Advancements in Regional Weather Modeling for South Asia Through the High Impact Weather Assessment Toolkit (HIWAT) Archive

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall, very interesting work.  I only have minor comments which should be clarified before publication.  I hope that I have the chance to use this dataset in the future.

 - The manuscript appears to abruptly end.  A summary or concluding paragraph about the potential impacts of this work and potential next steps may be appropriate.

 - What's a data granule? It sounds like a file, but I've never heard that term used as a unit of measure for data.

 - In the straight-line damaging wind parameters, are you looking at sub-hourly timescales?  Or is it post-processed from hourly data?

 - An internet search of UEFS didn't provide any direct results.  Could you provide more details and/or a citation for help readers understand this tool?

 - While your description of the technical configuration of HIWAT is detailed, because this is purporting to be an real-time operational system, I find myself wanting to know more about data assimilation. The weather observation infrastructure in the developed world can significantly contribute to the quality of their weather forecasts, and this information would help readers understand this work in that context.

 - Some more detail about the modules used in the WRF model would be appreciated: land surface model, microphysics scheme, etc.

 - While likely out of the scope of this paper, I can't help but want to see some form of validation of this model.  Does it resolve some of the high profile extreme weather events that have affected the region in the past few years?  Or are potential weaknesses which may have to be addressed or considered when using this model operationally?  Any information in this regard would help readers appreciate this work more fully, and potentially motivate them to use the dataset. 

Author Response

Editor and Reviewer Comments:

Review #1

 

Overall, very interesting work.  I only have minor comments which should be clarified before publication.  I hope that I have the chance to use this dataset in the future.

1.  - The manuscript appears to abruptly end.  A summary or concluding paragraph about the potential impacts of this work and potential next steps may be appropriate.

Thank you so much for your comment. Our author team agrees with the reviewers comment and have provided additional conclusion sections. 

(Line: 181-190)

Conclusions

For greater details regarding HIWAT model configuration and implementation within SOCRATES please see Gatlin et al. (2021) [10] and Case et al. 2023 [3]. Wherein damaging wind, large hail, lightning, a meteorologically rare Nepalese tornado, and landfalling tropical cyclone are provided as use cases demonstrating the efficacy of HIWAT to provide vital information to forecasters. Finally, the HIWAT system, focused on South Asia, has progressed since its nascent development stages into a reliable operational system. The longer-term operation of a 12-member ensemble HIWAT forecast system, based at ICIMOD Nepal, continues where vital forecast information is provided to hydrometeorological offices to support the need for high resolution weather forecasts. 

1.  - What's a data granule? It sounds like a file, but I've never heard that term used as a unit of measure for data.

Thank you so much for this comment. Yes, NASA Distributed Active Archive Centers (DAACs) including Global Hydrometeorology Resource Center Distributed Active Archive Center (GHRC) refer to individual files within a larger holding as a data granule in this case. We have updated the text to reduce the unnecessary jargon.

(Line: 125-127)

The HIWAT archive spans April 2, 2017, through October 2, 2022, with the full pre-monsoon and wet-monsoon months of March-September spanning 2018-2022. The total HIWAT GHRC DAAC archive is 1,239,382 files at 349.24 TB disk storage.

1.  - In the straight-line damaging wind parameters, are you looking at sub-hourly timescales?  Or is it post-processed from hourly data?

Thank you so much for this comment.  Yes, the HIWAT system does incorporate sub-hourly information into the straight-line damaging wind metric. During the ensemble member model simulations, the maximum wind speed at every grid point is updated each dynamic model timestep (~ every 30-60 seconds), such that at the hourly output time, the 2D maximum wind speed field is written to the file. This accounting method described in Kain et al. (2010) helps to preserve details of rapidly-evolving phenomena at finer scales than the file output interval frequency. This method is similarly applied to other rapidly-evolving phenomena such as the updraft helicity metric for denoting rotating thunderstorm updrafts, and vertically-integrated graupel for representing hail threat. This additional reference and updated text have been added to the manuscript for more clarity. Thank you again for the comment. 

 

Kain, J. S., S. R. Dembek, S. J. Weiss, J. L. Case, J. J. Levit, and R. A. Sobash, 2010: Extracting unique information from high-resolution forecast models: Monitoring selected fields and phenomena every time step. Wea. Forecasting, 25(5), 1536-1542, https://doi.org/10.1175/2010WAF2222430.1.

(Line: 107-112)

Additionally, HIWAT incorporate sub-hourly information at grid points that enable dynamic model timestep to inform the hourly outputs which preserves details of rapidly-evolving phenomena at finer scales (Kain et al. 2010) This approach is used for rapidly-evolving phenomena such as the updraft helicity metric for denoting rotating thunderstorm updrafts, and vertically-integrated graupel for representing hail threat. Finally, as a caveat, from 2021-2022 the number of ensembles was reduced from twelve to nine due to computing constraints/

1.  - An internet search of UEFS didn't provide any direct results.  Could you provide more details and/or a citation for help readers understand this tool?

Thank you so much for this comment. Our team leveraged the Unified Environmental Modeling System (UEMS) which more information can be found with the associated citations below. UEMS was largely based on the Weather Research and Forecasting community modeling framework. The UEMS was managed for over twenty  years by the NOAA/National Weather Service. However, this system has been reprioritized and no longer provides public support for the UEMS framework. As a result the UEMS web page as of 2024 is not long operational. This prompted our team to choose to no longer provide a direct link as in other publications. 

Gatlin, P.N., Case, J.L., Srikishen, J. and Adhikary, B., 2021. The high-impact weather assessment toolkit. Earth Observation Science and Applications for Risk Reduction and Enhanced Resilience in Hindu Kush Himalaya Region: A Decade of Experience from SERVIR, pp.231-250.

Case, J.L., Gatlin, P.N., Srikishen, J., Adhikary, B., Mannan, M.A. and Bell, J.R., 2023. Building thunderstorm resilience in the Hindu Kush Himalaya Region through probabilistic forecasts and satellite observations. Bulletin of the American Meteorological Society, 104(5), pp.E1105-E1131.

1. - While your description of the technical configuration of HIWAT is detailed, because this is purporting to be an real-time operational system, I find myself wanting to know more about data assimilation. The weather observation infrastructure in the developed world can significantly contribute to the quality of their weather forecasts, and this information would help readers understand this work in that context.

Thank you so much for this comment. As the focus of the Data paper was to provide more access, discoverability, and context of existing high quality research and operational HIWAT data hosted at the GHRC DAAC, our team provided limited in-text discussion to detailed modeling aspects.However, the additional conclusion section should help prompt readers to explore the methodology and applications further.

And to your comment directly, in our team's instantiation of the WRF/UEMS model for HIWAT, we did not explicitly conduct regional data assimilation at the initialization hour. However, by incorporating alternative Global Ensemble Forecasting System (GEFS) modeling members to initialize the different HIWAT members, we implicitly include the impacts of global data assimilation and ensemble perturbation techniques from the global ensemble system managed by the NOAA/NCEP's Environmental Modeling Center.

 

 

1.  - Some more detail about the modules used in the WRF model would be appreciated: land surface model, microphysics scheme, etc.

Thank you so much for this comment. Similar to the above response.

1.  - While likely out of the scope of this paper, I can't help but want to see some form of validation of this model.  Does it resolve some of the high profile extreme weather events that have affected the region in the past few years?  Or are potential weaknesses which may have to be addressed or considered when using this model operationally?  Any information in this regard would help readers appreciate this work more fully, and potentially motivate them to use the dataset. 

Thank you so much for your comment. Yes we believe this to be outside the scope of the current work. However, with the addition of the conclusion section we aim to direct readers to a robust article that provides detailed use cases with more information on applications and utility of the forecast.

We encourage the reviewer to refer to Case et al. (2023), in which a model validation comparing the HIWAT members to the GEFS members clearly demonstrate HIWAT's ability to simulate and resolve intense precipitation systems within the complex terrain of south-central Asia that the GEFS global ensemble is unable to resolve.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for sending your manuscript to us. It is a compilation of advanced meteorological products developed within the NASA SERVIR program. HIWAT focuses on integrating high-resolution weather forecasts and assessing extreme weather event likelihood within the Hindu Kush Himalaya (HKH), which includes Nepal, Bangladesh, Bhutan and northeastern India. Within the study, the researchers focus on post- and pre-monsoon (March:June and July:October) seasons, which are frequently characterized by high winds, intense rainfall, severe thunderstorms and extreme lightning occurrences. As for the manuscript itself, it is overall well-organized and the methods described follow a logical order. The use of a 4 km resolution is appropriate for representing severe thunderstorms and high intensity precipitation events in an area of significant orography. The use of Ensemble Forecasting allows for better uncertainty estimation; this is paramount as those more vulnerable areas may become susceptible to more extreme weather events. The implementation of an HPC cluster is an incredibly useful resource for complex calculations on daily initialization and in multi-member ensemble integration terms.

Lastly, I would like to make a few recommendations that would improve the manuscript: The forecasting model uses different parameterization schemes but the selection criteria are not explained in detail. It would be helpful to discuss why schemes were selected and not others, particularly when selection is based on sensitivity to meteorological characteristics of the region. If there have been comparisons of schemes available, it would be an important addition to the study. Although dealing with uncertainties is valid with a probabilistic approach using Ensemble Forecasting the manuscript is lacking discussion on the validation process that was performed. You have mentioned that the probabilistic outputs are compared with observed data, it would increase reader's confidence in model trustworthiness if possible. It would enhance the content to include performance metrics like Brier score, calibration, or ROC curves. While the 4 km resolution is moderate, higher resolution settings (i.e 1 km) would potentially help with detection of micro-meteorological events. Similarly, real-time satellite observations could be integrated into the forecasting method to improve accuracy and validation of forecasts. This could be a topic of discussion if it is possible to implement within your current timeframe, as well as discussing computational implications.

Author Response

Review #2

 

Thank you for sending your manuscript to us. It is a compilation of advanced meteorological products developed within the NASA SERVIR program. HIWAT focuses on integrating high-resolution weather forecasts and assessing extreme weather event likelihood within the Hindu Kush Himalaya (HKH), which includes Nepal, Bangladesh, Bhutan and northeastern India. Within the study, the researchers focus on post- and pre-monsoon (March:June and July:October) seasons, which are frequently characterized by high winds, intense rainfall, severe thunderstorms and extreme lightning occurrences. As for the manuscript itself, it is overall well-organized and the methods described follow a logical order. The use of a 4 km resolution is appropriate for representing severe thunderstorms and high intensity precipitation events in an area of significant orography. The use of Ensemble Forecasting allows for better uncertainty estimation; this is paramount as those more vulnerable areas may become susceptible to more extreme weather events. The implementation of an HPC cluster is an incredibly useful resource for complex calculations on daily initialization and in multi-member ensemble integration terms.

Lastly, I would like to make a few recommendations that would improve the manuscript: 

1. The forecasting model uses different parameterization schemes but the selection criteria are not explained in detail. It would be helpful to discuss why schemes were selected and not others, particularly when selection is based on sensitivity to meteorological characteristics of the region. If there have been comparisons of schemes available, it would be an important addition to the study. 

Thank you so much for your comment. Our author team agrees with the reviewers comment and we believe the addition of the conclusions section will help direct the reader to relevant HIWAT research and application articles that expound on these details. To keep within the scope of the Data journal discussion,  we now refer the reader to the foundational publications describing the HIWAT system in greater detail, which includes the motivation for choice of physical parameterization variability in the ensemble system. We should also note herein for the reviewer’s information that for physical parameterization choices, we aimed to (1) offer variability to create meaningful ensemble spread important to the onset and evolution of convective precipitation systems, and (2) that the schemes were relatively inexpensive computationally, so as to minimize the latency of the delivered output products.

(Line: 181-190)

Conclusions

For greater details regarding HIWAT model configuration and implementation within SOCRATES please see Gatlin et al. (2021) [10] and Case et al. 2023 [3]. Wherein damaging wind, large hail, lightning, a meteorologically rare Nepalese tornado, and landfalling tropical cyclone are provided as use cases demonstrating the efficacy of HIWAT to provide vital information to forecasters. Finally, the HIWAT system, focused on South Asia, has progressed since its nascent development stages into a reliable operational system. The longer-term operation of a 12-member ensemble HIWAT forecast system, based at ICIMOD Nepal, continues where vital forecast information is provided to hydrometeorological offices to support the need for high resolution weather forecasts. 

1. Although dealing with uncertainties is valid with a probabilistic approach using Ensemble Forecasting the manuscript is lacking discussion on the validation process that was performed. You have mentioned that the probabilistic outputs are compared with observed data, it would increase reader's confidence in model trustworthiness if possible. It would enhance the content to include performance metrics like Brier score, calibration, or ROC curves. While the 4 km resolution is moderate, higher resolution settings (i.e 1 km) would potentially help with detection of micro-meteorological events.

Thank you so much for your comment. Our author team agrees and aims to have the added conclusions eciton help direct the reader toward through resources, while keeping the length and breadth of the Data article streamlined. We therefore encourage the Reviewer to refer to Case et al. (2023), in which a model validation comparing the HIWAT members to the GEFS members clearly demonstrate HIWAT's ability to simulate and resolve intense precipitation systems within the complex terrain of south-central Asia that the GEFS global ensemble is unable to resolve.

 

1. Similarly, real-time satellite observations could be integrated into the forecasting method to improve accuracy and validation of forecasts. This could be a topic of discussion if it is possible to implement within your current timeframe, as well as discussing computational implications.

Thank you so much for your comment and a great suggestion for future work to explore. However, our author team has concluded this is to be outside the scope of this current work. We again ask that the Reviewer further explore the Gatlin et al. (2021) and Case et al. (2023) references, which provide case studies and validation of HIWAT against satellite datasets and ground observations, as well as quantitative precipitation estimates (i.e., GPM/IMERG-Final).

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have successfully addressed all of my previous concerns.  Thank you for all of the hard work and attention. 

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for resubmitting your manuscript entitled "Update on Regional Weather Modelling for South Asia Using the High Impact Weather Assessments Toolkit (HIWAT) Archive."

I would like to take this opportunity to acknowledge the significant improvements made to the manuscript with respect to version V2 in response to the comments and recommendations from the first review.

You have clearly attended to the main issues in the former review. In particular, the methods sections are much clearer, and especially with respect to the development and interpretation of the probabilistic products from the HIWAT system. This work provides a stronger and additional scientific basis for the study.

I would also like to remark on the addition of technical details that were omitted previously. In particular, consideration of implementing sub-hourly model data is a significant enhancement because it allows for the representation of rapidly evolving weather phenomena such as updraft helicity and column-integrated graupel better than previous iterations. It is also evident, that consideration of the HIWAT products temporal and spatial resolution have improved relevance and applicability for the operational user.

From a presentation perspective, the manuscript has made significant improvements. The document now has a greater structural integrity, the terminology choice appears consistent and appropriate throughout, meaning the document is now much easier to follow. The conclusion also has been enhanced with a clearer message concerning the ongoing operational implementation of HIWAT.

In summary, the manuscript represents a clear improvement over the previous version and is in a much more favorable position for publication.

Sincerely,