the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Nov 2024
Relationship between latent and radiative heating fields of Tropical cloud systems using synergistic satellite observations
Abstract. In order to investigate the relationship between latent and radiative heating (LH, RH), particularly within mesoscale convective systems (MCSs), we used synergistic satellite-derived data from active instruments. Given the sparse sampling of these observations, we expanded the Spectral LH profiles derived from the Tropical Rain Measurement Mission (TRMM-SLH) by applying artificial neural network regressions on Clouds from InfraRed Sounder data and meteorological reanalyses, following a similar approach as for the expansion of the RH profiles. The zonal averages of vertically integrated LH (LP) at 1:30 AM and PM LT align well with those from the full diurnal sampling of TRMM–SLH over ocean. For Upper Tropospheric (UT) clouds releasing large latent heat, the surface temperature has a larger impact on the atmospheric cloud radiative effect (ACRE) in dry than in humid environments, while for lower clouds, producing relatively small latent heat, humidity plays a large role in enhanced ACRE. The distribution of UT clouds in the LP–ACRE plane shows a very large spread in ACRE for small LP, which is gradually reduced towards larger LP. The mean ACRE per MCS increases with LP, ranging from 50 to 115 W m-2. As expected, the shapes of the LH profiles of mature MCSs show that larger, more organized MCSs have a larger contribution of stratiform rain than the smaller MCSs. Convective organization enhances the mean ACRE of the MCS by up to 20 W m-2.
- Preprint
(9383 KB) - Metadata XML
-
Supplement
(3964 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-3434', Anonymous Referee #1, 03 Dec 2024
Latent heating and radiative heating are two major diabatic heating sources of the atmosphere. However, the observation of global latent and radiative heating profiles are difficult to obtain. Active sensors onboard satellites provide a way to measure these profiles, such as the products from TRMM and GPM. However, TRMM and GPM satellites are low-orbit satellites thus the spatiotemporal coverage of the heating profile data are limited. The previous work from the author(s) have applied a ML technique to extend the radiative heating profile to greater coverage. This paper does similar thing but for latent heating (LH), with further analysis of LH profiles for different surface conditions, different environmental conditions and MCS characteristics. One of the main conclusions is that although the mean values show good agreement, the ML-expanded latent heating profiles show much smaller variability than the target data. This is different from the ML-expanded radiative heating profiles. The authors made a further analysis and suggested that the data are only robust for coarse-resolution larger than 2.5deg.
Although some results of this paper are interesting and this expanded dataset would be useful for scientific community, I feel that the values and limitations need to be carefully elaborated and better demonstrated. One major concern I have is that since the authors explicitly suggest that “our ML-expanded LH dataset is suitable at scales larger than about 2.5◦”, how robust can we trust the results for MCSs in the size of around 1x1deg, or 100kmx100km, as the authors show in Fig. 13 to 15?
Besides this question, I also have some specific comments listed below:
Line 24: as RH is usually referred as relative humidity, maybe change the abbreviation (e.g. Qrad)
Line 61: maybe need a little bit more introduction on CRE and ACRE, for example, from which data source did you obtain the CRE/ACRE data (clear-sky and full-sky radiation)? Sometimes the authors state CRE and radiative heating profiles together (even the units in Fig.5 have K/day and W/m2 on the same panel, this confuses me a bit), do they represent similar effect?
Fig.1: more details of the figure are needed: what do the different blue colors mean? what are the thin and thick swaths?
Line 160: maximum should be minimum
Line 162: what are the thresholds in your rain intensity categorization?
Line 164: How do you define UT clouds in this paper?
Line 192: I am not an expert of machine learning, but 20 iterations looks a lot to me if you don't see any improvement in the loss function. Is it true that it is still not overfitting after 20 iterations without improvement?
Line 247: Is there any explanation of the two peaks of LH profiles over ocean?
Fig. 5: which variables are the left and right y axes corresponding to?
Fig. 5: why is the SW radiative effect is almost zero? My direct intuition is that when cloud exists, shortwave radiative effect should be pretty negative.
Fig. 5: The ‘pink solid line’ mentioned in the caption is not shown in the figure.
Line 334: do they need to be neighboring grid cells or only by cloud height to be merged into the same system?
Line 335: I don’t quite understand this sentence "the size ... is computed... by the number...". Is there anything missing?
Fig. 9: this is another place that I don’t quite understand what this paper use, is it CRE or radiative heating?
Line 364: not sure I understand this statement. Are you comparing MCS intensity relations with MCS size and opacity? or comparing radiative heating profile with LH profiles?
Line 372-373: from where did you obtain this conclusion that LP is more reliable over ocean?
Fig. 14: this is an interesting result that LP-ACRE relation shows no distinction between developing, mature and dissipating stages, but is dependent on MCS sizes. Not sure how to understand/explain this.
Line 457: What are ‘the larger ones’ refer to?
Citation: https://doi.org/10.5194/egusphere-2024-3434-RC1 - AC1: 'Reply on RC1', Xiaoting Chen, 06 Mar 2025
-
RC2: 'Comment on egusphere-2024-3434', Anonymous Referee #2, 24 Jan 2025
Review of “Relationship between latent and radiative heating fields of Tropical
cloud systems using synergistic satellite observations”, by Chen et al.
General comments:
This study investigates the latent and radiative heating fields of tropical cloud systems using synergistic satellite observations. The artificial neural network (ANN) regression is used to generate 'observational data' based on limited satellite observations and meteorological reanalyses. This work could be useful for understanding tropical cloud systems, particularly mesoscale convective systems. Overall, the paper is well-written. However, the presentation and analysis need improvement before the paper is suitable for publication.
Recommendation: Major revision
Major comments:
1. Some conclusions are not well supported by the presented results. For example, in lines 515-518, it is stated that "The smaller CIRS-retrieved height may be interpreted as anvils of convective systems having descended at a later stage of their life cycle or as relatively thick clouds with diffusive tops, for which the retrieved (radiative) height may be deeper within the cloud because of very small ice water content in the upper part of the cloud." However, the authors do not provide any related analysis to support this conclusion.
2. Lines 144-145: “The rain intensity classification (no rain, light rain, heavy rain), determined by an ANN trained with precipitation data from CloudSat, considers light rain to be < 5 mm h−1 and heavy rain > 5 mm h−1 (Stubenrauch et al., 2023).”
- a ) The GPCP and TRMM precipitation data are widely used by the community. Could you provide a justification for why you chose to use CloudSat precipitation data? Also, please include the link to the CloudSat precipitation data in the Data Availability Statement.
- b) How did you handle data when the rain rate is exactly 5 mm h−1?
3. Could you please provide the definitions of UT and UT clouds in this study? How are the CRE and ACRE calculated, given that they represent the clouds' impact on radiative heating?
4. Different time periods are selected for analysis (e.g., Fig. 2 (2008-2013), Fig. 3 (2004-2013/2007-2010), Fig. 9 (2004-2018)). Why were these specific time periods chosen?
5. Line 336: “An MCS is defined as an UT cloud system with at least one convective core and the presence of precipitation.” How did you distinguish between an MCS and an isolated deep convective cloud system?
6. Line 293-295: “In summary, the increasing slopes of the relationship between LP of CIRS–ML and TRMM suggest that our ML-expanded LH dataset is suitable at scales larger than about 2.5° (with a slope of 0.7).”
- a) It would be better to present a plot with the 2.5° scale in Figure 6.
- b) Line 501: “with slopes of 0.68 and 0.76 for 2.5° and 5°, respectively.” Please check the slope for the 2.5° scale: is it 0.68 or 0.7?
7. Figure 11 shows negative ACRE when LP < 200 W/m² for precipitating UT clouds over the ocean. However, Figure 10b shows positive ACRE when LP < 200 W/m² for precipitating UT clouds over the ocean. Why is there a discrepancy?
Minor comments:
1. Line 101-102: “Its spectral coverage spans 2378 radiance channels within the wavelength range of 3.7–15.4 μm (650–2700 cm−1)”
What does “650–2700 cm−1” represent?
2. Line 214-215: “In all cases, the real data reveal a large variability between 600 and 800 hPa, which may be linked to the variability between stratiform and convective rain within the 0.5°.” Why?
3. Line 232: “The minor cooling observed at approximately 550 hPa is attributed to the melting process” Why does melting induce cooling in the radiative longwave (LW) heating profile?
4. Figure 5: Why does the vertically integrate LH still have the units of K/day? Which lines represent 9:30 AM/PM?
5. Figure 7: “10-year (2008–2018 JAN) averages” Should it be 11 years instead?
6. Figure 8: Can you show the plot of the difference between CIRS-ML and TRMM/GPCP? The title of the label bar should be LP instead of LH.
7. Line 348: “Figure 9 presents profiles of latent heating and radiative heating….” The label of Figure 9b indicates that the cloud radiative effect (CRE) is presented.
8. Line 440: circulation Stephens et al. (2024) -> circulation (Stephens et al. 2024).
9. Line 405 : “Humid environments increase the buoyancy of convective clouds, which allows clouds to reach greater heights (Holloway and Neelin, 2009), confirmed by Fig. 10b”. However, Fig. 10b does not provide any information about cloud heights.
10. Line 461: “20 km2” -> “20 W m-2”
11. “LP” is used to represent two different physical variables in one study.
- a) Line 60: “In our analyses, we use the following definitions: LH for latent heating profile, LP for vertically integrated LH”
- b) Line 453: “The relationship between precipitation intensity (LP) and radiative enhancement (ACRE)”
Citation: https://doi.org/10.5194/egusphere-2024-3434-RC2 - AC2: 'Reply on RC2', Xiaoting Chen, 06 Mar 2025
-
EC1: 'Comment on egusphere-2024-3434', Shaocheng Xie, 30 Jan 2025
Dear Authors,
I have received the third reviewer's report (attached). Please make sure to address these comments in your response as well.
Thank you
Shaocheng Xie
Editor
Atmospheric Chemistry and Physics
-
AC3: 'Reply on EC1', Xiaoting Chen, 06 Mar 2025
The authors would like to sincerely thank all three reviewers and the editor for their constructive and helpful comments, which have greatly improved the clarity of the manuscript.
A detailed response to all comments is provided in the attached file.
-
AC3: 'Reply on EC1', Xiaoting Chen, 06 Mar 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 200 | 65 | 14 | 279 | 35 | 7 | 13 |
- HTML: 200
- PDF: 65
- XML: 14
- Total: 279
- Supplement: 35
- BibTeX: 7
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 109 | 41 |
| China | 2 | 29 | 11 |
| France | 3 | 29 | 11 |
| Germany | 4 | 11 | 4 |
| Russia | 5 | 9 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 109