Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

Why coupling a Global Climate Model in WRF (CMIP6)

A

Alexsander

New member
Greatings,

I intend to regionalizate forecast climate change in order to capture small scale effects as topograhy and land use (dynamical downscaling). For this, i will use the WRF model to reducing spacial resolution (dinamycal downscaling).

So, my issue is couple a Global Climate Model in the WRF model (WPS).

According to WRF user guide (https://www2.mmm.ucar.edu/wrf/users/docs/user_guide_v4/v4.2/WRFUsersGuide_v42.pdf), user can coupling any synotic data in WRF since be construct a respective Vtable to decodification the meteorological and land-surface fields.

However, when i assessed avaliable Vtables (downloaded when the WRF instalation), i realize that different fields are described, depending Vtable.
In front of this, i concluded that maybe exist obligate meteorological parameters and not obligate meteorological parameters to initialize WRF.

I have seached by models that integrating last IPCC experiment (CMIP6), but i realize that Global Climate Models became avaliable different meteorological parameters.

Therefore i need know what meteorological and land-surface fields are essencial to inicializate WRF (“minimal meteorological and land-surface fields”) to seach in Global Climate Models output.

According to WRF user guide (https://www2.mmm.ucar.edu/wrf/users/docs/user_guide_v4/v4.2/WRFUsersGuide_v42.pdf), in pgs. 6-36 /6-37, the meteorological and land-surface fields requeried in intermediates files to incialize WRF are:

Parameters necessary to initialize WRF model:
Field name in intermediate file Units Description Notes
TT K 3-d air temperature -
RH % 3-d relative humidity Not needed if SPECHUMD is available
SPECHUMD kg kg-1 3-d specific humidity Not needed if RH is available
UU m s-1 3-d wind u-component -
VV m s-1 3-d wind v-component -
GHT m 3-d geopotential height -
PRESSURE Pa 3-d pressure Only needed for nonisobaric datasets
PSFC Pa Surface pressure -
PMSL Pa Mean sea-level pressure -
SKINTEMP K Skin temperature -
SOILHGT m Soil height -
TT K 2-meter air temperature -
RH % 2-meter relative humidity Not needed if SPECHUMD is available
SPECHUMD kg kg-1 2-meter specific humidity Not needed if RH is available
UU m s-1 10-meter wind u-component -
VV m s-1 10-meter wind v-component -
LANDSEA Fração Land-sea mask (0=water, 1=land) -
SMtttbbb m3 m-3 Soil moisture 'ttt' is the layer top depth in cm, and 'bbb' is the layer bottom depth in cm
STtttbbb K Soil temperature
SOILMmmm kg m-3 Soil moisture 'mmm' is the level depth in cm, not needed if SMtttbbb available
SOILTmmm K Soil temperature

After this background, my douts are:

1. Do any synotic data to inicializate WRF must have all this fields?
2. Is it possible inicializate WRF with less fields of this? Why different Vtables (downloaded when the WRF instalation) describe diferents fields for each Global Models?
3. Is to coupling a Global Model output to WRF (dynamical downscaling) necessary only ajust a respective Vtable to decodificate the fields?

Best Regards,

Alexsander
 
Hi Alexander,
In the WRF Users' Guide, pages 6-* should correspond to chapter 6, which is the WRFDA section. The fields may be the same, but to be sure, take a look at chapter 3, which specifies the input variables that are required to simulate WRF. Any type of input data you use will need to contain these mandatory fields. The various Vtables are created to interpolate the gribbed data based on the grib codes available in the dataset (which vary, depending on the input data type).
 
kwerner said:
Hi Alexander,
In the WRF Users' Guide, pages 6-* should correspond to chapter 6, which is the WRFDA section. The fields may be the same, but to be sure, take a look at chapter 3, which specifies the input variables that are required to simulate WRF. Any type of input data you use will need to contain these mandatory fields. The various Vtables are created to interpolate the gribbed data based on the grib codes available in the dataset (which vary, depending on the input data type).
Click to expand...

Greatings,

I have contacted Climate Centers, around the wourld, to talk about lacking data to initialize WRF model.

However I have had communications issues to ask about meteorological and land surface fields to Climate Centers.

First, there is not a standardization in terminology: meteorological and land surface fields are called by different terms by WRF manual and by CMIP6 repository (https://clipc-services.ceda.ac.uk//dreq/index/var.html). For instance, the “Land-sea mask” (0=water, 1=land) field for WRF Manual is called “sftlf” in the CMIP6 repository. Same for “10-meter wind u-component” / “10-meter wind v-component” (WRF Manual) and “Eastward Near-Surface Wind” / “Northward Near-Surface Wind” (CMIP6 repository) as well as 2-meter air temperature (WRF Manual) and “Near-Surface Air Temperature” (CMIP6 repository).

Second, there is not standardization in terms of output data. Distinct Models became available distinct meteorological and land surface fields. For instance, i need to historical meteorological fields to validate my modeling system, but some Climate Models became available historical meteorological data and another no (for Variant Label = “r1i1p1f1” e Table ID = “6hrPlevPt”).

Finally, i have realized there is a laking data between available by Climate Models and needed to initialize WRF Model. For instance, Max Planck Institute Models (MPI-ESM1-2-LR and MPI-ESM1-2-HR) do not have: “3-d pressure (Pa)”; “Surface pressure (Pa)”; “Skin temperature (K)”; “Soil height (m)”; “Land-sea mask (0=water, 1=land)” called “sftlf”; “Soil moisture” and “Soil temperature” (“SMtttbbb” and “STtttbbb” or “SOILMmmm” and “SOILTmmm”).

Based on this, I have the following doubts:

1) About “3-d pressure (Pa)”, the WRF Manual have mention to “3-d pressure (Pa)” “Only needed for nonisobaric datasets”. Can I run WRF without “3-d pressure (Pa)” field?

2) About “Skin temperature (K)”, I have read some materials and I have found:
My question is, as “Skin temperature (K)” is closely related to “2-meter air temperature (K)” and Climate Models provide “3-d air temperature” as well as “2-meter air temperature (K)”, is it possible calculate “Skin temperature (K)” based on “3-d air temperature” or on “2-meter air temperature (K)”?

3) About “Soil height (m)”, is a thickness of soil used to calculate flux budget (latent and sensible)? “...The WRF default soil data are in two layers: 0 to 30 cm for the top soil layer and 30 to 100 cm for the bottom soil layer...” (https://www.researchgate.net/public...n_for_the_Noah_LSM_SOIL_MAP_AND_SOIL_MOISTURE). So, the “Soil height (m)” default used by WRF is 0.1 m? The unique mention of “soil height” on CMIP6 repository is with a variable “slthick” (Thickness of Soil Layers). However, Unfortunately most of Climate Models do not became available “slthick”, and among this, only for a historical simulation (for instance CMIP6.CMIP.NCAR.CESM2.historical.r11i1p1f1.Efx.slthick.gn). Based on this:
  • Firstly, “Soil height (m)” (WRF Manual) is the “slthick” (CMIP6 repository)?
  • In hipotesys that “Soil height (m)” = “slthick”, treating on a fixed variable, can i use a variable of another Climate Model (for example CMIP6.CMIP.NCAR.CESM2.historical.r11i1p1f1.Efx.slthick.gn)?
  • 4) “Land-sea mask (0=water, 1=land)” (WRF Manual) called “sftlf” (CMIP6 repository) is too a land surface field that is not becoming available by most of Climate Models. My doubt is: can I use any “sftlf”, of different Climate Models with the same resolution, taking into account that is just a information if each cell is “land” or “water”?

  • 5) At last, i need understand what are soil data necessary to initialize WRF. CMIP6 repository have many data of moisture soil and only one to soil temperature. Assessing dimensions of soil variable, i selected these fields:
    • Temperature: “tsl” - Temperature of Soil [K]. Is this temperature (CMIP6 repository) the WRF Manual “Soil temperature (K)”?
    • Moisture: What is a WRF Manual Soil moisture (kg m-3)?
      • “mrsoLut”: Total Soil Moisture [kg m-2] [/“Soil height (m)”] ?
      • “mrsos”: Moisture in Upper Portion of Soil Column [kg m-2] [/“Soil height (m)”] ?
      • “mrsol”: Total Water Content of Soil Layer [kg m-2] [/“Soil height (m)”] ?
      • “mrsll”: Liquid Water Content of Soil Layer [kg m-2] [/“Soil height (m)”] ?
      • “mrsosLut”: Moisture in Upper Portion of Soil Column of Land-Use Tile [kg m-2] [/“Soil height (m)”] ?
      • “mrso”: Total Soil Moisture Content [kg m-2] [/“Soil height (m)”] ?
I would appreciate it and make myself available to take any doubts.

Gratefully,

Alexsander
 
Alexsander,
Please see my answers below to your first post:

(1) Do any synotic data to inicializate WRF must have all this fields?
The answer is yes. The following variables are required for successful WRF run:
TT K 3-d air temperature -
RH % 3-d relative humidity Not needed if SPECHUMD is available
SPECHUMD kg kg-1 3-d specific humidity Not needed if RH is available
UU m s-1 3-d wind u-component -
VV m s-1 3-d wind v-component -
GHT m 3-d geopotential height -
PRESSURE Pa 3-d pressure Only needed for nonisobaric datasets
PSFC Pa Surface pressure -
PMSL Pa Mean sea-level pressure -
SKINTEMP K Skin temperature -
SOILHGT m Soil height -
TT K 2-meter air temperature -
RH % 2-meter relative humidity Not needed if SPECHUMD is available
SPECHUMD kg kg-1 2-meter specific humidity Not needed if RH is available
UU m s-1 10-meter wind u-component -
VV m s-1 10-meter wind v-component -
LANDSEA Fração Land-sea mask (0=water, 1=land) -
SMtttbbb m3 m-3 Soil moisture 'ttt' is the layer top depth in cm, and 'bbb' is the layer bottom depth in cm
STtttbbb K Soil temperature
SOILMmmm kg m-3 Soil moisture 'mmm' is the level depth in cm, not needed if SMtttbbb available
SOILTmmm K Soil temperature

2. Is it possible inicializate WRF with less fields of this? Why different Vtables (downloaded when the WRF instalation) describe diferents fields for each Global Models?
The answer is 'NO'. You must have all the required fields. Different Vtables are provided for different datasets, but the variables specified in these tables are basically the same.

3. Is to coupling a Global Model output to WRF (dynamical downscaling) necessary only ajust a respective Vtable to decodificate the fields?
That is correct. If you try to extract required fields from a specific global dataset, you need to modify Vtable to make it work for that dataset.
 
Alexsander,
To answer your questions in the second post:
1) About “3-d pressure (Pa)”, the WRF Manual have mention to “3-d pressure (Pa)” “Only needed for nonisobaric datasets”. Can I run WRF without “3-d pressure (Pa)” field?

No. You cannot. This is because REAL need this information for vertical interpolation.

2) About “Skin temperature (K)”, I have read some materials and I have found:
My question is, as “Skin temperature (K)” is closely related to “2-meter air temperature (K)” and Climate Models provide “3-d air temperature” as well as “2-meter air temperature (K)”, is it possible calculate “Skin temperature (K)” based on “3-d air temperature” or on “2-meter air temperature (K)”?

You can use 2-m temperature to represent skin temperature, although this is not accurate.

3) About “Soil height (m)”, is a thickness of soil used to calculate flux budget (latent and sensible)? “...The WRF default soil data are in two layers: 0 to 30 cm for the top soil layer and 30 to 100 cm for the bottom soil layer...” (https://www.researchgate.net/public...n_for_the_Noah_LSM_SOIL_MAP_AND_SOIL_MOISTURE). So, the “Soil height (m)” default used by WRF is 0.1 m?
The unique mention of “soil height” on CMIP6 repository is with a variable “slthick” (Thickness of Soil Layers). However, Unfortunately most of Climate Models do not became available “slthick”, and among this, only for a historical simulation (for instance CMIP6.CMIP.NCAR.CESM2.historical.r11i1p1f1.Efx.slthick.gn). Based on this:
  • Firstly, “Soil height (m)” (WRF Manual) is the “slthick” (CMIP6 repository)?
Based on your description, I believe they have the same physical meaning.
  • In hipotesys that “Soil height (m)” = “slthick”, treating on a fixed variable, can i use a variable of another Climate Model (for example CMIP6.CMIP.NCAR.CESM2.historical.r11i1p1f1.Efx.slthick.gn)
Without knowledge of CMIP6.CMIP.NCAR.CESM2.historical.r11i1p1f1.Efx.slthick.gn, I am not quite sure.
  • 4) “Land-sea mask (0=water, 1=land)” (WRF Manual) called “sftlf” (CMIP6 repository) is too a land surface field that is not becoming available by most of Climate Models. My doubt is: can I use any “sftlf”, of different Climate Models with the same resolution, taking into account that is just a information if each cell is “land” or “water”?
I suppose that climate models should have land-sea mask information. You need to derive this field based on the information you have.

5) At last, i need understand what are soil data necessary to initialize WRF. CMIP6 repository have many data of moisture soil and only one to soil temperature. Assessing dimensions of soil variable, i selected these fields:
  • Temperature: “tsl” - Temperature of Soil [K]. Is this temperature (CMIP6 repository) the WRF Manual “Soil temperature (K)”?
  • Moisture: What is a WRF Manual Soil moisture (kg m-3)?
WRF needs soil moisture (volumetric) and soil temperature at at least 2 levels if you run with LSM schemes such as Noah, NoahMP, RUC etc.
 
Hi,

I am a bit new to WPS and using WRF, but I have some questions related to the topics in this thread.

I am trying to prepare WPS intermediate files using the python package pywinter; these will be used by metgrid to make the met_em files.

I am using data from the MPI-ESM1.2-HR "historical" experiment, with variant label "r1i1p1f1" and grid label "gn".

As has been mentioned earlier in this thread there seem to be many soil moisture variables available but only 1 soil temperature variable (tsl).

I have downloaded mrsol (Total Water Content of Soil Layer) which has 5 levels. However, tsl has only 1 level (depth=0.05cm).

How do I create soil temperature data at the levels required by WPS if there is only data on 1 level??!

I believe WPS is looking for ST and SM data at 10, 40, 100 and 200cm. If the data are given at levels different to these, do I need to do some sort of linear interpolation between the data levels to get the values at the levels needed by WPS?

Any advice would be greatly appreciated!!
 
Note that at least 2 levels of soll data are required if you run WRF with Noah, NoahMP, ACM, CLM, RUC surface schemes.
If you only have 1 levels of soil temperature, I am afraid that you cannot ru with the above schemes.
 
Alexsander said:
Greatings,

I intend to regionalizate forecast climate change in order to capture small scale effects as topograhy and land use (dynamical downscaling). For this, i will use the WRF model to reducing spacial resolution (dinamycal downscaling).

So, my issue is couple a Global Climate Model in the WRF model (WPS).

According to WRF user guide (https://www2.mmm.ucar.edu/wrf/users/docs/user_guide_v4/v4.2/WRFUsersGuide_v42.pdf), user can coupling any synotic data in WRF since be construct a respective Vtable to decodification the meteorological and land-surface fields.

However, when i assessed avaliable Vtables (downloaded when the WRF instalation), i realize that different fields are described, depending Vtable.
In front of this, i concluded that maybe exist obligate meteorological parameters and not obligate meteorological parameters to initialize WRF.

I have seached by models that integrating last IPCC experiment (CMIP6), but i realize that Global Climate Models became avaliable different meteorological parameters.

Therefore i need know what meteorological and land-surface fields are essencial to inicializate WRF (“minimal meteorological and land-surface fields”) to seach in Global Climate Models output.

According to WRF user guide (https://www2.mmm.ucar.edu/wrf/users/docs/user_guide_v4/v4.2/WRFUsersGuide_v42.pdf), in pgs. 6-36 /6-37, the meteorological and land-surface fields requeried in intermediates files to incialize WRF are:

Parameters necessary to initialize WRF model:
Field name in intermediate file Units Description Notes
TT K 3-d air temperature -
RH % 3-d relative humidity Not needed if SPECHUMD is available
SPECHUMD kg kg-1 3-d specific humidity Not needed if RH is available
UU m s-1 3-d wind u-component -
VV m s-1 3-d wind v-component -
GHT m 3-d geopotential height -
PRESSURE Pa 3-d pressure Only needed for nonisobaric datasets
PSFC Pa Surface pressure -
PMSL Pa Mean sea-level pressure -
SKINTEMP K Skin temperature -
SOILHGT m Soil height -
TT K 2-meter air temperature -
RH % 2-meter relative humidity Not needed if SPECHUMD is available
SPECHUMD kg kg-1 2-meter specific humidity Not needed if RH is available
UU m s-1 10-meter wind u-component -
VV m s-1 10-meter wind v-component -
LANDSEA Fração Land-sea mask (0=water, 1=land) -
SMtttbbb m3 m-3 Soil moisture 'ttt' is the layer top depth in cm, and 'bbb' is the layer bottom depth in cm
STtttbbb K Soil temperature
SOILMmmm kg m-3 Soil moisture 'mmm' is the level depth in cm, not needed if SMtttbbb available
SOILTmmm K Soil temperature

After this background, my douts are:

1. Do any synotic data to inicializate WRF must have all this fields?
2. Is it possible inicializate WRF with less fields of this? Why different Vtables (downloaded when the WRF instalation) describe diferents fields for each Global Models?
3. Is to coupling a Global Model output to WRF (dynamical downscaling) necessary only ajust a respective Vtable to decodificate the fields?

Best Regards,

Alexsander
Click to expand...
Hello Alexsander,
I find myself in a similar position. I am familiar with WRF model, but now I want to use it to do dynamic downscaling of climate data such as CMIP6 or CORDEX. Did you manage to achieve this? I see the main problem is to collect all the data variables needed for the model and prepare the Vtable since it doesn't exist for now.
I find articles that have used this methodology but don't describe this part of the workflow, so I know the theory but don't know where to start with the practical. Any help would be appreciated.
Jana
 
  • Like
Reactions: jjj
Can anyone working with CMIP6 data, please help me with the Vtable? I am finding some fields missing in the CMIP6 data with respect to wrf requirements.
 
As far as I know, the CMIP6 output files are netCDF format. Please let me know if I am wrong.
You don't need Vtable if you process netCDF format data.
 
You must log in or register to reply here.
Top