the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Apr 2025
A Reanalysis-Based Global Tropical Cyclone Tracks Dataset for the Twentieth Century (RGTracks-20C)
Abstract. Tropical cyclones (TCs) are among the deadliest disasters affecting human society, and their response to climate change has widely drawn attention from the public. However, assessing how historical TC activity changed with climate change has proven challenging due to incomplete TC records in the early years. Here, we introduce the Reanalysis-Based Global Tropical Cyclone Tracks Dataset for the Twentieth Century (RGTracks-20C) (Ye et al., 2024), a publicly available century- long global TC track dataset spanning from 1850–2014. The RGTracks-20C is reconstructed from the National Oceanic and Atmospheric Administration Twentieth Century Reanalysis using two independent TC tracking algorithms. Validation based on observations confirms that the RGTracks- 20C effectively captures the climatology and long-term variability of TC numbers, tracks, duration, and intensity across various ocean basins. A remarkable key strength of the RGTracks-20C is its ability to fill the missing intensity and location records of TCs observed in early years. This dataset serves as a valuable historical data reference for future research on climate change and TC-related disasters.
- Preprint
(3059 KB) - Metadata XML
-
Supplement
(1638 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2025-126', Christopher w. Landsea, 09 May 2025
Review of “A Reanalysis-Based Global Tropical Cyclone Tracks Dataset for the Twentieth Century (RGTrack-20C)” for Earth System Science Data (ESSD)
Recommendation: Accept after major revisions
Review “signed” by Chris Landsea
This paper is an innovative look at tropical cyclone activity via the model reanalysis products. This does have the potential for improving our understanding of tropical cyclone variability and trends over time. A very interesting aspect of the study is the ability of the reanalysis to fill in the gaps from historic storms to provide realistic-looking central pressure information, where little to none had existed before. There are a number of issues – a few of significant concern – that must be addressed before this reviewer can recommend publication. However, these issues should be not insurmountable and the paper should help advance the field once published.
Significant issues:
- Trends in the number of sea level pressure observations available to the model reanalysis: While the model reanalysis provides an objective homogeneous platform to compare tropical cyclones now versus the past, the observations going into it change over time. More sea level pressure measurement will allow for better detection of tropical cyclones as well as more complete representations of the intensity (maximum wind/minimum pressure). Please discuss/show how the number of pressure measurements from ships, stations, and buoys going into the reanalysis changes over time.
- IBTRACS’ central pressures into reanalysis: The IBTRACS’ central pressure values for tropical cyclones are included as an “observation” for the reanalysis to assimilate. This means that the reanalysis and IBTRACS are not independent and that the reanalysis should replicate much of the characteristics of IBTRACS. Please make sure that this point is clearly stated. Moreover, please indicate in the paper what year by basin that the central pressures were routinely included into IBTRACS. For example, the Northeast/North Central Pacific only began including pressures into IBTRACS starting in 1988. This has a profound effect on the ability of the reanalysis to detect the tropical cyclone activity. Again focusing upon the NE/NC Pacific, the low hit/high miss rate in in Figure 3, the very low numbers of TCs in Figure 7a before 1988, the very low duration of TCs in Figure 7b before 1988, and the low correlations in Table 3 are likely a consequence of these missing pressures before 1988. Please address these issues and also point out where other basins may have similar issues before central pressures were routinely included into IBTRACS.
Minor Points:
- Does duration include non-tropical cyclone stages (extratropical, pre-genesis low, remnant low)? It should not.
- Figure 9: It is noted that the reanalysis completely missed the Category 3 hurricane stage of Hurricane Andrew while it was over the Gulf of Mexico. This could be related to the very small size of system.
- Figure 4 and S4: Are the red and green curves right on top of each other on C) and D)? If so, please mention this explicitly.
- Figure 7 and 8: The authors should note and comment about where there are discrepancies in the slope of the long-term trend between the reanalysis and IBTRACS, such as Figure 6C for central pressure.
Citation: https://doi.org/10.5194/essd-2025-126-RC1 -
AC1: 'Reply on RC1', Guiling Ye, 17 Jun 2025
Thank you very much for your valuable comments and constructive suggestions on our research work. According to your suggestions, we have made substantial revisions to the manuscript. Particularly, we have discussed the changes in the assimilation of observational data into 20CRv3, as well as the starting years for providing intensity information in each ocean basin in IBTrACS, and conducted detailed comparisons between long-term tropical cyclone trends in RGTracks-20C and observational datasets. We hope we have addressed all your concerns.
Please find our answers in the attached file (.zip).
With kind regards,On behalf of all co-authors
-
RC2: 'Comment on essd-2025-126', Anonymous Referee #2, 15 May 2025
Review
A Reanalysis-Based Global Tropical Cyclone Tracks Dataset for the Twentieth Century (RGTracks-20C)
Essd-2025-126
General Statement:
The manuscript by Ye and the colleagues presents a reconstructed tropical cyclone (TC) dataset using two commonly applied tracker algorithms, OWZ and UZ. The authors evaluate the climatology and long-term variability of TC frequency, track patterns, duration, and intensity across various ocean basins, comparing their RGTracks-20C product with the IBTrACS dataset. They argue that a key advantage of RGTracks-20C lies in its ability to fill gaps in historical TC intensity and location records, particularly in the earlier part of the record. However, considering the lack of methodological novelty and the limitations associated with the underlying reanalysis dataset, I am not convinced that this study meets the publication standards of ESSD.
Comments:
- The authors appear to apply two widely used trackers (OWZ and UZ) without any evident modification or improvement. Notably, the UZ tracker has recently been enhanced through the integration of AI-based techniques, showing considerable performance gains over its original version (Han & Ullrich, 2025). A new TC dataset has been reconstructed using this improved approach. If the authors have indeed introduced any innovation or adaptation to the tracking algorithms, this should be clearly detailed in the manuscript. Otherwise, the study amounts to a straightforward application of existing (and arguably outdated) methods to the 20CRv3 dataset. From this perspective, the contribution seems largely computational and lacks substantive novelty.
- The limitations of the 20CRv3 reanalysis dataset have been discussed in detail by Emanuel (2024). Moreover, previous studies have already generated TC track datasets using reanalysis data, such as ERA5, in conjunction with various tracking algorithms (e.g., Bourdin et al., 2022). A comparative assessment of TC tracks derived from different reanalysis datasets (e.g., ERA5 vs. 20CRv3) using identical or similar trackers would have added substantial value by helping validate the reliability of each dataset and identify the more suitable one for historical TC reconstruction.
Reference:
Bourdin, S., Fromang, S., Dulac, W., Cattiaux, J., & Chauvin, F. (2022). Intercomparison of four algorithms for detecting tropical cyclones using ERA5. Geoscientific Model Development, 15(17), 6759-6786.
Emanuel, K. (2024). Limitations of reanalyses for detecting tropical cyclone trends. Nature Climate Change, 14(2), 143-145.
Han, Y., & Ullrich, P. (2025) The System for Classification of Low-Pressure Systems (SyCLoPS): An All-In-One Objective Framework for Large-Scale Data Sets. JGR Atmospheres, https://doi.org/10.1029/2024JD041287
Citation: https://doi.org/10.5194/essd-2025-126-RC2 -
AC2: 'Reply on RC2', Guiling Ye, 30 Jun 2025
Thank you very much for your valuable comments and constructive suggestions on our manuscript. You have raised two important points regarding our work: the novelty of employed approaches and the choice of reanalysis datasets used for historical TC reconstruction. We appreciate your careful consideration of our manuscript. We have accordingly provided a detailed response to your concerns (attached).
Best regards,
-
AC3: 'Reply on RC2', Guiling Ye, 13 Jul 2025
We sincerely appreciate your comprehensive evaluation and constructive feedback on our study. We note your concern regarding the novelty of the research methodology. Thus, in this follow-up response, we have implemented the System for Classification of Low-Pressure Systems (SyCLoPS) algorithm developed by Han and Ullrich (2025), as suggested by you, to our 20CRv3 dataset for comparative evaluation. Detailed comparative analysis and results are presented in the attached document.
Best regards,
Data sets
RGTracks-20C: v1.0.2 Jeremy Cheuk-Hin Leung https://doi.org/10.5281/zenodo.8410596
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 969 | 84 | 27 | 1,080 | 44 | 29 | 40 |
- HTML: 969
- PDF: 84
- XML: 27
- Total: 1,080
- Supplement: 44
- BibTeX: 29
- EndNote: 40
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
