how to use different initial soil moisture as the input

[Iamwrfuser]

Hi All,
I want to change the initial condition with only with different soil moisture, but I have no clue how to do it?
Could you give me a guide?
Thank you!
Mandy

[kwerner]

Hi Mandy,
Do you have some soil moisture data that you want to use, in addition to your other first-guess data, or are you wanting to manually modify data points to new values for soil moisture?

[ehsantaghizadeh]

Hi,
I had some problems like that. May I ask whether you solved yours?
I want to use SMOPS data (including just soil moisture) as initializing. I wrote its Vtable and run (ungribe.exe and metgrid.exe) correctly (kwerner helped me about it: posted as SMOPS Vtable).
After running ./real.exe I found not suitable output for SMOIS in wrfinput files.
Some points I should mention: GFS data linked per 6hr (00,06,12,18) but SMOPS data linked daily (only 00).
So, for GFS and ./ungrib.exe (interval_seconds = 21600 in namelist.wps), but for SMOPS and ./ungrib.exe (interval_seconds = 86400 in namelist.wps). While ./real.exe (interval_seconds = 21600, in namelist.input).
However, I've test for GFS daily linked (only 00 hr and interval_seconds = 86400), but there is still bad difference between with and without SMOPS initialization.
I've attached 4 files for:
1- Only GFS 6rh (00,06,12,18)
2- GFS 6 hr (00,06,12,18) and SMOPS 24 hr (00)
3- Only GFS 24 hr (00)
4- GFS 24 hr (00) and SMOPS 24 hr (00)

I hope I could describe my issue clearly.
I'll be thankful for any help.

Sincerely
Ehsan

[ehsantaghizadeh]

About my previous post, I should mention SMOPS data are daily. So I think using daily data as input for 00 hr, may resulted these bad SMOIS values, may be not. However, SM in met_em* files doesn't show very significant differences.
Also, GFS first layer soil moisture data are for 0-10 cm and SMOPS soil moisture data are for 0-5cm!
Any solution will be appreciated.

Sincerely
Ehsan

[kwerner]

Hi Ehsan,
Are there missing points where you need GFS data to be used for soil moisture?
Can you attach the namelist.wps file that you are using, and if possible, can you attach 2 (consecutive times) of your SMOPS data files? Thanks.

[ehsantaghizadeh]

Dear kwerner,
Thank you so much for your attention.
1- I don't think there is missing point.
2- I've attached namelist.wps, namelist.input, Vtable.SMOPS, guidance of SMOPS data (gribtab_SMOPS3.0.dat), GFSp25 and SMOPS daily data.

I did following instruction:

$ ln -s ungrib/Variable_Tables/Vtable.GFS Vtable
$ ./link_grib.csh /home/x_292026/98_6_15/data_GFSp25/gfs.0p25.201704*
$ vi namelist.wps
interval_seconds = 21600,
prefix = 'GFS'
$ ./ungrib.exe

$ ./link_grib.csh /home/x_292026/98_6_15/data_SMOPS/NPR_SMOPS_CMAP_D2017041*.gr2
$ ln -sf ungrib/Variable_Tables/Vtable.SMOPS Vtable
$ vi namelist.wps
interval_seconds = 86400,
prefix = 'SMOPS'
fg_name = 'GFS', 'SMOPS'
$ ./ungrib.exe

$ vi namelist.wps
interval_seconds = 21600,
$ vi METGRID.TBL.ARW
name=SM000010
change missing_value=-1.E30
to
missing_value=0.
$ ./metgrid.exe

$ cd ../WRF/run/
$ ln -sf ../../WPS/met_em.d0* .
$ ./real.exe
$ ./wrf.exe

Please let me know if I should prepare anything else.
I'll be thankful if I have your idea.

Sincerely
Ehsan

[ehsantaghizadeh]

I had forgotten to upload Vtable.SMOPS. I've attached it in this post.

[kwerner]

Hi,
Even though your SMOPS data aren't available as often as the GFS data, in your namelist.wps, you should still be able to set interval_seconds = 21600 when ungribbing your SMOPS data, and the data will be interpolated to 6 hour intervals. This should give you a GFS and SMOPS file every 6 hours, and then when you run metgrid, it will use the interpolated SMOPS data at each time, instead of filling in GFS data during the gaps, which is very likely causing your problem. Give that a try and see if it makes any difference.

[ehsantaghizadeh]

Dear kwerner,
Thank you for your nice reply.
I tried with interval_seconds = 21600. It seems it could reduce the problem of SMIOS in wrfinput files. However, about SM000010 in met_em files, for 06, 12 and 18 hr, which there isn't SMOPS data, it doesn't show a good result. Attached file is SM000010 for 12hr.

Sincerely
Ehsan

[kwerner]

Hi,
Can you attach the namelist.wps file you are using to run metgrid, along with the ungrib.log file for your GFS run? Thanks.

[ehsantaghizadeh]

Dear kwerner,
Pardon me to check your reply with a delay.
I've attached namelist.wps and ungrib.log with for run GFS. I hope these files be correct. However, the date in these files may differ with previous files.
Please let me know if I should prepare any other files.

Sincerely
Ehsan

[kwerner]

Hi Ehsan,
In the namelist you sent, in the &metgrid section, I only see that you are using GFS data:

Code: Select all

&metgrid
 fg_name = 'GFS',

If you are using both GFS and SMOPS data, then that line should read something like:

Code: Select all

&metgrid
 fg_name = 'GFS','SMOPS'

How are you actually using both datasets?

[ehsantaghizadeh]

Dear kwerner,
Pardon me to back too late. Actually, I left that job for a few weeks. I was in a course.
You are right. I had uploaded wrong namelist.wps. I've attached the correct one here.
First, I've used
&ungrib
prefix = 'GFS',
and run ungrib.exe
then I've used:
&ungrib
prefix = 'SMOPS',

&metgrid
fg_name = 'GFS', 'SMOPS'

As, you said before, I've used
interval_seconds = 21600
for both "GFS" and "SMOPS"
However, I've attached png file for SM000010 in met_em.d02.2017-04-13_00:00:00.nc
May I ask if you test it. Did you get right picture?

Sincerely
Ehsan

[kwerner]

Hi,
I apologize for the delay. I did run it using the SMOPS data you previously sent, along with GFS data. I didn't get the same image you are seeing. I'm attaching screenshots of my images (one with only GFS, and the other is with GFS+SMOPS). I'm also attaching the namelist.wps file I used.