This study presents a novel convective cloud data product called GATM, which distinguishes individual MCSs from complex organizations and combine GEOs satellite image with A-Train satellites, making it possible to investigate the anvil production of MCSs and its association mechanisms. Overall, I find the manuscript well written and clearly structured, and I believe it will be a valuable contribution to the literature. I only have minor comments as listed below.

Comments:

• When combining GEO satellite images and A-Train satellites, how would an MCS be counted if the A-Train satellite orbit and the center of the MCS are offset (e.g., the MCS1 scenario in Figure 1)? Are there any special procedures in place to address this situation, and would that introduce any uncertainty to the product?
• I like how the author discusses the two mechanisms with simple yet clear analyses, but I think the conclusions regarding the two mechanisms and their impact on diurnal cycle of MCSs could be slightly revised for improved clarity (e.g., L304-308). My takeaway is that both mechanisms contribute to anvil production of MCSs. Daytime MCSs are primarily dominated by the lapse rate mechanism with a small contribution from differential radiation mechanism, whereas nighttime MCSs are largely driven by clear-sky convergence with almost no contribution from vertical destabilization. Thus, daytime MCSs produce much stronger anvil clouds than nighttime MCSs, leading to the diurnal cycle.
• Figure 8: The conclusion presented in the diagram is unclear based on the accompanying text. While there is more anvil clouds in daytime MCSs compared to night-time MCSs, the existence of night-time MCSs still trap LW radiation and reduce outgoing LW radiation (compared to a clear-sky scenario), thereby contributing to a net warming effects on the Earth. How does “less anvil” lead to “more outgoing LW radiation”? Is the author suggesting less anvil and more outgoing LW radiation over time?

Minor comments:

• Fig 2c: If Figure 2c shows the accumulated anvil production of MCSs over the development/decay periods, should the unit be km^2 instead?
• L229 “Here, the hourly anvil producing efficiency refers to the hourly anvil area produced in the MCS decay process.”: Since the unit is in %, shouldn’t it be further divided by the mean value instead?

• How does the clear-sky region define when calculating clear-sky radiative cooling?
• L317 “imposes a strong forcing on the Earth radiative energy budget”: Forcing usually refers to external factors acting on the system, such as increasing carbon dioxide concentration, aerosols, etc. Recommend changing the term to something like “plays an important role in Earth’s energy budget’ or something similar.

• L376: It seems that the unit “W m-2 K-1” is incorrect as it’s inconsistent with the conclusion (L426, W m-2). In addition, it is somewhat unclear how this value should be compared with the high-cloud altitude and tropical anvil cloud area feedbacks from Sherwood et al (2020) and how the difference should be interpreted. This seems to be an apple-to-orange comparison as the unit differ, and the feedback values from Sherwood et al (2020) should already be weighted by the ratio of regional area to global area.
• L381: It would be helpful to provide an estimate of the phase shift in response to warming based on previous literatures, such as an order of magnitude or an estimated range.

Text:

• L107: should be “and” rather than comma (i.e., cold cores and cold centers).
• L159: Jule -> July
• L379: “The” should be lowercase

Review of “Anvil-radiation diurnal interaction: Shortwave radiative-heating destabilization driving the diurnal variation of convective anvil outflow and its modulation on the radiative cancellation”

In this manuscript, the author presents novel research into the lifecycle of deep convective clouds, and how their anvil cloud amounts cloud radiative effect vary with the diurnal cycle. This study builds on a relatively unexplored area of the dependence of anvil cloud radiative effect on the diurnal cycle but unlike previous studies (e.g. Nowicki & Merchant, 2004; Bouniol et al., 2021; Jones et al., 2024) focuses on the impact of diurnal differences in anvil cloud processes rather than shifts in the timing of convection. This research also expands on previous studies into how the diurnal cycle of radiative heating affects anvil development (e.g. Wall et al. 2020; Gasparini et al. 2022), provides additional evidence for these processes by placing observations of anvils within the diurnal cycle and considering how changes in these processes may result in climate feedbacks. To conduct this research, the author has developed a novel dataset that combines Lagrangian properties of mesoscale convective systems tracked using geostationary satellite images with collocated retrievals from Cloudsat, MODIS and CERES. This new dataset helps resolve some of the shortcomings with studies using only one source of observations, and may also provide further results in future studies.

The manuscript is generally well written and presented, and presents impactful results. The introduction provides a clear background to the problem, although could include a little more discussion of some more recent literature. The methods and results are concisely described, and the discussion and conclusions are well described.

Overall, I have three main criticisms of the manuscript:

1. Use of non-standard acronyms and terminology makes some parts of the manuscript confusing, and in addition parts of the methods and results sections could be reworded to improve clarity
2. There is a lack of discussion of sources of bias in the data and how this may effect the results, and some of the decisions made in designing the study, while valid, would be improved by adding a description explaining why these choices were made.
3. The results in section 3.2 do not consider the different parts of the anvil lifecycle as discussed in section 3.1, although deeper discussion of changes in anvil CRE throughout the lifecycle may be more suitable for a subsequent paper.

Comments:

Line 31: Referencing some of the recent evidence for alternative anvil CRE feedback mechanisms (e.g. Sokol et al., 2024; McKinnon et al. 2024; Raghuraman et al. 2024) could help reinforcement this point about novel feedback processes and the importance of this study

Line 44: The anvil cloud area feedback is commonly explained through the radiative Iris effect (Bony et al., 2016), although this is somewhat disputed. More recent research has highlighted the importance of anvil structure to anvil CRE (see previous comment for references)

Line 59: Suggestion “in the widely-used Euler-view grid data set” “in widely-used Eulerian gridded data”

Line 130: What happens to the regions between the definitions of cold-core and anvil with precip > 1mm/hour and < 6 mm/hour?

Line 143: Are the ToA fluxes provided by the Cloudsat algorithm (i.e. are 1D slices through observed systems) or are 2D fields provided by CERES (covering the whole anvil)?

Line 159: Note that removing systems that intersect the edge of the domain may cause a low bias in the number of larger and longer lived systems included in the statistics, but it is difficult to account for (see deWitt et al. 2024).

Figure 2: It may be clearer to split 2a into two plots, one showing the number of MCSs observed with each peak BT, and another showing the % contributions to total anvil coverage from developing and decaying anvils. It would be useful to show lines for the total duration and the anvil production in figures 2b/c

Line 180: It may be clearer to refer to “convective peak BT” rather than “convective peaking strength” throughout. While colder BTs are strongly correlated with more intense convection, they can also be due to meteological or other differences

Line 181: The division of MCS lifecycle is different to previous methods for separating convective lifecycles, such as the growing, mature, dissipating (e.g. Futyan and del Genio, 2007) or splitting developing and decaying based on the time which the anvil reaches its largest horizontal extent (e.g. Roca et al. 2017). The choice of coldest BT is valid, and could be a useful consideration regarding anvil CRE, but I would like to see a discussion on why this is chosen and how it differs from other methods of segmenting the lifecycle. In particular, the time of coldest BT will tend to occur earlier in the lifecycle than the largest area. Also, MCSs which last for multiple days may go through multiple cycles of invigoration with the diurnal cycle, resulting in multiple peaks of the anvil BT throughout the MCSs lifetime. Choosing a single time may lead to anomalies with these systems, although they are likely rare and so should not have significant effects on any of the statistics.

Line 192: “Anvil production” sounds like a rate, e.g. the increase in anvil area per hour, rather than the total amount. It may be clearer to refer to this quantity as accumulated anvil area or total anvil coverage

Line 196: While the colder peak BT anvils produce larger anvils, I’m not sure it can be stated that they are more efficient at producing anvils. I would consider efficiency as a measure of anvil area vs core area/updraft strength

Line 229: It is unclear what “hourly anvil producing efficiency” means. Is this just the mean anvil area over the decaying portion of the MCS lifetime? I.e. the red line in fig 2c divided by the lifetime?

Line 231: Would it be simpler to say that there is more variation in the average area of anvils than the anvil lifetime with the diurnal cycle?

Line 265 onwards (discussion of figure 4): Is the sampling of anvil heating rates uniform across the anvil? Compositing 1D CloudSat overpasses to produce statistics across 2d anvils can lead to biases. For example, if the overpass is directly through the core of the MCS, then a larger proportion of the core will be sampled than the anvil further from the core, leading to a cold bias in the measured fluxes. The preprint of Igel et al. (2014) discussed the statistics of this, but I don’t think that a final version was published. A simple way to check this would be to plot the number of CloudSat samples vs the distance to the MCS cold core; if the sampling is uniform then this should show a linear relationship.

Figure 4: This is a really nice figure. It might be clearer to use different colour scales for destabilisation and divergence to emphasise the differences shown

Line 305: This discussion could be enhanced by also including the average column heating rates, destabilisation and divergence for anvils in figure 5, along with the clear-sky rates, so that the differences can be seen more clearly

Section 4: How does the interaction between the MCS lifecycle and the timing of the CloudSat overpass affect the measured anvil CRE? E.g. for a Cloudsat overpass at 13:30 LT, an MCS than peaks at 12:00 LT will be in the decaying phase while one that peaks at 15:00 LT will be observed in the developing phase. How much does the difference in lifecycle (and possible difference in anvil area/structure) affect the measured CRE compared to the change seen over the diurnal cycle. This contribution may be beyond the scope of this manuscript to explore in depth, but should be mentioned in the text regardless.

Figure 6: Why is this shown in units of kJ, rather than W m^-2? Is this the sum of CRE for anvils over their entire area and lifetime? Could the area weighted ToA flux be used instead to show the CRE in than W m^-2? It may help the explanation of the secondary cancellation to show all of these components. i.e. show the total/average anvil area and lifetime for MCSs peaking at different local times

Line 329: In generally used terminology this is a radiative effect, not a radiative forcing. It may be clearer to change REF to something like “radiative energy contribution” or discuss the area weighted CRE as discussed in the previous comment

Line 339: Are these factors important for the secondary cancellation? Anvil temperature is important for LW cooling, and is important in the primary cancellation. The anvil area coverage is more important for considering feedbacks in anvil area. It might also be useful to note the importance of anvil structure in the radiative cancellation (e.g. Berry and Mace, 2014), which has been shown to be increasingly important for the anvil radiative feedbacks in recent years (e.g. Raghuraman et al. 2024, Sokol et al. 2024, McKim et al. 2024)

Line 340: This also depends on the present day diurnal cycle of DCCs, which is very different between land and ocean.

Figure 7: The colour scales could be adjusted to show differences more clearly

Line 376: Should the units in the value of 11 W m^-2 K^-1 be W m^-2 instead? It is difficult to compare this value to the values from Sherwood et al., as it is not the average global response to a temperature change. How much do you expect the diurnal cycle of anvils to respond to temperature change? Interestingly the value of 11 W m^-2 is similar to the value estimated by Nowick and Merchant (2004) for a 1 hour shift in timing of convection over land.

Technical corrections:

Line 24: These two sentences might be in the wrong order

Line 206: “decay process” -> “decay period”

Line 219: “The MCSs of the peak BT at 195 K have 5310 samples and warmer MCSs are more” rephrase for grammar/clarity  

Line 244/246: “thus to increase” -> “thus increasing”

Line 322: Unfinished sentence

Line 342: “re-disturbed” -> “redistributed”