Modified LMD mixing

Discussion of how to use ROMS on different regional and basin scale applications.

Moderators: arango, robertson

Post Reply
Message
Author
User avatar
m.hadfield
Posts: 521
Joined: Tue Jul 01, 2003 4:12 am
Location: NIWA

Modified LMD mixing

#1 Post by m.hadfield » Wed Oct 20, 2004 4:04 am

I am applying ROMS to physical/biological simulations on a domain encompassing New Zealand's Chatham Rise (cf my message entitled "Iron-limited biology") and obviously I want to get my seasonal cycle in mixed layer depth about right. I have tried the LMD scheme, as implemented in ROMS 2.1, with 6-hourly NCEP Reanalysis fluxes and find that mixed-layer depths in summer are generally too shallow and the annual amplitude in SST is too large, by 25-50%.

My questions are:
  • Has anyone looked cosely at the performance of ROMS's LMD scheme in realistic situations? How well did it work? Were summer mixed layers too shallow?
  • Has anyone implemented Smyth et al's modifications (Smyth et al 2002, Nonlocal fluxes and Stokes drift effects in the K-profile parameterization, Ocean Dynamics 52(3) 104-115)?

User avatar
shchepet
Posts: 185
Joined: Fri Nov 14, 2003 4:57 pm

#2 Post by shchepet » Tue Oct 26, 2004 10:52 pm

As you probably know, I made some revision of KPP codes in ROMS last summer (2003) with the view to (1) update it first to the state of NCAR model of Bill Large of 2003 as much as I can; and (2) move on with some new ideas about handling of Ekman Boundary Layers and other things, like Monin-Obukhow depth limitation.

This is done with a variable success, and exposed problems need to be addressed more thoroghly in future.

My approach is different from Hernan's in sense that I want to start with striped-down version of the code and make sure that simple things work before bringing more processes. (In contrast, Hernan tries to import the whole code of full complexity and then let users to decide what CPP switches to activate and what not to; I consider it not very safe way because a lot of parts of KPP never been used in high-resolution, eddy-rich environment). So I focus on shear layer instability on surface boundary layer, including Ekman regime (there is some clarity here at this point) and M-O limitation (more obscure; I just found and fixed a cause for nonphysical hysteresis, pretty much similar to Gokhan's 2003 update), as well as introduced surface wave mixing fix (the mixing coefficients do not asymptote to zero at surface, as mandated by Kolmogorov theory (wall law) but go to finite limit due to "surface wave breaking". That limit is scaled the same way as wind stress, which is kind of natural, so it boils down to modification of cubic shape function.

KPP generally captures the seasonal cycle of PBL in large scale model like Pacific. However, there are some questional features, if one looks for details.

The problem of too shallow HBL in summer and in Equator in well known and unfortunately in stands. I feel that the whole thing requires recalibration in terms of what are the critical Ri numbers. Bill Large also noticed too shallow HBL in Gulf of Alaska in summer.

I believe that too large annual amplitude of SST is just the consequence of the shallowness.

I was not aware of Smyth et al 2002, Nonlocal fluxes paper. I will look at it.

Overall, as paradoxically as it might sound, historically a lot of effort was placed into implementation of unstable boundary layer regime, as well as in criticism of Mellor--Yamada schemes for not handeling these processes right because of inherently local nature of TKE approach. These processes are represented by the nonlocal fluxes in KPP. But most complaints about KPP performance occur in the most basic situation of vanishing PBL under stabilizing boyancy forcing conditions, when all nonlocal fluxes are turned off.

User avatar
m.hadfield
Posts: 521
Joined: Tue Jul 01, 2003 4:12 am
Location: NIWA

#3 Post by m.hadfield » Wed Nov 03, 2004 4:00 am

Re, my original message in this thread...
I am applying ROMS to physical/biological simulations on a domain encompassing New Zealand's Chatham Rise (cf my message entitled "Iron-limited biology") and obviously I want to get my seasonal cycle in mixed layer depth about right. I have tried the LMD scheme, as implemented in ROMS 2.1, with 6-hourly NCEP Reanalysis fluxes and find that mixed-layer depths in summer are generally too shallow and the annual amplitude in SST is too large, by 25-50%.
:oops: I forgot to turn on the SOLAR_SOURCE option to enable pentration of solar radiation into the water column. (I made a mental note to myself to do it, but I mislaid the note!) With this option enabled, mixed layer depths look more reasonable and annual amplitude in SST (for which I have very good observed data) is about right.

So I'm a happy camper :) and my interest in modifying the LMD scheme has waned somewhat

Post Reply