the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Oct 2024
Data-driven rolling model for global wave height
Abstract. Significant Wave Height (SWH) is crucial for many human activities, such as marine navigation, offshore operations, and coastal management. Traditionally, SWH is modeled using numerical wave models, which, while accurate, are computationally intensive and constrained by incomplete physical representations of wave spectral evolution. This study introduces a simple global deep learning-based model for SWH, which uses the current SWH field and the wind field at the next time step as inputs to predict the SWH field at the next time step. This approach mirrors the rolling prediction strategy of numerical wave models. After training on a re-analysis dataset, the errors of the model diverge lightly with time when given a good initial field because no spectral information is used. However, after diverging for ~200 hours, the errors stabilize, remaining comparable to those of state-of-the-art numerical wave models. Additionally, the error divergence can be mitigated through the assimilation of altimeter measurements. This deep learning model can not only serve as an efficient surrogate for traditional numerical wave models but also provide a baseline for statistical modeling of global SWH due to its simplicity in inputs and outputs.
- Preprint
(8455 KB) - Metadata XML
-
Supplement
(2234 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2024-181', Anonymous Referee #1, 19 Nov 2024
Through adopting a rolling  strategy, an artificial intelligence model is established by using a U-Net architecture, and global SWH prediction is produced by the model after trained using ERA5 data.
Specific comments:Â
1. I just wonder if  the RMSEs in Figure 2c and Figure 4d are smaller than that of climatology.
2. For the sake of comparison, the colorbar scales in  Figures 6-7 should be the same as those used in Figures 4-5.
3. The units of variables on the vertical axis are missing in (b) and (c) in Figures 2-3.Citation: https://doi.org/10.5194/gmd-2024-181-RC1 -
AC1: 'Reply on RC1', Xinxin Wang, 22 Nov 2024
We thank reviewer 1 for these insightful and useful comments. Please find the reply in the attached pdf.
-
RC2: 'Reply on AC1', Anonymous Referee #1, 22 Nov 2024
The authors have satisfactorily addressed my previous concerns. I recommend acceptance for publication.
Citation: https://doi.org/10.5194/gmd-2024-181-RC2
-
RC2: 'Reply on AC1', Anonymous Referee #1, 22 Nov 2024
-
AC1: 'Reply on RC1', Xinxin Wang, 22 Nov 2024
-
CEC1: 'Comment on gmd-2024-181', Juan Antonio Añel, 28 Nov 2024
Dear authors,
Unfortunately, after checking your manuscript, it has come to our attention that it does not comply with our "Code and Data Policy".
https://www.geoscientific-model-development.net/policies/code_and_data_policy.htmlWe have detected several major flaws in your manuscript, and In this way, if you do not fix these problems in a prompt manner, we will have to reject your manuscript for publication in our journal. Actually, with such flaws your manuscript should have never been accepted in Discussions. Therefore, the situation with your manuscript is irregular, as we can not accept manuscripts in Discussions that do not comply with our policy.
First, you have archived your code on GitHub. However, GitHub is not a suitable repository for scientific publication. Our policy mentions it clearly. GitHub itself instructs authors to use other long-term archival and publishing alternatives, such as Zenodo, PANGAEA, FigShare, etc. Please, publish your code in one of the appropriate repositories and reply to this comment with the relevant information (link and a permanent identifier for it (e.g. DOI)) as soon as possible. Also, in the GitHub repository no license is listed. If you do not include a license the code remains your property and nobody can use it. Therefore, when uploading the model's code to the new repository you could want to choose a free software/open-source (FLOSS) license. We suggest the GPLv3 or similar ones. For the GPLv3 you simply need to include the file 'https://www.gnu.org/licenses/gpl-3.0.txt' as LICENSE.txt with your code. Also, you can choose other options: e.g., GPLv2, Apache License, MIT License, etc.
Second, in your manuscript you state "The WW3-ST6 dataset is available from Liu et al (2021)". Actually this statement is useless, as Liu et al. (2021) only say that the dataset is made available by contacting the authors by email, something explicitly forbidden by the policy of our journal, which clearly states that all the assets necessary to replicate a work must be published openly in a permanent repository before submission. Therefore, you must publish the WW3-ST6 dataset in a suitable repository, and again reply to this comment with the corresponding link and DOI.
Also, being this submission about a deep-learning model it is necessary that you provide a repository with the specific data that you have used to train your model, not generic mentions to the full datasets used. The same is of application for the output data. This information is necessary to replicate your work. The same rules on DOI and link that I commented above for the other datasets apply here.
Finally, you must include the modified 'Code and Data Availability' section in a potentially reviewed manuscript, with the new DOIs and links for the requested repositories. As I have said, given the lack of necessary information to correctly evaluate your manuscript, and that it was accepted in Discussions and undergone peer review despite it, you must address quickly all this issues.
Juan A. Añel
Geosci. Model Dev. Executive Editor  ÂCitation: https://doi.org/10.5194/gmd-2024-181-CEC1 - AC2: 'Reply on CEC1', Xinxin Wang, 29 Nov 2024
-
RC3: 'Comment on gmd-2024-181', Anonymous Referee #2, 25 Dec 2024
Major comments:
1. It is not clear what goal this AI SWH model would like to achieve. Is it to make good 0-10 days SWH predictions similar to traditional numerical wave models? If yes, we should use predicted wind fields from GFS (or IFS) as the forcing instead of the ERA5 reanalysis wind fields. In the current manuscript, one year of SWH series was generated by the AI SWH model. But it essentially functions like post-processing the ERA5 wind analysis to 0-6h short tem SWH forecasts. This is not the regular 0-10-day SWH forecasts we normally expect.
2. Section 2.4 and other relevant parts: (1) data assimilation will improve initial conditions, but it is not related to the establishment of this AI SWH model and it is suggested to remove this part from this manuscript; (2) the data assimilation method here is too simple. Nowadays, we would expect at least a 2DVAR method that considers the uncertainties of the background and the observations.
Comments:Â
Lines 48-49: "However, …" The statement is not accurate. The SWH prediction is both a forcing and an initial value problem. When focusing on the short-term (a few hours to a day) forecasts, the initial value will have a large impact. Zhou et al., 2021 and Ouyang et al., 2023 targetted 24 h and 3 days forecasts respectively.ÂLine 124-125: "Particularly, if a long series…" Could you clarify which part of Song and Jiang, 2023 drew this conclusion?
Lines 133-134: "We believe…": Evidence, instead of a subjective "belief", is expected here to demonstrate why U-Net is suited for this work.
Lines 135-136: "The processes of both…": In AI for NWP practices, lots of literature has demonstrated that AI struggles to resolve different scales at the same time (for example, smoothing to get better medium-range forecasts and hence not able to resolve smaller scales features). So how can it be assured that "The processes of both local wave generation by wind and wave propagation in space can be captured by convolutional kernels at different scales“? More discussions on this are needed.
Lines 139-142: It looks like there is a problem with the "-190 degree to 190 degree" trick: How do waves at 180 degrees propagate to -179 degrees in this method?
 Line 161-168: (1) How does the epoch ensemble method work? Need more theoretical discussions here. Generally, we would like to use all available data to train the best AI model (instead of splitting them into different epochs). (2) We can add different perturbations to generate ensembles or use ensemble wind forcings (such as from GEFS)  (3) The reduction of RMSE from the 4 ensembles shown in the manuscript may come from the smoothing effect.Fig 2: It looks like the "AI model with data assimilation" results are NOT free forecasts but analyses performed every 6 hours along with short-term (0-6h) forecasts.Â
Line 220: "This suggests that the simple AI model can function independently, at least, in certain scenarios." What does this mean and what scenarios it refers to? Need clarifications.
Fig. 3, 4, 5: Are the results in these figures from the AI model with or without data assimilation?
Line 378-379: "why the AI model can slightly outperform the NWM in these areas." If we want to draw this conclusion, we will need to run both the NWM and AI models with the same settings and compare them directly.
LInes 386-390: The initial value problem has a predictability limit (around 10 days for the wave forecasts). So one would not expect the impact of initial SWH will last beyond ~10 days.
Line 423: "An important advantage of the AI SWH model proposed here is its low computational cost compared to traditional NWMs.": It will be good if we can give concrete examples of the computational resources needed by the AI SWH model and the traditional NWMs.
Lines 445-446: "While training such a model would be challenging, it is not an impossible task, and the rapid advancements in AI may make this goal more achievable in the future": This sentence is too subjective, Consider revising.
Edit:Â
Line 304: "clearly evident" Â -> Â "evident"Â
Citation: https://doi.org/10.5194/gmd-2024-181-RC3 - AC4: 'Reply on RC3', Xinxin Wang, 31 Dec 2024
- AC3: 'Comment on gmd-2024-181', Xinxin Wang, 31 Dec 2024
Data sets
AI-Rolling-Wave-Height-Model Xinxin Wang https://github.com/YulKeal/AI-Rolling-Wave-Height-Model
Video supplement
Supplementary Movies S1 & S2 Xinxin Wang https://drive.google.com/drive/folders/1zo-HOSLDrLjMSCDiwWIf9I37pFYPG3Zt?usp=sharing
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 464 | 75 | 18 | 557 | 38 | 5 | 6 |
- HTML: 464
- PDF: 75
- XML: 18
- Total: 557
- Supplement: 38
- BibTeX: 5
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1