Variability of ice supersaturated regions at flight altitudes: evaluation of ERA5 reanalysis using IAGOS in situ measurements

Hildebrandt, Katarina Grubbe; Castino, Federica; Meijer, Vincent; Yin, Feijia

doi:10.5194/egusphere-2025-3048

Preprints

https://doi.org/10.5194/egusphere-2025-3048

Preprints

22 Aug 2025

| 22 Aug 2025

Variability of ice supersaturated regions at flight altitudes: evaluation of ERA5 reanalysis using IAGOS in situ measurements

Katarina Grubbe Hildebrandt, Federica Castino, Vincent Meijer, and Feijia Yin

Abstract. Contrail cirrus is one of the largest contributors to aviation's radiative forcing, which arises from long-lived persistent contrails. Avoiding persistent contrail formation has been suggested as a measure to reduce the climate impact of aviation, requiring accurate forecasts of ice supersaturated conditions, i.e. where the relative humidity over ice (RHi) exceeds 100 %. Numerical weather prediction (NWP) models often underestimate or do not account for ice supersaturation. This study evaluates ice supersaturated regions (ISSRs) in the ECMWF ERA5 reanalysis using In-service Aircraft for a Global Observing System (IAGOS) measurements over tropical and extratropical regions in the upper troposphere and lower stratosphere for the period 2011–2022. It considers cloudy and clear-sky conditions and how North Atlantic weather patterns affect the ERA5 ability to predict ISSRs. ERA5 generally underestimates ISSR occurrence due to a dry bias in RHi; the equitable threat score (ETS) is 0.2–0.4, indicating a weak to mediocre relationship with IAGOS. Clear-sky conditions result in an ETS of 0.05-0.18 and generally below 0.1 in cloudy conditions, indicating an almost random relationship. The latter is the result of the saturation adjustment used by the NWP model underlying the reanalysis. North Atlantic winter weather patterns appear to affect the ability of ERA5 to predict ISSRs, particularly along eastbound routes. This may result from varying ISSR distributions relative to the jet stream. North Atlantic summer weather patterns show little impact due to weaker teleconnection patterns. Overall, the underestimation of ISSRs in ERA5 is most critical in the upper troposphere, where their occurrence is highest.

Received: 26 Jun 2025 – Discussion started: 22 Aug 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 3793 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (3793 KB)

Supplement (1523 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

13 May 2026

Variability of ice supersaturated regions at flight altitudes: evaluation of ERA5 reanalysis using IAGOS in situ measurements

Katarina Grubbe Hildebrandt, Federica Castino, Vincent Meijer, and Feijia Yin

Atmos. Chem. Phys., 26, 6449–6470, https://doi.org/10.5194/acp-26-6449-2026,https://doi.org/10.5194/acp-26-6449-2026, 2026

Short summary

Katarina Grubbe Hildebrandt, Federica Castino, Vincent Meijer, and Feijia Yin

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2025-3048', Anonymous Referee #1, 10 Sep 2025

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3048/egusphere-2025-3048-RC1-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2025-3048-RC1
- AC1: 'Reply on RC1', Katarina Grubbe Hildebrandt, 15 Dec 2025
  
  We would like to thank reviewer #1 for their constructive comments on the manuscript. All the comments have been taken into account. We provide a point-to-point response in the attached document.
  
  Citation: https://doi.org/10.5194/egusphere-2025-3048-AC1
RC2:
'Comment on egusphere-2025-3048', Anonymous Referee #2, 29 Sep 2025

Review of „Variability of ice supersaturated regions at flight altitudes: evaluation of ERA5 reanalysis using IAGOS in situ measurements by Hildebrandt et al., 2025
The present study by Hildebrandt et al. uses the IAGOS water vapour data as a reference to investigate the ability of ERA5 to predict ice-supersaturated regions. The manuscript is generally good, with appropriate methodology for the comparisons of ERA5 and IAGOS. The authors provide an update and a more comprehensive evaluation, both temporally and spatially, compared to previous publications. The results found are described and compared with the results of the earlier studies. A deficit is that there is often only confirmation of the earlier findings without any real new insights being gained. Although this update is certainly of relevance, the manuscript would benefit from being shortened and centered on the relevant results. Thus, I would overall recommend publication but only after major revisions.

Major Comments:

This study demonstrates a significant investment of time and effort, application of recognized methods, and extensive knowledge of IAGOS and ERA5, making the foundation of the study solid. However, the analysis itself is overly descriptive. Often minor differences between the old and new results are shown, regardless of their significance. Therefore, I recommend shortening the text and focusing more on relevant aspects.
I don't think it makes much sense to calculate the ETS for many different situations when the poor agreement between the ISSRs found by ERA5 and IAGOS is due to the adjustment scheme under cloudy conditions. I would recommend repeating this part of the analysis by using a threshold of ERA5 RHi of 90 % or 95 % to determine ISSRs (at least under cloudy conditions).
I also have to question the calculation of ETS in the context of discussing different weather patterns. Before discussing the differences in ETS, it would be helpful to examine the fraction of ISSRs for these weather patterns as observed by IAGOS, and determine if there are any differences among them.
It is an interesting feature that in South Asia you find the minimum temperature 60 hPa below the thermal tropopause. The structure of the lapse rate and cold point tropopause in the tropics can be quite complicated but, normally, if there are multiple tropopauses the cold point tropopause is above the WMO thermal tropopause (also in Muhsin et al., 2018?). I suggest performing a literature review to learn more about this. See for example: https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/2011JD016637

Line 334 and 516: “Therefore, the moist bias appears to arise as a result of the saturation adjustment, which cannot take into account that the (wet) mode of the RHi PDF can be subsaturated in some regions.” In my opinion this should not lead to a moist bias but to an alignment, because the (false) saturation adjustment would just not happen. Please explain.

Specific comments:

Line 123: In IAGOS RHL > 100 % is flagged invalid.
Line 131: Why does sampling every minute result in 2.5% of all IAGOS measurements?
Line 182: You find a higher frequency of indeterminate and cloudy conditions compared to IAGOS because of the limitations of the BCP (see Petzold et al., 2017).
Line 202: Yes, physically, relative humidity partially depends on temperature, but the ICH sensor from IAGOS measures RHL, so, this is the water variable which is independent from the temperature.
Line 205: Please specify how big the ERA 5 cold bias in the extratropics is.
Line 212: “extratropical seasonal cycle“: in the Northern Hemisphere
Line 250: it does not provide good quality results in dry conditions in the lower stratosphere due to the loss of sensitivity as a result of the adiabatic compression effect (Konjari et al., 2025)
Line 302: Please rephrase: “The tropopause layer is not completely dry (Petzold et al., 2020; Reutter et al., 2020)”
Petzold et al. state: “a tropopause layer characterized by mean RHice of 60% almost independent of the season…”
Reutter et al. state: “In the tropopause layer, still a significant amount of the data is exceeding values of RHi > 100 %, both in the in situ data as well as in the ERA data set.”

Line 314: “For clear-sky conditions, there is a smaller probability of low RHi, with increased probability of higher RHi, and a more equal probability across all observed RHi values.” I don’t understand this sentence.
Line 341: typo: “ERA5 may predict less ISSRs compared to ERA5”
Line 364: Delete the word “but” in: “On the other hand, Reutter et al. (2020) found a maximum of 40% using IAGOS when considering all seasons, but with the maximum also occurring 30 hPa below the tropopause and a reduction of the ISSR fraction to zero in the lower stratosphere.
Line 367: “ISSR fraction is highly sensitive to…”: Maybe better without the word “highly”
Line 402: Could the reason for higher ETS in North America be, that there are more data assimilated for ERA5?
Line 407: Please rephrase: “the maximum difference between seasons is approximately 10%, at most. In North America, we find that the variation can be up to 20%.”
Line 410: “tend to find”

Citation: https://doi.org/10.5194/egusphere-2025-3048-RC2
- AC2: 'Reply on RC2', Katarina Grubbe Hildebrandt, 15 Dec 2025
  
  We would like to thank reviewer #2 for their constructive comments on the manuscript. All the comments have been taken into account. We provide a point-to-point response in the attached document.
  
  Citation: https://doi.org/10.5194/egusphere-2025-3048-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2025-3048', Anonymous Referee #1, 10 Sep 2025

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3048/egusphere-2025-3048-RC1-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2025-3048-RC1
- AC1: 'Reply on RC1', Katarina Grubbe Hildebrandt, 15 Dec 2025
  
  We would like to thank reviewer #1 for their constructive comments on the manuscript. All the comments have been taken into account. We provide a point-to-point response in the attached document.
  
  Citation: https://doi.org/10.5194/egusphere-2025-3048-AC1
RC2:
'Comment on egusphere-2025-3048', Anonymous Referee #2, 29 Sep 2025

Review of „Variability of ice supersaturated regions at flight altitudes: evaluation of ERA5 reanalysis using IAGOS in situ measurements by Hildebrandt et al., 2025
The present study by Hildebrandt et al. uses the IAGOS water vapour data as a reference to investigate the ability of ERA5 to predict ice-supersaturated regions. The manuscript is generally good, with appropriate methodology for the comparisons of ERA5 and IAGOS. The authors provide an update and a more comprehensive evaluation, both temporally and spatially, compared to previous publications. The results found are described and compared with the results of the earlier studies. A deficit is that there is often only confirmation of the earlier findings without any real new insights being gained. Although this update is certainly of relevance, the manuscript would benefit from being shortened and centered on the relevant results. Thus, I would overall recommend publication but only after major revisions.

Major Comments:

This study demonstrates a significant investment of time and effort, application of recognized methods, and extensive knowledge of IAGOS and ERA5, making the foundation of the study solid. However, the analysis itself is overly descriptive. Often minor differences between the old and new results are shown, regardless of their significance. Therefore, I recommend shortening the text and focusing more on relevant aspects.
I don't think it makes much sense to calculate the ETS for many different situations when the poor agreement between the ISSRs found by ERA5 and IAGOS is due to the adjustment scheme under cloudy conditions. I would recommend repeating this part of the analysis by using a threshold of ERA5 RHi of 90 % or 95 % to determine ISSRs (at least under cloudy conditions).
I also have to question the calculation of ETS in the context of discussing different weather patterns. Before discussing the differences in ETS, it would be helpful to examine the fraction of ISSRs for these weather patterns as observed by IAGOS, and determine if there are any differences among them.
It is an interesting feature that in South Asia you find the minimum temperature 60 hPa below the thermal tropopause. The structure of the lapse rate and cold point tropopause in the tropics can be quite complicated but, normally, if there are multiple tropopauses the cold point tropopause is above the WMO thermal tropopause (also in Muhsin et al., 2018?). I suggest performing a literature review to learn more about this. See for example: https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/2011JD016637

Line 334 and 516: “Therefore, the moist bias appears to arise as a result of the saturation adjustment, which cannot take into account that the (wet) mode of the RHi PDF can be subsaturated in some regions.” In my opinion this should not lead to a moist bias but to an alignment, because the (false) saturation adjustment would just not happen. Please explain.

Specific comments:

Line 123: In IAGOS RHL > 100 % is flagged invalid.
Line 131: Why does sampling every minute result in 2.5% of all IAGOS measurements?
Line 182: You find a higher frequency of indeterminate and cloudy conditions compared to IAGOS because of the limitations of the BCP (see Petzold et al., 2017).
Line 202: Yes, physically, relative humidity partially depends on temperature, but the ICH sensor from IAGOS measures RHL, so, this is the water variable which is independent from the temperature.
Line 205: Please specify how big the ERA 5 cold bias in the extratropics is.
Line 212: “extratropical seasonal cycle“: in the Northern Hemisphere
Line 250: it does not provide good quality results in dry conditions in the lower stratosphere due to the loss of sensitivity as a result of the adiabatic compression effect (Konjari et al., 2025)
Line 302: Please rephrase: “The tropopause layer is not completely dry (Petzold et al., 2020; Reutter et al., 2020)”
Petzold et al. state: “a tropopause layer characterized by mean RHice of 60% almost independent of the season…”
Reutter et al. state: “In the tropopause layer, still a significant amount of the data is exceeding values of RHi > 100 %, both in the in situ data as well as in the ERA data set.”

Line 314: “For clear-sky conditions, there is a smaller probability of low RHi, with increased probability of higher RHi, and a more equal probability across all observed RHi values.” I don’t understand this sentence.
Line 341: typo: “ERA5 may predict less ISSRs compared to ERA5”
Line 364: Delete the word “but” in: “On the other hand, Reutter et al. (2020) found a maximum of 40% using IAGOS when considering all seasons, but with the maximum also occurring 30 hPa below the tropopause and a reduction of the ISSR fraction to zero in the lower stratosphere.
Line 367: “ISSR fraction is highly sensitive to…”: Maybe better without the word “highly”
Line 402: Could the reason for higher ETS in North America be, that there are more data assimilated for ERA5?
Line 407: Please rephrase: “the maximum difference between seasons is approximately 10%, at most. In North America, we find that the variation can be up to 20%.”
Line 410: “tend to find”

Citation: https://doi.org/10.5194/egusphere-2025-3048-RC2
- AC2: 'Reply on RC2', Katarina Grubbe Hildebrandt, 15 Dec 2025
  
  We would like to thank reviewer #2 for their constructive comments on the manuscript. All the comments have been taken into account. We provide a point-to-point response in the attached document.
  
  Citation: https://doi.org/10.5194/egusphere-2025-3048-AC2

Peer review completion

AR – Author's response | RR – Referee report | ED – Editor decision | EF – Editorial file upload

AR by Katarina Grubbe Hildebrandt on behalf of the Authors (09 Jan 2026) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (25 Feb 2026) by Franziska Aemisegger

RR by Anonymous Referee #1 (09 Mar 2026)

RR by Anonymous Referee #2 (15 Mar 2026)

Suggestions for revision or reasons for rejection

The revised version of the study by Hildebrandt et al. has improved significantly. It is now shorter and more focused on new results. My questions from the first review have been answered and revised.
However, there are some minor issues in the new text passages that need to be addressed. Therefore, I recommend publication after minor revisions.

Major comments
Studying the equitable threat score (ETS) for different RHi threshold of ERA5 is a major improvement. It clearly shows that lowering the RHi threshold can increase the ability of ERA5 to predict ISSRs. However, I get the impression that the results (in which the ETS is usually quite low), are being over-interpreted to some extent.
As an example, I refer to line 337: “In North America and Europe, the best ETS at 30 hPa above tropopause was found at an RHi threshold of 75%; perhaps lowering the threshold even more could allow for more improvements.” I doubt that there is a physical explanation for the best ETS at a threshold of 75% (or even lower). The dry bias in ERA5 is not large enough to justify this threshold. I rather believe that this result (with also quite low values) has no physical meaning but is only found because of the very low number of ISSR occurrences in the LS. I would recommend not using thresholds below 85% unless you can explain why using lower thresholds makes physical sense.

I like the detailed investigation of the “sensitivity to definition of cloudy and clear-sky conditions”. This is very helpful to get an understanding of the differences between the different possibilities to distinguish between cloud free and cloudy conditions.
However, the explanation why the use of IAGOS ice crystal number concentration and not ERA5 CIWC should be preferred doesn’t convince me. The ice crystal number concentration from IAGOS BCP is not designed to measure subvisible clouds and, hence, known to miss the very thin clouds; ERA5 CIWC on the other hand is a very sensitive tool to even detect few and small cloud particles.
Starting at line 44 in the supplement it is stated: “[…] the CIWC thresholds established by Wang et al. (2025) and Petzold et al. (2025) tends to classify a measurement as in-cloud if there is any sign of the presence of ice clouds. Hence, if we want to observe ice supersaturation under clear-sky and cloudy conditions for contrail avoidance purposes, we recommend to make use of the ice crystal number concentration in IAGOS and the cloud cover in ERA5.”
In my opinion it is good to classify a measurement as in-cloud, if there is any sign of the presence of ice clouds. Do you mean that for contrail avoidance purposes it is not necessary to distinguish between clear-sky and subvisible clouds because in both cases aircraft flying in these regions could cause contrail cirrus with a warming effect? And that therefore the ice crystal number concentration from IAGOS BCP is sufficient? Then, please phrase it like this (line 429 and in the supplement). Incidentally, including regions with sub-visible cirrus would also be possible with CIWC by using a higher threshold.

Specific comments:
Line 39: “ MOZAIC […] which is currently part of IAGOS”: MOZAIC is the predecessor of IAGOS.
Line 85: “ extensive geographical and temporal area.”: extensive geographical and temporal range.
Line 93: I must apologize – I overlooked that during my initial review: Since you use the ice crystal number concentration a short description of the BCP should be given in the Data and methodology section.
Line 96: Please check the start date of CARIBIC
Line 116: typo: show => shown
Line 121: We considered….
Line 132: since the quality flag of RHi wasn’t well derived at the time of data collection.
Line 146: Rephrase: “Since the ARCO ERA5 dataset does not provide RHi, it is derived from specific humidity, temperature and the saturation water vapour pressure over ice are used to calculate RHi.”
Line 189: It would be good to include a sentence explaining why you are investigating the same seasons in tropical regions as in extratropical regions (same for RHi).
Line 192: S10? Please check all references to the figures in the supplement.
Line 204: “Less than 5% of our IAGOS measurements show ISS above this temperature threshold.” I recommend to delete this sentence.
Line 219: “Based on the correlation between RHi and cloud ice particles, the lower RHi could be related to little to no cloudy conditions in DJF (see Fig. S13).” I recommend deleting this sentence. There is no doubt that these two variables are correlated, and besides, the relationship works the other way round: high humidity leads to cloud formation.
Line 222: typo: as seen in as seen in
Line 226: typo: it could due to => it could be due to
Line 312: typo: shows => show
Line 329: “This could be due to lower mean RHi values and thus ISSRs in warmer months…” ISSRs are not necessarily correlated to the mean RHi values but rather to the high percentiles. I recommend deleting “mean” and inserting the word “less” before ISSRs.
Line 338: see major comments
Line 340: typo: than => that
Line 341 - 345: I don’t understand why cloudy conditions should be a limiting factor when decreasing the ERA5 RHi threshold to improve its ability to predict ISSRs. Wouldn't that be exactly the situation where decreasing the threshold would “repair” the saturation adjustment?
Line 380: Please clarify. Does the extratropical cold bias of 0.5 K in ERA5 compared to IAGOS refer to the upper troposphere or to an average across all altitudes?
Line 382: I suggest to add behind ERA5 “compared to IAGOS of 0.75- 1 K”.
Line 409: “In instances where the RHi is overestimated by ERA5, such as in South Asia for season JJA, the ISSR fraction is overestimated by ERA5.” I suggest to slightly alter the sentence: As expected, when the RHi is overestimated by ERA5, as it is in South Asia during the JJA season, the ISSR fraction is overestimated by ERA5 as well.
Line 432: see major comments
Line 438: delete: “which are weaker.”
Line 458: Maybe better: “For future work, it would be helpful to increase...”
Line 462: You could mention the saturation adjustment.
Line 502: same as above
Line 513: Why? Shouldn’t it be possible to decrease the RHi threshold for ISS (as long as you have ERA5 RHi).
Line 536: I don’t understand what you mean with “same region”.
Line 540: typo to => the
Line 561: delete: in the upper troposphere

Hide

ED: Publish subject to technical corrections (29 Mar 2026) by Franziska Aemisegger

AR by Katarina Grubbe Hildebrandt on behalf of the Authors (02 Apr 2026) Author's response Manuscript

Journal article(s) based on this preprint

13 May 2026

Variability of ice supersaturated regions at flight altitudes: evaluation of ERA5 reanalysis using IAGOS in situ measurements

Katarina Grubbe Hildebrandt, Federica Castino, Vincent Meijer, and Feijia Yin

Atmos. Chem. Phys., 26, 6449–6470, https://doi.org/10.5194/acp-26-6449-2026,https://doi.org/10.5194/acp-26-6449-2026, 2026

Short summary

Katarina Grubbe Hildebrandt, Federica Castino, Vincent Meijer, and Feijia Yin

Supplement

https://doi.org/10.5194/egusphere-2025-3048-supplement

Katarina Grubbe Hildebrandt, Federica Castino, Vincent Meijer, and Feijia Yin

Viewed

Total article views: 14,865 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
13,465	1,162	238	14,865	375	253	208

HTML: 13,465
PDF: 1,162
XML: 238
Total: 14,865
Supplement: 375
BibTeX: 253
EndNote: 208

Views and downloads (calculated since 22 Aug 2025)

Month	HTML	PDF	XML	Total
Aug 2025	5,003	75	15	5,093
Sep 2025	6,485	145	24	6,654
Oct 2025	305	110	30	445
Nov 2025	305	100	30	435
Dec 2025	432	149	70	651
Jan 2026	307	189	15	511
Feb 2026	207	126	26	359
Mar 2026	330	199	24	553
Apr 2026	77	59	2	138
May 2026	14	10	2	26

Cumulative views and downloads (calculated since 22 Aug 2025)

Month	HTML	PDF	XML	Total
Aug 2025	5,003	75	15	5,093
Sep 2025	6,485	145	24	6,654
Oct 2025	305	110	30	445
Nov 2025	305	100	30	435
Dec 2025	432	149	70	651
Jan 2026	307	189	15	511
Feb 2026	207	126	26	359
Mar 2026	330	199	24	553
Apr 2026	77	59	2	138
May 2026	14	10	2	26

Viewed (geographical distribution)

Total article views: 14,873 (including HTML, PDF, and XML) Thereof 14,873 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 23 May 2026

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (3793 KB)
Metadata XML

Short summary

We evaluate the regional and seasonal variability in the prediction of ice supersaturated region (ISSRS) in the ERA5 reanalysis using in situ measurements. ERA5 shows better ability to predict ISSRs in the extratropics, compared to the tropics, and in colder seasons, such as extratropical winter. While ERA5 generally underestimates the ISSR occurrence, we find an overestimation in tropical regions in seasons associated larger weather variability, such as South Asia in June, July and August.


Total:	0
HTML:	0
PDF:	0
XML:	0