Runoff mixing and spreading

[kieranricardo]

CM3 river runoff points (runs 23-09-2024 and earlier) are freshening rapidly. This is most likely because river mixing and river spreading options are not turned on in CM3 (and are used in CM2).

CM2 options:

&ocean_rivermix_nml
    debug_this_module = .false.
    river_diffuse_salt = .false.
    river_diffuse_temp = .false.
    river_diffusion_thickness = 0.0
    river_diffusivity = 0.0
    river_insertion_thickness = 40.0
    use_this_module = .true.
/

&ocean_riverspread_nml
    use_this_module = .true.
/

[kieranricardo]

CM3 river mixing options (used in cm3-run-24-09-2024-add-river-mixing):

DO_RIVERMIX = True
RIVERMIX_DEPTH = 40.0

River spreading does not appear to be implemented in MOM6.

[MartinDix]

From cm3-run-18-09-2024-no-zap-no-iceberg

Salinity averaged over several cells at the Amazon mouth decreases rapidly over the first 6 months and never recovers.

[kieranricardo]

Update, the river mixing is a part of the bulk layer mixing scheme which can only be turned on if USE_REGRIDDING=False.

[aekiss]

@kieranricardo can we use RIVERMIX with USE_REGRIDDING=True if we use EPBL instead of KPP? See here

[ofa001]

@aekiss we wanted to test EPBL set up, it looks as MOM6 have that as their preferred option, that may be the only one that works with vertical river mixing, does the horizontal mixing work as well.
Is this something that can be tested in OM3 as well as CM3.

[aekiss]

We'd intended to use KPP for OM3 configs for CM3 for CMIP7, and investigate EPBL in OM3 later - see https://forum.access-hive.org.au/t/namelist-configuration-discussion-meeting/1917/9
But if EPBL is the only way to get river mixing, we may need to expedite EPBL tests in OM3.

@willaguiar, have you looked at EPBL vs KPP in MOM6?

[aekiss]

GFDL OM5 doesn't use rivermix, even though they appear to be using EPBL (and KPP is off).

[willaguiar]

@willaguiar, have you looked at EPBL vs KPP in MOM6?

A little bit (in panan-01 compared to OM2-01, but have not compared KPP and ePBL within MOM6). As far as I could tell, EPBL usage didn't impact much on anything, other than perhaps making the MLD on the Antarctic shelf a couple tens of meters smaller compared to OM2. (MOM5_1m Vs MOM6_1m below). Other than that ePBL is suppose to be less sensitive than KPP to vertical resolution and to timestep.

[kieranricardo]

@aekiss and @willaguiar - great I'll give EPBL a try, do you know what MOM6 parameter options I should use? It looks like ENERGETICS_SFC_PBL= True (and USE_KPP= False) will turn the scheme on, but other than that it looks like the other PBL parameters and defaults are:

EPBL_IS_ADDITIVE= True
PERT_EPBL = False
MLE_USE_PBL_MLD = False

[willaguiar]

For panan, we use a slightly different set of EPBL parameters:

MLE_USE_PBL_MLD = True
ENERGETICS_SFC_PBL = True
EPBL_IS_ADDITIVE = False
PRANDTL_EPBL = 1.0     
EPBL_2018_ANSWERS = False      
EPBL_ORIGINAL_PE_CALC = True   

! === module MOM_energetic_PBL ===
EPBL_MSTAR_SCHEME = "OM4"
MSTAR = 0.0
MIX_LEN_EXPONENT = 1.0
MSTAR_CAP = 10.0
MSTAR_CONV_ADJ = 0.667
MSTAR2_COEF1 = 0.29
MSTAR2_COEF2 = 0.152
NSTAR = 0.06
TKE_DECAY = 0.01
ML_OMEGA_FRAC = 0.001
USE_MLD_ITERATION = True
EPBL_TRANSITION_SCALE = 0.01
USE_LA_LI2016 = True
EPBL_LANGMUIR_SCHEME = "ADDITIVE"
LT_ENHANCE_COEF = 0.044
LT_ENHANCE_EXP = -1.5
LT_MOD_LAC1 = 0.0
LT_MOD_LAC4 = 0.0
LT_MOD_LAC5 = 0.22
!EPBL_USTAR_MIN = 1.45842E-18

If it helps, I attached the full list of parameters used for the panan run.
MOM_parameter_doc.all.txt

[aekiss]

For reference, here are the GFDL OM5 settings

[kieranricardo]

Thanks @willaguiar and @aekiss!

[ofa001]

Hi @kieranricardo and @aekiss, I had a look at subroutine riverspread and how we used it in CM2 and ESM1.5, it looks like it might not have been active after all. It might have been either rivermix or riverspread was used not both in the end. When riverspread was used a long time ago there were field table settings on for river locations but none are now set and other settings are defaulted to zero values.

This may be way its not transferred across to MOM6 its fallen out of use, of the years., and they expect the model diffusion to do the spreading but it isnt? Why not? The plume is too buoyant? or do other parameters need to change.

[MartinDix]

CESM river spreading COSIMA/access-om3#231 (comment)