Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

How do met_rlx_bot and met_rlx_top work in FSDW?

QikaiYu

Qikai Yu
New Member
Hi all

I am interested in namelist variables that control which levels in the vertical the nudging should apply to.

The namelist variable, met_rlx_bot and met_rlx_top, their descriptions are (km) bottom/top of relaxation region of winds for offline waccm.

I feel very confused about these descriptions, and I have three questions below.



1. Do these namelist variables only work for waccm (i.e. FSDW compset), which means that these namelist varibales don't work in FSDCHM?

2. Why is the unit of these variables KM? it's very strange because vertical levels in the model are hybrid sigma coordinates, not Z coordinates (height coordinates). Using km as units does not give an intuitive sense of which level is nudged.

3. Why do these variables only control the winds? How is the temperature field controlled?

I really appreciate your great help! Thank you!
 

mmills

CSEG and Liaisons
Staff member
1. By default, winds and temperatures in SD-WACCM (FSDW compset) are nudged (by either 1% or 10% per timestep) to analyses from GEOS5 or MERRA at levels from the surface up to 50 km (met_rlx_bot). The nudging is relaxed between 50 and 60 km (met_rlx_top), above which there is zero nudging. This is because of the limitations of GEOS5 and MERRA, which are not available or are unreliable at higher altitudes. Because SD-CAMchem has a model top that is below 50 km, it is nudged throughout its vertical domain.

2. The relaxation boundaries are in kilometers, most likely for ease of use. There will be some variation in the precise nudging of several levels between 50 and 60 km, but the relaxation is gradual from full (1% or 10% per timestep) below 50 km, to zero nudging above 60 km. The levels are indexed starting with level 1 at the model top. This means that the level indices are different in WACCM than in CAM, which have different model tops. It is therefore not possible to restrict nudging based on the model indices.

3. The variables do control the temperature nudging as well as the wind nudging.
 

QikaiYu

Qikai Yu
New Member
Thank you very much for your help, Mike!

Recently I have read the source code and done lots of tests. I can basically understand your answer.

However, I found that the variables not only control the temperature and wind nudging, but also the PS fields. Besides, I found that the default FSDW compset also nudges many surface variables such as SHFLX QFLX TS (maybe even undge TAUX, TUAY, OCNFRAC, ICEFRAC, and PHIS). Am i right? Why these variables don't appear in FSDW compset description? Are there namelist variables can control or shut down them?


Forgive me for having so many questions. Again, I sincerely thank you for your help!
 

QikaiYu

Qikai Yu
New Member
1. By default, winds and temperatures in SD-WACCM (FSDW compset) are nudged (by either 1% or 10% per timestep) to analyses from GEOS5 or MERRA at levels from the surface up to 50 km (met_rlx_bot). The nudging is relaxed between 50 and 60 km (met_rlx_top), above which there is zero nudging. This is because of the limitations of GEOS5 and MERRA, which are not available or are unreliable at higher altitudes. Because SD-CAMchem has a model top that is below 50 km, it is nudged throughout its vertical domain.

2. The relaxation boundaries are in kilometers, most likely for ease of use. There will be some variation in the precise nudging of several levels between 50 and 60 km, but the relaxation is gradual from full (1% or 10% per timestep) below 50 km, to zero nudging above 60 km. The levels are indexed starting with level 1 at the model top. This means that the level indices are different in WACCM than in CAM, which have different model tops. It is therefore not possible to restrict nudging based on the model indices.

3. The variables do control the temperature nudging as well as the wind nudging.
Thank you very much for your help, Mike!

Recently I have read the source code and done lots of tests. I can basically understand your answer.

However, I found that the variables not only control the temperature and wind nudging, but also the PS fields. Besides, I found that the default FSDW compset also nudges many surface variables such as SHFLX QFLX TS (maybe even undge TAUX, TUAY, OCNFRAC, ICEFRAC, and PHIS). Am i right? Why these variables don't appear in FSDW compset description? Are there namelist variables can control or shut down them?


Forgive me for having so many questions. Again, I sincerely thank you for your help!
 

tilmes

Member
Hi Qikai Yu,
you are right that other variables are used in the FSDW nudging approach. To turn those off you have to modify the model code, in particular in metdata.F90.
However, we developed a new nudging method that may be of interest to you, and we already developed compsets for
CAMchem (32vertical levels), called FCnudged. Similar compsets may be developed for WACCM as well.
This type of nudging is described here:
and by default only nudges winds and temperatures, not the surface fluxes. This nudging requires met fields to be interpolated to the models' vertical grid.
You can add Q as well, and potentially other variables. This type of nudging allows more changes through the namelist settings.
WACCM has been run on 70levels and met data have been already interpolated, so it may be easy to use this.
Cheers, Simone
 

QikaiYu

Qikai Yu
New Member
Hi Qikai Yu,
you are right that other variables are used in the FSDW nudging approach. To turn those off you have to modify the model code, in particular in metdata.F90.
However, we developed a new nudging method that may be of interest to you, and we already developed compsets for
CAMchem (32vertical levels), called FCnudged. Similar compsets may be developed for WACCM as well.
This type of nudging is described here:
and by default only nudges winds and temperatures, not the surface fluxes. This nudging requires met fields to be interpolated to the models' vertical grid.
You can add Q as well, and potentially other variables. This type of nudging allows more changes through the namelist settings.
WACCM has been run on 70levels and met data have been already interpolated, so it may be easy to use this.
Cheers, Simone

Hi Simone,

Thanks a lot for your help. I am interested in only nudging shflx and qflx. I have set met_max_rlx=0 to turn U V T PS nudging off, this does work. However, I didn't find a proper way to turn TAUX TAUY TS nudging off by modifying code in metdata.F90. It's very strange that I have already commented the code associated with these variables, the nudging of TAUX\TAUY\TS still works. So I wonder if there is any other source code that needs to be modified. Thank you very much!
 

nuvolet

Toni Viudez
Member
Hi Qikai Yu,
you are right that other variables are used in the FSDW nudging approach. To turn those off you have to modify the model code, in particular in metdata.F90.
However, we developed a new nudging method that may be of interest to you, and we already developed compsets for
CAMchem (32vertical levels), called FCnudged. Similar compsets may be developed for WACCM as well.
This type of nudging is described here:
and by default only nudges winds and temperatures, not the surface fluxes. This nudging requires met fields to be interpolated to the models' vertical grid.
You can add Q as well, and potentially other variables. This type of nudging allows more changes through the namelist settings.
WACCM has been run on 70levels and met data have been already interpolated, so it may be easy to use this.
Cheers, Simone
Hi Simone,

I read the chain of replies from the beginning of this post and you suggest to use another compost (FCnudged), what kind go gris resolutions from the documentation are supported (CESM2.2.0 Grid Resolution Definitions).
We are trying to nudge MERRA2 using F2000climo but there're difference between levels, although our target would be nudge the model from some of its constituents in particular aerosols (AODs).

Thanks in advance
 
Top