Review of “ISWFM-NSCS v2.0: advancing the internal solitary wave forecasting model with background currents and horizontally inhomogeneous stratifications”

General comments:

Presented is an improved version of the internal solitary wave model. The efforts of the improvements have been made through several aspects: inclusion of background currents, horizontally inhomogeneous stratifications, and many others. The results are very encouraging. This paper documents these achievements towards a better internal solitary wave model.  The graphs are impressively beautiful. I would recommend it be accepted for publication after some clarifications and minor revision.

Specific comments:

The title gives readers an impression that this is a forecast model. However, it has not been implemented as a forecast model yet. The numerical experiments reported in this paper are hindcast modeling. I would suggest to remove the keyword “forecasting”.

Generally, the paper is well written by non-native English speakers.  However, there are some minor writing issues (e.g., wrong word choices) that could and should be fixed through revision. I list some here, but there may be more in the paper.

L17, “ISWFM-NSCS v2.0” can be replaced with “the new version”.

L20, “in ISWFM-NSCS v2.0” can be removed.

L21, “presented” should be changed to “used”.

L68, ”by comparing with numbers of” can be changed to “through”.

 L79, “by” should be changed to “by using” or “from”, and “a” should be “the”.

L87, “well” should be changed to “good”.

L138-139, it would be good to provide some examples, e.g., Liu et al. (2000, 2008).

Figure 5, the model name “MITgcm” should be changed to “ISWFM-NSCS v2.0” for a more specific case.

L145-148, it is mentioned that background currents are added and that velocity data are also extracted from the HYCOM.  However, it is not clear whether the HYCOM velocity data are used as the background currents.  If so, please clearly state it in this paragraph.  If not, then provide more information on what data are used for specifying the background currents or temperature and salinity fields in the ISWFM-NSCS v2.0.  (This information is found later in the manuscript, L320-321. But it would be better to state that earlier in this paragraph).

References:

Liu et al. (2000), Circulation in the South China Sea in summer of 1998, Chinese Science Bulletin 45, 1648-1655.

Liu et al. (2008), Patterns of upper layer circulation variability in the South China Sea from satellite altimetry using the self-organizing map, Acta Oceanologica Sinica 27 (Suppl), 129-144.

This paper presents an evolution of an already existing model focused on the internal solitary wave forcasting, known to be challengous for many physical reasons. Based on my own reading, the paper is clear, well written and present several important results that might interest the community. Three major evolutions are presented : modification of the eddy viscosity/diffusivity , add of a realistics non-homogeneous stratification and background current. The authors showed the impact of adding each component on the forecasting performances.

 However, my main concern is about the choice of a constant eddy viscosity/diffusity for their model. If the value is well suited for ISW as the authors explained, the values are far too large out of these phenomenon, probably rapdily destroing the stratification, a problem they already faced in the first version of their model. I would suggest using an adaptative turbulent closure scheme , like the classical  k-epsilon. At least, I would appreciate such a test or a discussion on that. Another important remark regarding the choice of the eddy viscosity/diffusivity, concerns the numerical diffusivity  which is neither taken into account nor discussed. I would appreciate the author to warn on this potential issue, when they explain the choice of the values based on in situ measurements. 

Minor suggestions :

The introduction is lacking details on the amplitude/wavelength of the ISW in the region. It's important to be fully convince that you resolution is enough to really capture this small-scale waves. In addition, I fell that it is lacking geographical information about the region. Maybe you can add the quoted names on Fig 2

Line 155 it is not clear if you are talking about 28 properties or 28 ISWs

Regarding the observations, even if they are already introduced in Gong et al 2023, I suggest to brively introduce them : which instrument, how long ...

In section 4, test the statistical significance of your differences. Especially in lines 310.

Fig 8. Please add the legend in the panels b and d

Fig 11 : It is quite hard reading this figure. It would suggest to move to a tab that summarize the averaged values should be much more readable. In particular, I recommend to separate the 15 first days and the 15 last days for your analysis.