the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 May 2025
Indonesia Tornado Database: Tornado Climatology of Indonesia
Abstract. The climatology of tornadoes, waterspouts, and funnel clouds for Indonesia is constructed into a new tornado database based on newspaper archives, news on the Internet, and social media (X, YouTube) covering the period of 1834–2024. We present the analysis over two periods: (i) historical and (ii) recent periods. The climatology includes the annual, monthly, diurnal, and geographical for tornado cases. Based on a review of 47,669 reports mentioning tornadoes, we identified 436 tornado cases with sufficient evidence to be classified as tornado events. In the recent period (2010–2024), the annual frequency of tornadoes was 16.20 cases/year, while in the historical period (1834–2009), the annual frequency was 1.10 cases/year. Tornadoes were mostly documented in Java Island, followed by Sumatra and Sulawesi. The monthly variability of tornadoes shows a maximum during November, followed by December and January. The peak of the diurnal cycle of tornado cases is between 1300 to 1700 Local Solar Time.
- Preprint
(1705 KB) - Metadata XML
-
Supplement
(450 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1857', Anonymous Referee #1, 09 Jul 2025
The manuscript presents the first comprehensive tornado climatology for Indonesia, compiling 436 events from 1834–2024 using multiple data sources. The work fills an important gap in our current knowledge, because unlike the US or Europe, Indonesia had no tornado database, so publishing an analysis of confirmed and probable cases is a valuable contribution. The manuscript is in general well-structured and mostly clear in writing. The inclusion of historical archives, Indonesian Met Service records, and social-media/news reports demonstrates effort and to some extent novelty. The climatological findings are potentially of interest for regional hazard assessment. At the same time, the authors should be very cautious in interpreting the rising trend historical vs. recent period as anything other than improved reporting. Overall, the manuscript is a good contribution to tornado climatology in Southeast Asia, but some contextual analysis is lacking (see below)
Specific Comments
1. The authors note that modern reports (2010–2024) dominate the database due to Internet/social-media proliferation. This implies strong observational bias (i.e., underreporting before the digital era). The spatial climatology likely reflects population density and media coverage rather than the true distribution of tornadoes. It might help to explicitly state how duplicate reports were merged (e.g., when one tornado generated multiple news reports). The use of “generative AI” (Gemini) to filter social media reports is useful, but a brief note on validation or potential biases of this method would strengthen confidence in the results. Lastly, ensure all non-English terms are clearly translated when first used.
2. The classification scheme (Table 1) follows Rauhala et al. (2012) and similar studies. One suggestion is to ensure the terms like “credible eyewitness” are well-defined. The wording in Table 1 could be tightened (e.g., “credible eyewitness who reported hearing a thunderous sound” should be “reported hearing thunder”). It might be worthwhile to note explicitly that without damage surveys, classification from media accounts is inherently uncertain, for example, some “probable” tornadoes may have been straight-line wind events. The authors should could consider cross-referencing the cases with meteorological data (e.g., radar, reanalysis).
3. Caution when interpreting a climate change signal in the occurrence of tornadoes. The spatial distribution shows a strong bias towards Java and very few reports from Maluku–Papua, which reflects both population and possibly data availability. This should be emphasised as a limitations. For example, in the Discussion section the authors could add that provinces with few reports may simply lack observers or press coverage. One further suggestions is to have a separate Discussion section.
4. The manuscript analyses two recent tornadoes (Rancaekek 2024 and Bogor 2018) using the EF-Scale and JEF-Scale. This shows that the same damage translates to a higher rating on the JEF-Scale (EF2 vs JEF3 for Rancaekek). The authors argue that differences in building practices cause this discrepancy, and suggest developing an Indonesian scale. However, two case studies are not enough to fully justify a new scale and the general conclusion.5. The Discussion mentions that most tornadoes occurred in the Nov–Mar season, aligning with the Austral summer monsoon. It also notes that effects of the MJO and ENSO are worth investigating. This is an important point as large-scale atmospheric modes strongly modulate convection over Indonesia. However, the manuscript does not analyse any of these factors quantitatively. A minimum improvement would be to cite previous studies that link convection in Indonesian to ENSO/MJO.
Technical Corrections
- line 26: the phrase “a tornado events” should be “tornado events.”
- lines 35-36: the term “basic climatology characteristics” should be “basic climatological characteristics.”
- line 37: replace “catalyst future tornado studies” with “catalyze future tornado studies.”
- line 47: “occuring” should be “occurring.”Citation: https://doi.org/10.5194/egusphere-2025-1857-RC1 - AC1: 'Reply on RC1', Irfans Firdaus, 19 Aug 2025
-
RC2: 'Comment on egusphere-2025-1857', Anonymous Referee #2, 11 Aug 2025
Overall: I’m always pleased to see additional information on tornadoes around the world. In the last 30 years, awareness has grown as to how widespread their occurrence is.
1. In the list of recently developed tornado databases, there is an additional one that should be included covering much of the former Soviet Union. Chernokulsky et al. (2020). Chernokulsky, A., and Coauthors, 2020: Tornadoes in Northern Eurasia: From the Middle Age to the Information Era. Mon. Wea. Rev., 148, 3081–3110, https://doi.org/10.1175/MWR-D-19-0251.1
2. The inclusion of some metadata about the quality of the report is critical. It dates back to the origin of the European Severe Weather Database. Any additional information, particularly about the quality of witnesses, would be helpful. The difference between the opinions of, say, the study’s authors upon seeing a tornado live compared to a child would be useful to know about.
3. The discussion that changes in frequency are likely due to non-meteorological factors is useful to include. Given the apparent relatively rare occurrence of tornadoes in Indonesia, it is unlikely that the attribution to meteorological changes will ever occur. Even in the United States, with more than 1000 tornadoes per year, it is difficult to pull out the non-meteorological effects.
4. I don’t see much of a need to develop a new damage scale. The sample size superimposed on the relatively rare nature of the events makes it hard to have much confidence in meaningful information. If it is possible to find relatively similar construction practices in other countries, it might be possible to get a significant sample, but I doubt it. Even in the United States, there are serious problems with ratings of tornadoes, e.g., Lyza et al. (2025) Lyza, A. W., H. E. Brooks, and M. J. Krocak, 2025: Where Have the EF5s Gone? A Closer Look at the “Drought” of the Most Violent Tornadoes in the United States. Bull. Amer. Meteor. Soc., https://doi.org/10.1175/BAMS-D-24-0066.1, in press.Citation: https://doi.org/10.5194/egusphere-2025-1857-RC2 - AC2: 'Reply on RC2', Irfans Firdaus, 19 Aug 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 543 | 95 | 16 | 654 | 24 | 12 | 27 |
- HTML: 543
- PDF: 95
- XML: 16
- Total: 654
- Supplement: 24
- BibTeX: 12
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1