Cryo-seismicity triggered by ice mass discharge through the Antarctic subglacial hydrographic network

Danesi, Stefania; Salimbeni, Simone; Borghi, Alessandra; Urbini, Stefano; Frezzotti, Massimo

doi:https://doi.org/10.5194/egusphere-2022-29

Preprints

https://doi.org/10.5194/egusphere-2022-29

Preprints

13 Apr 2022

Cryo-seismicity triggered by ice mass discharge through the Antarctic subglacial hydrographic network

Stefania Danesi¹, Simone Salimbeni¹, Alessandra Borghi¹, Stefano Urbini², and Massimo Frezzotti³ Stefania Danesi et al.

Show author details

¹Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Bologna, Italy
²Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Roma2, Roma, Italy
³Department of Science, Roma Tre University, Rome, Italy

Received: 06 Mar 2022 – Discussion started: 13 Apr 2022

Abstract. We analyze seismic time series collected during experimental campaigns in the area of the David Glacier, South Victoria Land, Antarctica, between 2003 and 2016. We observe hundreds of repeating seismic events, characterized by highly correlated waveforms (cross-correlation > 0.95), which mainly occur in the floating area between the grounding and the floating line of the ice stream. The joint analysis of seismic occurrences and observed local tide measurements seem to show that seismicity is not triggered by a seasonal, regular, periodic forcing such as the ocean tide, but more likely by transient irregular impulses. We consider possible environmental processes and their impact on the coupling between the glacier flow and the bedrock brittle failure. Our results suggest that clustered and repeated seismic events may be correlated with transient episodes of mass ice discharge (observed by satellite GRACE and GRACE-FO experiments) through the subglacial hydrographic system that originates upstream of the glacier and extends to the grounding zone, lubricating the interface with the bedrock.

Download & links

Preprint (PDF, 30288 KB)

Supplement (79 KB)

Stefania Danesi et al.

Status: final response (author comments only)

RC1:
'Comment on egusphere-2022-29', Anonymous Referee #1, 27 Apr 2022

The paper titled "Cryo-seismicity triggered by ice mass discharge through the Antarctic subglacial hydrographic network" by S. Danesi et al. has sufficient review of long term seismicity around the David Glacier of South Victoria Land, Antarctica, by comparing different environmental parameters of meteorological and tidal, satellite information by GRACE etc.. The discussed contents and results are enough credible in regards to the currrent avairable dataset by the authers. The full paper has enough sofisticated as already reviewed by another referees, but please make minor revisions as mentioned below.

page 2, line 62; Bannister and Kennett (2002); This article is not included in the Referneces. Please add.

Figure 2: We cannot identitfy the exact location of the ice-rader profile on the map above. Please add the profile location.

Figure 5. B) Ravels for Cl_0x, Cl_0y, Cl_z, etc. Please add the one-order number. We cannot identify which clasters are the ones.

Citation: https://doi.org/10.5194/egusphere-2022-29-RC1
- AC1: 'Reply on RC1', Stefania Danesi, 02 May 2022
  
  Reply to Referee 1:
  Dear Referee
  we are grateful for your positive comments.
  We have changed the manuscript in accord with your requests:
  page 2, line 62; Bannister and Kennett (2002); This article is not included in the Referneces. Please add.
  Yes, thank you. Done
  Figure 2: We cannot identitfy the exact location of the ice-rader profile on the map above. Please add the profile location.
  Done
  Figure 5. B) Ravels for Cl_0x, Cl_0y, Cl_z, etc. Please add the one-order number. We cannot identify which clasters are the ones.
  Done
  
  Citation: https://doi.org/10.5194/egusphere-2022-29-AC1
RC2:
'Comment on egusphere-2022-29', Ugo Nanni, 04 Jul 2022

The study by Danesi et al., present a new analysis of an extensive seismic dataset obtained in Antarctica. The methods is very well described an very clear, and the results are coherent and well discussed. They present a new catalogue of seismic events that they compare with environmental variables and glacier dynamics. They show that the main clusters dynamic is mainly linked to glacier dynamics, and discuss the implication of understadong subglacial condition.

I find this study very informative, but I have a difficult time to see (especially in the Introduction) the research question or the red thread of this study other that reanalysing this dataset. I would suggest to make the scope of the study a bit more clear in the introduction, as well as its outcomes in the conclusion. I think the purpose of this work could be made more clear.

Have you calculated the events magnitude, it would be interesting to see the values and compare it with other, for instance in the Whillans Ice Plain

How can you discriminate that the events you observe are basal events? It would be nice to have a bit more details on what can of seismicity you are observing.

Here are some minor comments:

L 68: what do you mean by extrem scenarios?

L132: do you mean earthquakes or icequakes?

L147: is it manual of automatic location?

L227-229: could you describe a bit more how the reader should read the figure 8, the wavelet analysis is not that common

Fig 9 is a bit difficult to read, maybe splitting it would help.

Best,

Ugo

Citation: https://doi.org/10.5194/egusphere-2022-29-RC2
- AC2:
  'Reply on RC2', Stefania Danesi, 11 Jul 2022
  Dear Ugo
  we are really grateful for your positive comments.
  In the following we show how we have changed the manuscript in accord with your requests:
  - I find this study very informative, but I have a difficult time to see (especially in the Introduction) the research question or the red thread of this study other that reanalysing this dataset. I would suggest to make the scope of the study a bit more clear in the introduction, as well as its outcomes in the conclusion. I think the purpose of this work could be made more clear.
  We have integrated the Introduction as follows:
  Miles et al., (2022) pointed out that the variability of the David ice discharge between 2005 and 2018 was associated to changes in ice shelf buttressing and the modulating effect of local glacier geometry.
  […]
  Assuming that the seismicity does not have a tectonic origin, but is rather activated by the dynamics of the glacier, we aim to provide a quantitative estimate of any causal processes of environmental origin that may have triggered the seismic events.
  In the Discussion we have changed the following lines:
  On the other hand, significant fluctuations in the David Glacier flow velocity, characterized by sudden transient increases in the ice speed (about 5-10%) with no regularity over time, were observed by Moon et al. (2021) and Miles et al. (2022). Moon et al. (2021) pointed out that the observed glacier velocities are inclined to increase during the Antarctic summer for at least three years during the 2016-2020 period probably due to the extension of the summer melt in ice surface and thus the increase of basal sliding. Besides, Miles et al. (2022) stated that the ice discharge change could be driven by both external forcing (e.g. katabatic wind), and also by internal ice sheet processes.
  We conclude with:
  Our result shows that continuous measurements of the seismic activity at ice/bedrock interface could be a helpful tool for the monitoring of ice sheet dynamics.
  The new item (Miles et al. 2022) was added to the References
  - Have you calculated the events magnitude, it would be interesting to see the values and compare it with other, for instance in the Whillans Ice Plain
  In Section 2.3.2 an estimate of earthquake magnitude was provided with no details about the method. Earthquake magnitudes were estimated in continuity with Danesi et al. 2007. We have added this piece of information in Section 2.3.2
  (Ml≤1.8, calculated as in Danesi et al. 2007)
  and, later:
  It is noteworthy that the magnitude estimate comes from a few numbers of observations, therefore it is affected by significant uncertainty.
  - How can you discriminate that the events you observe are basal events? It would be nice to have a bit more details on what can of seismicity you are observing.
  Yes, you are right. We have added these considerations in Section 2.3 .
  The seismic network has recorded thousands of events, most of which can be classified as icequakes, i.e. episodes of fracturing of the ice layer. A few hundred events can be classified as basal events occurring at the interface ice-bedrock, principally on the basis of some signal characteristics: the frequency content, the duration, the hypocenter depth, and smaller attenuation with distance. In particular, (i) the seismic spectrum of icequakes has maximum amplitudes at frequencies higher than 5-10 Hz, while the earthquake spectrum is shifted towards lower frequencies; (ii) the icequake signal can only be recorded in one or two nearby stations because it decays very quickly while basal events can be detected even at stations tens of km away; (iii) icequakes have a signal duration of a few seconds, while generally basal events have longer waveform durations. This work is focused on the analysis of basal events, therefore we neglect all signals classified as icequakes.
  Here are some minor comments:
  - L 68: what do you mean by extreme scenarios?
  We have changed into: It is worth noting that seismological analyses in harsh climate scenarios such as Antarctica cannot overlook the particular environmental conditions.
  L132: do you mean earthquakes or icequakes?
  
  In Section 2.3 we have added:
  This work is focused on the analysis of basal events, therefore we neglect all signals classified as icequakes.
  - L147: is it manual of automatic location?
  We manually re-located with NLL the events under study. We have added this piece of information.
  - L227-229: could you describe a bit more how the reader should read figure 8, the wavelet analysis is not that common
  We have added some indications about the interpretation of the cross-wavelet power spectrum:
  Interestingly, the resulting cross-wavelet power spectrum (Figure 8) implies that the inter-event time during the 5-weeks clustered seismicity is not correlated with the tide period, because the values are close to zero and the cross-wavelet power spectrum can be interpreted as the local covariance between two time-series.
  - Fig 9 is a bit difficult to read, maybe splitting it would help.
  OK - done
  
  Citation: https://doi.org/10.5194/egusphere-2022-29-AC2
RC3:
'Comment on egusphere-2022-29', Fabian Walter, 03 Aug 2022

This submission by Danesi et al. describes a study of cryoseismological sources associated with the flow of David Glacier, an ice stream draining the Eastern Antarctic Ice Sheet. The glacier has been subject to various studies, including seismological monitoring, which have suggested a relation to basal sliding modulated by tidal height at the ice-ocean interface. The present study may be a valuable contribution as it documents a shift in seismic source locations that may be related to water-enhanced sliding episodes.

In its current form, the manuscript is very difficult to digest and I have found an evaluation of the scientific quality nearly impossible. Key information seems to drown in varying numbers, poor figures and unclear explanations given in different parts of the manuscript. In the present form I cannot recommend this manuscript for publication unless it is extensively revised, explanations are improved and/or additional data analysis (though this may be brief) is given. Please see my comments below for further elaboration.

Fabian Walter.

MAJOR COMMENTS

The main problem I found with this manuscript is that I could not wrap my brain around the data set. The scientific interpretation seems to hinge on the “ice fall” and the “Cauldron area”. Neither of these places is labeled in any of the figures (unless I missed this). In addition, almost all clusters in Figure 5 have the same label. Consequently, when individual clusters or seismic events above or below the ice fall are described, the reader cannot verify or follow the explanations (see specific comments below). The cluster activity and lifetime can also not be verified or put into context. A range of dates specifying the data time window is given: 2003/04, 2005/06, 2015/16 (Line 85); 2003-presence (Lines 93-95); 2005-2017 (Line 96); “period 2003-2004” under study (Line 303), 2002-2016 (Figure 1) and perhaps some more. It is not clear which period ends up being the focus and of relevance to this study. This seems to produce some contradictory statements. For example, on Lines 194-196 the authors state that there existed clusters that were active between 2003 and 2016. On Lines 287ff the authors write that repeating seismicity was only observed between November and December 2003. In general, I was not able to understand which events have which characteristics (locations, waveform similarity, clustering behavior, correlation with forcing data, …).

If I understand correctly then Zoet et al. (2012 in Nature Geoscience) studied the same glacier and found a forcing by tidal amplitudes. However, the authors of the present manuscript did not attempt to look for this correlation, in fact they purposefully ignore such a correlation against the tidal phase and only investigate the tidal amplitude (Lines 223 ff). Why? It would be interesting to see which events/clusters behave like the Zoet events and which do not. If the authors cannot reproduce the Zoet results, then there may by some problems with the catalogue completeness (even though different seismometers were used for the two studies and the data epochs do not overlap). This deserves discussion.

To be honest, the discussion/conclusion called into question the few take-home messages I thought I had understood when reading the manuscript. First of all, there is a reference to seasonal “behavior” inferred from the seismic data. How is this seasonality backed up in the manuscript? Moreover, the discussion rejects the hypothesis of surface melt affecting basal conditions. Whereas I agree with this, why did the authors come up with this hypothesis in the first place? The temperature data argue against any surface melt. In this regard, it would be interesting to compare the activity of individual clusters that overlapped with the Moon et al. (2021) study to see if there is a correlation with the seasonal speedup.

Source parameters: The authors mention source magnitudes. How were they obtained? This is an important piece of information, because it seems that there is a magnitude overlap with the events by Zoet et al. (2012). Moreover, the size of these events should be discussed and compared to other stick-slip events beneath polar ice streams. It would also be interesting to analyze the waveforms more. Are there different P-phases (refraction through the crust; multiple reflections within the ice column; P-S conversions at the ice-bed interface; …)? See next comment. Furthermore, the authors seem to have access to a good azimuthal coverage of recording stations. From this, first motion polarities and focal mechanisms could be obtained that provide relevant information for comparison with ice flow.

Locations: Do the locations from the 1-D and 3-D velocity models differ substantially? If so, then this could be related to refracted phases. Moreover, the maps showing event locations have to include error bars. Note that the uncertainties for the relative location is rather low (100 m). How large is the area occupied by location maxima of events within one cluster? Larger or smaller than the error bars? An analysis of the relative locations could give an estimate of asperity size and perhaps even indicate source migration, which would be extremely informative for basal conditions characterizing the material interface at the rupture area (see, e.g., discussion in Gräff et al., 2021, in GRL).

The text is generally easy to follow, but there are many grammatical errors and phrases, which do not conform to scientific writing standards. It seems that the manuscript was put together in haste, with many orphan paragraphs, false punctuation, single-sentence paragraphs and incorrect terminology. I suggest a thorough re-read and revision from native English speaker.

SPECIFIC COMMENTS

Lines 27ff: Not all the named studies talk about failure mechanisms (e.g., the tremor sources). This has to be reworded.

Lines 35-49: Pointing out the slope break in the respective figures is extremely important here.

Lines 50ff: The hydraulic system is indicated in some figures. How was it inferred? This should be discussed better and perhaps be plotted in other figures, too.

Line 65: “significant correlation with data” is too unspecific.

Line 109: “integrate the lower number of bedrock reflections” is unclear.

Lines 116-117: This information belongs in the figure caption, not the main text.

Line 123: Does the “9.5 km depth” refer to below the surface or below the ice sole?

Lines 125-126: “we merge mean velocity values” is unclear.

Line 131: Specify “many full years”.

Line 142: What are “coherent P-picks”?

Lines 143-145: Why did you decide this way?

Line 147: “manual locations“: From the explanations above it seemed that the locations were automatic.

Lines 150-151: Indicate this location on the map.

Lines 170ff: You mention what you did “initially”, but what did you end up doing? What is the difference between the two correlation scans?

Lines 199ff and elsewhere: The authors tend to announce what they are about to do in upcoming sections. This is a waste of space.

Lines 211ff: The shift in inter-event times needs to be backed up with a figure.

Line 226: “by interpolation” is too unspecific.

Lines 233ff: This paragraph is unconnected from the rest of the text.

Lines 237ff: The correlation between trigger threshold changes and wind speed can be complicated. It can depend on how you parameterize wind speed, i.e. via the noise floor or peak. The latter would isolate gusts. If a noise floor (e.g., lower percentile), median or mean is chosen then you may miss the influence of gusts.

Line 247: “December 1st” of which year?

Line 267: “a resolution higher than the average value”: Unclear.

Lines 273ff: Is Smith et al. (2015) the right reference? It talks about Rutford Ice Stream, not David Glacier. “soaring up” has to be quantified.

Lines 281ff: A waveform record supporting the fluid resonance is needed. There could be other explanations for dominant frequencies. I do not see how the last sentence in this paragraph follows.

Lines 297ff: The discussion about sliding physics seems oversimplified. First of all, I suggest not only referring to regelation and plastic flow. This holds for hard beds, but there are other mechanisms in the context of soft beds. Zoet and Iverson (2020 in Science) should be discussed. Finally, is it really valid to consider a simple velocity threshold for seismogenesis? After all, it seems that large changes in basal hydraulic pressures were involved.

Lines 310ff, including the next paragraph: Ideally, some ice flow velocities should be shown here. Is there absolutely no data from 2003? If so, this statement should be made more concrete (there certainly exist flow velocity estimates) and backed up with references.

FIGURES

In general, panels should be labeled a, b, c, … or equivalently and not be referred to with “upper”, “lower“, “right”, “left”, …

Figure 1: Mark ice fall and Cauldron. Give a date for “now” in caption.

Figure 2: Mark Cauldron. Where was the radar transect taken? Annotate/label the radar transect (e.g., fast flowing ice vs. ice sheet).

Figure 3: Not sure if showing the surface in grey shade has an added value.

Figure 4: Is this from the 1-D or 3-D inversion? Show error bars.

Figure 5: How were the grounding/hydrostatic lines inferred? I would explain them in the text, too. How was the subglacial water flux determined? This is important information, by the way! Show error bars. Where are the cluster events? Are they masked by the number circles? This would be unfortunate, actual locations would be informative.

Figure 6: Fonts are too small. What are the different grey shades in the top right panel?

Figure 7: Fonts are too small. Which clusters do the events shown in the top panel belong to? Vertical green line is missing or hard to see.

Figure 8: How was significance determined/defined?

Figure 9: Fonts are too small.

Figure 10: The horizontal bars should be shown against some along-flow-line quantity like ice thickness or bedrock elevation. A reference for the mass change data should be given in the caption (in view of GRACE).

Citation: https://doi.org/10.5194/egusphere-2022-29-RC3
- AC3: 'Reply on RC3', Stefania Danesi, 14 Oct 2022
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/egusphere-2022-29/egusphere-2022-29-AC3-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2022-29-AC3

Stefania Danesi et al.

Supplement

https://doi.org/10.5194/egusphere-2022-29-supplement

Stefania Danesi et al.

Viewed

Total article views: 605 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
451	139	15	605	58	5	9

HTML: 451
PDF: 139
XML: 15
Total: 605
Supplement: 58
BibTeX: 5
EndNote: 9

Views and downloads (calculated since 13 Apr 2022)

Month	HTML	PDF	XML	Total
Apr 2022	168	53	4	225
May 2022	60	11	1	72
Jun 2022	27	7	1	35
Jul 2022	37	11	2	50
Aug 2022	38	24	5	67
Sep 2022	9	2	0	11
Oct 2022	30	13	2	45
Nov 2022	17	7	0	24
Dec 2022	17	6	0	23
Jan 2023	48	5	0	53

Cumulative views and downloads (calculated since 13 Apr 2022)

Month	HTML	PDF	XML	Total
Apr 2022	168	53	4	225
May 2022	60	11	1	72
Jun 2022	27	7	1	35
Jul 2022	37	11	2	50
Aug 2022	38	24	5	67
Sep 2022	9	2	0	11
Oct 2022	30	13	2	45
Nov 2022	17	7	0	24
Dec 2022	17	6	0	23
Jan 2023	48	5	0	53

Viewed (geographical distribution)

Total article views: 557 (including HTML, PDF, and XML) Thereof 557 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 30 Jan 2023

Short summary

Clusters of low-energy seismic events, concentrated in space and time, characterized by highly correlated waveforms (cross-correlation coefficient ≥ 0.95), occur at the floating area of a major ice stream in Antarctica (David Glacier, North Victoria Land). The transient injection of fluids from the David catchment into the regional subglacial hydrographic network, observed by GRACE measurements, is indicated as the main trigger for clustered and repeated seismic occurrences.


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