the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Combined GNSS reflectometry/refractometry for automated and continuous in situ surface mass balance estimation on an Antarctic ice shelf
Ladina Steiner
Holger Schmithüsen
Jens Wickert
Olaf Eisen
Abstract. Reliable in situ surface mass balance (SMB) estimates in polar regions are scarce due to limited spatial and temporal data availability. This study aims at deriving automated and continuous specific SMB time series for fast moving parts of ice sheets and shelves (flow velocity > 10 m a-1) by developing a combined Global Navigation Satellite Systems (GNSS) reflectometry and refractometry (GNSS-RR) method. In situ snow density, snow water equivalent (SWE), and snow deposition or erosion are estimated simultaneously as an average over an area of several square meters and independent on weather conditions. The combined GNSS-RR method is validated and evaluated regarding its applicability on a moving, high latitude ice shelf. A combined GNSS-RR system was therefore installed in November 2021 on the Ekström ice shelf (flow velocity≈150 m a-1) in Dronning Maud Land, Antarctica. Reflected and refracted GNSS observations from the site are post-processed to obtain snow accumulation (deposition and erosion), SWE, and snow density estimates with a 15 min temporal resolution. Results of the first 16 months of data show a high level of agreement with manual and automated reference observations from the same site. Snow accumulation is derived with an uncertainty of around 9 cm, SWE around 40 kg m−2 a−1, and density around 72 kg m−3.
This pilot study forms the base for extending observational networks with GNSS-RR capabilities, in particular in polar regions. Regional climate models, local snow modelling, and extensive remote sensing data products will profit from calibration and validation based on such in situ time series, especially if multiples of such sensors will be deployed over larger regional scales.
- Preprint
(3751 KB) - Metadata XML
- BibTeX
- EndNote
Ladina Steiner et al.
Status: final response (author comments only)
-
CC1: 'Comment on tc-2023-89', David Shean, 14 Jul 2023
Great to see this preprint in TCD! This looks like a really nice experimental setup, offering an important demonstration and evaluation of GNSS snow measurements in Antarctica.
There are a few additional relevant papers in the literature involving GNSS-Reflectometry for SMB of Antarctic ice shelves and ice streams:
Shean et al. (2017) GPS-derived estimates of surface mass balance and ocean-induced basal melt for Pine Island Glacier ice shelf, Antarctica, https://tc.copernicus.org/articles/11/2655/2017/tc-11-2655-2017.html
Seigfried et al (2017), Snow accumulation variability on a West Antarctic ice stream observed with GPS reflectometry, 2007–2017: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2017GL074039These papers offer some context, identify limitations, and provide further justification for the refractometry. I recommend you give them a read, and consider citing.
I have not had a chance to do a detailed review, but I believe your experimental setup provides the (potentially unique) opportunity to characterize potential bias in snow measurements extracted from GNSS-Reflectometry alone. While I agree that GNSS-RR is a better option (assuming field support and equipment resources are available), there is also value in opportunistic reflectometry-only approaches leveraging existing archives of GNSS data collected by receivers deployed for other purposes (like ice motion). In other words, if you didn't have the refractometry, is there still value in the reflectometry results from your reference antenna alone?
I may have missed it, but it would be valuable to report the depth of the tower (or "sensor mast") in the firn at the start of the experiment. One important question is the depth of "bonding" between the tower and the firn, and whether the receivers rigidly mounted to the tower are experiencing relative downward motion due to compaction within upper or deeper layers of the firn column. Twit Conway has anecdotes about GNSS antenna poles near South Pole station penetrating several 10s of cm through plywood sheets due to differential firn compaction rates.
I believe your upward-looking "rover" antennas are rigidly attached to the same tower as the reference antenna, with an assumption that the rovers remain fixed relative to the original "firn surface" from the start of the experiment. If the tower (and all receivers) are moving downward at the same rate as the original firn surface, all is well, but if not (due to the tower bonding in deeper layers), then your "rover" antennas will be pulled below the layer corresponding to the original firn surface, and your derived measurements will also include upper layers of firn instead of just new snow accumulation. Hopefully that makes sense. If you can demonstrate that this is not an issue, that would be a nice addition!
Citation: https://doi.org/10.5194/tc-2023-89-CC1 - AC3: 'Reply on CC1', Ladina Steiner, 25 Sep 2023
-
RC1: 'Comment on tc-2023-89', Ian Brown, 05 Aug 2023
General Comments:
In my opinion this is a valuable and well written contribution and the athors are to be congratulated: I very much enjoyed reading this manuscript.
The authors describe an experiment that aims to test combined GNSS reflectometry and refractometry to derive snow water equivalent (SWE), snow density and surface mass balance (SMB) at an Antarctic ice shelf site. The method described is largely automatic, high resolution in space and time, and relatively low cost. The research topic is very important as the authors note. The method offers an alternative to labour intensive and destructive manual measurements or (incomplete) automatic weather station approaches. It has great potential and the article is both highly relevant to the journal and of high quality.
Specific Comments:
The main challenge the authors faced in comparing different measurement series was the difference in the footprint of the individual measurements and, therefore, the impact of surface roughness on the results. As a reader I would like to see this issue discussed further, ideally supported by some data. The authors could add a more rigorous analysis of uncertainties. At present it is more or less covered by two sentences (section 5. line 252-257). I would also like to know if outliers were filtered from the laser data? Lasers are subject to errors caused by blowing snow at the snow surface.
Why do the high-end and low cost receiver time series diverge after August 2022? I think a hypothesis is needed (section 4.2, line 194). I note that according to figure 6 the low-cost data better tracks the manual density measurements after Aug. 2022 (section 4.3, line 228). This is not mentioned in the text and this omission should, in my option, be addressed. I would also expect the high-end receiver to better correlated with the laser data as the former seems directly under the latter while the low-cost antenna is further from the laser footprint.
Section 4.1: I would like to know if there are relationships between GNSS satellite zenith and aspect, and the errors. Is there are directional component, related to, for example, the effect of the mast and prevailing wind on surface roughness? Should tiltmeters and power monitoring be added to the setup?
Would a comparison of your results and accuracies from different measurements (manual density, SWE; surface height/accumulation from lasers and sonic rangers) be appropriate using the literature? Perhaps as a short paragraph at the end of the discussion to provide additional context.
Technical Corrections:
In several places, especially in sections 2 and 3 you use the term "ground" to describe the ice shelf surface. Please use the term "surface" as it is not strictly speaking ground.
In section 2.1 it would be useful to know which reference ellipsoid was used: was it the same for all instruments? If not could there be an effect on data quality?
p4. l80: could the arm for the buried GNSS antennas bend? Was it rigid?
p5. l10; eq. 1: should h not be δh if it is a difference or change?p.6 fig. 3. For the sake of completeness please explain the difference between cyan and mauve ellipses or what they represent (which antennas they represent). How were they calculated?
p7. eq. 3. Please remind the reader that m is SWE and b is accumulation. I have a short memory.
p9. fig. 4. (also figures 5 and 6): would it not be better to have sub-figures a) and b) across the top and c) and d) along the bottom of these figures? This might be pedantry, I know.
Citation: https://doi.org/10.5194/tc-2023-89-RC1 - AC1: 'Reply on RC1', Ladina Steiner, 25 Sep 2023
-
RC2: 'Comment on tc-2023-89', Maaria Nordman, 15 Aug 2023
The manuscript shows an interesting and well done study in the growing field of using GNSS for all kinds of environmental research and monitoring. The manuscript is well written and describes the set-up, methods and results mostly adequately. In general, I don’t see any major issues with the manuscript, but some details should be clarified.
What is the reference antenna? It might be the same Leica AS10 as the high-end rover, but it should be mentioned. Same antennas usually give results matching better, thus some of the difference between high-end and low-cost rovers might stem from this. Also the antenna calibration might be a thing to check, if you want to improve the results, at least no calibration tables were mentioned in the text.
How is the mast attached to the moving ice, i.e. how is the stability of the monument ensured? I was also wondering, along with a colleague earlier, how you can ensure that the mast or the lever is not deforming inside the ice?
One thing that could be added would be the number of satellites observed by the GNSS antennas. How many satellites are left when the elevation angle is limited for the reflectometry? And how many satellites there are generally available? This is one factor affecting the quality of the results as well and could maybe explain if there were problems with the data.
Some terminology is used abundantly, for example, line 129 “vertical Up (height) component U”. For a geodesist, one would be enough, up component, height, or vertical component. Also the symbol m seems to mean both mass (in equations) and the slope (in tables).
I agree with colleague on the points regarding 1) the order of figures 4 and 5, a and b on top row and c and d below is much easier to read, and 2) some explanation why the results seem to lie nicely on top of each other before August 2022 and diverge after that.
Citation: https://doi.org/10.5194/tc-2023-89-RC2 - AC2: 'Reply on RC2', Ladina Steiner, 25 Sep 2023
Ladina Steiner et al.
Ladina Steiner et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 253 | 109 | 23 | 385 | 6 | 6 |
- HTML: 253
- PDF: 109
- XML: 23
- Total: 385
- BibTeX: 6
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1