the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Aug 2026
Spatiotemporal Distribution of Tropopause Folds over the Arctic during Spring Season: A 24-Year Analysis (2000–2023)
Abstract. Springtime tropopause folds facilitate stratosphere–troposphere exchange and may influence Arctic tropospheric composition, yet their long-term variability and dynamical background over the Arctic remain insufficiently characterized. Here, we use ERA5 reanalysis data and a three-dimensional labeling algorithm to identify tropopause folds over the Arctic during March–May from 2000 to 2023, with particular focus on four monitoring sites: Alert, Barrow, Villum, and Zeppelin. The results show spatial heterogeneity and depth dependence. Total and shallow folds occur over relatively broad Arctic regions, with high-frequency centers around Greenland, the Greenland–Barents Sea region, and the North Pacific side of the Arctic. Medium and deep folds are less frequent and are more strongly confined to the lower-latitude Arctic periphery. This spatial structure is dynamically consistent with the 700–250 hPa layer-mean maximum Eady growth rate, which is larger along the Arctic margins than over the central Arctic. Station-based composites further indicate that deeper folds are associated with stronger 250 hPa geopotential height contrasts, larger horizontal GPH gradient magnitude (GZ), and stronger wind speed characteristics. Temporally, total fold frequency shows interannual positive trends at all four stations during 2000–2023, with the strongest increases at Alert and Villum, mainly contributed by shallow folds. A monthly decline is also found from March to May, with May frequencies only about 20 %–30 % of March values, whereas diurnal variations are weak. These results provide a climatological framework for understanding Arctic springtime tropopause folds and their potential implications for Arctic tropospheric composition.
- Preprint
(22619 KB) - Metadata XML
-
Supplement
(27 KB) - BibTeX
- EndNote
Status: open (until 06 Nov 2026)
- RC1: 'Comment on egusphere-2026-3196', Anonymous Referee #1, 02 Sep 2026 reply
-
RC2: 'Comment on egusphere-2026-3196', Anonymous Referee #2, 23 Oct 2026
reply
Chen et al. conduct a study on tropopause folds in the springtime Arctic, focusing on four locations (Alert, Barrow, Villum, and Zeppelin) over the period 2000–2023. Through analysis of ERA5 data, they suggest that the fold frequency in the Arctic decreases with latitude (explained by increasing baroclinicity with decreasing latitude). Pertaining to the fold characteristics of the four locations, it is found that at Alert and Villum there is a statistically significant increase in fold frequency, and the authors argue that the variations in 250 hPa height, height gradient, and wind patterns can account for this change. The said variations can also explain why folds are most common in March and least common in May over these locations.
Tropopause folds over the Arctic are relatively less studied and are a worthwhile topic. However, in my evaluation, there is very little novelty and there are no significant results in this study. For instance, the abstract states, "Station-based composites further indicate that deeper folds are associated with stronger 250 hPa geopotential height contrasts, larger horizontal GPH gradient magnitude (Gz), and stronger wind speed characteristics." The same statement applies everywhere outside of the tropics. In addition, the text attempts to relate fold trends to trends in height, height gradient, and wind speed, but a comparison of Figure 8 and Table 2 reveals that the results are inconclusive (I explain this in the major comments below). The relationship between fold frequency trends and trends in height, height gradient, and wind speed is mainly descriptive, and no correlation is evaluated.
Aside from the above concerns, the purpose of selecting the four sites is not clear, as no observations from those sites are used in the analysis (aside from correlating them with ERA5 to suggest that ERA5 is reliable), if I understand correctly.
With this work as is, I do not see clear ways to revise this work such that it will present significant findings without a complete overhaul. Given the above mentioned issues, and some incomplete statistics in the work (such as statistical testing on some trends but not on others), after serious consideration I regret to recommend rejection of this manuscript.
Major comments
This work picks four locations (Alert, Barrow, Villum, and Zeppelin) in the Arctic and evaluates the fold occurrence and trends over these regions. The authors use surface station data at these locations to evaluate the reliability of the ERA5 data. However, all of the fold characteristics and GPH/wind values are upper-air data, so the correlation between station surface measurements and ERA5 data is not meaningful.
Lines 128–129: "In this study, the maximum vertical resolution of the ERA5 dataset officially provided by ECMWF on pressure levels (i.e., 37 levels) was used to identify the folds." This statement is not true, as ERA5 on 137 model levels is available.
Regarding the Škerlak et al. (2015) algorithm: in their methods, they deploy a "Lagrangian filter" (their Section 2.5) to filter out high-PV air that is diabatically generated. They do this to eliminate cases where high-PV air from the surface in polar regions merges with the stratosphere to create a fake fold. Can you explain whether this Lagrangian filter is implemented, and if not, what measures are taken to filter out false-positive detections?
Lines 187–195: There is a hotspot of total/shallow/medium TFs in central Greenland, but this is not noted in the discussion. It is the largest hotspot of medium TFs. If the authors choose to discuss hotspots, the one over central Greenland should be mentioned as well.
Lines 202–204: "The upper tail of the depth distribution also contracts poleward: the 90th and 95th percentiles are approximately 212.5 and 262.5 hPa, respectively, between 60 and 75◦ N, but decrease to about 187.5 and 237.5 hPa at 80◦ N and to roughly 162.5–170.8 and 195.8–197.5 hPa near 89–90◦ N." It seems the stated values do not match the plot; for instance, between 60 and 75° N, the 90th and 95th percentiles in the plot are roughly 260 and 340 hPa. The stated values for 90° N look off as well. Please check and revise.
Figures 9 and 10: Please perform statistical testing and hatch the statistically significant areas.
Table 2: Could you label the trends that are statistically significant, as you have done for Figure 8?
Lines 306–320 and Table 2: I believe the findings based on Table 2 are overstated and need to be revised. Overall, I think it is very hard to draw any significant conclusions from the numbers in Table 2. For instance, Alert exhibited a significant fold trend in Figure 8, but its Gz and wind speed trends are small. Yet in the discussion you state that "This contrast further suggests that changes in the horizontal GPH gradient, rather than changes in the absolute GPH alone, are more directly related to the dynamical environment controlling fold occurrence." This applies to Villum and not Alert.
Figure 12: What is the blue patch in (b) and (c)?
Figure 12: Especially in (d–f), the station text labels block the underlying contours, and we cannot see what Gz looks like at the stations. Perhaps change the station labels to different symbols and add a legend outside the plots.
Line 343: "and more regionally variable changes in wind speed." What does this mean?
Minor suggestions
Figures 6 and 7: If I understand correctly, the four plots in the left column are the same? If so, consider showing just one.
Figures 6 and 7: The image in Figure 6 has a title saying "Figure 4. Multi-year …"; the number is wrong. The same applies to Figure 7. I think it is better not to state the figure number in the title.
Table S1: "Annual correlation coefficients for temperature, wind components, and pressure at Barrow and Zeppelin." Please state the correlated data sources (reanalysis vs. surface measurements).
Table 1 and Lines 224–235: Can you clarify what is meant by "GPH differences" in this table? Does the plus/minus indicate the range of GPH at each station, or is it the largest difference in the map composite of GPH for the fold periods at each station? Because in this paragraph you mention spatial contrasts, I interpret it to be the latter, though I am not sure. Please clearly define this quantity.
Lines 236–245: As above, please clearly explain what the wind speed ranges mean (is it the range directly over the station or something else?).
Figure 9: I suggest individual color scales for each plot so that we can see the medium/deep ones clearly.
Line 289: "Trends in medium and deep folds are weaker and more spatially fragmented." To me, the medium fold trend pattern is very similar to those of the total and shallow folds, so I do not think one can say that it is more fragmented.
Figure 10: Please improve the legibility of the labels.
Citation: https://doi.org/10.5194/egusphere-2026-3196-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 181 | 145 | 33 | 359 | 60 | 41 | 38 |
- HTML: 181
- PDF: 145
- XML: 33
- Total: 359
- Supplement: 60
- BibTeX: 41
- EndNote: 38
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Please see the attached file.