Soil temperature elevation adjustment

[francescomaicu]

Dear WRF Supporters,
I'm running WRF forced by ERA5.
I downloaded the grib separately for 3D and 2D variables https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels?tab=overview.
In the log file of real.exe I get this warning message:

 no soil temperature elevation adjustment, soil height too high =   46964.70

for several height, and also this message:

no soil temperature elevation adjustment, diff of soil height and terrain =   3371.954

I do not understand where is the problem, if there is any.
I used the Vtable.ERA-interim.pl file to ungrib the data, and I attached the log of the program rd_intermediate.exe of one of the intermediate 2D file and 3D file.
I also attached the log of the command wgrib -v era5_3D_presslev.grib | grep 20200915 and wgrib -v era5_2D_presslev.grib | grep 20200915.
PLease can someone give me an explanation of what is happening?
Thank you very much!

 

[kwerner]

Hi,
Can you attach the full error log file from your real.exe run, along with your namelist.input and namelist.wps files? I'd like to take a look at those. Also, just in case, take a look at this FAQ that discusses using ERA5 data. Thanks!

 

[francescomaicu]

Dear Kelly,
I uploaded all the needed files.
Real.exe runs correctly up to the end of the needed period (1 day for shortness).
One problem is that it seems that in ERA5 the RH at the surface (level 200100) is missing in the dataset, and real.exe writes in the rsl.err that is trying to interpolate from the closest level, and that can be OK.
But in the rsl.out, you will see lots of messages "no soil temperature elevation adjustment, diff of soil height and terrain" and it seems to me quite weird that the program run successfully.
I will read carefully your post.
Thanks a lot

 

[kwerner]

Hi,
Thank you for sending those. In this module (adjust_soil_temp_new), the difference between the WRF model topography and the first guess is evaluated. If the WRF model topo was different, then a -6.5 K/km adjustment to the soil temperature is applied.

The difference of the two fields is usually a few Kelvins, which isn't a big deal. But if the delta z is large (in this case OVER 3000 m different), then a message is printed. I agree that it's odd that the model doesn't automatically stop here, but it's good that you looked at it. Likely the initial soil temperature looks bad, with places where the adjustment happens, and a few grid cells where it does not. I suggest taking a closer look at the soil temperature fields from the input data.

 

[francescomaicu]

Dear Kelly,
thanks a lot for your help.
I found that module, but I could not understand properly what it was doing. I guess it deals with the vertical interpolation of some variable that is missing.
Looking at your previous post (ERA5), I realized that I did not download the topography (together with the lansea) so that is the reason with the difference was soo big when trying to vertically interpolate.
I did not use the ERA5 invariants (topography and land-sea) because I downloaded along with the data.
By the way I found another weird message in the REAL:

Noah: bad soil moisture at i, j =

I see in the code that there is a check on soil moisture if it is less that 0.005 , the warning is issued and the variable is set at 0.005.
This happens because in my case I found small negative values of SM000007 in the northern African region in the grib file from both ERA5 and ECMWF.
Do you know more about this evenience and its consequences?

 

[kwerner]

Did you get the negative soil moisture values even after using the invariant ERA5 data? Including that file has corrected this problem for previous users.

 

[francescomaicu]

Did you get the negative soil moisture values even after using the invariant ERA5 data? Including that file has corrected this problem for previous users.

Kelly, I did't check that sorry.
I was trying to use ERA5, but with the SST from ECMWF for comparison with previous simulations.
Using the ERA5 invariants, METGRID stopped because SST was coming along with its own LANDSEA mask, and I got an error because METGRID cannot menage two different LANDSEA masks.
Including the ERA5 LANDSEA mask in its own intermediate files, and the ECMWF LANDSEA mask in the SST intermediate files (for a better interpolation), does not create any problem.
I'll make a check using the full ERA5 dataset with the invariants and let you know.
Thanks a lot

 

[qingling]

Did you get the negative soil moisture values even after using the invariant ERA5 data? Including that file has corrected this problem for previous users.

I have same question about the negative soil moisture values even after using the invariant ERA5 data, How do you solve this problem？