Analysis of the Occurrence Frequency of Seedable Clouds on the Korean Peninsula for Precipitation Enhancement Experiments
Round 1
Reviewer 1 Report
This paper analyzes the occurrence frequency and distribution of seedable clouds around the Korean Peninsula by using COMS cloud product. I recommend this paper to be published after minor revision. 1. In abstract, the authors say ‘Cloud products from the Communication, Ocean, and Meteorological Satellite (COMS), including cloud fraction, cloud top height, cloud top temperature, cloud phase, cloud top pressure, and cloud optical thickness were used’, and in section 2, ‘The data used for analysis consisted of cloud fraction, cloud top height, cloud top temperature, and cloud phase information…’. So, what is the data source for rainfall intensity? And did the authors classify cloud phase by themselves, or use the cloud phase from COMS product ? Section 2 should be rewritten in a clearer way. And please give the uncertainties of the COMS cloud retrievals, and also some discussions on the possible influences of the uncertainties on the results in this paper. 2. Line 106-107: Abbreviation should be defined the first time it is used in the main text. What is the definition of the cloud fraction used in the paper? 3. For passive sensor, the upper mid- or high- level thick cloud can block the detection of the overlaid low cloud. Is there any influence of this issue on the seasonal statistical results derived in this paper?Author Response
This paper analyzes the occurrence frequency and distribution of seedable clouds around the Korean Peninsula by using COMS cloud product. I recommend this paper to be published after minor revision.
We appreciate the valuable opinions and comments of the reviewer, thanks to which we could revise our manuscript to largely enhance its quality. We have made the following revisions in accordance with the additional comments of the reviewer.
1. In abstract, the authors say ‘Cloud products from the Communication, Ocean, and Meteorological Satellite (COMS), including cloud fraction, cloud top height, cloud top temperature, cloud phase, cloud top pressure, and cloud optical thickness were used’, and in section 2, ‘The data used for analysis consisted of cloud fraction, cloud top height, cloud top temperature, and cloud phase information…’. So, what is the data source for rainfall intensity?
We have revised some sentence follow as:
“Cloud products from the Communication, Ocean, and Meteorological Satellite (COMS), including cloud fraction, cloud top height, cloud top temperature, cloud phase, cloud top pressure, and cloud optical thickness were used.” to “Cloud products from the Communication, Ocean, and Meteorological Satellite (COMS), including cloud fraction, cloud top height, cloud top temperature, cloud phase, cloud top pressure, cloud optical thickness, and rainfall intensity were used.”
and “The data used for analysis consisted of cloud fraction, cloud top height, cloud top temperature, and cloud phase information collected from hourly data between 0900 and 1800 LST observed over three years from December 2016 to November 2019.” to “The data used for analysis consisted of cloud fraction (CF), cloud top height (CTH), cloud top temperature (CTT), cloud phase, cloud top pressure (CTP), cloud optical thickness (COT), and rainfall intensity (RI) information collected from hourly data between 0900 and 1800 LST observed from December 2016 to November 2019.”.
And did the authors classify cloud phase by themselves, or use the cloud phase from COMS product ? Section 2 should be rewritten in a clearer way. And please give the uncertainties of the COMS cloud retrievals, and also some discussions on the possible influences of the uncertainties on the results in this paper.
We have revised Section 2 and have added uncertainties of the COMS cloud products:
“In this study, the COMS cloud products used for the determination of seedable clouds with uncertainties were calculated as follows: Cloud detection using infrared channel brightness temperature (BT) was estimated using the split window and threshold methods, and the probability of detection (POD) was found to be above 88% compared with moderate resolution imaging spectroradiometer (MODIS) data [43]. The cloud phase was estimated via the threshold method using the infrared channel BT of water vapor, which showed an average POD of 80% based on the MODIS cloud phase [44]. Additionally, in accordance with the reflectivity of the 0.6-µm channel, the COT was calculated using a simulated look-up table method depending on the different atmospheric conditions. The RMSE obtained on the basis of MODIS COT were about 3 and 5 in the water and ice phases, respectively [45]. For optically thick clouds (COT > 10), CTT and CTP were estimated using BT10.8, whereas for optically thin clouds, they were estimated using ratio of BT10.8 and BT6.7. The uncertainty of CTT and CTP was 3 K and 50 hPa, respectively [31,46]. Based on regression method using the infrared channel BT, RI was calculated, and in accordance with SSM/I, it showed RMSE below 1 mm [47]. The abovementioned verification values also included artificial errors resulting from the matching of the spatiotemporal resolution of COMS and MODIS data for verification [48–50].”
2. Line 106-107: Abbreviation should be defined the first time it is used in the main text.
We have checked the entire manuscript to ensure that all the abbreviations have been defined at their first mention.
What is the definition of the cloud fraction used in the paper?
We have added the definition of the cloud fraction in the revised manuscript as follows:
“The CF represents the percentage of the clouds detected within 7 × 7 grids around the reference grid [34]. Since the uncertainty of cloud detection increased as CF decreased toward the edge of the cloud [35], sufficient CF, i.e., exceeding 80%, was defined as a condition in this study. Additionally, for a cloud to be used for seeding, it must have a suitable CTH and little or no natural precipitation.”
For passive sensor, the upper mid- or high- level thick cloud can block the detection of the overlaid low cloud. Is there any influence of this issue on the seasonal statistical results derived in this paper?
We have added discussion on the issue as follows:
“Because of the nature of passive sensors in geostationary satellites, low-level atmospheric information of clouds observed by the satellite cannot be determined. As a result, lower-level clouds with a top height exceeding 4 km were not detected. Joiner et al. [51] conducted a multilayer cloud detection study using MODIS data. They found that on an average, mid-latitudes generated about 11% of multilayered clouds, and that the occurrence frequency of multilayered clouds on land was higher than that on the sea, i.e., 12% and 10%, respectively. Therefore, in this study, the occurrence frequency of seedable clouds was analyzed by detecting clouds with limited height; however, it can be assumed that in practice, the occurrence frequency of seedable clouds is higher or similar than the occurrence frequency observed in this study even given the uncertainty shown above. Additionally, the average CTH of seedable clouds detected under limited CTH (~4 km) may be lower than seedable clouds detected under unlimited CTH. Similarly, CTT may be higher.”
Reviewer 2 Report
The manuscript titled “Analysis of the occurrence frequency of seedable clouds on the Korean Peninsula for precipitation enhancement experiments” examined spatial-temporal distribution of the occurrence frequency (OF) of seedable clouds, and characteristics of these seedable clouds (in different phases and types) around the Korean Peninsula based on cloud products of COMS. The main conclusion of this study is that seedable cloud OF shows remarkable seasonal variation: Sc exhibits the highest OF in all seasons, followed by Cu, As, and Ac in a decreasing order. The authors further show that western and eastern seas around the Korean Peninsula and mountain areas on land are suitable for the precipitation enhancement experiments (PEE). In general, I think this manuscript of the current version is not acceptable due to following reasons. I suggest major revision for reconsideration.
My main concerns:
There is a lack of uncertainty analysis, which I think is very important. Specifically, firstly, the authors need to present the accuracy in variables from the COMS cloud products like cloud fraction, cloud top height, temperature, and pressure; secondly, the authors need to include sensitivity tests on threshold values appeared in the seedable cloud detection algorithms (Figure 1) to answer if the derived OF of feedable cloud fraction is sensitive to these thresholds. At least, you should have discussions on uncertainty of your results. Similarly, I suggest to add discussions on the accuracy in categorizing clouds into water, ice, and mixed phases based on two infrared channels of COMS.
In addition, analysis or discussion on other elements (such as residence time of the seeding material, cloud dynamics and microphysics, cloud environment, presence of nature aerosols, and their interactions) that are important for PEE design is missing. It is unsolid to conclude which area is suitable for PEE, because the presence of feedable clouds of different types, cloud top temperature and pressure are only some of those elements that need to considered when designing PEE. I think it would be helpful to add analysis on some of those elements, especially on cloud characteristics such as liquid or ice water path for clouds of different types, cloud droplet size, and cloud optical thickness (maybe also on that of nature aerosol and that of meteorological fields near the cloud base).
Furthermore, your conclusion regarding the cloud top height is limited by the fact that clouds whose top height exceeds 4 km were not detectable by COMS, which very likely makes the reported average cloud top height of feedable clouds biased toward low values. This should be clearly pointed out.
Finally, results are not fully discussed. Most of the time, the authors simply present the results, but didn’t give explanations.
Minor comments:
Line 14: Spell out LST.
Line 21: Better to be specific: “high”.
Line 29: “Rising surface”: sea surface level?
Line 36-38: Better to be explicit: why “two-thirds annual precipitation in the summer leads to the difficulties...”.
Line 77: Change “using numerical simulation results” to “using numerical simulations”. Delete “with various methods”.
Line 78: What do you mean by saying “The latter”?? “model simulation” is definitely not observational data!
Line 80: Delete “basis research data”, and change “for analyzing” to “to analyze”.
Line 88-89: Can you list some examples?
Line 90: In this section, need more texts to describe the datasets you used. How to retrive CTT, CTH, and so on.
Line 99: Delete “over three years”.
Line 101: Change to “aircraft needs to fly…”.
Line 106: Specify what nature it is.
Line 112-123: Rewrite. This sentence is unclear. Cannot understand your meaning.
Line 117-118: Dry ice as the nucleus of ice?
Line 124: “brightness temperature (BT)”.
Line 124-127: Need a reference to support the criterions.
Line 145-146: “inflow of clouds generated at sea”. Reference or figure?
Line 155: Figure 2: Change the scale. Don’t need to be as large as 100%.
Line 157: CTH already defined before.
Line 173-177: This should be given in the methodology section.
Line 228: Delete “As”.
Line 240-242: Change the unit “K” to “oC”.
Line 306: Change to “seedable clouds of different types”.
Line 306-308: To be explicit: based on exactly what you get this conclusion.
Line 324-325: Why don’t you use products from GK-2A and/or Himawari-8, considering their high spatio-temporal resolution?
Line 310-326: I don’t think this paragraph well fit the conclusion section.
Author Response
The manuscript titled “Analysis of the occurrence frequency of seedable clouds on the Korean Peninsula for precipitation enhancement experiments” examined spatial-temporal distribution of the occurrence frequency (OF) of seedable clouds, and characteristics of these seedable clouds (in different phases and types) around the Korean Peninsula based on cloud products of COMS. The main conclusion of this study is that seedable cloud OF shows remarkable seasonal variation: Sc exhibits the highest OF in all seasons, followed by Cu, As, and Ac in a decreasing order. The authors further show that western and eastern seas around the Korean Peninsula and mountain areas on land are suitable for the precipitation enhancement experiments (PEE). In general, I think this manuscript of the current version is not acceptable due to following reasons. I suggest major revision for reconsideration.
We appreciate the valuable opinions and comments of the reviewer, thanks to which we could revise our manuscript to largely enhance its quality. We have made the following revisions in accordance with the additional comments of the reviewer.
My main concerns:
There is a lack of uncertainty analysis, which I think is very important. Specifically, firstly, the authors need to present the accuracy in variables from the COMS cloud products like cloud fraction, cloud top height, temperature, and pressure; secondly, the authors need to include sensitivity tests on threshold values appeared in the seedable cloud detection algorithms (Figure 1) to answer if the derived OF of feedable cloud fraction is sensitive to these thresholds. At least, you should have discussions on uncertainty of your results. Similarly, I suggest to add discussions on the accuracy in categorizing clouds into water, ice, and mixed phases based on two infrared channels of COMS.
We have added uncertainties of the COMS cloud products:
“In this study, the COMS cloud products used for the determination of seedable clouds with uncertainties were calculated as follows: Cloud detection using infrared channel brightness temperature (BT) was estimated using the split window and threshold methods, and the probability of detection (POD) was found to be above 88% compared with moderate resolution imaging spectroradiometer (MODIS) data [43]. The cloud phase was estimated via the threshold method using the infrared channel BT of water vapor, which showed an average POD of 80% based on the MODIS cloud phase [44]. Additionally, in accordance with the reflectivity of the 0.6-µm channel, the COT was calculated using a simulated look-up table method depending on the different atmospheric conditions. The RMSE obtained on the basis of MODIS COT were about 3 and 5 in the water and ice phases, respectively [45]. For optically thick clouds (COT > 10), CTT and CTP were estimated using BT10.8, whereas for optically thin clouds, they were estimated using ratio of BT10.8 and BT6.7. The uncertainty of CTT and CTP was 3 K and 50 hPa, respectively [31,46]. Based on regression method using the infrared channel BT, RI was calculated, and in accordance with SSM/I, it showed RMSE below 1 mm [47]. The abovementioned verification values also included artificial errors resulting from the matching of the spatiotemporal resolution of COMS and MODIS data for verification [48–50]. Because of the nature of passive sensors in geostationary satellites, low-level atmospheric information of clouds observed by the satellite cannot be determined. As a result, lower-level clouds with a top height exceeding 4 km were not detected. Joiner et al. [51] conducted a multilayer cloud detection study using MODIS data. They found that on an average, mid-latitudes generated about 11% of multilayered clouds, and that the occurrence frequency of multilayered clouds on land was higher than that on the sea, i.e., 12% and 10%, respectively. Therefore, in this study, the occurrence frequency of seedable clouds was analyzed by detecting clouds with limited height; however, it can be assumed that in practice, the occurrence frequency of seedable clouds is higher or similar than the occurrence frequency observed in this study even given the uncertainty shown above. Additionally, the average CTH of seedable clouds detected under limited CTH (~4 km) may be lower than seedable clouds detected under unlimited CTH. Similarly, CTT may be higher.”
In addition, analysis or discussion on other elements (such as residence time of the seeding material, cloud dynamics and microphysics, cloud environment, presence of nature aerosols, and their interactions) that are important for PEE design is missing. It is unsolid to conclude which area is suitable for PEE, because the presence of feedable clouds of different types, cloud top temperature and pressure are only some of those elements that need to considered when designing PEE. I think it would be helpful to add analysis on some of those elements, especially on cloud characteristics such as liquid or ice water path for clouds of different types, cloud droplet size, and cloud optical thickness (maybe also on that of nature aerosol and that of meteorological fields near the cloud base).
Thank you for your suggestion. This study aims to detect seedable clouds and analyzes the frequency of their occurrence using satellite cloud products. Therefore, we used COMS cloud products data. The effective cloud radius (Re) was also obtained from COMS data. Unfortunately, the limitation is that the calculation of Re within the cloud area is inaccurate and it is associated with very high levels of uncertainty. COMS Re data is not used in many studies. Therefore, LWP could not be calculated. And we think that the cloud optical thickness can be inferred depending on the cloud type based on ISCCP classification (thin: 0–3.6, middle: 3.6–23, thick: 23–379).
Furthermore, your conclusion regarding the cloud top height is limited by the fact that clouds whose top height exceeds 4 km were not detectable by COMS, which very likely makes the reported average cloud top height of feedable clouds biased toward low values. This should be clearly pointed out.
We have revised the manuscript as follows:
“Additionally, the average CTH of seedable clouds detected under limited CTH (~4 km) may be lower than seedable clouds detected under unlimited CTH. Similarly, CTT may be higher.”
Finally, results are not fully discussed. Most of the time, the authors simply present the results, but didn’t give explanations.
This study focuses on analyzing seedable cloud occurrence around the Korean Peninsula and aims to detect the suitable areas for cloud seeding experiments. Since discussing the results in the results section and presenting the same discussions in the discussion section will result in needless duplication, explanations were only included in the discussion and conclusion sections.
Minor comments:
Line 14: Spell out LST.
We have added spell “local standard time (LST)”
Line 21: Better to be specific: “high”.
We have revised this sentence “We determined that low-level clouds primarily occur around the Korean Peninsula, and the occurrence frequency of stratiform clouds was high for water phase seedable clouds. The occurrence frequency of cumuliform clouds was also high for ice phase seedable clouds.” to “We determined that low-level clouds primarily occur around the Korean Peninsula, and the occurrence frequency of stratiform clouds was highest for water phase seedable clouds, while the occurrence frequency of cumuliform clouds was highest for ice phase seedable clouds.”
Line 29: “Rising surface”: sea surface level?
We have revised the sentence “Rising surface and atmospheric temperatures due to global warming are causing climate change worldwide.” to “The rise in the temperature of the earth surface and its atmosphere owing to global warming is causing climate change worldwide.”
Line 36-38: Better to be explicit: why “two-thirds annual precipitation in the summer leads to the difficulties...”.
We have revised this sentence “Two-thirds of South Korea's average annual precipitation normally occurs in the summer, leading to difficulties in securing water resources [5–7].” to “Two-thirds annual precipitation in summer leads to the difficulties in securing water resources in South Korea [5–7].”
Line 77: Change “using numerical simulation results” to “using numerical simulations”. Delete “with various methods”.
Line 78: What do you mean by saying “The latter”?? “model simulation” is definitely not observational data!
We have revised this sentence “To verify and analyze precipitation enhancement experiments, studies have been performed using numerical simulation results and ground/space observational data with various methods. The latter have included model simulation, radar, lidar, and satellite [23–27].” to “To verify and analyze precipitation enhancement experiments, studies have been performed using numerical simulations [23] and ground/space observational data (radar, lidar, and satellite) [24-27].”
Line 80: Delete “basis research data”, and change “for analyzing” to “to analyze”.
We have revised this sentence “Koshida et al. [28] and Morrison et al. [29] used satellite data as basic research data for analyzing cloud occurrence and characteristics.” to “Koshida et al. [28] and Morrison et al. [29] used satellite data as to analyze cloud occurrence and characteristics.”
Line 88-89: Can you list some examples?
We have revised this sentence “This research can be utilized by future studies in addition to helping secure future water resources.” to “This research can be utilized by future studies such as relieving drought and preventing forest fire in addition to helping secure future water resources.”
Line 90: In this section, need more texts to describe the datasets you used. How to retrive CTT, CTH, and so on.
We have added the sentence to describe the datasets as follows:
“In this study, the COMS cloud products used for the determination of seedable clouds with uncertainties were calculated as follows: Cloud detection using infrared channel brightness temperature (BT) was estimated using the split window and threshold methods, and the probability of detection (POD) was found to be above 88% compared with moderate resolution imaging spectroradiometer (MODIS) data [43]. The cloud phase was estimated via the threshold method using the infrared channel BT of water vapor, which showed an average POD of 80% based on the MODIS cloud phase [44]. Additionally, in accordance with the reflectivity of the 0.6-µm channel, the COT was calculated using a simulated look-up table method depending on the different atmospheric conditions. The RMSE obtained on the basis of MODIS COT were about 3 and 5 in the water and ice phases, respectively [45]. For optically thick clouds (COT > 10), CTT and CTP were estimated using BT10.8, whereas for optically thin clouds, they were estimated using ratio of BT10.8 and BT6.7. The uncertainty of CTT and CTP was 3 K and 50 hPa, respectively [31,46]. Based on regression method using the infrared channel BT, RI was calculated, and in accordance with SSM/I, it showed RMSE below 1 mm [47].”
Line 99: Delete “over three years”.
We have revised this sentence “The data used for analysis consisted of cloud fraction, cloud top height, cloud top temperature, and cloud phase information collected from hourly data between 0900 and 1800 LST observed over three years from December 2016 to November 2019.” to “The data used for analysis consisted of cloud fraction (CF), cloud top height (CTH), cloud top temperature (CTT), cloud phase, cloud top pressure (CTP), cloud optical thickness (COT), and rainfall intensity (RI) information collected from hourly data between 0900 and 1800 LST observed from December 2016 to November 2019.”
Line 101: Change to “aircraft needs to fly…”.
We have revised this sentence “It is dangerous to conduct precipitation enhancement experiments at night because atmospheric aircraft fly around the top and bottom of clouds.” to “It is dangerous to conduct precipitation enhancement experiments at night because atmospheric aircraft needs to fly around the top and bottom of clouds.”
Line 106: Specify what nature it is.
We have revised this sentence “This was necessary because the data characteristics may not be continuous due to the nature of satellite data.” to “This was necessary because the data characteristics may not be continuous due to the grid data calculated by each grid of satellite data.”
Line 112-123: Rewrite. This sentence is unclear. Cannot understand your meaning.
We have rewritten the sentence as follows:
“In precipitation enhancement experiments using an atmospheric aircraft, the seeding material serves as an artificial cloud seed that facilitates the development of rain clouds, thereby resulting in rainfall (snowfall). The precipitation development process varies with temperature; therefore, the seeding materials for cold and warm clouds are different [37,38]. For cold clouds, dry ice or silver iodide (AgI), which generate ice, are seeded to cause snowfall or rainfall. In warm clouds, hygroscopic calcium chloride (CaCl2) or sodium chloride (NaCl) are seeded to cause rainfall [39–41]. The seedable clouds can be classified as ice or water phase clouds based on satellite images [31]. If a mixed phase is observed, it is classified as an ice-cloud phase if the CTT is below –5 °C, whereas it is classified as a water-cloud phase if the CTT is –5 °C or higher [25]. The occurrence frequency of seedable clouds can then be analyzed accordingly [28].”
Line 117-118: Dry ice as the nucleus of ice?
Usually, AgI is used; however, at times, dry ice is also used when conducting seeding experiments.
Line 124: “brightness temperature (BT)”.
We have added this spell.
Line 124-127: Need a reference to support the criterions.
We have revised this sentence and added the reference.
Line 145-146: “inflow of clouds generated at sea”. Reference or figure?
We added the reference as follows:
“Jeju Island is surrounded by the sea; therefore, the occurrence frequency of clouds here is approximately 28%. This is about 17% higher than other land areas due to the inflow of clouds generated at sea [53].”
Ito, J., and Niino, H. Atmospheric Kármán Vortex Shedding from Jeju Island, East China Sea: A Numerical Study. Monthly Weather Review, 2016, 144(1), 139-148.
Line 155: Figure 2: Change the scale. Don’t need to be as large as 100%.
We have replot the scale in Figure 2.
Line 157: CTH already defined before.
We have removed this spell.
Line 173-177: This should be given in the methodology section.
We rewritten this section to methodology section as follows:
“In addition, the occurrence frequency of seedable clouds occurred on land and sea and occurrence frequency according to cloud types were analyzed. Cloud types were classified using CTP and the COT calculated from COMS according to the method specified by the International Satellite Cloud Climatology Project (ISCCP) [42]. Using this method, cloud types were classified according to height (low-level: 1,000–680 hPa, mid-level: 680–440 hPa, high-level: 440–50 hPa) and optical thickness (thin: 0–3.6, middle: 3.6–23, thick: 23–379).”
Line 228: Delete “As”.
We have revised this sentence “The As mid-level cloud …” to “The mid-level As cloud …”
Line 240-242: Change the unit “K” to “oC”.
We have revised this unit.
Line 306: Change to “seedable clouds of different types”.
We have revised this sentence “This demonstrates that the frequency distribution of seedable clouds was very similar.” to “This demonstrates that the frequency distribution of seedable clouds of different types was very similar.”
Line 306-308: To be explicit: based on exactly what you get this conclusion.
We have revised this sentence “In particular, precipitation enhancement experiments were found to be most suitable areas for seedable clouds occurring in the northern and southern regions of the Korean Peninsula in the winter.” to “In particular, the northern and southern sea regions of the Korean Peninsula in winter were found to be most suitable areas for conducting precipitation enhancement experiments because of the high occurrence frequency of seedable clouds.”
Line 324-325: Why don’t you use products from GK-2A and/or Himawari-8, considering their high spatio-temporal resolution?
COMS has been in service since 2012; however, given that the mission ended on March 31, 2020, this system will no longer provide data. Therefore, as an alternative, GK-2A was launched in November 2018, and has been providing data since June 2019. Himawari-8 cloud products are not provided by KMA. Therefore, the only valid data for the analysis of seedable cloud around the Korean Peninsula in the last three years were those provided by COMS cloud products.
Line 310-326: I don’t think this paragraph well fit the conclusion section.
We agree with you. The paragraph can be removed; however, we intend to leave it as a reference for future research on various seeding experiments.
Reviewer 3 Report
1) The idea to utilize satellite data for assessment seedable clouds frequency is not new, but is more useful for big areas comparing to other instruments (ground radars, in-situ aircraft measurements, etc.).
2) Strictly speaking, the data obtained for 2016-2019 may not be representative for 2020 - 2021 due to annual variability. So I would appreciate some explanation of why authors used 3 years data statistics in their analysis ? Is it just a limitation of dataset or something else?
3) The English text style looks good and understandable.
4) There are minor grammatic errors:
Line 411, in word huidelines, correct - guidelines
Line 118 - "...is seeded to cause...". I am not sure, but in my opinion correct is "are seeded to cause..." as we mention two types of seeding materials - dry ice and silver iodide.
5) References made to the literature are sufficiently high-quality and reflect the current state of this field of science
6) I would suggest to correct the phrase "...and helps to develop the cloud" in Line 115 for something like "..which might help cloud to develop" or "...which might change microphysical and macrophysical characteristics of cloud", etc. So the magic word here should be may/might/can/could as we cannot claim that seeding will lead to cloud development.
7) In Fig. 2 and table 3 authors operate with frequencies in percent. It is not very useful comparing to absolute values. I suggest to add somewhere the absolute reference value.
8) The main question that arises for me in this study is whether the criteria taken in the classification of suitable clouds (Lines 103-113) are really adequate. Can you give some clarification that the criteria adopted by Prof. M. Murakami and his team for Japan must be true for Korean Peninsula? (1-3 sentences)
Author Response
1) The idea to utilize satellite data for assessment seedable clouds frequency is not new, but is more useful for big areas comparing to other instruments (ground radars, in-situ aircraft measurements, etc.).
We appreciate the valuable opinions and comments of the reviewer, thanks to which we could revise our manuscript to largely enhance its quality. We have made the following revisions in accordance with the additional comments of the reviewer.
2) Strictly speaking, the data obtained for 2016-2019 may not be representative for 2020 - 2021 due to annual variability. So I would appreciate some explanation of why authors used 3 years data statistics in their analysis ? Is it just a limitation of dataset or something else?
COMS has been in service since 2012; however, given that the mission ended on March 31, 2020, this system will no longer provide data. Therefore, as an alternative, GK-2A was launched in November 2018, and has been providing data since June 2019. Himawari-8 cloud products are not provided by KMA. Therefore, the only valid data for the analysis of seedable cloud around the Korean Peninsula in the last three years were those provided by COMS cloud products.
3) The English text style looks good and understandable.
4) There are minor grammatic errors:
Line 411, in word huidelines, correct - guidelines
We have revised this word.
Line 118 - "...is seeded to cause...". I am not sure, but in my opinion correct is "are seeded to cause..." as we mention two types of seeding materials - dry ice and silver iodide.
We have revised this sentence “In cold clouds, dry ice or silver iodide (AgI), which serves as the nucleus of ice, is seeded to cause snowfall or rainfall.” to “For cold clouds, dry ice or silver iodide (AgI), which serves as the nucleus of ice, are seeded to cause snowfall or rainfall.”
5) References made to the literature are sufficiently high-quality and reflect the current state of this field of science
6) I would suggest to correct the phrase "...and helps to develop the cloud" in Line 115 for something like "..which might help cloud to develop" or "...which might change microphysical and macrophysical characteristics of cloud", etc. So the magic word here should be may/might/can/could as we cannot claim that seeding will lead to cloud development.
We have revised this sentence “In precipitation enhancement experiments using atmospheric aircraft, the seeding material serves as an artificial cloud seed and helps to develop the cloud while causing rainfall (snowfall).” to “In precipitation enhancement experiments using atmospheric aircraft, the seeding material serves as an artificial cloud seed and which might help cloud to develop while causing rainfall (snowfall).”
7) In Fig. 2 and table 3 authors operate with frequencies in percent. It is not very useful comparing to absolute values. I suggest to add somewhere the absolute reference value.
We added the values in Section 2 as follows:
“The seasonal samples of data for each grid are average 2,618 (winter), 2,697 (spring), 2,730 (summer), and 2,687 (fall), respectively.”
8) The main question that arises for me in this study is whether the criteria taken in the classification of suitable clouds (Lines 103-113) are really adequate. Can you give some clarification that the criteria adopted by Prof. M. Murakami and his team for Japan must be true for Korean Peninsula? (1-3 sentences)
In the study conducted by Koshida et al. (2012), CF, RI, and cloud top temperature were considered to be appropriate conditions for the detection of seedable clouds in Korean Peninsula, as shown in Section 2. However, in this previous study, cloud top height was not considered. Therefore, in this study, cloud top height was used as a condition for the detection of seedable clouds by referring to the height considered during the seeding experiment.
Round 2
Reviewer 2 Report
Line 135-160 (newly added texts): Some of these should be put in the discussion section. It will be better also to add how these uncertainties affect your results.
Line 143: RMSE needs to be spelled out.
I am confused by the percentage in these texts as well: for example, “The probability of detection (POD) was found to be above 88%”. Do you mean that it can successfully detect 88% of the total cloud or that 88% of the “clouds” are real clouds and the left 12% are mistakenly interpreted as “clouds”? I think it may be helpful to define the POD.
Author Response
We are grateful to the reviewers for their careful reading of the manuscript and their thoughtful comments.
Line 135-160 (newly added texts): Some of these should be put in the discussion section. It will be better also to add how these uncertainties affect your results.
We have added the sentence follow as:
“The COMS cloud products used to detect seedable clouds in this study include uncertainty for each product, as shown in Section 2. This uncertainty leads to false detection of seedable clouds. However, geostationary satellites cannot detect lower-level clouds as they are underneath the upper clouds.In other words, lower-level clouds underneath the clouds exceeding the cloud top height, set in this study as seedable clouds, cannot be detected. Joiner et al. [51] reported that on average, approximately 11% of these multi-layered clouds occur in the middle latitudes. In other words, the frequency of the undetected seedable clouds in the lower-level underneath the upper cloud may be higher than the seedable clouds not detected due to the uncertainty of COMS cloud products. Therefore, the occurrence frequency of the seedable clouds analyzed may be higher or similar. However, land- and sea-based as well as regional differences may occur [51,58].”
Line 143: RMSE needs to be spelled out.
We have added the full name.
I am confused by the percentage in these texts as well: for example, “The probability of detection (POD) was found to be above 88%”. Do you mean that it can successfully detect 88% of the total cloud or that 88% of the “clouds” are real clouds and the left 12% are mistakenly interpreted as “clouds”? I think it may be helpful to define the POD.
We have added the sentence regarding the POD.
“POD is the rate at which clouds are detected as clouds or clear sky (Ncloud/(Ncloud+Nclear)100(%)).”
