Vertical Velocity

[Bex Ross]

Hi All,

I just wanted to do a quick survey of the group:

How many of you use vertical velocity in your simulations?

If you do: 

Do you trust it?

If you don't: 

Do you use vertical diffusivity instead? Do you think vertical diffusivity is an adequate substitute?

I'm still trying to decide for my own studies, and since I am using HYCOM global (GLBa0.08), vertical velocity does not come as a standard variable and I would need to take time computing it myself, so I'm hoping other peoples experience will help guide me, to work out whether it is worth making the effort to compute that for at least 20 odd depth layers, per day, for 5 years of data. I mentioned this to some of you at OSM 2014, but just wanted to open it to the floor... 

All the best,

Bex

[Gabriel Lucas]

Dear Bex Ross,

Firstly, I want to thank you for your support in the connectivity-modeling-system community. As someone just learning CMS for his PhD, many of your posts in the community were extremely helpful.

After 10 years from your question about using or not manually derived vertical velocities, I would like to return the question to you, if there's no problem. How it went so far? Have you been using it for your past or recent works? 

I'm facing the same dilemma here, since the hydrodynamic fields I intend to use for my research also does not include the w component. Initially, I was thinking in disregarding it, but after diving a little bit on this topic here in the CMS community I became more willing to try it. The computational effort in deriving it from original data is no problem, but I do have bigger concerns regarding the error propagation in my simulation. I would really appreciate if you could give your thoughts about it based on your experience.

Thank you very much!

Best,

Gabriel

[Rebecca Ross]

Hi Gabriel,

Sorry for a slow reply, I am not checking the forum these days as my job for the last 6 years has not involved using CMS/dispersal modelling, so I have only just noticed your kind message and unfortunately my answers will be out of date. My hope is that you will find a few others still around who can answer though! Also this question is not entirely CMS related so you could ask it in other places too (hydrodynamic model forums, other particle tracker forums). 

From memory, when I did calculate the vertical velocities they were negligible - the hydrodynamic model had no focus on them, so it is hard to tell whether that was realistic or not. Ideally you would do some groundtruthing of this, but that data is not easy to come by either. So it might not be worth the effort anyway. You could try it for a couple of fields and see what sort of values you get to help you decide if it is worth worrying about or not. It definitely seems like a field of research that needs more research. I cant remember if you have the option for a random vertical kick in the CMS or not - if you do, that could also help in testing how different dispersal patterns might be if there were stonger w velocities involved. But ultimately I agree that it feels like too much guess work resulting in possibly additive errors in the models, so is probably not worth it unless you have a) groundtruthing data to help verify they are needed, b) a model that would generate non negligible values anyway.

But maybe you have made all your decisions already now anyway?

Either way, it is a good question to ask and hopefully a few others can answer too!

All the best,

Bex

--
You received this message because you are subscribed to the Google Groups "connectivity-modeling-system-club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/connectivity-modeling-system-club/2e1dee02-e7c3-4829-aba6-b075522a7a32n%40googlegroups.com.

[Robin Faillettaz]

Hi Gabriel, Bex, 

Just to add to Bex’s answer, there is a w-turb module in the CMS that adds some randomness to the vertical movement of the particles, regardless of currents. Yet the outcomes are highly sensitive to the amont of vertical diffusivity that you implement, so you’ll probably need to run a sensitivity analysis, unless you have field values. 

Regarding the vertical velocity, you may contact Ana Vaz directly (she’s working at the NOAA), I think she’s the one who implemented the w-vel recalculation in the CMS and she has used w-velocities quite extensively. 

Best,