Upper-atmosphere responses to the 2022 Hunga Tonga–Hunga Ha′apai volcanic eruption via acoustic gravity waves and air–sea interaction

Li, Qinzeng; Xu, Jiyao; Gusman, Aditya Riadi; Liu, Hanli; Yuan, Wei; Liu, Weijun; Zhu, Yajun; Liu, Xiao

doi:https://doi.org/10.5194/acp-24-8343-2024

Articles | Volume 24, issue 14

https://doi.org/10.5194/acp-24-8343-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/acp-24-8343-2024

© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 24, issue 14

Research article

|

25 Jul 2024

Research article |

| 25 Jul 2024

Upper-atmosphere responses to the 2022 Hunga Tonga–Hunga Ha′apai volcanic eruption via acoustic gravity waves and air–sea interaction

Qinzeng Li, Jiyao Xu, Aditya Riadi Gusman, Hanli Liu, Wei Yuan, Weijun Liu, Yajun Zhu, and Xiao Liu

Download

Final revised paper (published on 25 Jul 2024)
Preprint (discussion started on 27 Oct 2023)

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-2429', Anonymous Referee #1, 21 Nov 2023
In the current manuscript, Li et al. have used the very high temporal and spatial resolution of the ground-based OH and OI airglow network to look at the upper atmosphere response to the HTHH ( Hunga Tonga-Hunga Ha’apai) Volcanic Eruption occurred in 2022, which has been an active and important research topic after the eruption. OH airglow (? In the Figure 1) measurements have detected five group atmospheric waves in the ground-based airglow network (the first wave packet is observed after 8 hours). Then they have obtained the phase speed and amplitude over 10 hours period (?). The wave front can be also observed by OI airglow at higher altitude ~250 km. The authors also noticed that short period of sudden surface pressure changes over Xinlong station, which were caused by the pressure disturbances that spread out in the form of Lamb waves due to HTHH eruption. The authors also carried out tsunami model simulations using two sources (one is the localised source, termed TTIVE, the other considers the atmospheric pressure changes, termed as TTAPW) to investigate the propagating speed to China due to different tsunami sources used in the model. Then they discussed the air-sea interactive process on the upper atmospheric responses to HTHH eruption. They also did some related important diagnosis about the vertical wavelength and ray path of the waves etc. Finally the provided a nice figure describing the dynamic coupling process between air and sea via acoustic gravity waves. Overall, the paper has a clear structure and the motivation is clear. This study is important and interesting.
However, sometimes it is unclear to me because the figures are not clear enough or some descriptions/explanations are vague (see the detailed comments below). It requires some clarification.
I have two major concerns: One is the tsunami model simulations by assuming two sources. I can not see any information about the model. Is that is a global model or regional model? What are the necessary/important input required to model tsunami? How is the model performances to represent the tsunami due to HTHH volcanic eruption?
The other is about the inconsistent use of meteorological data product when the authors did some diagnosis (for example vertical wave numbers, lines 120-125). The authors use temperature profiles from SABER (20-100km) and the wind from both ERA5 and HWM-14. I know ERA5 is only up to 85 km and the authors would like the profile of vertical wave number up to 100 km. Which altitude range use ERA5, then HWM-14. How realistic is the HWM-14 modelled wind changes (since this model is an empirical model which only uses geomagnetic Ap index) due to the HTHH volcanic eruption? What will be the difference when you use both ERA5 wind and temperature profiles below some certain altitude compared with the current result?
In the Introduction, the authors have introduced HTHH volcanic eruption triggered broad spectrum atmospheric disturbances including different waves (Lines 41-44), but the second paragraph mainly focuses on Lamb waves. Why are other waves ignored?

Specified comments:

Line 54, it reads weird “from GNSS TEC analysis”. Better to change what “analysis” method used based on “GNSS TEC”. The acronym “GNSS TEC” should be described when they first appear in the text.

Lines 58-60. Can you provide some explanation about the “atmospheric pressure wave”?

Line 60, AGWs is defined in the Abstract for the first time but not in the main text. Better to define it here.

Line 61, please delete “the height of”.

Line 64, the authors need to make clear “that arrived before the tsunami”, for example when/where Tsunami occurred, how many hours earlier etc.

Line 65. Again, it is very vague “sea wave and GW almost simultaneously in Chile”. I guess the authors still talk about tsunami, but should make it clear if this is the same Tsunami as mentioned in Line 64.

Line 66, though readers know 3D is three dimensional, better to add this after 3D.

Lines 69-70, I am confused with “AGWs on the mesopause airglow radiation”.

Line 73. What is “convention tsunami”?

Line 75, What is “this typical type tsunami”?

Lines 58-59, 77-78. Again, It is not clear why “tsunamis induced by the atmospheric pressure wave (TIAPW)” is more important than “tsunamis directly induced by the 2022 Tonga volcano eruption (TITVE)”.

Line 83, change “(air-water-air-coupling process)” to “through air-water-air-coupling process”?

Line 86, why “Double layer airglow network” since the authors mentioned “multi-layer” in Line 87?

Line 93, please add the latitude/longitude information about “Xinglong Station”.

Line 94-95. Are all the airglow (OH, OI, 557nm) having the same resolution as “The temporal resolution is 1 min and the spatial 95 resolution is 1 km”?

Line 95. Can you add some reference for the “standard star map”? Or some reference how the calibration works “with the help of standard star map” here.

Line 96, Again, some reference is needed for “removed by differential method”.

Sections 2.2, the model description is not clear at all! Do you use WACCM-X model which has been mentioned several times in the whole paper? This part just asks the readers to read two cited papers and has never mentioned the details what the conditions to be able to simulate the tsunami after HTHH. How readers know the model results robust and reliable? Can you explain why the second type of tsunami sources is better or more realistic than the first localised source? Readers can not judge it because there is no detailed information about the “Tsunami simulation model”.

Lines 110-111, can you describe the variables in equations 1) and 2)?

Line 112. It is weird that “There is no real-time temperature data available in this study”.

Lines 112-117, I am confused why temperature is mentioned here, which should be moved after Lines 124-125.

Line 123. Why use the “HWM-14” model? Is this model suitable to study the HTHH?

Lines 131-149 and Figure 1. The description of Figure 1 is too general. Can you make a detailed list of the airglow (which one? Lines 133-135 only mentioned OH airglow)? The figure is very hard to follow, I have to zoom it 4 times to look carefully but it is still unclear for the wave packet in each sub-figures. Therefore it is hard to judge.

Line 158, It looks to me that the reference using the subscription 4 should not the 4^th reference in the list.

Line 159, please add some references here for the surface pressure changes.

Lines 159-161. It reads the logical is not clear. As mentioned Lamb wave is almost non dispersive which has purely horizontal motion. So the second sentence to Describe Figure 3 of atmospheric waves from the ionosphere to surface will cause confusion. It would be better to move or delete the first sentence.

Line 189. This is not pressure profile, it is time series of surface pressure.

Line 192, please change “position” to location.

Lines 204-205. “are very consistent with the simulated tsunamis near the coast”? Where have you shown the simulation result?

Line 210. Can you describe Figure 5 in detail? Just saying snapshots is not enough since the contour unit is in cm.

Lines 216-218. What the evidence (where the results, assuming Figure 5 but it is vague in its description) to support these?

Lines 241-243. How do you estimate the wavelength? Which one (TIAPW, TITVE) or both are consistent with the derived from the airglow network which was mentioned in Lines 204-205?

Line 255-256. If wind comes from ERA5 (which has hourly product), why use SABER temperature?

Figure 8c. Why the ray tracing calculation shows similar gradient (for example, sharp gradient around 60 km) for different sampling points A1, A2 and A3?

Line 288. “If AGWs observed by the airglow network satisfy the dispersion relation” reads weird, since the figures/main text are trying to persuade the readers the current work from the multi-layer airglow observation network has detected the information of AGWs, then lines 117-118 from the reference gives the dispersion relationship of AGWs and there is a proximate relationship in Lines 279-281…

Line 319: SWITVE should be TTIVE?

Line 334. Similar comment as above.

Lines 335, re-order this sentence. Better move “in the mesopause region” after AGWs?

Line 345, why “directly”? remove it.

Figure 10. Can you explain more about this possible mechanism? It is too general. It is well known how important the coupling processes among ocean/land/atmosphere to study the whole atmosphere etc.
Citation: https://doi.org/10.5194/egusphere-2023-2429-RC1
- AC1: 'Reply on RC1', QINZENG LI, 28 Feb 2024
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2429/egusphere-2023-2429-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-2429-AC1
RC2:
'Comment on egusphere-2023-2429', Anonymous Referee #2, 05 Jan 2024

The manuscript did a very comprehensive study covering a very complex sea-air interaction through the link: volcano eruption - atmospheric pressure waves - tsunamis - gravity waves - mesosphere airglow. This manuscript provides important and likely first observations of volcano-related Lamb and gravity waves in the mesosphere airglow. In general, the authors provide evidence and arguments to support their conclusions. But I think the provided evidence is a little weak, more like a correlation between two, not robust enough to prove the generation relation. The Tonga volcano and related tsunami/waves have been well observed and reported; there should be good cross-verification between this study and previous studies. It is fine that the presented results are different from those previous studies, but you need to provide explanations. As the manuscript mentioned, many presented results are the first ever observed, so more careful reasoning and analysis are needed to support the observations. Five waves (Lamb wave L0, L1, and three gravity waves) were observed and reported in this study; I have concerns about each of them.

Lamb waves have a vertical phase front below the thermosphere or zero vertical wave numbers. This means what you observe in the mesosphere should look the same or very similar to what is observed in the stratosphere. Many studies have reported the Lamb wave in the brightness temperature or IR radiation from geostationary satellites such as GOES (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2023GL106097), Himawari-8 (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022GL098324). So, I would expect that the wave pattern in the mesosphere OH airglow looks very similar to these studies, like a solitary wave with one strong leading wave with much weaker trailing waves. What you show in the OH airglow images are stable wave trains? You show the surface pressure for the L0/L1 mode; both seem to be like solitary waves. (You should show the time derivative of the pressure to emphasize the wave.) Are you able to get the Himawari-8 brightness temperature and compare it with your OH airglow results, then you will have the full link of surface-> stratosphere->mesosphere.

Regarding the L0/L1 mode (wave #1 and #2), this manuscript seems to be one of the few that report the weaker L1 mode. You need a better estimate of the wave parameters including wavelength, period, and amplitude (airglow intensity fluctuations to be compared with L0), to justify the observations. Also, is there any evidence for the ~3-hour separation between L0 and L1? Wright et al. 2022 report what they called primary and secondary Lamb waves (not sure if they match L0/L1), but the time difference is about 60-min. This study (https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2023GL103809) presents L1 mode also in the mesosphere. From wind measurements, the Lamb wave L1 mode period (2-hr), wavelength (1400 km), and wave pattern (a solitary wave) seem not to match what is presented in the airglow. And very interesting, they do not see L0 mode and argue that L0 mode is likely a higher-frequency wave and got averaged out.

For tsunami-generated gravity waves (#3—#5 in this manuscript), they have been studied in many model simulations. (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022JA030301, https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020JA028309, https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016JD025673). However, it seems the ocean surface vertical displacements need to be large enough (10 cm or even higher) to produce 1% airglow fluctuations, Table 3 of Inchin et al. 2022). Otherwise, the wave signature might not be detectable in the mesosphere airglow. This is another reason that you need to carefully estimate the observed wave amplitudes in airglow fluctuation percentage. Another drawback of this part is the tsunami simulations were not verified, such as by DART buoy wave height measurements as well as many published studies.

I am very skeptical about the gravity waves being tsunamis generated if the surface wave height is in the order of 2 cm, as shown in Figure 6. Weak tsunamis likely won’t generate atmospheric gravity waves. It might even be the opposite; those weak tsunamis are possibly excited by atmospheric gravity waves. (https://www.nature.com/articles/s41586-022-04926-4 and https://www.nature.com/articles/s41598-022-25854-3). This Figure (https://www.nature.com/articles/s41586-022-04926-4/figures/10) demonstrates the observed tsunamis are very well simulated. The order of magnitudes of the tsunami's surface height is somewhat similar to the 2 cm for two sites, 21416 and 23219, both similar distances to the East China Sea from the Tonga volcano.

Many other studies that report the Tonga eruption-related waves did a statistical analysis to verify the observed perturbations on that day were not random or just background (they exist even if there is no volcano eruption). Basically, they take a long-term average of several days to get a background perturbation. In your case, make sure you do not see those similar waves every day. If you observe something different before and after that day, then you can tell what you observe on that day is eruption-related.

For the reason above, the selected snapshot-style airglow images at several time steps are not very robust evidence for such volcano/tsunami-related waves. Keograms or movies can better demonstrate the arrival of the waves. Also, you need to add the distance to the volcano on the map.

Section 3.2, Figures 5 and 6 present results about the tsunami simulation, but not many details are included, are they verified?

All figures and corresponding discussions: Add the lapse time since the volcano eruption in all the figures, which could better help to understand the evolution of the wave pattern.

21, airglow imaging system network or airglow imager network.

22, I anticipate this ‘phase speed’ is horizontal phase speed or total phase speed.

24, confusing, L0 and L1 wavefront is vertical, then L0 wavefront mode tilts forward.

42-44, if you explicitly mention ’this volcanic eruption,’ I would expect that those waves are all reported from this eruption, so what is the purpose of those old references, even if they are very classic?

54, GNSS TEC, better to give the full name, at least for TEC.

73, conventional tsunami, sounds unclear here.

74, only two studies, maybe because they are rare.

83, better use one extra sentence to describe the term in ().

90-91, I guess there are two filters used on the same lens; basically, it is the same equipment. It is a little confusing to declare two airglow networks. Also, clearly state what types of airglow are observed.

105, what is “moving change pressure”

107, I would expect to see a brief here about the tsunami simulation and some more details like section 2.3.

125, is there any detrending or filtering on the SABER temperature before it is used in the calculation?

Line 141, do not use the alphabet o for a degree; use the math symbol.

Section 3.1 and Figure 1, better give each station a name or a number, so you could clearly state the wave front was visible in which station.

Figure 1, in sub-Figure 6 (middle row, right), P1-P6 are labeled; what are they? In sub-Figures 8 and 9, what are early and late wave packets?

Figure 1: the labels are often obscured by the airglow images on the map, hard to read. There are plenty of blank areas between airglow images where you can put those labels.

Figure 1: what happened in sub-Figure 1-6 (top and middle rows) for the station over the 110E and 20S, on that island?

Figure 1: for the station over the 120E, 25S, it seems it does not provide much information due to weather.

Figure 1, after I checked all the information on the figure, I went to find if any supplementary materials, such as videos showing the motion of the 5 wave packets, are available. Unfortunately, I did not find any. Can the authors make some videos from the airglow images?

149-155, how are those wave phase speeds and amplitudes estimated? For reference, what is the speed of sound at the same altitudes?

Figure 2: you need to clarify the P1-P6 points, even if you show them in Figure 1. Why only information of wave #1-#3 is shown, not wave #4-#5.

Figure 2: wave amplitudes in the airglow intensity need a better qualification using the percentage of the airglow intensity fluctuations rather than arbitrary units.

159-161, this statement about vertical distribution is too much for what you showed in Figure 3, where only wave information from three layers is shown, and only Lamb waves.

167, are you able to verify the results by estimating the speed * travel time to be the distance between two sites?

179-185, the scale height between two types of waves only accounts for how fast the wave energy attenuates with respect to altitude; you need wave amplitudes at the source level to compare the observed wave amplitudes at the mesosphere.

Figure 3: the band in OI airglow image is very similar to the overexposure due to clouds or reflections. Still, I would like to see some supporting evidence, like a Keogram or a video. It is better to use the derivative of the pressure to isolate the waves.

224, many blank spaces are missing; better proofreading is needed.

Figure 4: yellow lines mark other wavefronts, which is the wave #3.

Figure 5: up to 5 cm wave height for both types of tsunamis, see comments at the beginning.

249, how do you calculate the m2, which formula was used?

Citation: https://doi.org/10.5194/egusphere-2023-2429-RC2
- AC2: 'Reply on RC2', QINZENG LI, 28 Feb 2024
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2429/egusphere-2023-2429-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2023-2429-AC2

Peer review completion

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

AR by QINZENG LI on behalf of the Authors (05 Mar 2024) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (25 Mar 2024) by John Plane

RR by Anonymous Referee #2 (13 May 2024)

Suggestions for revision or reasons for rejection

The authors carefully addressed the comments from the reviewers and made substantial changes to the manuscript. I appreciate that the authors spent time creating the movies. They are very helpful in interpreting the waves. What is the exposure time or temporal resolution? Section 2.1 states that it is 1 minute. However, the movie presented in the supplementary materials shows it is roughly 3 or 4 minutes from the time stamp on the movie. Did some images get skipped? This is important because it is related to the determined wave period of 6 minutes. Please confirm the time resolution.

What is the yellow box in Figure 2? Did you choose this area to apply the 2D FFT to get the wavenumber and phase?

To compare Figure 5(e) one-to-one, it is better to zoom it into the same map area as Figure (a-c).

Checking the airglow movies and looking at the coherence and spatial coverage of the wave pattern among continuous images, I am convinced that wave packets 3 to 5 should be related to some larger-scale wave sources toward SE, the oceanic area. If further ray-tracing simulation can prove they are tracked to be related to a tsunami by independent simulation, then it is fair to say those waves are tsunami-generated.

Back to wave packets #1 and #2 about Lamb wave mode L0 and mode L1. It seems this is still an open question about whether Lamb waves can reach the mesosphere and be detected there.

https://acp.copernicus.org/articles/24/4851/2024/acp-24-4851-2024.html
https://angeo.copernicus.org/articles/41/197/2023/angeo-41-197-2023.html

These two papers, from the same author, used Meteor radar data to analyze the wave signature in the MLT region. It seems they could not confirm that the Lamb wave reached the upper atmosphere. The following is a quote from the paper:

Based on our observations, we can almost rule out that the primary lamb wave that was caused by the volcanic eruption reached the upper atmosphere. Thus, the thermospheric/ionospheric observations are likely the result of multistep vertical coupling processes as described in Becker and Vadas (2018), Vadas and Becker (2018), and Vadas et al. (2018). However, the wind measurements indicated several other signals exhibiting a similar morphology, and the lifetimes of other peaks in the winds could not be linked to the lamb wave excited by the Hunga Tonga–Hunga Ha′apai eruption.

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2023GL103809

However, this paper supports the L1 Lamb wave signature in the meteor radar, aka mesosphere.

It is well known that the airglow imaging system is more likely to capture high-frequency waves, especially if you use the time-difference method. The meteor radar would average out those high-frequency waves.

So the key question here: Do we believe the airglow imagers captured higher frequency waves are lamb waves?

You have estimated the following wave parameters: 300 m/s phase speed, 6-min period, 100 km wavelength.

The 300 m/s wave speed seems to be good in the acoustic range. The 6-minute period is also good and might explain why they are not captured by the meteor radar. My main concern is on the 100-km wavelength. If we carefully check Liu et al. 2023, the WACCM-X simulation indeed presented continuous wave trains above 31 km and a solitary wave pattern below. However, the wavelength does not vary much between solitary waves and wave trains. For wave packet #1 presented in this study, the wavelength estimated from the Himawari-8 satellite brightness temperature should be much larger than 100 km, as well as the OI airglow at about 250 km altitude. What kind of background condition between the stratosphere and mesosphere (30-80 km) would lead to waves of several hundred km wavelengths “breaking” into a much smaller scale?

From Figure 10, the winds between 40-80 km are mostly eastward and southward, so the Lamb waves propagate against the background winds. Could the wind explain something about the wavelength?

Wave #2 (Lamb L1) comes together with the gravity waves; it looks very much like gravity waves, except for the slight phase front orientation differences.

Also, it is disappointing that the wave #1 and #2 features are mostly captured by stations near 40N and 120E, making them appear like localized features. The wave signatures are very vague in the two coastal stations, so it is hard to tell the actual scale of the waves.

In summary, I am still unconvinced that waves #1/#2 are Lamb waves, but I could not provide clear arguments for why they are not. If this paper were purely about gravity waves #3-#5, it would be solid and comprehensive. I do not want to stop the manuscript from being published, but I would recommend the authors down-tune the conclusive remarks about wave #1/#2 since they are still debatable.

Hide

RR by Anonymous Referee #1 (13 May 2024)

ED: Publish subject to minor revisions (review by editor) (20 May 2024) by John Plane

AR by QINZENG LI on behalf of the Authors (30 May 2024) Author's response Author's tracked changes Manuscript

ED: Publish as is (31 May 2024) by John Plane

AR by QINZENG LI on behalf of the Authors (04 Jun 2024) Manuscript

Short summary

The 2022 Hunga Tonga–Hunga Ha’apai (HTHH) volcanic eruption not only triggered broad-spectrum atmospheric waves but also generated unusual tsunamis which can generate atmospheric gravity waves (AGWs). Multiple strong atmospheric waves were observed in the far-field area of the 2022 HTHH volcano eruption in the upper atmosphere by a ground-based airglow imager network. AGWs caused by tsunamis can propagate to the mesopause region; there is a good match between atmospheric waves and tsunamis.