the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Aug 2026
The Weather Extremes Archive (WEXA)
Abstract. The Weather Extremes Archive (WEXA) is a global, event-based dataset designed to systematically detect, track, and document persistent weather extremes. In its initial release, WEXA focuses on heat waves and cold spells identified from 2-m air temperature in ERA5 reanalysis data from 1979 to present. Cold and warm extremes are defined using seasonally varying 10th and 90th percentile thresholds derived from a fixed 1991–2020 climatological baseline. Periods of at least three consecutive extreme days are grouped into coherent spatiotemporal heat waves or cold spell events using three-dimensional connected component labelling. WEXA delivers a global catalogue of all detected individual heat wave and cold spell events, providing standardized diagnostics of their timing, duration, spatial extent, and intensity. The archive is openly accessible at https://doi.org/10.5281/zenodo.20550615 and https://wexa.jamstec.go.jp, permitting the exploration of event statistics and visualizations through an interactive analysis tool.
- Preprint
(2467 KB) - Metadata XML
-
Supplement
(40178 KB) - BibTeX
- EndNote
Status: open (until 31 Oct 2026)
- RC1: 'Comment on essd-2026-434', Ion Marinică, 26 Sep 2026 reply
-
RC2: 'Comment on essd-2026-434', Anonymous Referee #2, 26 Oct 2026
reply
The paper presents a global dataset of heatwave and cold spell spatiotemporal objects. The notion of defining spatiotemporal extreme event footprints is well-established in the literature, and the study does not present significant advances in this respect. Related to this, I appreciated the authors’ openness in recognising that they use a relatively simple existing algorithmic approach for defining the spatiotemporal objects, arguing that its simplicity makes it preferable to more advanced existing algorithms. I see the benefits of the study as chiefly being: presenting a dataset that saves the community the effort of applying publicly available algorithms to publicly available data; and providing a web interface for accessing and visualising the data, including regular updating. On balance, I do not find this sufficient to recommend further consideration in Earth System Science Data. I list below the major concerns that led me to this conclusion.
1. Unclear relevance for the scientific community. Many scientific users do not find the task of applying a clustering algorithm to a climate dataset particularly complex, as evidenced by the widespread use of spatiotemporal extreme event characterisation in the literature. While the dataset presented in this paper could be useful for some, I therefore see the benefit of the work as being chiefly restricted to non-specialist users with limited technical skills. The authors to some extent recognise this by citing decision-makers, media and the public as potential end-users.
2. Unclear value proposition for non-specialist users. Assuming that the paper indeed targets non-specialist users, and that this is considered suitable for ESSD, I still find the value proposition unconvincing. For these users, the web interface and dataset provide limited flexibility, as there is no evident way of tailoring the definition of the extreme events to the users’ individual needs in the online tool. The value of the dataset for non-specialist users is further limited by the fact that the pre-computed diagnostics are available from 2020 onward and the interactive analysis from 1979 onward. Users who are able to download the data and apply the algorithm to go further back in time probably do not need this dataset in the first place, and the users who could actually find it of benefit are left with a limited data selection.
3. Unintuitive web interface. As I discuss above, I find the web interface to be one of the most useful aspects of the dataset. However, its use should be clearly documented in the paper, and in the webpage itself. Right now, once a user accesses the web interface there is minimal guidance on how to use it. None of the visible tabs display a user manual or similar, and it is not always easy to guess what the different functions do. Take the example of “Duration (days)”. I would intuitively understand this to mean the duration of the extreme events in days, but then I see that the scale extends to 3000, so this is maybe not the case. If I then slide the indicator to “100”, I get a pop-up reading “3” (I intended to show a screenshot but am unable to include it in the online comment). So is this 100 days or 3 days? And if 100 days, what does it mean? There are also illegible labels for the years, as they overlap each other. Analogous considerations apply to various other aspects of the webpage. If the target audience for this is indeed decision-makers and the media, the useability of the webpage must be systematically improved.
4. Lack of separation between land and sea regions. The events in the database seem to extend without separation across land and sea regions. However, the magnitudes, characteristics and impacts of temperature extremes over land and sea differ drastically. In the introduction, the authors angle their contribution strongly towards understanding of extremes and assessment of their socio-economic impacts. Having continuous events extending across land and sea regions is suboptimal for both purposes.
5. Difficulty in interpreting events. Paraphrasing the authors’ own statement, it is meaningful to treat separate regions as part of the same event, as they are typically generated by a common large-scale atmospheric circulation feature and therefore likely have similar formation and dissipation processes. However, when one sees the Northwestern Pacific heatwave of 2021 extending across the North Atlantic, including several small disconnected areas, one is led to question the authors’ claim. There is a vast literature on this specific heatwave. Is there any evidence that atmospheric heatwaves over the North Atlantic were an integral part of the heatwave event? The same goes for some of what the authors call “mega-events”. There are events in the dataset that span across hemispheres, and persist on multi-season to interannual timescales. I struggle to see a justification for considering these as single events, regardless of whether one takes a physical or an impacts-based perspective. The only justification for considering such episodes as single events is a statistical one. The authors indeed seem to follow this line of reasoning, outlining that in a warmer climate there can be several concurrent heatwaves that according to their algorithm are viewed as a single event. The authors to some extent shift the onus of how to deal with such events to the users. I see two major issues with this. First, I consider the dataset as mainly targeting non-specialist users, and I would argue that leaving them with having to deal with "mega-events" further decreases the value of the dataset for these users. Second, the framing in the introduction is of the dataset as a tool to advance understanding of temperature extremes and assess their impacts. Taking a purely statistical view and claiming that it is acceptable to class globally-spanning disconnected extreme temperature episodes in a single one because they form a connected structure goes directly against these considerations.
Other minor comments:
Some of the figures could be enhanced or merged. For example, Figs. 4 and 5 would fit naturally together, as would Figs. 7 and 8. Alternatively the latter two figures could show examples from two different heatwaves.
Table 1. I would suggest having a joint table for both heatwaves and cold spells.
l. 37 Missing space: “(ETCCDI),which”
ll. 263-264 “To find this event, users can select, for instance, the 50 strongest heat wave events that have affected Canada in June 2021.” Do the authors actually mean this, i.e. that there are more than 50 distinct heatwave events affecting Canada in June 2021 and that the visualisation selects the 50 most intense?
Citation: https://doi.org/10.5194/essd-2026-434-RC2
Data sets
Weather Extremes Archive Patrick Martineau, Ingo Richter, Ayako Yamamoto, Masami Nonaka, and Hidehiro Fujio https://doi.org/10.5281/zenodo.20550615
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 399 | 145 | 149 | 693 | 138 | 186 | 147 |
- HTML: 399
- PDF: 145
- XML: 149
- Total: 693
- Supplement: 138
- BibTeX: 186
- EndNote: 147
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
September 23, 2026
Dear friends,
Thank you for the opportunity to review the paper
The Weather Extremes Archive (WEXA)
Patrick Martineau, Hidehiro Fujio, Ingo Richter, Ayako Yamamoto, and Masami Nonaka
https://doi.org/10.5281/zenodo.20550615
This is an excellent paper, and its publication is welcome, as it allows the scientific community to gain an overview of what is currently encompassed by the term "extreme weather phenomena."
Regarding this term—"extreme weather phenomena," currently defined as "weather events that far exceed the usual climatic limits of a specific region; they are unusual, severe, and can endanger lives, the environment, or property"—I would like to offer the following comment:
Based on the intensity scale of a specific type of weather phenomenon, there are only two extremes for each type: "the most intense recorded to date in the entire data archive" and "the least intense recorded to date in the entire data archive." For instance, regarding temperature at a specific location on Earth, these would be the highest and lowest temperatures recorded across the entire data archive. Naturally, these limits provide an overview of the variability range of weather phenomena within a specific region or across the globe. Thus, extreme weather events are phenomena that occur infrequently (rare or exceptionally rare events—such as "Noah's Flood," which may not be merely a figment of the imagination or a legend).
Many media outlets speak of "extreme weather events" almost daily, creating the impression that they occur frequently or even on a daily basis. When phenomena occur frequently or daily, they are obviously not extreme weather events but rather ordinary or "normal" weather phenomena for the region in question.
Therefore, we need to reach a consensus on a precise definition for these phenomena—a step essential for a proper understanding of the situation and for eliminating the ambiguity inherent in the current definition (e.g., how severe? how intense? how long did it last? what damage did it cause?).
Instead of analyzing the synoptic, dynamic (and potentially geological or astronomical) causes behind such exceptionally intense phenomena, almost everyone is content with the simple explanation that they are caused by "climate change."
"Climate change" is a term fraught with ambiguity, stemming initially from imprecise expressions such as "night turns into day" (and vice versa), "the cold season gives way to the warm season" (and vice versa), "the seasons change," and so on.
Yet, these are not changes; this is simply how things have always been and how they will always be for as long as the Earth exists.
It is more accurate to classify these processes under "climate variability"—a form of variability characterized by a crucial property: it sustains life on Earth and operates within broad yet finite limits.
Many people (including scientists) fail to distinguish between two fundamentally different yet seemingly similar processes: "variability" and "change."
To clarify certain long-standing ambiguities in climate science, I must point out that Earth's climate involves no processes of "change," but rather only... variability.
Processes of change are characteristic of ephemeral entities: humans, living beings, plants, etc.
Childhood, youth, maturity, and old age are stages of processes of change.
The clearest example of change is death.
Any entity that evolves through change disappears; it is, therefore, ephemeral.
Variability signifies an evolution within certain limits—maintaining specific conditions within a broad range—and is characteristic of eternal entities.
For example, Earth's climate maintains the conditions necessary for life within a broad range, even if not everywhere.
If processes of change had occurred in Earth's climate, life would have vanished long ago, and not even a single virus would exist today.
The existence of life on Earth is clear proof that there are no processes of change.
I have published the mathematical proof demonstrating that "there are no processes of change in the climate of Earth, nor in that of any planet or star in the Universe."
Fig. 1. The distinction between variability (top) and change (bottom). (Marinică et al., 2024)
At the same time, I recall that the principle of variability (not necessarily climatic) is "what is, has been before and will be again"—first formulated by Ecclesiastes (about 2,400 years ago)—whereas the principle of change is "what is, has never been before and will never be again." The principle of change is the logical negation of the principle of variability, and vice versa. They cannot coexist within the same entity. There is no third possibility. It is a mistake for some to claim that a single entity can exhibit both variability and change. Defining "change" merely as a specific, statistically significant value of meteorological parameters constitutes a misuse of statistics. By this same logic—and through such statistical misuse—one could "prove" the existence of extraterrestrials among us, and so forth.
I have no wish to offend or contradict anyone. However, it is absolutely essential that we develop precise, unambiguous terminology to guide future research—research that must rectify the vast array of errors accumulated to date, not only in "climate science" but in other disciplines as well. These fields often rush to blame humanity for all the ills on Earth, thereby providing ill-intentioned individuals with a pretext to argue for the "necessity of reducing the human population."
I commend your work and wish you great success in all your endeavors; may integrity light your path.
With deep respect and high regard,
Ion Marinică