the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 May 2022
Snow stratigraphy observations from Operation IceBridge surveys in Alaska using S/C band airborne ultra-wideband FMCW radar
Fernando Rodriguez-Morales
Carl Leuschen
John Paden
Daniel Gomez-Garcia
Emily Arnold
Abstract. During the concluding phase of the NASA Operation IceBridge (OIB), we successfully completed two airborne measurement campaigns (in 2018 and 2021, respectively) using a compact S/C band radar installed on a Single Otter aircraft and collected data over Alaskan mountains, ice fields, and glaciers. We observed snow strata in ice facies, wet-snow/percolation facies and dry snow facies from radar data. This paper reports seasonal snow depths derived from our observations. We found large variations in seasonal radar-inferred depths assuming a constant relative permittivity for snow equal to 1.89. The majority of the seasonal depths observed in 2018 were between 3.2 m and 4.2 m, and around 3 m in 2021. We also identified the transition areas from wet-snow facies to ice facies for multiple glaciers based on the snow strata and radar backscattering characteristics. Our analysis focuses on the measured strata of multiple years at the caldera of Mount Wrangell to estimate the local snow accumulation rate. We developed a method for using our radar readings of multi-year strata to constrain the uncertain parameters of interpretation models with the assumption that most of the snow layers detected by the radar at the caldera are annual accumulation layers. At a 2004 ice core and 2005 temperature sensor tower site, the locally estimated average snow accumulation rate is ~2.89 m w. e. a-1 between the years 2002 and 2021. Our estimate of the snow accumulation rate between 2005 and 2006 is 2.82 m w. e. a-1, which matches closely to the 2.75 m w. e. a-1 inferred from independent ground-truth measurements made the same year. We also found a linear increasing trend of 0.011 m w. e. a-1 per year between the years 2002 and 2021. With this trend, we extrapolated the snow accumulation back to 1992 and obtained an average accumulation rate of 2.74 w. e. a-1 between the years 1992 and 2004, which agrees well with the value of 2.66 w. e. a-1 for the same period determined from the ice core data retrieved at the caldera in 2004. The results reported here verified the efficacy of our method, its assumption, and the interpretation models.
-
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
(1966 KB)
-
Supplement
(898 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1966 KB) -
Supplement
(898 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Jilu Li et al.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-368', William D. Harcourt, 06 Aug 2022
Overview
This paper uses radar measurements from OIB campaigns in summer 2021 over Alaska to 1) track annual snow layers in radargrams, 2) estimate the snow accumulation rate over Mt Wrangell, and 3) analyse snow strata within ice facies. The authors develop a radar age-depth model from Clark et al. (1989) to quantify the two-way travel time of the radar wave through the snow and ice, and constrain the associated parameters using a cost function that aims to minimise the difference between two snow depositional ages to annual increments (i.e. 1 year). The modelled age-depth relationship fits the derived data sets very well and hence provides confidence that the subsequent estimated accumulated rate is sufficiently accurate, with the caveat that local surface processes such as wind redistribution are not be completely accounted for. The key result from a glaciological perspective is shown in Figure 8 which shows increasing accumulation rates between 2004 to 2021.
General Comments
The paper is well-written overall and provides a very detailed account of the methods used to constrain the parameters in the age-depth model and the subsequent extraction of key variables such as annual snow accumulation rate. The results will be of significant interest to glaciologists and hydrologists interested in understanding glacier mass balance and its impacts on catchment hydrology. Whilst the technical details of the paper are well-described, the glaciological interpretation of the data set is under-developed. In particular, I think the paper would benefit from a discussion about surface mass balance processes and how these have changed over time e.g. what are the processes underpinning the increasing accumulation rates in Figure 8. Are there any regional SMB measurements that you can compare to?
I’ve noted some relatively minor technical corrections below which are mostly areas of clarification. If the authors can integrate these suggestions into a revised manuscript I believe the paper will be ready for publication.
Technical Corrections (References to line numbers in preprint)
Abstract
L9-L10: This sentence should come straight after your introductory line. Then you can launch into a d description of your findings including observing snow strata in ice facies.
Introduction
L25: Not sure what is meant by ‘Earth’s ecosystem’. Maybe just ‘the Earth’s climate system’.
L25-26: Suggest change to: ‘in the Earth’s climate system and respond rapidly to changes in climate which impacts regional hydrology and the local economy.’
L29-L33: This is a long sentence and can be shortened. Focus on the global trend and then specify exactly the mass loss from Alaskan glaciers as an example. Where possible, avoid lots of clauses as it breaks up the flow of the sentence.
L34: Start new sentence here: ‘…Hill et al., 2015).The changes in glacier discharge…’
L35: ‘home to important’
L39: Maybe spaceborne?
L41: ‘However,”
L51-53: Worth stating that ground based measurements are also used for satellite validation of snow products.
L55: ‘at a glacier-scale’
L58: ‘within temperate firn’
L63: ‘with a 6-GHz’
L64: Reference Figure 1 here
L66: ‘using snow pit measurements to 10 m depth’
L71: ‘across a broader spatial region than compared to the 2018 campaign (Li et al., 2019)’
Data collection and processing
L79: ‘over 8 days covering 5115 linear km’
L80-88: It would be better to briefly describe and LiDAR system and discuss the radar antenna installation in a little more detail rather than referring to a previous paper. A table of critical radar system parameters would also be useful.
L81: ‘altitude above ground level (AGL)’
L83-84: To understand this the reader would also need to know the ADC sampling frequency, which can go into a table of parameters.
L85: State the vertical resolution before and after changing the bandwidth.
L90: Change brown to colour to distinguish from red; black might work?
L92-93: State spatial coverage in km2?
L99: What was the magnitude of the correction applied to the radar system delay?
L101-104: It would be helpful to provide a little more detail on these processing steps e.g. general outline of how the processing performed, performance improvement and the reason for using each step. Does the order matter? Similar to the deconvolution, were any of these steps applied differently to previous campaigns?
Results analysis and discussions
L112: “above sea level”
L113: ‘focus on the analysis of
L114: ‘discuss observations along the transition from the accumulation to the ablation zone along
L122: I assume by ‘flattening’ you mean normalised to surface elevation? If so, was this from the lidar data?
L124-126: Both years have multi-modal peaks largely ranging between 1-6 m. Better to state this and the means of each individual distribution. This would also reveal the lack of a third peak in the 2021 data. Why might this be? More melt?
L133: What month were the 1994 measurements taken and are you able to quantify differences in air temperature between that study and this one?
L139: Change ‘massive’ to ‘large’
L141: ‘researchers have been drawn to study glacier-volcano interactions
L144: ‘are also both’
L145: ‘covers a 4.2 km by 2.7 km area.’
L155: ‘subsurface layers’
L169: ‘shows a plot of the flight line’
L180: It might be beneficial to have a short sentence explain what is meant by an ‘interpretation model’.
182-189: It’s very hard to differentiate the notation for density and pressure. Maybe change the notation for pressure to capital P for readability?
L210: I agree with the assumption of steady-state conditions. Maybe also state that based on S3 there is also a skew towards more positive differences which could imply more snow accumulation in winter 2021.
L227: As far as I can see you haven’t stated how the layers were picked – manual, semi-automatic or automatic?
L228: What density values were used to calculate the permittivity, kg/m3?
L229: 1.127 km east, west, north or south?
L251: ‘values of the cost’
L257-258: Exactly how is the value of J applied to calculate the depositional ages of the tracked layers?
L259-261: Not entirely clear why these are accumulation layers – they broadly fit into the sequence of annual integer increments…
L269: “Therefore, our purpose” (i.e. because of the shift identified in the previous paragraph, only accumulation rates can be determined)
L279-309: These are interesting results and their glaciological intepretations should be assessed further. Why is there a rising trend in accumulation? How does this relate to glacier mass balance? Is there any evidence for melt on ice internal layers and radar backscatter? It’s worth highlighting in this section that you are interpretating radargrams from the dry snow facies to illustrate that melt layers are unlikely to be present.
L327-332: This description would benefit from some annotations of Figure 10a, particularly highlighting the broad locations of the facies.
Figures
Figure 1: A little difficult to see the flight lines. Could you have a small inset panel for the region and then extent indicators showing the two main regions surveyed?
Figure 3: A legend stating what the blue and red dots represent would be helpful.
Figure 4: Could you also annotate the location of the surface for clarity?
Figure 9: Better to state the elevation of the snow surface in panel d.
Citation: https://doi.org/10.5194/egusphere-2022-368-RC1 - AC1: 'Reply on RC1', Jilu Li, 21 Oct 2022
-
RC2: 'Comment on egusphere-2022-368', HP Marshall, 19 Sep 2022
This is a valuable, well executed study, demonstrating the ability of ultra-wideband airborne FMCW radar to measure annual accumulation layers on mountain glaciers, as well as to distinguish between dry snow and percolation/wet snow facies, and ice facies. The methodology is well described, and the combination of the radar observations with a one-dimensional physically based depth-age model is solid. The authors show good agreement between the radar-derived accumulation rate estimates and some limited field observations, lending confidence to the approach. The results are significant both in terms of the radar observation advances, and the implications of the measured accumulation rates. This will be an important paper, and just needs a bit more detail and sensitivity analysis before publication.
General comments:
1) A few more details on the processing approach (i.e. what FFT windowing method was used, what kind of horizontal filtering was applied) would be helpful to add.
2) The sensitivity of the results to the chosen permitivity should be evaluated. A relative seasonal snow permittivity of 1.89 was used, based on work on other glaciers - how does this compare to this study site? If field measurements are not available from the time of these flights, how much do results change for a conservative range of seasonal snow density/permittivity?
3) How were the layers picked? Manually? Semi-automatically with control points? Automatically? This needs more detail.
4) The model appears to produce a depth-age scale, but also a depth-density result. How does this compare to the ice core data? The authors state that although the model was tuned to Greenland, it represents the firn well - it would be useful to show this with field observations from this site, which should be available from the ice core. For example, Greenland gets a huge amount of wind influence, and the authors state that the main site is not wind effected. Its possible the assumed surface density in the model, for example is a bit too high.
5) This is the most important general comment -- the interpretation of the linear increase in accumulation rate is the most important result from a glaciological/snow science point of view, but needs a bit more work to test this trend, and explain why it might be happening. Is this linear increase over the past several decades expected based on regional climate models? Other glacier observations? Even if we had a linear increase the last 2 decades, why would we expect this to be the case in the decade previous (which was used to extrapolate to the early 1990s)?
I'm honestly a bit worried that this trend is caused by an assumption of constant density, or some artifact in the model. Can you back this linear increase in accumulation up with any other regional data from field observations or models? Or how about the ice core - does that show a linear increase in accumulation rate? I would expect the chemistry in the ice core would have resulted in a depth-age scale, so accumulation time series should be available from that?
Related to this - your model assumes a steady state - that the accumulation rate is balanced by the melt/densification and flow divergence. You show this with the little change in surface elevation between 2018 and 2021. Then what does the increasing accumulation over time then imply? Greater basal melt to balance this increase? Or a volume flux or densification rate increase with time?
Is it possible that a bias in the densification model is leading to the linear increase in estimated accumulation? I think a sensitivity analysis is needed here, along with error bars for the plot of accumulation vs time.
Detailed suggestions/edits in the attached annoted PDF.
Great paper, I look forward to the final version!
cheers,
HP
- AC2: 'Reply on RC2', Jilu Li, 21 Oct 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-368', William D. Harcourt, 06 Aug 2022
Overview
This paper uses radar measurements from OIB campaigns in summer 2021 over Alaska to 1) track annual snow layers in radargrams, 2) estimate the snow accumulation rate over Mt Wrangell, and 3) analyse snow strata within ice facies. The authors develop a radar age-depth model from Clark et al. (1989) to quantify the two-way travel time of the radar wave through the snow and ice, and constrain the associated parameters using a cost function that aims to minimise the difference between two snow depositional ages to annual increments (i.e. 1 year). The modelled age-depth relationship fits the derived data sets very well and hence provides confidence that the subsequent estimated accumulated rate is sufficiently accurate, with the caveat that local surface processes such as wind redistribution are not be completely accounted for. The key result from a glaciological perspective is shown in Figure 8 which shows increasing accumulation rates between 2004 to 2021.
General Comments
The paper is well-written overall and provides a very detailed account of the methods used to constrain the parameters in the age-depth model and the subsequent extraction of key variables such as annual snow accumulation rate. The results will be of significant interest to glaciologists and hydrologists interested in understanding glacier mass balance and its impacts on catchment hydrology. Whilst the technical details of the paper are well-described, the glaciological interpretation of the data set is under-developed. In particular, I think the paper would benefit from a discussion about surface mass balance processes and how these have changed over time e.g. what are the processes underpinning the increasing accumulation rates in Figure 8. Are there any regional SMB measurements that you can compare to?
I’ve noted some relatively minor technical corrections below which are mostly areas of clarification. If the authors can integrate these suggestions into a revised manuscript I believe the paper will be ready for publication.
Technical Corrections (References to line numbers in preprint)
Abstract
L9-L10: This sentence should come straight after your introductory line. Then you can launch into a d description of your findings including observing snow strata in ice facies.
Introduction
L25: Not sure what is meant by ‘Earth’s ecosystem’. Maybe just ‘the Earth’s climate system’.
L25-26: Suggest change to: ‘in the Earth’s climate system and respond rapidly to changes in climate which impacts regional hydrology and the local economy.’
L29-L33: This is a long sentence and can be shortened. Focus on the global trend and then specify exactly the mass loss from Alaskan glaciers as an example. Where possible, avoid lots of clauses as it breaks up the flow of the sentence.
L34: Start new sentence here: ‘…Hill et al., 2015).The changes in glacier discharge…’
L35: ‘home to important’
L39: Maybe spaceborne?
L41: ‘However,”
L51-53: Worth stating that ground based measurements are also used for satellite validation of snow products.
L55: ‘at a glacier-scale’
L58: ‘within temperate firn’
L63: ‘with a 6-GHz’
L64: Reference Figure 1 here
L66: ‘using snow pit measurements to 10 m depth’
L71: ‘across a broader spatial region than compared to the 2018 campaign (Li et al., 2019)’
Data collection and processing
L79: ‘over 8 days covering 5115 linear km’
L80-88: It would be better to briefly describe and LiDAR system and discuss the radar antenna installation in a little more detail rather than referring to a previous paper. A table of critical radar system parameters would also be useful.
L81: ‘altitude above ground level (AGL)’
L83-84: To understand this the reader would also need to know the ADC sampling frequency, which can go into a table of parameters.
L85: State the vertical resolution before and after changing the bandwidth.
L90: Change brown to colour to distinguish from red; black might work?
L92-93: State spatial coverage in km2?
L99: What was the magnitude of the correction applied to the radar system delay?
L101-104: It would be helpful to provide a little more detail on these processing steps e.g. general outline of how the processing performed, performance improvement and the reason for using each step. Does the order matter? Similar to the deconvolution, were any of these steps applied differently to previous campaigns?
Results analysis and discussions
L112: “above sea level”
L113: ‘focus on the analysis of
L114: ‘discuss observations along the transition from the accumulation to the ablation zone along
L122: I assume by ‘flattening’ you mean normalised to surface elevation? If so, was this from the lidar data?
L124-126: Both years have multi-modal peaks largely ranging between 1-6 m. Better to state this and the means of each individual distribution. This would also reveal the lack of a third peak in the 2021 data. Why might this be? More melt?
L133: What month were the 1994 measurements taken and are you able to quantify differences in air temperature between that study and this one?
L139: Change ‘massive’ to ‘large’
L141: ‘researchers have been drawn to study glacier-volcano interactions
L144: ‘are also both’
L145: ‘covers a 4.2 km by 2.7 km area.’
L155: ‘subsurface layers’
L169: ‘shows a plot of the flight line’
L180: It might be beneficial to have a short sentence explain what is meant by an ‘interpretation model’.
182-189: It’s very hard to differentiate the notation for density and pressure. Maybe change the notation for pressure to capital P for readability?
L210: I agree with the assumption of steady-state conditions. Maybe also state that based on S3 there is also a skew towards more positive differences which could imply more snow accumulation in winter 2021.
L227: As far as I can see you haven’t stated how the layers were picked – manual, semi-automatic or automatic?
L228: What density values were used to calculate the permittivity, kg/m3?
L229: 1.127 km east, west, north or south?
L251: ‘values of the cost’
L257-258: Exactly how is the value of J applied to calculate the depositional ages of the tracked layers?
L259-261: Not entirely clear why these are accumulation layers – they broadly fit into the sequence of annual integer increments…
L269: “Therefore, our purpose” (i.e. because of the shift identified in the previous paragraph, only accumulation rates can be determined)
L279-309: These are interesting results and their glaciological intepretations should be assessed further. Why is there a rising trend in accumulation? How does this relate to glacier mass balance? Is there any evidence for melt on ice internal layers and radar backscatter? It’s worth highlighting in this section that you are interpretating radargrams from the dry snow facies to illustrate that melt layers are unlikely to be present.
L327-332: This description would benefit from some annotations of Figure 10a, particularly highlighting the broad locations of the facies.
Figures
Figure 1: A little difficult to see the flight lines. Could you have a small inset panel for the region and then extent indicators showing the two main regions surveyed?
Figure 3: A legend stating what the blue and red dots represent would be helpful.
Figure 4: Could you also annotate the location of the surface for clarity?
Figure 9: Better to state the elevation of the snow surface in panel d.
Citation: https://doi.org/10.5194/egusphere-2022-368-RC1 - AC1: 'Reply on RC1', Jilu Li, 21 Oct 2022
-
RC2: 'Comment on egusphere-2022-368', HP Marshall, 19 Sep 2022
This is a valuable, well executed study, demonstrating the ability of ultra-wideband airborne FMCW radar to measure annual accumulation layers on mountain glaciers, as well as to distinguish between dry snow and percolation/wet snow facies, and ice facies. The methodology is well described, and the combination of the radar observations with a one-dimensional physically based depth-age model is solid. The authors show good agreement between the radar-derived accumulation rate estimates and some limited field observations, lending confidence to the approach. The results are significant both in terms of the radar observation advances, and the implications of the measured accumulation rates. This will be an important paper, and just needs a bit more detail and sensitivity analysis before publication.
General comments:
1) A few more details on the processing approach (i.e. what FFT windowing method was used, what kind of horizontal filtering was applied) would be helpful to add.
2) The sensitivity of the results to the chosen permitivity should be evaluated. A relative seasonal snow permittivity of 1.89 was used, based on work on other glaciers - how does this compare to this study site? If field measurements are not available from the time of these flights, how much do results change for a conservative range of seasonal snow density/permittivity?
3) How were the layers picked? Manually? Semi-automatically with control points? Automatically? This needs more detail.
4) The model appears to produce a depth-age scale, but also a depth-density result. How does this compare to the ice core data? The authors state that although the model was tuned to Greenland, it represents the firn well - it would be useful to show this with field observations from this site, which should be available from the ice core. For example, Greenland gets a huge amount of wind influence, and the authors state that the main site is not wind effected. Its possible the assumed surface density in the model, for example is a bit too high.
5) This is the most important general comment -- the interpretation of the linear increase in accumulation rate is the most important result from a glaciological/snow science point of view, but needs a bit more work to test this trend, and explain why it might be happening. Is this linear increase over the past several decades expected based on regional climate models? Other glacier observations? Even if we had a linear increase the last 2 decades, why would we expect this to be the case in the decade previous (which was used to extrapolate to the early 1990s)?
I'm honestly a bit worried that this trend is caused by an assumption of constant density, or some artifact in the model. Can you back this linear increase in accumulation up with any other regional data from field observations or models? Or how about the ice core - does that show a linear increase in accumulation rate? I would expect the chemistry in the ice core would have resulted in a depth-age scale, so accumulation time series should be available from that?
Related to this - your model assumes a steady state - that the accumulation rate is balanced by the melt/densification and flow divergence. You show this with the little change in surface elevation between 2018 and 2021. Then what does the increasing accumulation over time then imply? Greater basal melt to balance this increase? Or a volume flux or densification rate increase with time?
Is it possible that a bias in the densification model is leading to the linear increase in estimated accumulation? I think a sensitivity analysis is needed here, along with error bars for the plot of accumulation vs time.
Detailed suggestions/edits in the attached annoted PDF.
Great paper, I look forward to the final version!
cheers,
HP
- AC2: 'Reply on RC2', Jilu Li, 21 Oct 2022
Peer review completion
Journal article(s) based on this preprint
Jilu Li et al.
Data sets
CReSIS Snow Radar Data Products John Paden, Jilu Li, Carl Leuschen https://data.cresis.ku.edu/data/snow/
Jilu Li et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 475 | 119 | 14 | 608 | 45 | 2 | 3 |
- HTML: 475
- PDF: 119
- XML: 14
- Total: 608
- Supplement: 45
- BibTeX: 2
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1966 KB) - Metadata XML
-
Supplement
(898 KB) - BibTeX
- EndNote
- Final revised paper