the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Dec 2024
Analysis of raindrop size distribution from the double moment cloud microphysics scheme for monsoon over a tropical station
Abstract. Accurate precipitation forecasting hinges on the representation of microphysical processes within numerical models. A key approach to understanding these processes is through the analysis of hydrometeor drop size distribution (DSD). The characteristics of DSD bulk parameters:-Mass Weighted Mean Diameter (Dm) and the Normalized Number Concentration parameter (Nw), are estimated from the double moment cloud microphysical scheme (CASIM: Cloud-Aerosol Interacting Microphysics) employed in the operational convection permitted model of National Centre for Medium-Range Weather Fore- casting (NCUM-R). The observations from the Joss-Valdvogel Disdrometer (JWD) and the Global Precipitation Mission – Dual Frequency Precipitation Radar (GPM-DPR) are analyzed for providing essential validation. An algorithm for separating the monsoon precipitation into convective and stratiform types in NCUM-R and a new parameter estimation module to obtain DSD parameters from the CASIM are established in the study. The model exhibits agreement with the characteristics of the DSD of raindrops with Dm ranging from 0.5 mm to 2.5 mm marking the majority of the monsoon precipitation events. However, the underestimation when it comes to the larger drops (with Dm > 3.25 mm and Rainrate >= 8 mm/hr) demands a reassessment in microphysical parameterizations. The advanced autoconversion parameterization scheme applied in CASIM favored the growth of large drops compared to the existing scheme. The enhanced growth of larger drops is reflected in the increased accuracy in the prediction of extreme precipitation associated with a convective event. The current study underscores the importance of refining microphysical parameterizations to improve the accuracy of precipitation forecasts offering a pathway for enhanced model performance in future operational forecasting systems.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(1436 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1436 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-3538', Anonymous Referee #1, 04 Jan 2025
-
AC2: 'Reply on RC1', K s Apsara, 10 Mar 2025
Thank you for the valuable comments. We have carefully addressed all the feedback, and our responses are detailed in the attached document, which includes figures to support our explanations. Please find the attached PDF containing our point-by-point replies.
-
AC2: 'Reply on RC1', K s Apsara, 10 Mar 2025
-
RC2: 'Comment on egusphere-2024-3538', Anonymous Referee #2, 22 Jan 2025
This manuscript is generally nonsense.
Lethal problem:
There is a basic concept error about autoconversion process. The authors state that "The autoconversion process is a primary microphysical process in which the cloud droplets collect the raindrops to form a bigger drop." This is totally wrong. According to glossary of meteorology from AMS, autoconversion means "The initial stage of the collision–coalescence process whereby cloud droplets collide and coalesce to form drizzle drops". This process doesnot directly produce big raindrop at all.
Major comments:
The comparison of DSD parameters from JWD, DPR, and NCUM-R after stratiform-convective separation is even not an apple-to-apple comparison. Three separation algorithms are based on different criteria and physical concept. The same criteria should be used, such as a simple reflectivity threshold of 35 dBZ.
Figure 7: Within the Dm-Nw framework, the precipitation rate R can be directly calculated for a given shape parameter μ. The differences simply come from the sample error and truncation error, which have no physical meaning.
Citation: https://doi.org/10.5194/egusphere-2024-3538-RC2 -
AC1: 'Reply on RC2', K s Apsara, 10 Mar 2025
We sincerely thank the reviewer for pointing out the misstatement regarding the autoconversion process. Upon review, we acknowledge that our original statement was incorrect and does not align with the widely accepted definition provided by the AMS Glossary of Meteorology.
While autoconversion is not a discrete process observed in the real-world continuum of collision-coalescence, it is a conceptual parameterization used in bulk microphysical models. In these models, autoconversion represents the flux of mass and number across a size threshold, which separates the cloud water from rainwater species. Specifically, the parameterization simulates the coalescence of small cloud droplets into drizzle or rain-sized particles, facilitating the conversion of cloud water species into rainwater species. Beyond this threshold, processes such as self-collection within the rain species further increase the mean size of raindrops.
Implications for our Study:
○ Autoconversion remains an essential process to test and evaluate in microphysical schemes, as it plays a critical role in the partitioning of water species and impacts precipitation characteristics in bulk models.
○ This process, while idealized, allows models to account for the continuum of droplet growth and serves as a proxy for the early stages of the
collision-coalescence process
○ The evaluation and sensitivity experiments conducted in this study were not intended to enhance the growth of raindrops but rather to assess how a more advanced parameterization impacts the representation of the drop size distribution (DSD) compared to the existing approach.
We will revise the text in our manuscript as follows to accurately reflect the concept:
"In bulk microphysical models, the autoconversion process is a parameterized mechanism that simulates the transition of cloud water species to rainwater species due to the coalescence of cloud droplets. It represents the flux of mass and number across a size threshold, distinguishing clouds from rain particles. While this process is an idealization, it is crucial for modeling precipitation and requires careful evaluation."
We thank the reviewer for bringing this to our attention and enabling us to improve the scientific accuracy of our manuscript.
Major comments:
1) The comparison of DSD parameters from JWD, DPR, and NCUM-R after stratiform-convective separation is even not an apple-to-apple comparison. Three separation algorithms are based on different criteria and physical concept. The same criteria should be used, such as a simple reflectivity threshold of 35 dBZ.
We acknowledge the concern raised by the reviewer about the ‘apple to apple’ comparability of stratiform and convective separation criteria used in different datasets. While using the uniform reflectivity threshold might provide consistency, at the same time, it overlooks the unique strengths and limitations of each dataset.
1) JWD data lacks a direct calculation of reflectivity measurements; it relies on the rain rate and drop size to obtain reflectivity. So, using rain rate thresholds which is a derived parameter used to classify stratiform and convective rain, is more meaningful and sensible.
2) GPM-DPR, which has radar-based reliable capabilities, uses multiple criteria using both the presence of a bright band and the horizontal method, both evaluate the reflectivity thresholds across the vertical profiles and horizontal gradients, which is not possible to evaluate from JWD as it is a ground observation instrument. (reference: https://doi.org/10.1175/JTECH-D-16-0016.1)
3) NCUM-R is a non-hydrostatic model that uses vertical velocity as a prognostic variable, hence for tropics, it can provide a comprehensive classification of convective and stratiform rain from the embedded convective system. This along with the combined rain rate criteria marks a more reliable classification of rain. (Houze Jr, R. A.: Stratiform precipitation in regions of convection: A meteorological paradox? Bulletin of the American Meteorological Society, 78, 2179–2196, 1997.)
In terms of the above-mentioned points, using data-specific separation criteria is relevant in the context of each system as:
○ Each dataset has unique capabilities and limitations so using the same criteria (35dBZ) does not promise to fulfill the intention of the study.
○ The study tried to use the best possible method for each dataset, as the motive was to more accurately separate the convective and stratiform precipitation, hence evaluating the drop size distribution pattern represented by each dataset and analyzing it further in terms of microphysical processes.
○ Recent work by Peinó et al (https://doi.org/10.3390/rs16142594) used data-specific criteria for similar classifications for their objectives.
2) Figure 7: Within the Dm-Nw framework, the precipitation rate R can be directly calculated for a given shape parameter μ. The differences simply come from the sample error and truncation error, which have no physical meaning.
Thank you for your valuable comment. A similar concern was raised by reviewer 1, and in response, we conducted a detailed analysis to assess the impact of truncation on the mass-weighted mean diameter (Dm). Our approach follows the CASIM particle size distribution (PSD) framework, where Dm is derived as the ratio of the fourth and third moments. Using the regularized incomplete gamma function, we quantified the truncation effect at Dcut, which corresponds to the lowest droplet size detected by JWD.
The results indicate that while truncation influences individual moments, their ratio remains largely unchanged. Given that our histograms of Dm already start above 0.5 mm, the bias introduced by truncation is less than 10%. Consequently, the shape parameter (μ) remains unaffected as Dm is realistic, and the discrepancies between JWD and model-derived Dm are not a result of truncation.
Additionally, sample errors in Dm can be related to errors in the precipitation rate (R) within the Dm-Nw framework. However, since Dm is relatively insensitive to truncation and its effect on μ is negligible, the analysis-based Dm remains valid. The detailed derivation is attached.
-
AC1: 'Reply on RC2', K s Apsara, 10 Mar 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-3538', Anonymous Referee #1, 04 Jan 2025
Comments on “Analysis of raindrop size distribution from the double moment cloud microphysics scheme for monsoon over a tropical station”
General comments:
This paper examines the double moment cloud physics scheme used in the National Centre for Medium-Range Weather Forecasting model with ground-based disdrometer observations. The result that autoconversion schemes affect the mass-weighted mean diameter (Dm) is interesting. However, as the measurement error of the Joss-Waldvogel disdrometer (JWD) on the number concentration of raindrops is not taken into account, there are some questions about the evaluation of the simulation results. I recommend a major revision before publication in ACP.Major comments:
(1) From equations (12) and (15), the 0th and the 3rd moments of DSD are used to estimate Dm in this study. However, it is questionable whether JWD can accurately measure the 0th moment (number concentration of raindrops). The dead time problem and the cut-off at 0.3mm are the reasons for this. According to equation (12), underestimation of the number concentration of raindrops by JWD leads to underestimation of λ, which in turn leads to overestimation of Dm. It can therefore be assumed that the difference in Dm between the JWD and the simulation is due to errors in the JWD measurement. The DSDs from JWD should be corrected, for example, using the method of Raupach et al. (2019).(2) P12 Lines 267-269: The authors state that “the model shows agreement with the JWD and GPM for raindrops with a maximum frequency of occurrence of Dm between 1 mm and 2 mm”, but the model seems to overestimate Nw compared to the JWD. It is desirable to have a quantitative comparison between the simulation and the observation.
Minor comments:
(3) Equation (8): 103 should be in the numerator because the unit of LWC is [g/m3] and the unit of ρw is [kg/m3].(4) Fig. 7c: The plots for convective precipitation are not visible due to overlap.
References
Raupach, T. H., Thurai, M., Bringi, V. N., and Berne, A. 2019: Reconstructing the Drizzle Mode of the Raindrop Size Distribution Using Double-Moment Normalization, J. Appl. Meteorol. Clim., 58, 145–164.Citation: https://doi.org/10.5194/egusphere-2024-3538-RC1 -
AC2: 'Reply on RC1', K s Apsara, 10 Mar 2025
Thank you for the valuable comments. We have carefully addressed all the feedback, and our responses are detailed in the attached document, which includes figures to support our explanations. Please find the attached PDF containing our point-by-point replies.
-
AC2: 'Reply on RC1', K s Apsara, 10 Mar 2025
-
RC2: 'Comment on egusphere-2024-3538', Anonymous Referee #2, 22 Jan 2025
This manuscript is generally nonsense.
Lethal problem:
There is a basic concept error about autoconversion process. The authors state that "The autoconversion process is a primary microphysical process in which the cloud droplets collect the raindrops to form a bigger drop." This is totally wrong. According to glossary of meteorology from AMS, autoconversion means "The initial stage of the collision–coalescence process whereby cloud droplets collide and coalesce to form drizzle drops". This process doesnot directly produce big raindrop at all.
Major comments:
The comparison of DSD parameters from JWD, DPR, and NCUM-R after stratiform-convective separation is even not an apple-to-apple comparison. Three separation algorithms are based on different criteria and physical concept. The same criteria should be used, such as a simple reflectivity threshold of 35 dBZ.
Figure 7: Within the Dm-Nw framework, the precipitation rate R can be directly calculated for a given shape parameter μ. The differences simply come from the sample error and truncation error, which have no physical meaning.
Citation: https://doi.org/10.5194/egusphere-2024-3538-RC2 -
AC1: 'Reply on RC2', K s Apsara, 10 Mar 2025
We sincerely thank the reviewer for pointing out the misstatement regarding the autoconversion process. Upon review, we acknowledge that our original statement was incorrect and does not align with the widely accepted definition provided by the AMS Glossary of Meteorology.
While autoconversion is not a discrete process observed in the real-world continuum of collision-coalescence, it is a conceptual parameterization used in bulk microphysical models. In these models, autoconversion represents the flux of mass and number across a size threshold, which separates the cloud water from rainwater species. Specifically, the parameterization simulates the coalescence of small cloud droplets into drizzle or rain-sized particles, facilitating the conversion of cloud water species into rainwater species. Beyond this threshold, processes such as self-collection within the rain species further increase the mean size of raindrops.
Implications for our Study:
○ Autoconversion remains an essential process to test and evaluate in microphysical schemes, as it plays a critical role in the partitioning of water species and impacts precipitation characteristics in bulk models.
○ This process, while idealized, allows models to account for the continuum of droplet growth and serves as a proxy for the early stages of the
collision-coalescence process
○ The evaluation and sensitivity experiments conducted in this study were not intended to enhance the growth of raindrops but rather to assess how a more advanced parameterization impacts the representation of the drop size distribution (DSD) compared to the existing approach.
We will revise the text in our manuscript as follows to accurately reflect the concept:
"In bulk microphysical models, the autoconversion process is a parameterized mechanism that simulates the transition of cloud water species to rainwater species due to the coalescence of cloud droplets. It represents the flux of mass and number across a size threshold, distinguishing clouds from rain particles. While this process is an idealization, it is crucial for modeling precipitation and requires careful evaluation."
We thank the reviewer for bringing this to our attention and enabling us to improve the scientific accuracy of our manuscript.
Major comments:
1) The comparison of DSD parameters from JWD, DPR, and NCUM-R after stratiform-convective separation is even not an apple-to-apple comparison. Three separation algorithms are based on different criteria and physical concept. The same criteria should be used, such as a simple reflectivity threshold of 35 dBZ.
We acknowledge the concern raised by the reviewer about the ‘apple to apple’ comparability of stratiform and convective separation criteria used in different datasets. While using the uniform reflectivity threshold might provide consistency, at the same time, it overlooks the unique strengths and limitations of each dataset.
1) JWD data lacks a direct calculation of reflectivity measurements; it relies on the rain rate and drop size to obtain reflectivity. So, using rain rate thresholds which is a derived parameter used to classify stratiform and convective rain, is more meaningful and sensible.
2) GPM-DPR, which has radar-based reliable capabilities, uses multiple criteria using both the presence of a bright band and the horizontal method, both evaluate the reflectivity thresholds across the vertical profiles and horizontal gradients, which is not possible to evaluate from JWD as it is a ground observation instrument. (reference: https://doi.org/10.1175/JTECH-D-16-0016.1)
3) NCUM-R is a non-hydrostatic model that uses vertical velocity as a prognostic variable, hence for tropics, it can provide a comprehensive classification of convective and stratiform rain from the embedded convective system. This along with the combined rain rate criteria marks a more reliable classification of rain. (Houze Jr, R. A.: Stratiform precipitation in regions of convection: A meteorological paradox? Bulletin of the American Meteorological Society, 78, 2179–2196, 1997.)
In terms of the above-mentioned points, using data-specific separation criteria is relevant in the context of each system as:
○ Each dataset has unique capabilities and limitations so using the same criteria (35dBZ) does not promise to fulfill the intention of the study.
○ The study tried to use the best possible method for each dataset, as the motive was to more accurately separate the convective and stratiform precipitation, hence evaluating the drop size distribution pattern represented by each dataset and analyzing it further in terms of microphysical processes.
○ Recent work by Peinó et al (https://doi.org/10.3390/rs16142594) used data-specific criteria for similar classifications for their objectives.
2) Figure 7: Within the Dm-Nw framework, the precipitation rate R can be directly calculated for a given shape parameter μ. The differences simply come from the sample error and truncation error, which have no physical meaning.
Thank you for your valuable comment. A similar concern was raised by reviewer 1, and in response, we conducted a detailed analysis to assess the impact of truncation on the mass-weighted mean diameter (Dm). Our approach follows the CASIM particle size distribution (PSD) framework, where Dm is derived as the ratio of the fourth and third moments. Using the regularized incomplete gamma function, we quantified the truncation effect at Dcut, which corresponds to the lowest droplet size detected by JWD.
The results indicate that while truncation influences individual moments, their ratio remains largely unchanged. Given that our histograms of Dm already start above 0.5 mm, the bias introduced by truncation is less than 10%. Consequently, the shape parameter (μ) remains unaffected as Dm is realistic, and the discrepancies between JWD and model-derived Dm are not a result of truncation.
Additionally, sample errors in Dm can be related to errors in the precipitation rate (R) within the Dm-Nw framework. However, since Dm is relatively insensitive to truncation and its effect on μ is negligible, the analysis-based Dm remains valid. The detailed derivation is attached.
-
AC1: 'Reply on RC2', K s Apsara, 10 Mar 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 834 | 181 | 24 | 1,039 | 17 | 34 |
- HTML: 834
- PDF: 181
- XML: 24
- Total: 1,039
- BibTeX: 17
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Kadavathu Sreekumar Apsara
Jayakumar Aravindakshan
Anurose Theethai Jacob
Saji Mohandas
Paul Field
Hamish Gordan
Thara Prabhakaran
Mahen Konwar
Vijapurap Srinivasa Prasad
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1436 KB) - Metadata XML
Comments on “Analysis of raindrop size distribution from the double moment cloud microphysics scheme for monsoon over a tropical station”
General comments:
This paper examines the double moment cloud physics scheme used in the National Centre for Medium-Range Weather Forecasting model with ground-based disdrometer observations. The result that autoconversion schemes affect the mass-weighted mean diameter (Dm) is interesting. However, as the measurement error of the Joss-Waldvogel disdrometer (JWD) on the number concentration of raindrops is not taken into account, there are some questions about the evaluation of the simulation results. I recommend a major revision before publication in ACP.
Major comments:
(1) From equations (12) and (15), the 0th and the 3rd moments of DSD are used to estimate Dm in this study. However, it is questionable whether JWD can accurately measure the 0th moment (number concentration of raindrops). The dead time problem and the cut-off at 0.3mm are the reasons for this. According to equation (12), underestimation of the number concentration of raindrops by JWD leads to underestimation of λ, which in turn leads to overestimation of Dm. It can therefore be assumed that the difference in Dm between the JWD and the simulation is due to errors in the JWD measurement. The DSDs from JWD should be corrected, for example, using the method of Raupach et al. (2019).
(2) P12 Lines 267-269: The authors state that “the model shows agreement with the JWD and GPM for raindrops with a maximum frequency of occurrence of Dm between 1 mm and 2 mm”, but the model seems to overestimate Nw compared to the JWD. It is desirable to have a quantitative comparison between the simulation and the observation.
Minor comments:
(3) Equation (8): 103 should be in the numerator because the unit of LWC is [g/m3] and the unit of ρw is [kg/m3].
(4) Fig. 7c: The plots for convective precipitation are not visible due to overlap.
References
Raupach, T. H., Thurai, M., Bringi, V. N., and Berne, A. 2019: Reconstructing the Drizzle Mode of the Raindrop Size Distribution Using Double-Moment Normalization, J. Appl. Meteorol. Clim., 58, 145–164.