Calculating uncertainty for the RICE ice core continuous flow analysis water isotope record

Keller, Elizabeth D.; Baisden, W. Troy; Bertler, Nancy A. N.; Emanuelsson, B. Daniel; Canessa, Silvia; Phillips, Andy

doi:https://doi.org/10.5194/amt-11-4725-2018

Articles | Volume 11, issue 8

https://doi.org/10.5194/amt-11-4725-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-11-4725-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 11, issue 8

Research article

|

13 Aug 2018

Research article |

| 13 Aug 2018

Calculating uncertainty for the RICE ice core continuous flow analysis water isotope record

Elizabeth D. Keller, W. Troy Baisden, Nancy A. N. Bertler, B. Daniel Emanuelsson, Silvia Canessa, and Andy Phillips

Download

Final revised paper (published on 13 Aug 2018)
Supplement to the final revised paper
Preprint (discussion started on 09 Jan 2018)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review Keller et al: Calculating uncertainty for the RICE ice core continuous flow analysis water isotope record', Anonymous Referee #1, 13 Mar 2018
- RC2: 'typo in review', Anonymous Referee #1, 13 Mar 2018
  - AC2: 'Authors' response to RC#2', Elizabeth Keller, 31 May 2018
- AC1: 'Authors' response to RC#1-3', Elizabeth Keller, 31 May 2018
RC3: 'Review, Keller et al.', Anonymous Referee #2, 05 Apr 2018
- AC1: 'Authors' response to RC#1-3', Elizabeth Keller, 31 May 2018

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Elizabeth Keller on behalf of the Authors (31 May 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (04 Jun 2018) by Frank Keppler

RR by Anonymous Referee #3 (23 Jun 2018)

Suggestions for revision or reasons for rejection

Review of “Calculating uncertainty for the RICE ice core continuous flow analysis water isotope record” by Keller E. D. et al. for publication in AMT.

General:
The manuscript deals with an important and very valuable topic to evaluation the uncertainty of a measuring system for water stable isotopes. Indeed the continuous determination of these parameters has increased considerably over the last years due the availability of high precision laser-based instruments. Especially in view of the upcoming new ice core projects in particular the search for the oldest ice, the uncertainty determination has one of the highest priorities together with the smoothing and smearing of isotope signals due to the melting process and the passing through the analyser cavity cell. Unfortunately, both issues were not included in this manuscript, which is indeed a drawback of this manuscript.

Yet, I suggest publication once the following points have been addressed:
Major points:

- Based on the two reviewers and the corresponding replies from the authors the manuscript has significantly been improved. I very much acknowledge the comments given by the other reviewers and I agree with them. In particular, it is important to be concise and correct with the description of the calibration approach. This has significantly been improved in the revised version. Yet, there are still a couple of remaining points see below.

- Due to time constraints I could not check all the raised issues, in particular by reviewer 1, in detail, but as far as I did so the replies by the authors are acceptable.

Specific remarks:
p. 1, l. 26: replace the equal sign with the word equal, otherwise you should use a second R for hydrogen ratio.
p.1, l. 26: either use per mil or the per mil symbol
p. 4, l. …260 m in 2014 (delete point)
p. 4, l. 2-3: The following sentence is not clear to me, so what has been matched to 1 Hz?. “The change in 2014 was made to match the depth recording rate in both years (1 Hz).”
p. 4, l. 28: Why a ratio? This is the H2O concentration.
p. 4, l. 33: sigma, s not defined. (at least as far as I could see)
p. 6, l. 5: I can understand that different batches can have different isotope ratios. Each batch has been analysed and calibrated through three secondary standards to the official IAEA standards. If this has been done, why were you not concerned about the large differences? Why do you not report the uncertainty of the assigned values of the standards in use (RICE, ITASE, WS1) for 2014, see table 1.
p. 7, l. 1: I would rather use “assigned” than “true” values.
p. 7, l. 8: Why do you make this assumption? You also could take the calculated value from the regression. What were your considerations? How would it affect the results when the regression approach would be used for the offset b.
p. 7, l. 15: Here is something wrong for the definition of f, t should read (t-t1). Therefore, f become either 1 for t=t2 or 0 for t=t1.
p. 7, l. 25: I would list it in the same order as introduced just before!!!!
p. 8, l. 2: believe is not science. What are arguments for your way of thinking regarding this interactions?
p. 8, l. 13: Again there seems to be an error here. Tau is the averaging time, but tau, n is tau x n times, which is the total observation time. Here in the formula you should use tau and not tau, n and say that (j+1) and j are dependent on tau.
p. 9, eq. 7: I would add the base of the logarithm in use which is e, i.e. you are using the logarithm naturalis.
p. 9, l. 8: How do you do that time-weighted average?
p. 9, l.9-10: This makes sense but how have you implemented it, mathematically.
p. 9, l. 16: but this moving window uses only independent values, that means it jumps by 15 seconds from i to i+1. Is my statement correct? If so, you may add a clarification here.
p. 10, l. 2-5: Why unsystematic, it could also be a systematic error due to a mis-assigned standard value. In other research fields the WS1 is often called a target value, which checks the quality of calibration procedure applied.
p. 10, l. 7-8: I would change the two words here from the “measured corrected value” to “corrected measured value”.
p. 11, l. 1-2: This means that you have used all three standards that were in use in 2013 completely? This sounds strange.
p. 11, l. 5-6: Well you can do this by calibrating with WS1 and ITASE and use RICE as target.
p. 11, l. 13: The first sentence of this paragraph can be deleted, since it is the same message as the first sentence of the paragraph above. These two paragraphs can be combined.
p. 11, l. 21: anectodal: This is not a science expression
p. 11, l. 26-27: …but you exactly tried this in this manuscript. I would skip this last sentence.
p. 12, l. 3: Nothing has been said about the smoothing and smearing of the ice data due to the melting procedure and due to the cavity volume smoothing. This could be as important as the error estimation done here.
p. 12, l. 9-10: One could check the linear approach of Allan deviation error based on repeated sequences of the three standards in use but use only part of the standards for calibration and check the drift of the instrument through this approach. Hence, the time dependence (linearity as assumed by the authors) of such an approach could be investigated.
Figures:
Caption Figure 4: Allan error (use Allan variance error as used in 2.4.1)
Figure 6: I would change the x-axis and show a smaller spacing.
Table 1: Why are there no uncertainties given for the standards in 2014.
Table S2: add the ITASE values as well!!
Figure S3: Add the uncertainty for the green star values (or say that it is within the symbol size).
Figure S3: This is indeed a problem. Why is the green star far from the measurements? Has the system been setup differently during the green star measurement and the actually field data measurements.
Figure S1: Has the green value been determined in the lab or during the field campaign. Check only the difference between the WS1, RICE and ITASE. This might help.

Hide

ED: Publish subject to minor revisions (review by editor) (01 Jul 2018) by Frank Keppler

AR by Elizabeth Keller on behalf of the Authors (21 Jul 2018) Author's response Manuscript

ED: Publish subject to technical corrections (25 Jul 2018) by Frank Keppler

Short summary

We describe a systematic approach to the calibration and uncertainty estimation of a high-resolution continuous flow analysis (CFA) water isotope (δ²H, δ¹⁸O) record from the Roosevelt Island Climate Evolution (RICE) Antarctic ice core. Our method establishes robust uncertainty estimates for CFA δ²H and δ¹⁸O measurements, comparable to those reported for discrete sample δ²H and δ¹⁸O analysis. The resulting mean total errors for the record are 0.74 ‰ and 0.21 ‰ for δ²H and δ¹⁸O, respectively.