the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Jun 2025
Integrating Fireline Observations to Characterize Fire Plumes During Pyroconvective Extreme Wildfire Events: Implications for Firefighter Safety and Plume Modeling
Abstract. Firefighter entrapments occur when wildfires suddenly transition into extreme wildfire events (EWEs). These transitions are often caused by pyroconvective fire-atmosphere coupling, triggered by a combination of high fire intensity and atmospheric vertical thermodynamic structure. Pyroconvection indices calculated using coarse atmospheric modeling data crudely detect these dynamic transitions due to highly localized atmospheric processes and changes in atmospheric conditions caused by the fire. Consequently, fire managers may remain unaware that fire behavior intensification due to fire-atmosphere coupling is outdating the safety protocols in place. This study presents a new in-plume profiling methodology to improve the assessment of fire-atmosphere interaction dynamics in real-time. As proof of concept, we analyzed 156 in-plume sondes launched during the 2021–2025 fire seasons in Spain, Chile, Greece, and The Netherlands. As a strategy to measure the coupling fire-atmosphere, we propose simultaneously launching two radiosondes: one to measure ambient conditions and another to capture data within the plume updraft. Comparing these profiles, we measure in-situ and in-real time the modification of state variables by the fire-atmosphere interaction. These new observations and methodology improve our assessment of pyroconvection dynamics demonstrating practical implications that support their use by incident management teams. It has the potential to enhance awareness of possible near-accidents and tactical failures during extreme pyroconvective wildfire events. Additionally, it offers a comprehensive observational dataset to improve pyroconvection nowcasting and advance research on fire-atmosphere interaction.
-
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
(2758 KB)
-
Supplement
(660 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2758 KB) - Metadata XML
-
Supplement
(660 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-1923', Anonymous Referee #1, 03 Jul 2025
Overview
This preprint presents a novel and practical methodology for characterising pyroconvective wildfire plume dynamics using dual radiosonde soundings (in-plume and ambient). The study spans 156 field launches across four countries between 2021 and 2025 and offers both operational and scientific insights into plume development, real-time hazard awareness, and fire-atmosphere interaction modelling.
The manuscript is timely, rigorously detailed, and bridges a rare and valuable gap between operational field constraints and mesoscale meteorology. The work is distinguished by its applied innovation, extensive empirical validation, and potential to substantially inform firefighter safety procedures.
Strengths1. Novel Methodology with Operational Value
- The use of paired in-plume and ambient radiosonde profiles is both innovative and cost-effective, rendering it feasible for deployment during active wildfires.
- The operational integration into tactical decision-making workflows sets this study apart from traditional simulation-based or laboratory-bound research on pyroconvection.
2. Robust Field Campaign
- With 156 launches covering a diverse range of vegetation types, meteorological conditions, and terrain profiles, the dataset represents an impressive empirical foundation.
- The inclusion of both prescribed burns and uncontrolled wildfires increases the method’s general applicability.
3. Validation Through Multi-Modal Comparison
- Plume-top altitudes inferred from vertical velocity profiles were validated against radar echotop data, which significantly strengthens confidence in the method.
- Application of parcel theory for forecasting potential plume development (e.g., pyroCu onset) is methodologically sound and well-executed.
4. Classification Framework
- The six-category plume prototype typology (based on ABL height, LCL height, and wind shear layers) is operationally intuitive and scientifically coherent.
5. Actionable Outcomes
- Several case studies (e.g., Martorell and Santa Ana) demonstrate that tactical decisions informed by the sonde data likely contributed to risk mitigation. This real-world applicability is a major strength.
Weaknesses / Areas for Improvement
1. Clarification of Balloon Typology (Ref. Line 145, Table 1)
- The classification "professional high altitude balloon" is misleading. A more accurate term would be “operational radiosonde systems”, as both small and large balloons may be used professionally. For instance, an MW51 Vaisala ground station combined with an RS41 sonde and a Totex TA50 or TA100 balloon could easily reach 400 hPa level, albeit with higher helium demand.
- Table 1 also incorrectly claims that such systems are incapable of simultaneous launches. In fact, MW51 systems can track up to four sondes concurrently and have portable variants.
- Furthermore, the claim that professional radio-sounding systems are “not safe for aerial resources” is inaccurate. All operational weather balloons (regardless of size) comply with aviation safety regulations. In contrast, marking helicopters as inherently "safe" for aerial operations is misleading, as such platforms require strict coordination with air traffic control and firefighting aviation assets. These inaccuracies should be revised to reflect standard aviation safety protocols.
2. Reliability of Windsonde Data During Descent (Ref. Line 190)
- The authors should clearly state that measurements during descent are generally considered less reliable, even for professional sondes. Windsonde systems have not been formally validated for descent-phase data collection.
- As shown in Bessardon et al. (2016, Kumasi campaign), ground-based reference measurements of pressure and temperature were used to calibrate Windsonde outputs. It is unclear whether a similar calibration procedure was applied in this study.
- Moreover, the same research highlights concerns regarding wind speed and direction accuracy, particularly during turbulent or shear-laden environments. The authors should explicitly discuss whether and how these limitations were addressed.
- Known instrumental limitations of Windsonde include weak response to rapid humidity and temperature changes, and systematic underestimation of altitude (up to ~40 m). These should be acknowledged and addressed to justify continued use of this platform.
3. Quantitative Predictive Success Rate
The manuscript would benefit from a summary table or appendix quantifying:
- How many launches successfully entered the plume core?
- How many failed or produced partial profiles?
- In what proportion of cases did fire development escalate to Extreme Wildfire Events (EWEs)?
- In how many instances did the radiosonde data lead to a tactical change (e.g., crew withdrawal)?
4. Reproducibility: Launch Schedule and Decision Criteria
- For reproducibility and model intercomparison, the authors should provide a complete launch schedule overview, including exact timestamps and GPS coordinates of each sonde release.
- Additionally, the criteria used to determine the moment and location of launch should be explicitly stated (e.g., wind indicators, visual cues, forecast thresholds).
- These criteria appear to be field-operational in nature, but formalising them would help transfer the method to other contexts.
5. Real-Time Workflow and Decision Chain
The article does not describe the complete operational workflow from launch to decision. Clarifying the following would significantly improve transparency:
- How is data transmitted to and from the operations centre?
- Who is expected to perform the data analysis (on-site, centralised, or remote team)?
- What additional data sources are used (e.g., satellite, radar, fireline reports)?
- What is the end-to-end latency between launch and actionable tactical insight?
- Are any supporting information systems or software platforms (e.g., for visualisation or alerting) required or recommended?
6. Sonde Sampling Bias
- The authors acknowledge that sondes may not always enter the plume core, which may skew thermal and vertical velocity readings. Further statistical quantification of this sampling uncertainty would be beneficial.
7. Terminology
- Some terminology (e.g., “θv spike”, “fireABL”, “S parcel”) may not be immediately clear to the broader meteorological or fire-behaviour audience. A glossary or summary table of variables and acronyms is recommended.
Suggested Revisions
- Replace ambiguous or inaccurate entries in Table 1, particularly regarding balloon classifications, safety, and simultaneous sounding capability.
- Add a summary table of all launches with fire name, time, coordinates, outcome, and tactical decision if applicable.
- Include error metrics for vertical velocity-derived plume tops compared to radar.
- Clarify how and when real-time analysis was conducted, by whom, and how long it took.
- Address the known technical limitations of Windsondes and justify their use despite weaknesses.
- Provide access to a full launch log and reproducibility protocol, including selection criteria for launch timing and location.
Citation: https://doi.org/10.5194/egusphere-2025-1923-RC1 -
AC1: 'Reply on RC1', Marc Castellnou, 01 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1923/egusphere-2025-1923-AC1-supplement.pdf
-
CC1: 'Comment on egusphere-2025-1923', Tomàs Artés Vivancos, 07 Aug 2025
I would like to congratulate to the authors for this work that contributes with data analsysis for plume analsysis in real time in real scenarios for decision-making support.The article proposes a sounding methodology to characterize fire plumes during wildfire events. Although it does not provide a detailed protocol, it clearly emphasizes the importance of in-situ data collection using a feasible and affordable approach to better understand and assess plume behavior and atmospheric dynamics during a fire. I find this work highly valuable, as fire analyses are often based solely on forecast data. This study offers a practical method for estimating fire plume dynamics in areas where radar systems are prohibitively expensive. Additionally, the real-time data collection and analysis performed during the fire enhance its value for emergency management, moving beyond the traditional post-event analysis approach.
The article presents very useful information and includes clear application examples. It is well-written and well-structured. However, readers should be familiar with fire and atmospheric terminology, and ideally be up to date with the current state of research on fire plume dynamics and fire typologies.
The article outlines a method for obtaining atmospheric soundings from an ongoing fire, including measurements not only of the ambient atmosphere but also of the head, flank, and rear of the fire. While the work does not go into detail about the balloon sounding launch procedure or the success rate—points also raised by an anonymous reviewer-it focuses on the sounding data collected from different locations and their use for analyzing plume state and potential, which is useful for operational decision-making.
The authors acknowledge the limitations of their approach and provide recommendations for how to work with the data presented.
This article makes an important contribution to operational practices and this research field by characterizing fire plumes and assess transitions between different fire plume behaviors.
I would recomment to the authors to apply at least minor comments when they apply.
The article focuses on integrating sounding data to characterize fire plumes, as stated in the title, and the content aligns well with this aim. The authors address the inherent uncertainty in the sounding data, particularly due to the balloons moving within the plume's indraft and temperature peaks, which affects the success rate of non-ambient soundings. However, there is no quantitative assessment of the uncertainty for each variable derived from the sounding methodology, nor an analysis of how sensitive the plume top estimation is to those uncertainties—beyond potential temperature and rising speed. I understand this could be the focus of a separate study, which would require a large number of soundings and additional instruments for validation in different contexts with prescibed fires. Still, the article highlights the importance of potential temperature and vertical velocity as key factors in plume-top estimation, with other variables being more relevant when using the parcel method for potential plume-top prediction.
Minor Comments:
Update model name and cite missing reference: MESO-NH/Forefire – Jean-Baptiste Filippi.
There is no need to justify the balloon radiosounding method beyond its affordability and operational simplicity. Other technologies like high-altitude atmospheric balloons, Doppler radar, UAVs, and helicopter-mounted sensors are capable of real-time data collection as well.
Table 1 may be unnecessary, but I understand that in real-world field campaigns, time constraints, costs, and complexity support the value of the information presented.
It would be helpful to use a consistent temperature unit throughout the article.
"S parcel increase by3°C" → should be "S parcel increases by 3°C" (missing space and verb correction).
"estimating the dilution plume height, is adequate" → remove the comma: "estimating the dilution plume height is adequate."
In Section 3.3.3, regarding the Casablanca III fire in Chile: it states the fire grew to 12,073 ha between February 8–10, 2023, but also says it had already blown up on February 2, reaching 363,000 ha. Including fire spread metrics would help readers better understand the relationship between plume dynamics and fire spread behavior.
"Fire behavior was initially expected to calm in the early evening, but there was a 60% chance of intensification during the day-to-night transition due to the advection of drier air from the SW" — I assume this is for context, and the 60% probability comes from a weather forecast ensemble.
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC1 -
AC2: 'Reply on CC1', Marc Castellnou, 13 Aug 2025
We would like to thank the reviewer for taking the time to review the paper and for providing valuable comments that have helped us clarify and improve the proposed methodology and its explanation.
-
CC3: 'Reply on AC2', Tomàs Artés Vivancos, 18 Aug 2025
The authors provided an answer to my comment. Actually, the answer was much more detailed that I expected. In addition, they addressed every comment.
Regarding the uncertainty of the different variables measured with the radiosonde, I was not expecting the authors to take new measurements, at all. Apologies for the misunderstanding. The article in its initial version already provided statistics about the measurements. My comment was about how variables like RH and potential temperature may affect the parcel models for potential top heights. I am not expecting the authors to address that fact since it could be a completely new work.
I congratulate the authors for the work that describes an important contribution to plume science dynamics within science and operational fields bringing together two worlds that are often too far away.
I consider this comment closed. Thanks
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC3
-
CC3: 'Reply on AC2', Tomàs Artés Vivancos, 18 Aug 2025
-
AC2: 'Reply on CC1', Marc Castellnou, 13 Aug 2025
-
CC2: 'Comment on egusphere-2025-1923', Cristina Montiel Molina, 17 Aug 2025
This is an impressive paper which presents a comprehensive and timely study of pyroconvective fire-atmosphere coupling. It is based on a novel practical methodology and a very detailed and accurate dataset assessment, to obtain a precise characterization of fire plumes during pyroconvective events.
The main highlight of this study is the design and demonstration of an original and sound inductive methodology to reduce incertainty and improve efficiency of emergency systems at extreme wildfire events. This is a badly needed contribution to the effects of climate change on wildfire risk that are creating the most complex and critical emergency scenarios.
On the basis of an updated and complete state-of the art, the first strenght of this study is to design and test a novel data-gathering method of pyroconvection dynamics in-situ and in real-time. Furthermore, the global dimension and multi-scalar approach of the research design enhances its consistency in the context of global climate change. Another strenght of this study is to define and to apply an accurate pyroconvention prototype system.
The obtained results provide a completely innovative understanding and modeling of pyroconvection dynamics. This contribution has a strong potential to build-up a powerful tool of information management enabling to anticipate defense actions to control the most challenging wildfire events and to mitigate firefighters and civil population vulnerability.
The realistic and critical approach of the discussion section is also outstanding. The authors acknowledge the limitations of the methodology and consider the importance of contextual factors for a right assessment of measurements.
The conclussions on fire-atmosphere interaction are a badly needed contribution for a climate change adapted new model of proactive wildfire risk management beyond the current paradigm of existing fuel models and fire weather indexes.
It’s an excellent scientific paper that could have a huge scientific and operational impact. I don’t have any concrete suggestion for improvement further than those already done by previous reviewers. I congratulate the authors on this magnificent research.
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC2 -
CC4: 'Comment on egusphere-2025-1923', Peio Oria Iriarte, 31 Aug 2025
This paper presents a valuable contribution to wildfire science by integrating fireline observations to enhance the understanding of fire plume behavior during pyroconvective extreme wildfire events. The authors effectively highlight the importance of detailed, real-time observational data in characterizing complex fire-induced phenomena, which has significant implications for improving plume modeling accuracy. The study's focus on firefighter safety underscores the practical relevance of accurately predicting plume dynamics to inform operational decision-making. Overall, the work advances both the scientific understanding of extreme wildfire behavior and its application to emergency response and safety management, offering meaningful insights for researchers, practitioners, and policymakers involved in wildfire mitigation and response efforts.
Having atmospheric profile data in the environment and inside the convective plume is of undeniable value, especially since it is emphasized that the indices proposed in recent years to classify pyroconvective activity do not seem entirely satisfactory. It is also important to highlight the significant contribution of integrating numerical weather prediction model data, as these can help to underscore the models' own limitations. Would it be interesting to incorporate atmospheric profiles from models with higher vertical and/or horizontal resolution, such as AROME? On the other hand, meteorological radar measurements also hold great relevance, as they allow for validation the estimates of plume tops. In summary, this is a substantial work that sheds light on research in such a critical field, given the key role that pyroconvective activity plays in the escalation of megafires.
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC4 - AC3: 'Reply on CC4', Marc Castellnou, 08 Sep 2025
-
CC5: 'Comment on egusphere-2025-1923', Jongdae Kim, 06 Sep 2025
Review
This paper is an ambitious and very timely study that introduces a novel methodology for
in-plume radiosonde profiling during wildfires. The paper offers new insights into
fire–atmosphere interactions and pyroconvection dynamics, which is an active area of research
that needs further exploring. The indicators of potential transitions to extreme behavior help
identify plume top heights and characterize pyroconvection prototypes, which are notable
contributions that will help fire management better understand fire behavior. The authors also
provide one of the most extensive datasets of simultaneous ambient and in-plume profiles to
date. While the paper offers key contributions and advancements, the authors must refine the
write-up to be more consistent and coherent, improve the visualization of figures, and clarify
ambiguous points in the text to enhance readability. Please find comments and suggestions
below:
Clarify/highlight key contributions:
● I suggest restructuring the early stages of the paper to highlight key contributions and
advancements more clearly in the beginning. For example, line 466: “Observing the
plume top dilution just below the Lifting Condensation Level (LCL) in real-time is a
unique and valuable aspect of this methodology.”
● The title and abstract imply that the paper holds important implications for firefighter
safety. This is also foreshadowed in Lines 79-80 but not directly addressed in the paper.
Currently, it seems that safety is only discussed in terms of data collection. I recommend
including a sub-section or paragraph that discusses how the paper’s methods and
results can be used for firefighter safety.
Improving the visualization of main figures:
● Figure 1: The proportion of the observations are difficult to differentiate. I suggest adding
a text label denoting the relative proportion for each country/region (e.g., ME, 44.73%,
AE, 3.29%, SA, 51.98%). Also, since the symbols are difficult to see, it would help to see
the distribution of observations by fire type and by country/region (e.g., additional
stacked bar plots)
● Figure 3: Please update the horizontal axis labels so that they are consistent (i.e., Panel
a): θ (K) and Panel e): Temperature (oC))
● Figure 4: Please update the bottom table of the figure. I suggest to rename the contents
of “Fire event”, expand “Reg” (i.e., region), and write the fuel model code instead of as a
number. Also, under “Ambient Hour”, there are two times for Granja d’Escarp.
● Figure 8: I suggest updating the legend for plot a). First, I suggest using the full names of
the prototypes. Second, I suggest changing the font formatting of the legend and axis
labels to be more visible. Please also edit the x-axis label. Third, I suggest
re-ordering/modifying the visibility of the symbols (e.g., Re-order so symbols appear on
top or change the transparency). Fourth, for plot b), please edit the typo in the legend to
“radar echotop”
● Figures 9-12: Please edit the spacing of the plots in a) and b). Currently, the x-axis labels
are overlapping. Also, please edit the “Fire event” names,
● Figure 12: In the legend, the time formatting is incorrect (Check AM and PM) and the line
symbol for “Fire spread 19pm to 2am” is missing.
Clarification
● In Line 171, why should the soundings be taken “no more than 1 hour apart”? Is there
any supporting citation or any reasoning?
● Line 424: Why was the estimated plume top defined as the maximum height where radar
reflectivity was equal or higher than 12 dBz?
● Line 449: It is difficult to recognize the “excess of 7K”. I suggest explaining the surface
temperature values or labeling them to make this observation clearer
● Line 454: The authors claim that “the theoretical undiluted updraft height, estimated
using the MU parcel method (black dashed arrow), is located at 980 m AGL”. However,
Figure 9A shows that the black dashed arrow lies above 1000 m AGL. There are many
lines and colors in the plots which make the figure difficult to understand. While the
figure captions seem well-explained, the authors should clearly define each line and
color as well as provide ample reasoning for deciding on specific values (e.g., 980m
AGL) in the text to enhance readability.
Minor Comments
● Explicitly refer to ICON-EU as the atmospheric model
● Please add the short-form abbreviations of each prototype in Table 3. I suggest adding
them inside brackets below each prototype under the column “Pyroconvective Prototype”
● Please explain the state variables recorded by the in-plume (updraft) sonde or explicitly
refer to Table 2. The authors state that the in-plume sonde is classified based on its
position (head, flank, rear). Does this imply that, ideally, users should launch individual
sondes at each position to address turbulence experienced near the head direction of
the fire?
● Please use consistent terminology in the paper. For instance, in Figure 7, I suggest using
“Ambient” and “In-plume” (instead of “Environment”).
● Please provide a citation or reference to “the previous analysis” on Line 420Citation: https://doi.org/10.5194/egusphere-2025-1923-CC5 - AC4: 'Reply on CC5', Marc Castellnou, 23 Sep 2025
-
RC2: 'Comment on egusphere-2025-1923', Jean-Baptiste Filippi, 26 Oct 2025
The paper presents an operational and research-oriented observation method for in-plume radiosonde profiling during extreme wildfire events. It combines direct fireline observations with atmospheric soundings to quantify fire-atmosphere effects and evaluate pyroconvective transitions in real time. The authors should be commended for their long-term field effort—150 sondes over multiple fire seasons and continents—and for demonstrating the feasibility of affordable, lightweight instrumentation for operational plume monitoring. The work addresses a long-standing gap between model-based indices of pyroconvection and field observations available to incident managers. Congratulations for the work, it is obviously a very valuable field work analyzed here, there are no other consistent direct observations dataset of so many plumes to my knowledge.The paper is generally well structured, clearly written, and sound. It provides valuable insight into how in-plume thermodynamic profiles can be used to characterize the Atmospheric Boundary Layer (ABL), plume dilution, and potential transitions from dry to moist pyroconvection. The dataset has high potential value for model validation (e.g., Micro-HH, Meso-NH/ForeFire, WRF-Sfire) and for improving fire awareness protocols in operations. The figures are instructive, and the field documentation is impressive. The manuscript will interest both fire scientists and operational meteorologists.Some minor comments: is there any way to perform quantitative uncertainty associated with in-plume sondes ? representativeness in turbulent regimes and the sensitivity of plume-top estimates, maybe discuss that (a radiosonde is a single point in space / time).Also, among 150+ events, is there any availability on some on weather radar data ? a brief comparison of measured plume heights with radar would help to contextualize accuracy beyond the few examples shown.Minor:Ensure consistent notation for potential temperature (θ) and virtual potential temperature (θv).line 467 - “opyroCu” with “pyroCu”This is a well-executed and highly relevant contribution that bridges operational practice and research in pyroconvection monitoring. I recommend minor revision to address the small editorial and methodological clarifications listed above. Once revised, the manuscript will constitute an important reference for both field operations and coupled fire-atmosphere modeling.Citation: https://doi.org/
10.5194/egusphere-2025-1923-RC2 -
AC5: 'Reply on RC2', Marc Castellnou, 02 Nov 2025
We appreciate the reviewer's positive and encouraging feedback. We are pleased that the reviewer recognizes the significance of our long-term field effort and the potential of our dataset for both research and operational applications. Our primary goal has been to connect model-based indices with real-time field observations, and we are glad that this contribution is valued by both communities.
We will address all specific comments and suggestions raised by the reviewer in the following point-by-point responses in the supplement PDF
-
AC5: 'Reply on RC2', Marc Castellnou, 02 Nov 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-1923', Anonymous Referee #1, 03 Jul 2025
Overview
This preprint presents a novel and practical methodology for characterising pyroconvective wildfire plume dynamics using dual radiosonde soundings (in-plume and ambient). The study spans 156 field launches across four countries between 2021 and 2025 and offers both operational and scientific insights into plume development, real-time hazard awareness, and fire-atmosphere interaction modelling.
The manuscript is timely, rigorously detailed, and bridges a rare and valuable gap between operational field constraints and mesoscale meteorology. The work is distinguished by its applied innovation, extensive empirical validation, and potential to substantially inform firefighter safety procedures.
Strengths1. Novel Methodology with Operational Value
- The use of paired in-plume and ambient radiosonde profiles is both innovative and cost-effective, rendering it feasible for deployment during active wildfires.
- The operational integration into tactical decision-making workflows sets this study apart from traditional simulation-based or laboratory-bound research on pyroconvection.
2. Robust Field Campaign
- With 156 launches covering a diverse range of vegetation types, meteorological conditions, and terrain profiles, the dataset represents an impressive empirical foundation.
- The inclusion of both prescribed burns and uncontrolled wildfires increases the method’s general applicability.
3. Validation Through Multi-Modal Comparison
- Plume-top altitudes inferred from vertical velocity profiles were validated against radar echotop data, which significantly strengthens confidence in the method.
- Application of parcel theory for forecasting potential plume development (e.g., pyroCu onset) is methodologically sound and well-executed.
4. Classification Framework
- The six-category plume prototype typology (based on ABL height, LCL height, and wind shear layers) is operationally intuitive and scientifically coherent.
5. Actionable Outcomes
- Several case studies (e.g., Martorell and Santa Ana) demonstrate that tactical decisions informed by the sonde data likely contributed to risk mitigation. This real-world applicability is a major strength.
Weaknesses / Areas for Improvement
1. Clarification of Balloon Typology (Ref. Line 145, Table 1)
- The classification "professional high altitude balloon" is misleading. A more accurate term would be “operational radiosonde systems”, as both small and large balloons may be used professionally. For instance, an MW51 Vaisala ground station combined with an RS41 sonde and a Totex TA50 or TA100 balloon could easily reach 400 hPa level, albeit with higher helium demand.
- Table 1 also incorrectly claims that such systems are incapable of simultaneous launches. In fact, MW51 systems can track up to four sondes concurrently and have portable variants.
- Furthermore, the claim that professional radio-sounding systems are “not safe for aerial resources” is inaccurate. All operational weather balloons (regardless of size) comply with aviation safety regulations. In contrast, marking helicopters as inherently "safe" for aerial operations is misleading, as such platforms require strict coordination with air traffic control and firefighting aviation assets. These inaccuracies should be revised to reflect standard aviation safety protocols.
2. Reliability of Windsonde Data During Descent (Ref. Line 190)
- The authors should clearly state that measurements during descent are generally considered less reliable, even for professional sondes. Windsonde systems have not been formally validated for descent-phase data collection.
- As shown in Bessardon et al. (2016, Kumasi campaign), ground-based reference measurements of pressure and temperature were used to calibrate Windsonde outputs. It is unclear whether a similar calibration procedure was applied in this study.
- Moreover, the same research highlights concerns regarding wind speed and direction accuracy, particularly during turbulent or shear-laden environments. The authors should explicitly discuss whether and how these limitations were addressed.
- Known instrumental limitations of Windsonde include weak response to rapid humidity and temperature changes, and systematic underestimation of altitude (up to ~40 m). These should be acknowledged and addressed to justify continued use of this platform.
3. Quantitative Predictive Success Rate
The manuscript would benefit from a summary table or appendix quantifying:
- How many launches successfully entered the plume core?
- How many failed or produced partial profiles?
- In what proportion of cases did fire development escalate to Extreme Wildfire Events (EWEs)?
- In how many instances did the radiosonde data lead to a tactical change (e.g., crew withdrawal)?
4. Reproducibility: Launch Schedule and Decision Criteria
- For reproducibility and model intercomparison, the authors should provide a complete launch schedule overview, including exact timestamps and GPS coordinates of each sonde release.
- Additionally, the criteria used to determine the moment and location of launch should be explicitly stated (e.g., wind indicators, visual cues, forecast thresholds).
- These criteria appear to be field-operational in nature, but formalising them would help transfer the method to other contexts.
5. Real-Time Workflow and Decision Chain
The article does not describe the complete operational workflow from launch to decision. Clarifying the following would significantly improve transparency:
- How is data transmitted to and from the operations centre?
- Who is expected to perform the data analysis (on-site, centralised, or remote team)?
- What additional data sources are used (e.g., satellite, radar, fireline reports)?
- What is the end-to-end latency between launch and actionable tactical insight?
- Are any supporting information systems or software platforms (e.g., for visualisation or alerting) required or recommended?
6. Sonde Sampling Bias
- The authors acknowledge that sondes may not always enter the plume core, which may skew thermal and vertical velocity readings. Further statistical quantification of this sampling uncertainty would be beneficial.
7. Terminology
- Some terminology (e.g., “θv spike”, “fireABL”, “S parcel”) may not be immediately clear to the broader meteorological or fire-behaviour audience. A glossary or summary table of variables and acronyms is recommended.
Suggested Revisions
- Replace ambiguous or inaccurate entries in Table 1, particularly regarding balloon classifications, safety, and simultaneous sounding capability.
- Add a summary table of all launches with fire name, time, coordinates, outcome, and tactical decision if applicable.
- Include error metrics for vertical velocity-derived plume tops compared to radar.
- Clarify how and when real-time analysis was conducted, by whom, and how long it took.
- Address the known technical limitations of Windsondes and justify their use despite weaknesses.
- Provide access to a full launch log and reproducibility protocol, including selection criteria for launch timing and location.
Citation: https://doi.org/10.5194/egusphere-2025-1923-RC1 -
AC1: 'Reply on RC1', Marc Castellnou, 01 Aug 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-1923/egusphere-2025-1923-AC1-supplement.pdf
-
CC1: 'Comment on egusphere-2025-1923', Tomàs Artés Vivancos, 07 Aug 2025
I would like to congratulate to the authors for this work that contributes with data analsysis for plume analsysis in real time in real scenarios for decision-making support.The article proposes a sounding methodology to characterize fire plumes during wildfire events. Although it does not provide a detailed protocol, it clearly emphasizes the importance of in-situ data collection using a feasible and affordable approach to better understand and assess plume behavior and atmospheric dynamics during a fire. I find this work highly valuable, as fire analyses are often based solely on forecast data. This study offers a practical method for estimating fire plume dynamics in areas where radar systems are prohibitively expensive. Additionally, the real-time data collection and analysis performed during the fire enhance its value for emergency management, moving beyond the traditional post-event analysis approach.
The article presents very useful information and includes clear application examples. It is well-written and well-structured. However, readers should be familiar with fire and atmospheric terminology, and ideally be up to date with the current state of research on fire plume dynamics and fire typologies.
The article outlines a method for obtaining atmospheric soundings from an ongoing fire, including measurements not only of the ambient atmosphere but also of the head, flank, and rear of the fire. While the work does not go into detail about the balloon sounding launch procedure or the success rate—points also raised by an anonymous reviewer-it focuses on the sounding data collected from different locations and their use for analyzing plume state and potential, which is useful for operational decision-making.
The authors acknowledge the limitations of their approach and provide recommendations for how to work with the data presented.
This article makes an important contribution to operational practices and this research field by characterizing fire plumes and assess transitions between different fire plume behaviors.
I would recomment to the authors to apply at least minor comments when they apply.
The article focuses on integrating sounding data to characterize fire plumes, as stated in the title, and the content aligns well with this aim. The authors address the inherent uncertainty in the sounding data, particularly due to the balloons moving within the plume's indraft and temperature peaks, which affects the success rate of non-ambient soundings. However, there is no quantitative assessment of the uncertainty for each variable derived from the sounding methodology, nor an analysis of how sensitive the plume top estimation is to those uncertainties—beyond potential temperature and rising speed. I understand this could be the focus of a separate study, which would require a large number of soundings and additional instruments for validation in different contexts with prescibed fires. Still, the article highlights the importance of potential temperature and vertical velocity as key factors in plume-top estimation, with other variables being more relevant when using the parcel method for potential plume-top prediction.
Minor Comments:
Update model name and cite missing reference: MESO-NH/Forefire – Jean-Baptiste Filippi.
There is no need to justify the balloon radiosounding method beyond its affordability and operational simplicity. Other technologies like high-altitude atmospheric balloons, Doppler radar, UAVs, and helicopter-mounted sensors are capable of real-time data collection as well.
Table 1 may be unnecessary, but I understand that in real-world field campaigns, time constraints, costs, and complexity support the value of the information presented.
It would be helpful to use a consistent temperature unit throughout the article.
"S parcel increase by3°C" → should be "S parcel increases by 3°C" (missing space and verb correction).
"estimating the dilution plume height, is adequate" → remove the comma: "estimating the dilution plume height is adequate."
In Section 3.3.3, regarding the Casablanca III fire in Chile: it states the fire grew to 12,073 ha between February 8–10, 2023, but also says it had already blown up on February 2, reaching 363,000 ha. Including fire spread metrics would help readers better understand the relationship between plume dynamics and fire spread behavior.
"Fire behavior was initially expected to calm in the early evening, but there was a 60% chance of intensification during the day-to-night transition due to the advection of drier air from the SW" — I assume this is for context, and the 60% probability comes from a weather forecast ensemble.
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC1 -
AC2: 'Reply on CC1', Marc Castellnou, 13 Aug 2025
We would like to thank the reviewer for taking the time to review the paper and for providing valuable comments that have helped us clarify and improve the proposed methodology and its explanation.
-
CC3: 'Reply on AC2', Tomàs Artés Vivancos, 18 Aug 2025
The authors provided an answer to my comment. Actually, the answer was much more detailed that I expected. In addition, they addressed every comment.
Regarding the uncertainty of the different variables measured with the radiosonde, I was not expecting the authors to take new measurements, at all. Apologies for the misunderstanding. The article in its initial version already provided statistics about the measurements. My comment was about how variables like RH and potential temperature may affect the parcel models for potential top heights. I am not expecting the authors to address that fact since it could be a completely new work.
I congratulate the authors for the work that describes an important contribution to plume science dynamics within science and operational fields bringing together two worlds that are often too far away.
I consider this comment closed. Thanks
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC3
-
CC3: 'Reply on AC2', Tomàs Artés Vivancos, 18 Aug 2025
-
AC2: 'Reply on CC1', Marc Castellnou, 13 Aug 2025
-
CC2: 'Comment on egusphere-2025-1923', Cristina Montiel Molina, 17 Aug 2025
This is an impressive paper which presents a comprehensive and timely study of pyroconvective fire-atmosphere coupling. It is based on a novel practical methodology and a very detailed and accurate dataset assessment, to obtain a precise characterization of fire plumes during pyroconvective events.
The main highlight of this study is the design and demonstration of an original and sound inductive methodology to reduce incertainty and improve efficiency of emergency systems at extreme wildfire events. This is a badly needed contribution to the effects of climate change on wildfire risk that are creating the most complex and critical emergency scenarios.
On the basis of an updated and complete state-of the art, the first strenght of this study is to design and test a novel data-gathering method of pyroconvection dynamics in-situ and in real-time. Furthermore, the global dimension and multi-scalar approach of the research design enhances its consistency in the context of global climate change. Another strenght of this study is to define and to apply an accurate pyroconvention prototype system.
The obtained results provide a completely innovative understanding and modeling of pyroconvection dynamics. This contribution has a strong potential to build-up a powerful tool of information management enabling to anticipate defense actions to control the most challenging wildfire events and to mitigate firefighters and civil population vulnerability.
The realistic and critical approach of the discussion section is also outstanding. The authors acknowledge the limitations of the methodology and consider the importance of contextual factors for a right assessment of measurements.
The conclussions on fire-atmosphere interaction are a badly needed contribution for a climate change adapted new model of proactive wildfire risk management beyond the current paradigm of existing fuel models and fire weather indexes.
It’s an excellent scientific paper that could have a huge scientific and operational impact. I don’t have any concrete suggestion for improvement further than those already done by previous reviewers. I congratulate the authors on this magnificent research.
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC2 -
CC4: 'Comment on egusphere-2025-1923', Peio Oria Iriarte, 31 Aug 2025
This paper presents a valuable contribution to wildfire science by integrating fireline observations to enhance the understanding of fire plume behavior during pyroconvective extreme wildfire events. The authors effectively highlight the importance of detailed, real-time observational data in characterizing complex fire-induced phenomena, which has significant implications for improving plume modeling accuracy. The study's focus on firefighter safety underscores the practical relevance of accurately predicting plume dynamics to inform operational decision-making. Overall, the work advances both the scientific understanding of extreme wildfire behavior and its application to emergency response and safety management, offering meaningful insights for researchers, practitioners, and policymakers involved in wildfire mitigation and response efforts.
Having atmospheric profile data in the environment and inside the convective plume is of undeniable value, especially since it is emphasized that the indices proposed in recent years to classify pyroconvective activity do not seem entirely satisfactory. It is also important to highlight the significant contribution of integrating numerical weather prediction model data, as these can help to underscore the models' own limitations. Would it be interesting to incorporate atmospheric profiles from models with higher vertical and/or horizontal resolution, such as AROME? On the other hand, meteorological radar measurements also hold great relevance, as they allow for validation the estimates of plume tops. In summary, this is a substantial work that sheds light on research in such a critical field, given the key role that pyroconvective activity plays in the escalation of megafires.
Citation: https://doi.org/10.5194/egusphere-2025-1923-CC4 - AC3: 'Reply on CC4', Marc Castellnou, 08 Sep 2025
-
CC5: 'Comment on egusphere-2025-1923', Jongdae Kim, 06 Sep 2025
Review
This paper is an ambitious and very timely study that introduces a novel methodology for
in-plume radiosonde profiling during wildfires. The paper offers new insights into
fire–atmosphere interactions and pyroconvection dynamics, which is an active area of research
that needs further exploring. The indicators of potential transitions to extreme behavior help
identify plume top heights and characterize pyroconvection prototypes, which are notable
contributions that will help fire management better understand fire behavior. The authors also
provide one of the most extensive datasets of simultaneous ambient and in-plume profiles to
date. While the paper offers key contributions and advancements, the authors must refine the
write-up to be more consistent and coherent, improve the visualization of figures, and clarify
ambiguous points in the text to enhance readability. Please find comments and suggestions
below:
Clarify/highlight key contributions:
● I suggest restructuring the early stages of the paper to highlight key contributions and
advancements more clearly in the beginning. For example, line 466: “Observing the
plume top dilution just below the Lifting Condensation Level (LCL) in real-time is a
unique and valuable aspect of this methodology.”
● The title and abstract imply that the paper holds important implications for firefighter
safety. This is also foreshadowed in Lines 79-80 but not directly addressed in the paper.
Currently, it seems that safety is only discussed in terms of data collection. I recommend
including a sub-section or paragraph that discusses how the paper’s methods and
results can be used for firefighter safety.
Improving the visualization of main figures:
● Figure 1: The proportion of the observations are difficult to differentiate. I suggest adding
a text label denoting the relative proportion for each country/region (e.g., ME, 44.73%,
AE, 3.29%, SA, 51.98%). Also, since the symbols are difficult to see, it would help to see
the distribution of observations by fire type and by country/region (e.g., additional
stacked bar plots)
● Figure 3: Please update the horizontal axis labels so that they are consistent (i.e., Panel
a): θ (K) and Panel e): Temperature (oC))
● Figure 4: Please update the bottom table of the figure. I suggest to rename the contents
of “Fire event”, expand “Reg” (i.e., region), and write the fuel model code instead of as a
number. Also, under “Ambient Hour”, there are two times for Granja d’Escarp.
● Figure 8: I suggest updating the legend for plot a). First, I suggest using the full names of
the prototypes. Second, I suggest changing the font formatting of the legend and axis
labels to be more visible. Please also edit the x-axis label. Third, I suggest
re-ordering/modifying the visibility of the symbols (e.g., Re-order so symbols appear on
top or change the transparency). Fourth, for plot b), please edit the typo in the legend to
“radar echotop”
● Figures 9-12: Please edit the spacing of the plots in a) and b). Currently, the x-axis labels
are overlapping. Also, please edit the “Fire event” names,
● Figure 12: In the legend, the time formatting is incorrect (Check AM and PM) and the line
symbol for “Fire spread 19pm to 2am” is missing.
Clarification
● In Line 171, why should the soundings be taken “no more than 1 hour apart”? Is there
any supporting citation or any reasoning?
● Line 424: Why was the estimated plume top defined as the maximum height where radar
reflectivity was equal or higher than 12 dBz?
● Line 449: It is difficult to recognize the “excess of 7K”. I suggest explaining the surface
temperature values or labeling them to make this observation clearer
● Line 454: The authors claim that “the theoretical undiluted updraft height, estimated
using the MU parcel method (black dashed arrow), is located at 980 m AGL”. However,
Figure 9A shows that the black dashed arrow lies above 1000 m AGL. There are many
lines and colors in the plots which make the figure difficult to understand. While the
figure captions seem well-explained, the authors should clearly define each line and
color as well as provide ample reasoning for deciding on specific values (e.g., 980m
AGL) in the text to enhance readability.
Minor Comments
● Explicitly refer to ICON-EU as the atmospheric model
● Please add the short-form abbreviations of each prototype in Table 3. I suggest adding
them inside brackets below each prototype under the column “Pyroconvective Prototype”
● Please explain the state variables recorded by the in-plume (updraft) sonde or explicitly
refer to Table 2. The authors state that the in-plume sonde is classified based on its
position (head, flank, rear). Does this imply that, ideally, users should launch individual
sondes at each position to address turbulence experienced near the head direction of
the fire?
● Please use consistent terminology in the paper. For instance, in Figure 7, I suggest using
“Ambient” and “In-plume” (instead of “Environment”).
● Please provide a citation or reference to “the previous analysis” on Line 420Citation: https://doi.org/10.5194/egusphere-2025-1923-CC5 - AC4: 'Reply on CC5', Marc Castellnou, 23 Sep 2025
-
RC2: 'Comment on egusphere-2025-1923', Jean-Baptiste Filippi, 26 Oct 2025
The paper presents an operational and research-oriented observation method for in-plume radiosonde profiling during extreme wildfire events. It combines direct fireline observations with atmospheric soundings to quantify fire-atmosphere effects and evaluate pyroconvective transitions in real time. The authors should be commended for their long-term field effort—150 sondes over multiple fire seasons and continents—and for demonstrating the feasibility of affordable, lightweight instrumentation for operational plume monitoring. The work addresses a long-standing gap between model-based indices of pyroconvection and field observations available to incident managers. Congratulations for the work, it is obviously a very valuable field work analyzed here, there are no other consistent direct observations dataset of so many plumes to my knowledge.The paper is generally well structured, clearly written, and sound. It provides valuable insight into how in-plume thermodynamic profiles can be used to characterize the Atmospheric Boundary Layer (ABL), plume dilution, and potential transitions from dry to moist pyroconvection. The dataset has high potential value for model validation (e.g., Micro-HH, Meso-NH/ForeFire, WRF-Sfire) and for improving fire awareness protocols in operations. The figures are instructive, and the field documentation is impressive. The manuscript will interest both fire scientists and operational meteorologists.Some minor comments: is there any way to perform quantitative uncertainty associated with in-plume sondes ? representativeness in turbulent regimes and the sensitivity of plume-top estimates, maybe discuss that (a radiosonde is a single point in space / time).Also, among 150+ events, is there any availability on some on weather radar data ? a brief comparison of measured plume heights with radar would help to contextualize accuracy beyond the few examples shown.Minor:Ensure consistent notation for potential temperature (θ) and virtual potential temperature (θv).line 467 - “opyroCu” with “pyroCu”This is a well-executed and highly relevant contribution that bridges operational practice and research in pyroconvection monitoring. I recommend minor revision to address the small editorial and methodological clarifications listed above. Once revised, the manuscript will constitute an important reference for both field operations and coupled fire-atmosphere modeling.Citation: https://doi.org/
10.5194/egusphere-2025-1923-RC2 -
AC5: 'Reply on RC2', Marc Castellnou, 02 Nov 2025
We appreciate the reviewer's positive and encouraging feedback. We are pleased that the reviewer recognizes the significance of our long-term field effort and the potential of our dataset for both research and operational applications. Our primary goal has been to connect model-based indices with real-time field observations, and we are glad that this contribution is valued by both communities.
We will address all specific comments and suggestions raised by the reviewer in the following point-by-point responses in the supplement PDF
-
AC5: 'Reply on RC2', Marc Castellnou, 02 Nov 2025
Peer review completion
Journal article(s) based on this preprint
Data sets
In-Plume Radiosonde Fireline Observations Marc Castellnou Ribau et al. https://doi.org/10.5281/zenodo.15264835
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,187 | 199 | 48 | 1,434 | 73 | 24 | 41 |
- HTML: 1,187
- PDF: 199
- XML: 48
- Total: 1,434
- Supplement: 73
- BibTeX: 24
- EndNote: 41
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Marc Castellnou Ribau
Mercedes Bachfischer
Marta Miralles Bover
Borja Ruiz
Laia Estivill
Jordi Pages
Pau Guarque
Brian Verhoeven
Zisoula Ntasiou
Ove Stokkeland
Chiel Van Herwaeeden
Tristan Roelofs
Martin Janssens
Cathelijne Stoof
Jordi Vilà-Guerau de Arellano
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2758 KB) - Metadata XML
-
Supplement
(660 KB) - BibTeX
- EndNote
- Final revised paper