ABOUT wrf+roms

[bwendu]

Sorry to bother you again, I’m simulating typhoons using ROMS+WRF, The simulations can complete successfully, but I have some questions:

1、I noticed many studies pointed out the importance of spray in typhoon simulations, so I’m wondering if spray is considered in WRF(COAWST), if not, could you give me some suggestions to add that.
2、I’m focusing on the effect of tides during typhoon events in the East China Sea, so I ran a typhoon case whose track is from the northwest pacific and then moves towards northwest. I conducted experiments with and without tides. In the experiment with tides, the initial field was set as the history file from a long-term ROMS simulation corresponding to the specific time. In the experiment without tides, the initial field was set as the average field over three days around the initial time. Since tidal effects are very weak in the open ocean, it was expected that the typhoon's path and intensity in both experiments would be very similar, with differences gradually emerging as tides become influential near the coast. However, the results showed that tidal effects played a role much earlier than anticipated. Given that WRF is highly sensitive to the initial field, are there any key points in the setup of the ocean initial field in COAWST that I might have overlooked?
3、ROMS requires some time to self-adjust and reach stability. Based on the initial field given by the above method, how long does it take for the COAWST model to reach stability? The results of WRF are short-term effective, while the long-term stable results of ROMS are effective. So, for typhoons, when should I set the start time, and what is the time range of meaningful results?

Looking forward to receiving your any suggestions!

[jcwarner-usgs]

1. we had spent some time on sea spray, but i dont think we published or disseminated that code. maybe this can help:
   Duan, Z., Wang, Y., & Li, Y. (2023). Effects of sea spray on the simulated tropical cyclone development: Dependence on surface drag coefficient parameterization. Journal of Geophysical Research: Atmospheres, 128, e2022JD037299. https://doi.org/10.1029/2022JD037299

2. i am surprised as well, that tides would be that important. WRF is really only getting the SST. So how do the tides effect SST?

3. these 'how long til stability' questions are common. it depends on the situation. I dont have an exact answer.
   I would start the ROMS WRF simulation before the typhoon is in the domain. but that can be difficult. maybe have larger domains where the storm starts away from the area of interest, and then have nests at the coast. If you are just looking at storm impacts, then the longer term stability may not be so much of an issue.

[bwendu]

So glad receiving your reply and thanks for the reference and suggestions. maybe i didn't clarify the third question clearly, so I would like to seek further suggestions.

If I am focusing on the typhoon process from September 7th to 12th, I want to know whether it is better to set the start time on the 1st or earlier rather than the 6th or 7th. Additionally, how long does the spin-up time usually take for COAWST? It seems that WRF only requires 12 hours to a day, but ROMS takes longer to adjust the ocean dynamics, with spin-up appearing to need a week or even more. So i'm confused about how long spin-up time required when coupling the two models.

Furthermore, regarding the validity period of the data, the reasons why I'm confused are as follows:
WRF heavily relies on initial conditions. Based on previous experiences and other studies, it appears that only the the first few days data is reliable for simulating heavy changing processes like typhoons. Therefore, I wonder if it is reasonable to run a longer simulation after coupling. (In other words, if ROMS and WRF are running simultaneously, and the optimal data period of WRF overlaps with the spin-up period of ROMS, I am concerned about whether this affects the reliability of the simulation.)

i'm not sure if my understanding is correct, if not, hope any corrections!

[jcwarner-usgs]

These are all great questions, but i do not have an exact answer. Every case is different.
You will need to do some sensitivity tests to convince yourself on how the initial conditions are affecting your simulation.
Suggest you read the literature a bit and see what others have done.

[bwendu]

Thank you very much for your timely reply and valuable advice! I will follow your suggestions and conduct some tests!