SICOPOLIS-AD v1: an open-source adjoint modeling framework for ice sheet simulation enabled by the algorithmic differentiation tool OpenAD

Logan, Liz C.; Narayanan, Sri Hari Krishna; Greve, Ralf; Heimbach, Patrick

doi:https://doi.org/10.5194/gmd-13-1845-2020

Articles | Volume 13, issue 4

https://doi.org/10.5194/gmd-13-1845-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-13-1845-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 13, issue 4

Development and technical paper

|

09 Apr 2020

Development and technical paper |

| 09 Apr 2020

SICOPOLIS-AD v1: an open-source adjoint modeling framework for ice sheet simulation enabled by the algorithmic differentiation tool OpenAD

Liz C. Logan, Sri Hari Krishna Narayanan, Ralf Greve, and Patrick Heimbach

Download

Final revised paper (published on 09 Apr 2020)
Preprint (discussion started on 08 Aug 2019)
Supplement to the preprint

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Referee comment on SICOPOLIS-AD', Laurent Hascoet, 03 Sep 2019
- AC1: 'Response to reviewer 1, Laurent Hascoet', Patrick Heimbach, 01 Dec 2019
RC2: 'Referee comment re: SICOPOLIS-OpenAD', Lizz Ultee, 04 Oct 2019
- AC2: 'Response to reviewer 2, Lizz Ultee', Patrick Heimbach, 01 Dec 2019

Peer-review completion

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

AR by Patrick Heimbach on behalf of the Authors (01 Dec 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (17 Dec 2019) by Alexander Robel

RR by Lizz Ultee (02 Jan 2020)

Suggestions for revision or reasons for rejection

General comments:
I thank the authors for their careful response to both reviews, and I commend them for several edits to the manuscript that have improved it over the original submission. The revised manuscript is clearer and more accessible for general readers.

The manuscript presents a promising contribution to ice sheet modelling practice. Pending a few typographical corrections noted below, I support its publication. I look forward to interesting uncertainty quantification studies using the SICOPOLIS-AD tool.

Specific comments:
The authors have, at my suggestion, added text on manuscript P5 to separate previous applications of adjoint sensitivities in ice sheet modelling. I agree with the authors that the purpose of this manuscript is not to provide a review of the topic. However, my original suggestion was to discuss the previous work in more detail so that *any notable contrasts with the present study* (i.e. the novelty of this manuscript) might be drawn out. In my opinion, the new text on P5, L22-33 does better than the original to point interested readers to related work, but it does not quite achieve the aim of highlighting the scientific contribution of this manuscript.

I apologize if my original suggestion was unclear. If the work will be revised again (which I leave to the discretion of the editor/authors), the authors might consider just a little more wordsmithing on this paragraph to better integrate the discussion of previous work with signposting toward the manuscript’s novelty. In this present version, the authors do add some phrasing later in the manuscript that helps highlight novelty and applications. This is acceptable.

Technical corrections:
P10, Table 1: To reinforce the point made in the text, I suggest revising the caption to note that “Point-wise ice thickness, surface, and basal temperatures compare well, with a % deviation of less than 10%.” Patch-wide comparisons have considerably higher % deviation.
P16, L11: The sentence “They invite…science applications” is a great addition, but it is missing a period.
P17, L10: remove “do" in the revised phrase “…have much more potential to do have a lasting impact…”
P17, L28: “and compute” for verb tense agreement with preceding “We demonstrate…”
P17, L29: Missing comma(s) in “Examining, ascertaining, and understanding…"

Hide

RR by Mauro Perego (20 Jan 2020)

ED: Publish subject to minor revisions (review by editor) (21 Jan 2020) by Alexander Robel

AR by Patrick Heimbach on behalf of the Authors (25 Feb 2020) Manuscript

ED: Publish as is (03 Mar 2020) by Alexander Robel

AR by Patrick Heimbach on behalf of the Authors (09 Mar 2020)

Short summary

A new capability has been developed for the ice sheet model SICOPOLIS (SImulation COde for POLythermal Ice Sheets) that enables the generation of derivative code, such as tangent linear or adjoint models, by means of algorithmic differentiation. It relies on the source transformation algorithmic (AD) differentiation tool OpenAD. The reverse mode of AD provides the adjoint model, SICOPOLIS-AD, which may be applied for comprehensive sensitivity analyses as well as gradient-based optimization.