the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Feb 2024
A Newly Digitised Ice-penetrating Radar Data Set Acquired over the Greenland Ice Sheet in 1971–1979
Abstract. We present an ice-penetrating radar data set acquired over the Greenland ice sheet by aircraft during the years 1971, 1972, 1974, 1978, and 1979. The data set comprises over 177,000 km of flight lines and contains a wealth of information on the state of the Greenland ice sheet including information on ice thickness and englacial properties. During data collection in the 1970s, the data were recorded on optical film rolls and in this manuscript, we document the digitization of these film rolls and their associated geographical information. Our digitization of the data enables interaction with and analysis of the data and facilitates comparison with modern-day radar observations. The complete data set in full resolution is available at the Stanford Digital Repository (https://doi.org/10.25740/wm135gp2721) with the associated technical reports (Karlsson et al., 2023). Part of the data set has been quality checked and is available as low-resolution JPG files at The Technical University of Denmark’s data repository, DTU Data, with associated geographical information (DOI pending). Here, we release the full data sets to enable the larger community to access and interact with the data.
- Preprint
(8556 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2023-442', Joseph MacGregor, 25 Mar 2024
“A Newly Digitised Ice-penetrating Radar Data Set Acquired over the Greenland Ice Sheet in 1971-1979” by N.B. Karlsson et al.
Joseph A. MacGregor
25 March 2024
I am perhaps overly predisposed to appreciation of studies such as this one, which seek to resurrect long-ago datasets into a modern form so that they may provide insight into long-term ice-sheet behavior. Nevertheless, this MS is an excellent example thereof. The data are lucidly described, and the process of sorting out the wheat from the chaff in the various campaigns is rigorous. The examples provided of the radargrams are clear, and the suggested process for accessing the data, which I reproduced, is good. I have a couple larger concerns/perspectives that may or may not be resolvable within the context of an ESSD submission, and some other minor comments.
Larger interrelated concerns:
- Conspicuously missing from the list of ice thickness compilations on L32 for which these 1970’s DTU data have been used is Morlighem et al. (2017, GRL), i.e., BedMachine v3. This is the only ice thickness compilation in modern use, is regularly updated (now v5) and is the de facto standard for Greenland. In the past, it has not been clear to me whether these 1970’s data were included in BedMachine. The legend of Figure 1a in Morlighem et al. (2017) contained just enough ambiguity to make that unclear. However, the shapefiles provided at the DTU repository provided an unambiguous answer to this question when compared against the dataid field of BedMachine v5: they are not. While the authors fairly discuss many potential applications of the dataset, the most conservative assessment of their near-term value is ice thickness measurements in otherwise unsurveyed regions. I assume that the 3 nautical miles uncertainty, equivalent to 5.6 km or ~37 BedMachine 150-m grid cells, is the primary reason why, but I don’t know for sure. A direct assessment and discussion of this issue seems essential.
- The georeferencing synthesis ought to be highlighted sooner/better in the MS as it is fundamental to any interpretation. Further, why provide a large set of shapefiles when a substantially better and simpler one-file format is available, i.e., GeoPackage? See http://switchfromshapefile.org/ for relevant arguments. Further, it is not obvious from the filename or description in the MS that the surface and bed elevations are also included in those shapefiles, enabling calculation of the ice thickness directly. Providing an example QGIS style file for such a calculation, rather than forcing every user to generate their own QGIS style expression, would be helpful.
- If the 3 nautical mile uncertainty can indeed be considered an upper limit, are there places where the uncertainty is known to be lower and could this be added as a field to the georeferencing information?
Minor comments:
L32: “mappings” seems superfluous
44: comma missing after “signal”?
44: “unprecedented” seems like not the right word, given that Schroeder et al. (2019, PNAS) achieved something similar for Antarctica, and other pre-digital archival records have been digitized to similar beneficial effect (e.g., Kjeldsen et al., 2016, Nature)
47: comma missing after VXE 6?
62: simplify to “of two-way traveltime”? not sure why parenthetical statement chosen here
106: is *lower* than the ones scanned at DTU
110: “film originals” meant?
135: “from scratch” is a useful colloquialism but perhaps worth contextualizing with something like the following: “without requiring physical access to the film itself”.
182-3: This statement is fair but seems like an obtuse reference to the efforts of Schroeder et al. (2022, JGlac) on the Antarctic film record, without citing that study. Also, why would a correction for apparent internal reflection power, similar to Gades et al. (2000, JGlac) for Siple Dome, not work in this context? IRP variations are noise to some, but not necessarily a dealbreaker?
Table 1: The description of the “missing” flights is good. A minor perspective from my OIB days that may help understand what happened: It may be that the flight numbers were sometimes numbered not by science concerns but as a sequence in the broader campaign, which included flight numbers for transits between Thule AFB (now Pituffik SB) and Søndrestrøm AB (now Kangerlussuaq). It does seem that flights based from both locations, given their geometry, but I didn’t check if that happened in the same year. In other words, the flight numbers may have been set by the air crew, not the scientists, once the campaign began.
For Tables 1 and 2, it might be helpful to add columns with percentages of total know flight lines scanned or QC’d where appropriate, especially for Table 2 so that it requires a bit less referencing to Table 1.
Figure 6:
- Label individual panels a/b/c/d
- For panel c (flight line map), the uncertainty of the DTU line should be shown as background fill.
Citation: https://doi.org/10.5194/essd-2023-442-RC1 -
RC2: 'Review of Karlsson et al. (2024; essd-2023-442)', Julien Bodart, 02 Apr 2024
Dear Editor and Authors,
Please find attached my review of Karlsson et al. (ESSD manuscript number: essd-2023-442) with title “A Newly Digitised Ice-penetrating Radar Data Set Acquired over the Greenland Ice Sheet in 1971-1979”.
With best wishes,
Julien Bodart
==========================================================================================================
Summary and overall comment
In their paper, Karlsson et al. discuss the digitisation and data release that they undertook on legacy airborne radar data acquired over the Greenland Ice Sheet in the 1970s. They discuss in great detail the methods used to digitise and quality check the data in order to provide a robust dataset that can be used by scientists to observe decadal-scale changes over the Greenland Ice Sheet. This work is extremely valuable, as satellite data does not go back long enough to observe such decadal changes. In combination with the release of modern radar data, the release of legacy data as done here now enables such comparisons over multi-decadal timescales, as previously conducted over the Antarctic Ice Sheet by Schroeder et al., 2019, amongst others.
I had a really great time reading the paper and diving into the legacy reports and data that are provided alongside the manuscript. I found the paper well written and for the most part (except see general comment below regarding the methods), well structured. I also found the figures very useful and easy to interpret, even to a non-radar audience. The dataset is robust, easily accessible, and relatively well described in the paper.
As it stands, I would recommend very minor revisions to the paper to address the points highlighted below. I look forward to seeing this important paper and associated dataset published in Earth System Science Data.
==========================================================================================================
General comments
- The abstract mentions that data are available at both high resolution (through Stanford Repository) and low resolution (through DTU). It was confusing to me reading this why you would provide both at different locations, so I wonder whether something could be done in the abstract to explain, in a very brief way, why such difference exists. Otherwise, readers reading this for the first time may, like me, spend a good part of the paper wondering why this was done (until one reaches Section 3.1 and 3.2 where it is explained why the two formats exist).
- I would recommend moving Section 3.2 much higher in the manuscript. This links to my previous comment regarding the confusion about the low and high resolution files, and also because until the reader reaches Sections 3.1-3.2, multiple mentions of different resolutions, files, and methods used (e.g. QC of the data) are provided without appropriate context. This made the reading a little difficult to follow at first, and I spent a bit of time having to untie some knots and connect the dots by scrolling up and down the paper several time until I could understand the methods used (e.g. in the georeferencing). See my line-by-line comments below for more clarification.
==========================================================================================================
Line-by-line comments:
- Line 5: “Our digitisation of the data” – replace by “Our data digitisation”.
- Line 26: “and the NASA Operation IceBridge began in 2009.” – I found the end of this sentence a bit abrupt. Consider adding: “began in 2009 for approximately ten years in order to bridge the gap between NASA’s ICESat satellite missions (see MacGregor et al., 2021)”. Reference: https://doi.org/10.1029/2020RG000712
- Line 28: “is the easily accessible data format” – rephrase for “is the ease of access to the data”. I would argue that the words “easily accessible” can be nuanced here, primarily because even though the data is easy to find, the data format of the OIB surveys (i.e. matlab files) do not technically comply with the FAIR data standards, since MATLAB is a proprietary software.
- Line 28-29: “where the immediate digitisation of the data made it possible to carry out large-scale analyses of ice properties” – this sentence is a bit confusing to me. Could you rephrase as follows: “ease of access to the data, which in turn enabled large-scale analyses of ice properties.”
- Line 33: “radargrams (“Z-scopes”)”. Add mention of Fig. 3 here when mentioning Z scopes for the first time here.
- Line 34: Add new paragraph between Line 33 and Line 34
- Line 43: Add Schroeder et al. 2019 reference here after “West Antarctica between 1978 and 2009.”
- Lines 43-44: starting with “The high quality of our newly“- I would recommend to move this sentence to Lines 34-35 to replace the current sentence there (which starts with “The digitisation will make it possible to expand”). I think it’s a much better introductory sentence which would link up well with the following sentences and I’m not sure it has much value where it currently is at the end of the paragraph.
- Table 1: Perhaps explain briefly in the caption why some “Flight lines scanned” are marked as “Unknown”. And refer to the section which discusses this.
- Table 1 caption: “Please refer to Table 2 for a list of” – this is the first instance since the abstract that the quality check is mentioned, and it clarifies a little bit what has been done. However, I think it’s a bit too low in the manuscript and should be explained briefly when first mentioned (see general comment above).
- Line 62: “(or rather in two-way travel time)” – replace “or rather in” with “in radar”.
- Lines 73-77: I found this paragraph to read a bit out of place. I think it’s important to mention, but it should be placed further up in the introduction section in my opinio.
- Line 98: “landmarks” – can you be more specific? i.e.: “such as basal channels and bedrock features”
- Line 110: “on the data website” – which one? The Stanford repository or the DTU? Be specific
- Lines 115-118: I’m not sure I understand the difference between data that has been qced with geographic information (80%; orange line) and data that has been qced and georeferenced (red line)? Is the difference the sentences following this (Lines 118-121)? This is not super clear to me and is worth expanding or clarifying.
- Table 2: The caption mentions cross-polarised flights were conducted, but no mention of this is present in the text. I also see that there is a file “1978_fl111_crosspol”. I have two comments on this: (1) could you provide more details on the extent of those cross-polarised flights that were conducted as part of these 1970s surveys discussed in the paper? And (2) what is meant in the caption by “Geographical information is not yet available for the cross-polarized test flight” – the word “yet” implies this is work in progress? If so, how could users access this information in the future?
- Line 140: what is meant by “users will report”? Maybe clarify this point.
- Lines 149-150: Consider reversing the order of the variables mentioned in the text so that they match the csv header order (year, flightline number, and cbd number)
- Section 3.2: I found this section really useful, but it comes much too late in the paper. As my comments above (general and line-by-line) indicate, I was a bit confused with the structure of the text, particularly at the start when mentioning different resolutions in the files and different methods used (i.e. the qaqc which was applied only on a subset of the data). I had to switch back and forth between different sections of the paper to link them together. As a result, I think Section 3.2 should be moved before Figure 2 and Table 2, at least. This is particularly true for Lines 148-150 which should also definitely be moved before Section 2.2.1, at least.
- Line 153: “For selected flight…” – why not for all? Is it because of this sentence in the caption of Table 1 higher up in the paper: “particularly for the years 1971 and 1972 where the film rolls were not imprinted with year and flight number”? If so, please specify this again as this sentence can be overlooked at the start of the manuscript.
- Line 155: Ah ok! So this is what the red lines are. This partially addresses my comment above. Again, I would recommend moving this section higher up and make sure Figure 2b follows this explanation. I would also recommend adding at the end of Line 132 that this sentence is for the orange line in Figure 2b (if I understood completely!). I also now note that the caption of Table 2 explains the difference between qced and georeferenced or qced with geographic information, but this comes after having read this section multiple times and trying to tie the knots together.
- Line 163: “regarding frequency. The film rolls” – join both sentences together; i.e. “regarding frequency, as the film rolls”
- Line 181: Figure 8 is mentioned before Figure 7 in the text.
- Line 187 and end of this paragraph: Consider adding “It is worth noting that such issues can also be present in modern radar datasets”. I think it’s important to note that these can be common issues that are not necessarily down to your specific dataset, or the acquisition methods used at the time.
- Figure 6: Provide sub-panel numbering for each sub-figure. Link these too in the caption (i.e. “modern day radar data (a) and the recently digitised DTU data (b)”).
==========================================================================================================
Citation: https://doi.org/10.5194/essd-2023-442-RC2
Data sets
A Newly Digitised Ice-penetrating Radar Dataset Acquired over the Greenland Ice Sheet in 1971-1979. N. Karlsson et al. https://doi.org/10.25740/wm135gp2721
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 331 | 79 | 24 | 434 | 18 | 20 |
- HTML: 331
- PDF: 79
- XML: 24
- Total: 434
- BibTeX: 18
- EndNote: 20
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1