First layer of SPECHUMD from METGRID

A

afernandezody

Member
Hello,
Looking at the content of met_em files (from 'ncdump') shows the first layer of SPECHUMD to be filled with -1.
...
246.2546, 246.2881, 246.3213, 246.3541, 246.3867, 246.4188, 246.4504,
246.4817, 246.5126, 246.5432, 246.5737, 246.6039, 246.6337, 246.6629,
246.6914, 246.7195, 246.7469, 246.7738, 246.8 ;

SPECHUMD =
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
0.001452274, 0.001457454, 0.001463542, 0.001470226, 0.001477271,
0.001484462, 0.001492279, 0.001500457, 0.001508427, 0.001515602,
0.001521687, 0.001527058, 0.00153211, 0.001537248, 0.001542665,
0.001548078, 0.001553534, 0.001559084, 0.001564765,
0.00143931, 0.001444378, 0.001450233, 0.001456614, 0.001463291,
0.001470073, 0.001477385, 0.001485013, 0.001492409, 0.001499048,
0.001504583, 0.001509378, 0.001513925, 0.001518741, 0.001524113,
0.001529712, 0.001535484, 0.001541369, 0.001547318,
...
Click to expand...
All the other variables look like OK and the simulations have been completing, However, I am wondering if I'm doing anything wrong or if this might be causing any accuracy loss in the predictions,
 
No, I don't think you did anything wrong.

SPECHUMD is an optional variable not required by WPS and WRF. SO it is totally fine with missing values of SPFH.
 
The documentation states that SPECHUMD is not needed if RH is available and viceversa. This brings 2 questions:
Q1 - If both SPECHUMD and RH are available in the ungribbed or intermediate files and, consequently, METGRID generates the met_em files including both variables: Does WRF give a higher priority to RH over SPECHUMD and ignores the latter?
Q2 - If RH were not available in the ungribbed or intermediate files, I guess that you'd need to ensure that METGRID is not generating spurious SPECHUMD values (e.g. -1) in the met_em files, as they could potentially wreck the WRF simulations or at least deteriorate performance. Am I reading this scenario correctly?
 
Please see my answers below:
afernandezody said:
The documentation states that SPECHUMD is not needed if RH is available and viceversa. This brings 2 questions:
Q1 - If both SPECHUMD and RH are available in the ungribbed or intermediate files and, consequently, METGRID generates the met_em files including both variables: Does WRF give a higher priority to RH over SPECHUMD and ignores the latter?
Click to expand...
You are right that WRF always uses RH first. If no RH is found, it turns to use SPECHUMD.
afernandezody said:
Q2 - If RH were not available in the ungribbed or intermediate files, I guess that you'd need to ensure that METGRID is not generating spurious SPECHUMD values (e.g. -1) in the met_em files, as they could potentially wreck the WRF simulations or at least deteriorate performance. Am I reading this scenario correctly?
Click to expand...
No, I don't think so. mergrid only uses data from intermediate files. It doesn't "generate" any data itself.
 
I didn't mean to suggest that METGRID would generate data out of nothing. What I didn't express fully if that if the intermediate file does not have RH data but it DOES have SPECHUMD data, there would have to be some extra caution on the first layer generated by METGRID. The thing is that I'm working around the issue (by letting WRF choose RH over SPECHUMD) but still don't understand why METGRID is filling the 1st layer of SPECHUMD with -1.
 
You must log in or register to reply here.
Back
Top