Review of

"The Met Office operational wave forecasting system: the evolution of the Regional and Global models"

a paper by Valiente and six co-authors, under review for the journal GMD

Summary

The paper describes and evaluates four wave models run by the UK Met Office.

General comments.

The manuscript would be excellent as a technical report, but has been submitted as a journal article, so I review it in that context.

It is not yet suitable as a journal article. It is too long, and too much of the content will not be of much interest to readers beyond the author list.

The paper describes and evaluates four different models. This is too ambitious for a paper, unless it is a glossy "overview" type of paper. 

The abstract is too long.

The last section is too long. The reader should be able to find quickly what has been accomplished in the paper, without having to sort through a lot of discussion. Discussion and conclusions should be separated, and the conclusions should be a bald summary of what was found/accomplished. 

The English is excellent grammatically, and the paper is well-proofed (I found only one typo) but the writing appears rushed and haphazard in places. The sentence connectors do not always work (e.g. see how the word "hence" is used). There are references to things that are not explained (and should be explained or not mentioned). And, paradoxically, too much detail in many places, with much of the text unbearably laden. Lines 245-250 are a good example.

I recommend that the authors start with a list of interesting and novel findings from their work, and use that as a basis for the paper.

It may help to think about how much time an average reader will devote to reading the paper. One hour? Two? Start there, and then consider how to best present the most useful information. All six co-authors should read the manuscript carefully and think about how the text can be more considerate of the readers' situation. Find ways to draw attention to the most useful parts, so that they are not missed by a reader who only wants to spend a very short time with the paper (e.g. 10 minutes).

I can easily imagine that there is something accomplished here which is suitable for a peer-reviewed journal article. For example, the impact of currents on model accuracy. I hope that the authors can demonstrate this. I also hope that the authors recognize that many of us (S&E professionals) spend a lot of time writing, but not everything we write should necessarily be submitted for peer review. Some can be unpublished. Some can be published in technical reports. Some can be submitted for peer-review. As authors, we need to down-select and curate, and we should be careful not to impose unnecessary burden on the volunteer review system.

I estimate that the paper is between 13000 and 13500 words from abstract through conclusions. I suggest that 8500 words be used as a upper limit, though of course, a good paper could be shorter or longer based on the quantity of interesting findings and useful information.

The model description (section 2.2) needs to be checked by someone that is deeply familiar with the model features being described. I got the impression that the writer of this section is very familiar with the model's technical implementation and operation but less familiar with at least some of these features.

This paper gives an overview of the wave modelling and forecasting model suite at the Met Office. The paper is well written and describes the various components in some detail. My main concern is that the wave-current model AMM15 is only very cursorily described. I would want to see a more detailed discussion of why the model seems to have improved directional and wave periods but less accurate significant wave height in regions with strong currents. More specifically I would want to see a more in-depth discussion of why the authors still think it is worth running the forecast model.