the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Jun 2026
Dynamical and Microphysical Interactions in a Coastal Bow-Echo Producing Extreme Rainfall
Abstract. On 2 August 2020, a coastal bow-echo mesoscale convective system (MCS) produced severe rainfall and damaging winds over South Korea, resulting in casualties and property losses. Forecasting rapidly developing coastal bow echoes remains challenging due to limited understanding of the interactions between mesoscale dynamics and microphysical processes. Here, we analyze these interactions using improved multi-Doppler wind retrievals and polarimetric radar observations. The system evolved into a leading convective–trailing stratiform structure, reinforced by a rear-inflow jet (RIJ) that enhanced low-level convergence and shaped bowing segments. Feedbacks between RIJ-driven downdrafts, convective updrafts, and hydrometeor recycling sustained precipitation and prolonged the system’s lifetime after landfall. In particular, mixed-phase hydrometeors in stratiform clouds were advected into the leading convective line, where they enhanced and maintained deep convection. These dynamic–microphysical interactions governed storm organization and rainfall efficiency, explaining the persistence of heavy precipitation in the coastal zone. Beyond advancing process understanding, our results highlight the role of land–sea contrasts in shaping mesoscale circulations that intensify convection and provide observational benchmarks for improving forecasts and hazard resilience in coastal regions.
- Preprint
(6207 KB) - Metadata XML
-
Supplement
(2253 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-1878', Anonymous Referee #1, 12 Jul 2026
-
AC1: 'Reply on RC1', Seung Hee Kim, 22 Sep 2026
General comments
Mesoscale Convective Systems are one of the convective modes most associated with adverse weather around the World. The high organization of this type of structure causes MCS to last for a longer time than other structures. The particularity of the bow echoes structures is still more complex and requires more research, like that provided here. In general, the manuscript is well-conducted, but I miss some important comments regarding the data provided, the definition of the phenomena, the figures' description, and, what is one of the main issues, the references are outdated in general. Most of them are prior to 2010, and many updated references can give more relevance to some of the affirmations.
We appreciate the reviewer’s careful evaluation of our manuscript and the constructive comments and suggestions. We have revised the manuscript accordingly. In particular, we have updated the Introduction with more recent literature and clearer definitions of the principal bow-echo dynamical features, improved the presentation of the storm life cycle and several figures, and clarified the interpretation and limitations of the retrieved radar quantities. A point-by-point response is provided below.
Specific comments
L27-28: The references are outdated. I can understand that the Authors use some of the classical ones, but there are more updated works that can be perfectly referenced. Same in L30 and the following paragraphs of the Introduction.
I suggest briefly defining the main concepts: RIJs, bow echoes, pressure gradients, cold pool, etc.
We agree that the original Introduction relied primarily on classical studies when describing bow-echo dynamics. We retained these foundational references while adding more recent studies to provide an updated context for bow-echo organization and associated mesoscale circulations, including recent observational and climatological studies and the review by Zhou et al., (2023).
We also revised the opening paragraph of the Introduction to define the principal dynamical concepts used throughout the manuscript. Bow echoes are now described in terms of their characteristic bow-shaped radar-reflectivity structure and leading-convective–trailing-stratiform organization. Brief descriptions of the rear-inflow jet (RIJ), storm-generated cold pool, and associated buoyancy and pressure gradients were also added to provide the physical context for the subsequent analyses.
The revised text states:
“Within the MCS spectrum, bow echoes are particularly hazardous owing to their fast propagation, damaging surface winds, and potential to produce flash flooding and tornadoes (Fujita, 1978; Houze, 2004; Wakimoto, 2001; Xu et al., 2024; Yang et al., 2025; Zhou et al., 2023). Bow echoes are organized convective systems characterized by an arched or bow-shaped radar reflectivity structure, typically consisting of a leading convective line and a trailing stratiform precipitation region (Coniglio et al., 2004; Przybylinski, 1995; Weisman, 1993). Their evolution reflects interactions among convective updrafts, rear-inflow jets (RIJs), storm-generated cold pools, and mesoscale pressure gradients (Rotunno et al., 1988; Schenkman and Xue, 2016; Weisman, 1992). …”
Figure 1: Include a larger map of Eastern Asia to show the location of the area of study to external readers.
Following the reviewer’s suggestion, we have expanded Fig. 1 by adding an East Asia locator map to provide broader geographical context for international readers. The original national-scale and WISSDOM-domain maps are retained as the subsequent panels. The figure caption has also been revised accordingly.
About the use of different weather radar networks and bands: (1) provide more information about each system (e.g. time and spatial resolution, number of elevations, maximum range, minimum and maximum height); (2) Are all double-polarisation systems? (3) How difficult is it to combine the three technologies? I need more information about integrating different data types.
We expanded Sect. 2.1 and Table 1 to provide additional information on the radar systems used in the multiple-Doppler wind retrieval. The analysis incorporates 11 radars operating at S, C, and X bands. All radars have dual-polarization capability; however, only reflectivity and radial velocity were used as inputs to the WISSDOM wind synthesis. We now provide the native range resolution, volume-scan interval, min/max height, max range, and elevation for the different radar systems and have expanded Table 1 to summarize the radar specifications.
We also clarified how observations from radar systems with different wavelengths and sampling characteristics were combined. Prior to WISSDOM retrieval, each radar dataset was independently quality controlled and interpolated onto a common Cartesian grid. Differences in spatial resolution, observational range, and scan timing among the S-, C-, and X-band radars were accommodated during the interpolation and temporal synchronization procedures, with WISSDOM analyses produced at 30 min intervals. Because the wind synthesis uses only reflectivity and radial velocity, wavelength-dependent differences in polarimetric variables do not directly enter the retrieval. The immersed boundary method incorporated in WISSDOM was additionally used to represent the lower boundary over complex terrain (Liou et al., 2012). Previous applications have demonstrated the capability of this framework for integrating multiple radar observations and retrieving three-dimensional wind fields over complex terrain (Tsai et al., 2018, 2022, 2023, 2025).
The corresponding text has been added to Sect. 2.1, and the radar specifications have been expanded in Table 1.
Reference:
Liou, Y.-C., Chang, S.-F., and Sun, J.: An Application of the Immersed Boundary Method for Recovering the Three-Dimensional Wind Fields over Complex Terrain Using Multiple-Doppler Radar Data, Mon. Weather Rev., 140, 1603–1619, https://doi.org/10.1175/MWR-D-11-00151.1, 2012.
Tsai, C.-L., Kim, K., Liou, Y.-C., Lee, G., and Yu, C.: Impacts of Topography on Airflow and Precipitation in the Pyeongchang Area Seen from Multiple-Doppler Radar Observations, Mon. Weather Rev., 146, 3401–3424, https://doi.org/10.1175/MWR-D-17-0394.1, 2018.
Tsai, C.-L., Kim, K., Liou, Y.-C., Kim, J.-H., Lee, Y., and Lee, G.: Orographic-induced strong wind associated with a low-pressure system under clear-air condition during ICE-POP 2018, J. Geophys. Res. Atmos., 127, e2021JD036418, https://doi.org/10.1029/2021JD036418, 2022.
Tsai, C.-L., Kim, K., Liou, Y.-C., and Lee, G.: High-resolution 3D winds derived from a modified WISSDOM synthesis scheme using multiple Doppler lidars and observations, Atmos. Meas. Tech., 16, 845–869, https://doi.org/10.5194/amt-16-845-2023, 2023.
And the most directly relevant citation for the reviewer’s specific question about combining different radar wavelengths is:
Tsai, C.-L., Kim, K., Liou, Y.-C., and Lee, G.: Advantages of using multiple Doppler radars with different wavelengths for three-dimensional wind retrieval, Atmos. Meas. Tech., 18, 6371–6392, https://doi.org/10.5194/amt-18-6371-2025, 2025.
Before Figure 4, I suggest adding a new figure showing the life cycle of the bow-echo system at low levels.
We agree that showing the low-level evolution of the system provides useful context for the subsequent kinematic analysis. We therefore added a new six-panel figure (Fig. 4) showing radar reflectivity at 2 km altitude from 1200 to 1700 LST at hourly intervals. The figure illustrates the evolution of the convective system from its offshore stage through organization and landfall to its subsequent inland evolution and weakening.
We also added a corresponding paragraph at the beginning of Sect. 3.2 describing this morphological evolution. The new figure provides the broader life-cycle context for the subsequent quantitative analysis using convective height, echo-top height, and convective/stratiform areal coverage.
Figure 4: continue the vertical red dashed line in the bottom panel. It seems that the development and part of the system maturity are not included. Why?
We revised the time-series figure so that the vertical dashed line indicating the mature-to-decay transition extends through both panels. We also removed the “mature” and “decay” labels to avoid implying that an exact onset time of the mature stage was objectively determined.
Instead, we focus on the transition from the mature to decay stage near 1600 LST. Before 1600 LST, the convective and echo-top heights remained relatively stable overall despite short-term fluctuations, whereas both quantities decreased persistently after 1600 LST. The newly added low-level reflectivity sequence in Fig. 4 provides additional morphological context for the earlier development and organization of the bow echo. The manuscript and Fig. 5 caption were revised accordingly.
L232: “At 2 km altitude, system-relative southwesterly RIJ winds reached about 13 m s-1. Although this magnitude is slightly weaker than the >15 m s-1 values reported …” Could this issue be caused by some radar limitation? (e.g. distance to the echo, strong reflectivity prior to the area of interest, etc)
We agree that radar sampling and retrieval-related uncertainties could influence the estimated RIJ magnitude and therefore cannot be completely excluded. However, the present observations do not allow us to determine whether the difference between the approximately 13 m s-1) RIJ in this case and values exceeding 15 m s-1 reported in some previous studies is specifically attributable to radar limitations. In addition, RIJ magnitude can vary among bow-echo systems, geographical settings, and throughout their evolution.
We therefore removed the direct quantitative comparison with RIJ magnitudes reported in previous studies to avoid overinterpreting this relatively small difference. Instead, the revised manuscript focuses on the evolution of the RIJ within this event. The revised text states:
“At 2 km altitude, system-relative southwesterly RIJ winds reached about 13 m s-1). The rear inflow was concentrated behind the southern bowing segment and strengthened as the system approached the coastline, consistent with the increasing curvature of the convective line (Meng et al., 2012; Smull and Houze, 1985; Weisman, 1993).”
By 1430 LST, the retrieved RIJ intensified to nearly 15 m s-1, providing a within-event measure of RIJ strengthening as the bow echo became more clearly developed.
Which is the cause of the absence of low reflectivity echoes leading to the more intense area in the southern part of the System, observed in panels (a) to (c) of Figure 5?
We examined this feature using the KMA (Korea Meteorological Administration) Hybrid Surface Rainfall (HSR) observations to determine whether the apparent reduction in weak echoes could be associated with radar coverage. The HSR observations show the same spatial pattern, including the gradual decrease in reflectivity adjacent to the intense southern portion of the convective line. In addition, no radar-coverage gap occurs over this region. A coverage limitation would be expected to produce a more abrupt truncation of the radar echo, whereas the observed reflectivity decreases continuously toward the surrounding echo-free region.
We therefore interpret this feature as part of the observed precipitation structure of the convective system rather than an artifact associated with radar coverage. The newly added low-level HSR reflectivity sequence in Fig. 4 also shows this structure during the evolution of the system. Because this feature does not alter the principal kinematic interpretation, no additional manuscript text was added specifically for this feature.
Figure 6: add in any part of the graph where the field is reflectivity. Besides, to provide support to Figures 7 and 8, it would be interesting to mark the convective and stratiform regions.
We revised the vertical cross-section figure, now shown as Fig. 8, to make the shaded field more explicit by identifying it as radar reflectivity (dBZ; color shading). We also added labels indicating representative trailing stratiform and leading convective regions in panel (d).
These labels are based on the cross-sectional reflectivity structure and are intended to guide interpretation of the kinematic structure. The objective convective–stratiform classification used in the subsequent microphysical analyses was derived independently from the KMA HSR observations using the Powell et al. (2016) method. Therefore, the labels in Fig. 8 should not be interpreted as exact boundaries from the HSR-based classification.
Figure 7: How do you estimate the melting level?
The representative melting-layer height was estimated from the temperature profile observed at sounding station 47199, whose location is indicated by the black square in Fig. 1. Specifically, the altitude of the 0 °C level in the sounding was used as the representative melting-layer height for the polarimetric analysis.
We added this information to Sect. 2.2 and clarified the corresponding figure caption. The revised Methods text is:
“The representative melting-layer height used in the polarimetric analysis was estimated from the 0 °C level in the temperature profile observed at sounding station 47199, whose location is indicated by the black square in Fig. 1.”
The Fig. 9 caption now states that the black dashed line indicates the representative melting-layer height estimated from the sounding-derived 0 °C level at station 47199.
Citation: https://doi.org/10.5194/egusphere-2026-1878-AC1 -
AC2: 'Reply on RC1', Seung Hee Kim, 22 Sep 2026
General comments
Mesoscale Convective Systems are one of the convective modes most associated with adverse weather around the World. The high organization of this type of structure causes MCS to last for a longer time than other structures. The particularity of the bow echoes structures is still more complex and requires more research, like that provided here. In general, the manuscript is well-conducted, but I miss some important comments regarding the data provided, the definition of the phenomena, the figures' description, and, what is one of the main issues, the references are outdated in general. Most of them are prior to 2010, and many updated references can give more relevance to some of the affirmations.
We appreciate the reviewer’s careful evaluation of our manuscript and the constructive comments and suggestions. We have revised the manuscript accordingly. In particular, we have updated the Introduction with more recent literature and clearer definitions of the principal bow-echo dynamical features, improved the presentation of the storm life cycle and several figures, and clarified the interpretation and limitations of the retrieved radar quantities. A point-by-point response is provided below.
Specific comments
L27-28: The references are outdated. I can understand that the Authors use some of the classical ones, but there are more updated works that can be perfectly referenced. Same in L30 and the following paragraphs of the Introduction.
I suggest briefly defining the main concepts: RIJs, bow echoes, pressure gradients, cold pool, etc.
We agree that the original Introduction relied primarily on classical studies when describing bow-echo dynamics. We retained these foundational references while adding more recent studies to provide an updated context for bow-echo organization and associated mesoscale circulations, including recent observational and climatological studies and the review by Zhou et al., (2023).
We also revised the opening paragraph of the Introduction to define the principal dynamical concepts used throughout the manuscript. Bow echoes are now described in terms of their characteristic bow-shaped radar-reflectivity structure and leading-convective–trailing-stratiform organization. Brief descriptions of the rear-inflow jet (RIJ), storm-generated cold pool, and associated buoyancy and pressure gradients were also added to provide the physical context for the subsequent analyses.
The revised text states:
“Within the MCS spectrum, bow echoes are particularly hazardous owing to their fast propagation, damaging surface winds, and potential to produce flash flooding and tornadoes (Fujita, 1978; Houze, 2004; Wakimoto, 2001; Xu et al., 2024; Yang et al., 2025; Zhou et al., 2023). Bow echoes are organized convective systems characterized by an arched or bow-shaped radar reflectivity structure, typically consisting of a leading convective line and a trailing stratiform precipitation region (Coniglio et al., 2004; Przybylinski, 1995; Weisman, 1993). Their evolution reflects interactions among convective updrafts, rear-inflow jets (RIJs), storm-generated cold pools, and mesoscale pressure gradients (Rotunno et al., 1988; Schenkman and Xue, 2016; Weisman, 1992). …”
Figure 1: Include a larger map of Eastern Asia to show the location of the area of study to external readers.
Following the reviewer’s suggestion, we have expanded Fig. 1 by adding an East Asia locator map to provide broader geographical context for international readers. The original national-scale and WISSDOM-domain maps are retained as the subsequent panels. The figure caption has also been revised accordingly.
About the use of different weather radar networks and bands: (1) provide more information about each system (e.g. time and spatial resolution, number of elevations, maximum range, minimum and maximum height); (2) Are all double-polarisation systems? (3) How difficult is it to combine the three technologies? I need more information about integrating different data types.
We expanded Sect. 2.1 and Table 1 to provide additional information on the radar systems used in the multiple-Doppler wind retrieval. The analysis incorporates 11 radars operating at S, C, and X bands. All radars have dual-polarization capability; however, only reflectivity and radial velocity were used as inputs to the WISSDOM wind synthesis. We now provide the native range resolution, volume-scan interval, min/max height, max range, and elevation for the different radar systems and have expanded Table 1 to summarize the radar specifications.
We also clarified how observations from radar systems with different wavelengths and sampling characteristics were combined. Prior to WISSDOM retrieval, each radar dataset was independently quality controlled and interpolated onto a common Cartesian grid. Differences in spatial resolution, observational range, and scan timing among the S-, C-, and X-band radars were accommodated during the interpolation and temporal synchronization procedures, with WISSDOM analyses produced at 30 min intervals. Because the wind synthesis uses only reflectivity and radial velocity, wavelength-dependent differences in polarimetric variables do not directly enter the retrieval. The immersed boundary method incorporated in WISSDOM was additionally used to represent the lower boundary over complex terrain (Liou et al., 2012). Previous applications have demonstrated the capability of this framework for integrating multiple radar observations and retrieving three-dimensional wind fields over complex terrain (Tsai et al., 2018, 2022, 2023, 2025).
The corresponding text has been added to Sect. 2.1, and the radar specifications have been expanded in Table 1.
Reference:
Liou, Y.-C., Chang, S.-F., and Sun, J.: An Application of the Immersed Boundary Method for Recovering the Three-Dimensional Wind Fields over Complex Terrain Using Multiple-Doppler Radar Data, Mon. Weather Rev., 140, 1603–1619, https://doi.org/10.1175/MWR-D-11-00151.1, 2012.
Tsai, C.-L., Kim, K., Liou, Y.-C., Lee, G., and Yu, C.: Impacts of Topography on Airflow and Precipitation in the Pyeongchang Area Seen from Multiple-Doppler Radar Observations, Mon. Weather Rev., 146, 3401–3424, https://doi.org/10.1175/MWR-D-17-0394.1, 2018.
Tsai, C.-L., Kim, K., Liou, Y.-C., Kim, J.-H., Lee, Y., and Lee, G.: Orographic-induced strong wind associated with a low-pressure system under clear-air condition during ICE-POP 2018, J. Geophys. Res. Atmos., 127, e2021JD036418, https://doi.org/10.1029/2021JD036418, 2022.
Tsai, C.-L., Kim, K., Liou, Y.-C., and Lee, G.: High-resolution 3D winds derived from a modified WISSDOM synthesis scheme using multiple Doppler lidars and observations, Atmos. Meas. Tech., 16, 845–869, https://doi.org/10.5194/amt-16-845-2023, 2023.
And the most directly relevant citation for the reviewer’s specific question about combining different radar wavelengths is:
Tsai, C.-L., Kim, K., Liou, Y.-C., and Lee, G.: Advantages of using multiple Doppler radars with different wavelengths for three-dimensional wind retrieval, Atmos. Meas. Tech., 18, 6371–6392, https://doi.org/10.5194/amt-18-6371-2025, 2025.
Before Figure 4, I suggest adding a new figure showing the life cycle of the bow-echo system at low levels.
We agree that showing the low-level evolution of the system provides useful context for the subsequent kinematic analysis. We therefore added a new six-panel figure (Fig. 4) showing radar reflectivity at 2 km altitude from 1200 to 1700 LST at hourly intervals. The figure illustrates the evolution of the convective system from its offshore stage through organization and landfall to its subsequent inland evolution and weakening.
We also added a corresponding paragraph at the beginning of Sect. 3.2 describing this morphological evolution. The new figure provides the broader life-cycle context for the subsequent quantitative analysis using convective height, echo-top height, and convective/stratiform areal coverage.
Figure 4: continue the vertical red dashed line in the bottom panel. It seems that the development and part of the system maturity are not included. Why?
We revised the time-series figure so that the vertical dashed line indicating the mature-to-decay transition extends through both panels. We also removed the “mature” and “decay” labels to avoid implying that an exact onset time of the mature stage was objectively determined.
Instead, we focus on the transition from the mature to decay stage near 1600 LST. Before 1600 LST, the convective and echo-top heights remained relatively stable overall despite short-term fluctuations, whereas both quantities decreased persistently after 1600 LST. The newly added low-level reflectivity sequence in Fig. 4 provides additional morphological context for the earlier development and organization of the bow echo. The manuscript and Fig. 5 caption were revised accordingly.
L232: “At 2 km altitude, system-relative southwesterly RIJ winds reached about 13 m s-1. Although this magnitude is slightly weaker than the >15 m s-1 values reported …” Could this issue be caused by some radar limitation? (e.g. distance to the echo, strong reflectivity prior to the area of interest, etc)
We agree that radar sampling and retrieval-related uncertainties could influence the estimated RIJ magnitude and therefore cannot be completely excluded. However, the present observations do not allow us to determine whether the difference between the approximately 13 m s-1) RIJ in this case and values exceeding 15 m s-1 reported in some previous studies is specifically attributable to radar limitations. In addition, RIJ magnitude can vary among bow-echo systems, geographical settings, and throughout their evolution.
We therefore removed the direct quantitative comparison with RIJ magnitudes reported in previous studies to avoid overinterpreting this relatively small difference. Instead, the revised manuscript focuses on the evolution of the RIJ within this event. The revised text states:
“At 2 km altitude, system-relative southwesterly RIJ winds reached about 13 m s-1). The rear inflow was concentrated behind the southern bowing segment and strengthened as the system approached the coastline, consistent with the increasing curvature of the convective line (Meng et al., 2012; Smull and Houze, 1985; Weisman, 1993).”
By 1430 LST, the retrieved RIJ intensified to nearly 15 m s-1, providing a within-event measure of RIJ strengthening as the bow echo became more clearly developed.
Which is the cause of the absence of low reflectivity echoes leading to the more intense area in the southern part of the System, observed in panels (a) to (c) of Figure 5?
We examined this feature using the KMA (Korea Meteorological Administration) Hybrid Surface Rainfall (HSR) observations to determine whether the apparent reduction in weak echoes could be associated with radar coverage. The HSR observations show the same spatial pattern, including the gradual decrease in reflectivity adjacent to the intense southern portion of the convective line. In addition, no radar-coverage gap occurs over this region. A coverage limitation would be expected to produce a more abrupt truncation of the radar echo, whereas the observed reflectivity decreases continuously toward the surrounding echo-free region.
We therefore interpret this feature as part of the observed precipitation structure of the convective system rather than an artifact associated with radar coverage. The newly added low-level HSR reflectivity sequence in Fig. 4 also shows this structure during the evolution of the system. Because this feature does not alter the principal kinematic interpretation, no additional manuscript text was added specifically for this feature.
Figure 6: add in any part of the graph where the field is reflectivity. Besides, to provide support to Figures 7 and 8, it would be interesting to mark the convective and stratiform regions.
We revised the vertical cross-section figure, now shown as Fig. 8, to make the shaded field more explicit by identifying it as radar reflectivity (dBZ; color shading). We also added labels indicating representative trailing stratiform and leading convective regions in panel (d).
These labels are based on the cross-sectional reflectivity structure and are intended to guide interpretation of the kinematic structure. The objective convective–stratiform classification used in the subsequent microphysical analyses was derived independently from the KMA HSR observations using the Powell et al. (2016) method. Therefore, the labels in Fig. 8 should not be interpreted as exact boundaries from the HSR-based classification.
Figure 7: How do you estimate the melting level?
The representative melting-layer height was estimated from the temperature profile observed at sounding station 47199, whose location is indicated by the black square in Fig. 1. Specifically, the altitude of the 0 °C level in the sounding was used as the representative melting-layer height for the polarimetric analysis.
We added this information to Sect. 2.2 and clarified the corresponding figure caption. The revised Methods text is:
“The representative melting-layer height used in the polarimetric analysis was estimated from the 0 °C level in the temperature profile observed at sounding station 47199, whose location is indicated by the black square in Fig. 1.”
The Fig. 9 caption now states that the black dashed line indicates the representative melting-layer height estimated from the sounding-derived 0 °C level at station 47199.
Citation: https://doi.org/10.5194/egusphere-2026-1878-AC2
-
AC1: 'Reply on RC1', Seung Hee Kim, 22 Sep 2026
-
RC2: 'Comment on egusphere-2026-1878', Anonymous Referee #2, 17 Aug 2026
Journal: Natural Hazards and Earth System Sciences (NHESS)
Manuscript ID: egusphere-2026-1878
Title: Dynamical and Microphysical Interactions in a Coastal Bow-Echo Producing Extreme Rainfall
Authors: Jong-Hoon Jeong, Seung Hee Kim, Su-Bin Oh, Jeong-Eun Lee, Chia-Lun Tsai, and Gyuwon Lee
Recommendation: Minor Revision
General Assessment
This manuscript provides a high-quality observational investigation of a landfalling bow-echo mesoscale convective system (MCS) over the western coast of South Korea on 2 August 2020. By integrating high-resolution 3D multi-Doppler wind retrievals from the WISSDOM synthesis system with S-band polarimetric radar microphysical retrievals, the authors systematically document the kinematic evolution (rear-inflow jet, low-level convergence, and asymmetric mesovortices) alongside vertical microphysical distributions (LWC, IWC, D_m, and N_t). The proposed feedback between the trailing stratiform ice-phase reservoir and the leading convective line via a hydrometeor-recycling pathway offers valuable insight into how coastal MCSs sustain high precipitation efficiency after landfall.
The paper is well-written, logically organized, and suitable for publication in Natural Hazards and Earth System Sciences. I recommend a Minor Revision to address a few methodological clarifications, presentation enhancements, and minor technical corrections outlined below.
Specific Comments & Suggestions
1. Conceptual Schematic for Dynamic–Microphysical Coupling
The proposed hydrometeor-recycling mechanism—where ice particles generated in the leading convective line are advected into the trailing stratiform region and partially re-ingested into the convective updraft via the upper branch of the rear-inflow jet (RIJ)—is physically sound and well supported by the CFADs and vertical profiles (Figs. 7 and 8). However, because observational radar data cannot track individual trajectories directly, this conceptual model would be much more impactful if summarized visually.
- Suggestion: Consider adding a conceptual schematic diagram (or incorporating schematic arrows into an expanded panel of Fig. 6) that explicitly illustrates the 2D/3D airflow streams (front-to-rear flow, RIJ, recirculation), hydrometeor phase zones (graupel/large drops in convective core vs. aggregated ice aloft in stratiform), and the recycling pathway. Providing a clear conceptual schematic (similar in purpose to Fig. 15 of Park et al., 2021) would help synthesize these dynamic–microphysical interactions into an easily digestible visual summary for readers.
- Reference for context: Park, C., S.-W. Son, and J.-H. Kim, 2021: Role of baroclinic trough in triggering vertical motion during summertime heavy rainfall events in Korea. Journal of Atmospheric Sciences, 78(6), 1801–1817, https://doi.org/10.1175/JAS-D-20-0216.1.
2. Surface Cold Pool Characterization & Coastal AWS Observations
The manuscript highlights the role of the descending RIJ and low-level convergence in shaping the bowing line and maintaining deep convection near the coast. In classical bow-echo dynamics (e.g., RKW theory), the balance between cold pool-induced circulation and environmental low-level shear plays a central governing role.
- Suggestion: While 850 hPa shear and LLJ structures are presented (Fig. 3b), the surface thermodynamic footprint (Delta T, dewpoint drop, or surface pressure perturbations) of the cold pool during landfall (13:30–15:00 LST) is not explicitly shown. Including surface station time series or spatial mesonet/AWS analysis along the coastline would provide valuable observational grounding for how the cold pool interacted with onshore flow to slow system propagation (28 m s^-1) while sustaining coastal convergence. In particular, further analyzing cold pool evolution and its interaction/intensification with the descending RIJ using coastal AWS observations (e.g., drawing on the observational/mesoscale analysis framework in Byeon et al., 2024) would add substantial physical clarity to the manuscript.
- Reference for context: Byeon, K., J.-H. Kim, and Y.-J. Park, 2024: Synoptic and mesoscale mechanisms of reported tornado-like gust wind event in Korea using high-resolution numerical simulation. Atmosphere-Korea, 34(4), 397–415, https://doi.org/10.14191/ATMOS.2024.34.4.397.
3. Uncertainty and Resolution Constraints of Retrieved Vertical Velocity (w)
In Figure 6, vertical velocity (w) reaches values exceeding 5m s^-1, successfully capturing convective updrafts and bounded weak echo regions (BWER).
- Suggestion: Since vertical velocity derived from multi-Doppler variational synthesis is highly sensitive to mass continuity integration, lower boundary conditions, radar spatial sampling, and terrain-induced flow distortion (even when employing the Immersed Boundary Method), please add a brief discussion in Section 2.1 addressing the vertical resolution limits, smoothing constraints, and uncertainty bounds associated with the retrieved w field. Citing foundational variational multi-Doppler and WISSDOM error analysis literature (e.g., Gao et al., 1999; Liou and Chang, 2009; Liou et al., 2012; Potvin et al., 2012) would provide helpful context regarding how boundary layer errors and divergence integration weightings affect vertical motion magnitudes.
Minor / Technical Corrections
- Section Outlining and Numbering Sequence:
In Section 3 (Results), the manuscript transitions from Section 3.1 (Synoptic environment, p. 8, line 168) directly to Section 3.3 (Kinematic structure of the bow-echo MCS, p. 10, line 203), followed by Section 3.2 (Microphysical properties of the bow-echo MCS, p. 13, line 291). Please correct the subsection numbering sequence to follow numerical order (3.1 - 3.2 - 3.3). - Table 1 Platform Formatting:
In Table 1 (p. 20), please verify that radar station identifiers (e.g., KWK, BRI, KSN, GDK vs. SKWK, SBRI, SKSN, SGDK used in Section 2.1 line 100) are fully consistent across the text and table. - Figure 3 Caption Formatting Error:
In the caption of Figure 3 (p. 9–10, line 200), there appears to be a minor formatting artifact in the units for relative vorticity (10-5 s-1 200 , shading). Please clean up the text formatting in the figure caption.
Citation: https://doi.org/10.5194/egusphere-2026-1878-RC2 -
AC3: 'Reply on RC2', Seung Hee Kim, 22 Sep 2026
General Assessment
This manuscript provides a high-quality observational investigation of a landfalling bow-echo mesoscale convective system (MCS) over the western coast of South Korea on 2 August 2020. By integrating high-resolution 3D multi-Doppler wind retrievals from the WISSDOM synthesis system with S-band polarimetric radar microphysical retrievals, the authors systematically document the kinematic evolution (rear-inflow jet, low-level convergence, and asymmetric mesovortices) alongside vertical microphysical distributions (LWC, IWC, D_m, and N_t). The proposed feedback between the trailing stratiform ice-phase reservoir and the leading convective line via a hydrometeor-recycling pathway offers valuable insight into how coastal MCSs sustain high precipitation efficiency after landfall.
The paper is well-written, logically organized, and suitable for publication in Natural Hazards and Earth System Sciences. I recommend a Minor Revision to address a few methodological clarifications, presentation enhancements, and minor technical corrections outlined below.
We appreciate the reviewer’s positive assessment of the manuscript and the constructive suggestions. We have revised the manuscript to clarify the interpretation and limitations of the multiple-Doppler wind retrievals, improve the presentation of the dynamic–microphysical interactions, and correct the technical issues identified by the reviewer. Detailed responses are provided below.
Specific Comments & Suggestions
- Conceptual Schematic for Dynamic–Microphysical Coupling
The proposed hydrometeor-recycling mechanism—where ice particles generated in the leading convective line are advected into the trailing stratiform region and partially re-ingested into the convective updraft via the upper branch of the rear-inflow jet (RIJ)—is physically sound and well supported by the CFADs and vertical profiles (Figs. 7 and 8). However, because observational radar data cannot track individual trajectories directly, this conceptual model would be much more impactful if summarized visually.
- Suggestion:Consider adding a conceptual schematic diagram (or incorporating schematic arrows into an expanded panel of Fig. 6) that explicitly illustrates the 2D/3D airflow streams (front-to-rear flow, RIJ, recirculation), hydrometeor phase zones (graupel/large drops in convective core vs. aggregated ice aloft in stratiform), and the recycling pathway. Providing a clear conceptual schematic (similar in purpose to Fig. 15 of Park et al., 2021) would help synthesize these dynamic–microphysical interactions into an easily digestible visual summary for readers.
- Reference for context:Park, C., S.-W. Son, and J.-H. Kim, 2021: Role of baroclinic trough in triggering vertical motion during summertime heavy rainfall events in Korea. Journal of Atmospheric Sciences, 78(6), 1801–1817, https://doi.org/10.1175/JAS-D-20-0216.1.
We agree that a conceptual schematic could provide a useful synthesis of the proposed dynamic–microphysical interactions. Following the reviewer’s suggestions, we have added a schematic diagram (Fig.11) in the revised manuscript.
- Surface Cold Pool Characterization & Coastal AWS Observations
The manuscript highlights the role of the descending RIJ and low-level convergence in shaping the bowing line and maintaining deep convection near the coast. In classical bow-echo dynamics (e.g., RKW theory), the balance between cold pool-induced circulation and environmental low-level shear plays a central governing role.
- Suggestion:While 850 hPa shear and LLJ structures are presented (Fig. 3b), the surface thermodynamic footprint (Delta T, dewpoint drop, or surface pressure perturbations) of the cold pool during landfall (13:30–15:00 LST) is not explicitly shown. Including surface station time series or spatial mesonet/AWS analysis along the coastline would provide valuable observational grounding for how the cold pool interacted with onshore flow to slow system propagation (28 m s^-1) while sustaining coastal convergence. In particular, further analyzing cold pool evolution and its interaction/intensification with the descending RIJ using coastal AWS observations (e.g., drawing on the observational/mesoscale analysis framework in Byeon et al., 2024) would add substantial physical clarity to the manuscript.
- Reference for context:Byeon, K., J.-H. Kim, and Y.-J. Park, 2024: Synoptic and mesoscale mechanisms of reported tornado-like gust wind event in Korea using high-resolution numerical simulation. Atmosphere-Korea, 34(4), 397–415, https://doi.org/10.14191/ATMOS.2024.34.4.397.
We conducted an additional analysis using KMA Automatic Weather Station (AWS) observations to examine the near-surface thermodynamic and wind response associated with passage of the bow-echo system. Among the AWS stations within the analysis domain, station 469, which recorded the largest daily accumulated rainfall on 2 August 2020, was selected for detailed analysis. Its location is indicated by the purple star in Fig. 4e, and a new time-series figure (Fig. 6) shows wind speed, wind direction, temperature, relative humidity, and rainfall from 1330 to 1800 LST.
The observations show a distinct surface response beginning shortly after 1610 LST. The onset of precipitation was accompanied by a rapid temperature decrease of approximately 4–5 °C, an increase in relative humidity to above 90 %, and an abrupt increase in surface wind speed to more than 10 m s-1. These concurrent changes are consistent with passage of precipitation-cooled convective outflow and provide observational evidence of a surface cold pool associated with the bow-echo system.
We note that station 469 is located inland and that the strongest thermodynamic response at this station occurred after the coastal landfall period. We therefore use these observations to document the cold-pool signature during the post-landfall mature-to-decay evolution rather than to quantitatively attribute the earlier reduction in propagation speed to either the cold pool or descending RIJ. Accordingly, the manuscript states that the factors controlling the post-landfall deceleration cannot be determined directly from the present observations.
We added the AWS analysis in Sect. 3.2 before the detailed wind analysis and added a brief description of the AWS data selection in Sect. 2.1.
- Uncertainty and Resolution Constraints of Retrieved Vertical Velocity (w)
In Figure 6, vertical velocity (w) reaches values exceeding 5m s^-1, successfully capturing convective updrafts and bounded weak echo regions (BWER).
- Suggestion:Since vertical velocity derived from multi-Doppler variational synthesis is highly sensitive to mass continuity integration, lower boundary conditions, radar spatial sampling, and terrain-induced flow distortion (even when employing the Immersed Boundary Method), please add a brief discussion in Section 2.1 addressing the vertical resolution limits, smoothing constraints, and uncertainty bounds associated with the retrieved w field. Citing foundational variational multi-Doppler and WISSDOM error analysis literature (e.g., Gao et al., 1999; Liou and Chang, 2009; Liou et al., 2012; Potvin et al., 2012) would provide helpful context regarding how boundary layer errors and divergence integration weightings affect vertical motion magnitudes.
We revised Sect. 2.1 to clarify the uncertainty and resolution constraints associated with the retrieved vertical velocity. In the WISSDOM variational framework, vertical velocity is less directly constrained by Doppler radial-velocity observations than the horizontal wind components and is therefore more sensitive to radar sampling geometry, mass-continuity constraints, lower-boundary conditions, and spatial smoothing. We also clarify that limited radar overlap and terrain-related beam blockage can introduce greater uncertainty in localized vertical-motion structures.
The wind synthesis was performed on a Cartesian grid with 500 m horizontal and 250 m vertical spacing. Because independent vertical-velocity observations were unavailable for this event, a case-specific quantitative uncertainty range for (w) could not be determined. We therefore interpret the vertical-motion field primarily from spatially coherent structures and their temporal evolution rather than from individual local extrema.
We also revised the Results discussion to describe the feature as a spatially coherent convective updraft, with retrieved vertical velocity locally exceeding 5 m s-1, rather than relying solely on the local maximum magnitude.
Reference:
Potvin, C. K., Wicker, L. J., and Shapiro, A.: Assessing Errors in Variational Dual-Doppler Wind Syntheses of Supercell Thunderstorms Observed by Storm-Scale Mobile Radars, J. Atmos. Oceanic Technol., 29, 1009–1025, https://doi.org/10.1175/JTECH-D-11-00177.1, 2012.
Minor / Technical Corrections
- Section Outlining and Numbering Sequence:
In Section 3 (Results), the manuscript transitions from Section 3.1 (Synoptic environment, p. 8, line 168) directly to Section 3.3 (Kinematic structure of the bow-echo MCS, p. 10, line 203), followed by Section 3.2 (Microphysical properties of the bow-echo MCS, p. 13, line 291). Please correct the subsection numbering sequence to follow numerical order (3.1 - 3.2 - 3.3).
We corrected the subsection numbering in Sect. 3. The Results now follow the numerical sequence: Sect. 3.1, Synoptic environment; Sect. 3.2, Kinematic structure of the bow-echo MCS; and Sect. 3.3, Microphysical properties of the bow-echo MCS. The corresponding section references throughout the manuscript were also updated.
- Table 1 Platform Formatting:
In Table 1 (p. 20), please verify that radar station identifiers (e.g., KWK, BRI, KSN, GDK vs. SKWK, SBRI, SKSN, SGDK used in Section 2.1 line 100) are fully consistent across the text and table.
We checked the radar station identifiers throughout the manuscript and Table 1 and revised them for consistency. The radar-band prefixes are now used consistently for the S- (SGDK, SKWK, SBRI, and SKSN), C- (CIIA and CSAN), and X-band radars (XMIL, XDJK, XSRI, XKOU, and XYOU). We also checked the corresponding station identifiers in Sect. 2.1 and the figure labels to ensure consistent notation throughout the manuscript.
- Figure 3 Caption Formatting Error:
In the caption of Figure 3 (p. 9–10, line 200), there appears to be a minor formatting artifact in the units for relative vorticity (10-5 s-1 200 , shading). Please clean up the text formatting in the figure caption.
We corrected the formatting error in the Fig. 3 caption. The relative-vorticity units are now given as (10-5 s-1).
Citation: https://doi.org/10.5194/egusphere-2026-1878-AC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 133 | 51 | 32 | 216 | 32 | 24 | 27 |
- HTML: 133
- PDF: 51
- XML: 32
- Total: 216
- Supplement: 32
- BibTeX: 24
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
Specific comments