore relation, even after subselection? That is, what does Lcbcore have to do with the response to stability? Also, could the authors motivate more clearly what they suppose the feedback is from the strength of the capping inversion to the clouds at cloud base? From the profiles in figure 10, it seems that the local stability of that layer is comparable between LO and MO, so what is the non-local influence from the inversion?

Line 432: Can the authors conclude so strongly that the increased wcbcore in MO clouds supports their larger radiative impact? It seems logical that such systems, due to their increased dynamical contributions to cloud-base mass flux, produce deeper systems with larger coverage, but the result presented just contrasts MO and LO clouds, and these could differ in numerous other ways. For instance, if the MO cloud layers are indeed substantially moister, perhaps a lack of entrainment drying is decisive in deepening them. I do think attributing radiative impact to dynamical efficiency is an interesting direction for the authors to pursue further (in future work), perhaps with simple plume models rising through the observed environments.

Line 440: How do these observations square with the observations from the same field campaign, presented by George et al., (2022), which observe moister subcloud layers when the column is ascending at mesoscales (and so presumably a more organised cloud cluster is passing)? (The authors do not need to answer in the text, it just strikes me as a curious contrast)

Line 475: A central finding of this study is that the references in this paragraph have underappreciated the role of variability in vertical velocity when explaining variability in mass fluxes. Yet all these studies seem to have been rather successful at explaining mass flux variability without varying vertical velocity. It would be very interesting to see the authors reflect on how large the errors that the previous studies made were, and where the discrepancies come from (Did they not look? Did they have different tools? Sample different regions?).

Line 479: Revisiting Vogel et al., (2022), it seems to me that there is plenty of variability even in their mass flux-cloud coverage relation (their fig. 3c); could that be directly attributable to variability in mass flux being explained by variability in vertical motion, rather than in cloud cover?