Prediction of monthly Arctic sea ice concentrations using satellite and reanalysis data based on convolutional neural networks

Kim, Young Jun; Kim, Hyun-Cheol; Han, Daehyeon; Lee, Sanggyun; Im, Jungho

doi:https://doi.org/10.5194/tc-14-1083-2020

Articles | Volume 14, issue 3

https://doi.org/10.5194/tc-14-1083-2020

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

https://doi.org/10.5194/tc-14-1083-2020

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

Articles | Volume 14, issue 3

Research article

|

25 Mar 2020

Research article |

| 25 Mar 2020

Prediction of monthly Arctic sea ice concentrations using satellite and reanalysis data based on convolutional neural networks

Young Jun Kim, Hyun-Cheol Kim, Daehyeon Han, Sanggyun Lee, and Jungho Im

Download

Final revised paper (published on 25 Mar 2020)
Preprint (discussion started on 05 Aug 2019)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'Referee comment on "Prediction of monthly Arctic sea ice concentration using satellite and reanalysis data based on convolutional neural networks"', Anonymous Referee #1, 03 Oct 2019
- AC1: 'Authors's response to referee #1', Jungho Im, 15 Nov 2019
RC2: 'Review', Anonymous Referee #2, 14 Oct 2019
- AC2: 'Authors' responses to the referee #2', Jungho Im, 15 Nov 2019

Peer-review completion

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

ED: Reconsider after major revisions (further review by editor and referees) (26 Nov 2019) by David Schroeder

AR by Jungho Im on behalf of the Authors (26 Nov 2019) Author's response Manuscript

ED: Referee Nomination & Report Request started (29 Nov 2019) by David Schroeder

RR by Anonymous Referee #2 (07 Jan 2020)

RR by Anonymous Referee #3 (24 Jan 2020)

Suggestions for revision or reasons for rejection

This manuscript proposes convolutional neural networks model for one-month prediction of Arctic sea ice concentration (SIC) fields. The model is developed by using observed satellite SIC fields and oceanic/atmospheric information from reanalysis datasets. This work is an interesting addition to the suite of existing statistical models for prediction of Arctic sea ice, the topic of great interest and importance.

Reading the paper, past reviews and responses, I conclude that there are several key areas where manuscript needs improvement, such as providing more complete list of references, clarifying suitability of developed techniques for predicting sea ice extent as well as on longer lead times, see comments below.

1. Lines 40-45. should mention also these statistical models:

Wang, L., Yuan, X., & Li, C. (2019). Subseasonal forecast of Arctic sea ice concentration via statistical approaches. Climate Dynamics, 52(7), 4953–4971.

Kondrashov, D., M. D. Chekroun, and M. Ghil (2018). Data-adaptive harmonic decomposition and prediction of Arctic sea ice extent, Dynamics and Statistics of the Climate System, 3(1).

2. Authors need to clarify why they have restricted their results to one-month lead time, i.e. for monthly data it is basically one-step prediction. Can the presented method be used for predicting on longer lead times, i.e. several months ahead, such as for summertime prediction in Sea Ice Outlook (see next)? It is not clear why not, and that begs the question why results are shown for one-month only, thus creating an impression of incomplete study.

3. It would be very helpful to illustrate how prediction of SIC translates into the one for sea ice extent (SIE), and in particular for summertime prediction of September minimum of pan-Arctic SIE which is the main focus of Sea Ice Outlook community effort (Stroeve et al. 2014). Adding and showing results for one-month prediction (i.e. from August) and observed September pan-Arctic SIE, as well as skill in comparison with RF and baseline prediction model, would be illuminating.

Hide

ED: Publish subject to minor revisions (review by editor) (04 Feb 2020) by David Schroeder

AR by Jungho Im on behalf of the Authors (11 Feb 2020) Author's response Manuscript

ED: Publish as is (26 Feb 2020) by David Schroeder

AR by Jungho Im on behalf of the Authors (27 Feb 2020)

Short summary

In this study, we proposed a novel 1-month sea ice concentration (SIC) prediction model with eight predictors using a deep-learning approach, convolutional neural networks (CNNs). The proposed CNN model was evaluated and compared with the two baseline approaches, random-forest and simple-regression models, resulting in better performance. This study also examined SIC predictions for two extreme cases in 2007 and 2012 in detail and the influencing factors through a sensitivity analysis.