Modifying temperature in the input has no effect on simulated 2 meter temperature

[Liliputian_123]

Dear WRF experts,

I modified the 2 meter temperature in input ECMWF data in my WRF runs.
Specifically, I added +1 degree and +3 degrees. However, the results seem to be insensitive to this (see figure below).

I would like to ask:
1. How is 2 meter temperature computed by the WRF model?
2. Am I correct to just change the 2 meter temperature in the input file or it should be something else?

Thank you in advance for any help that you can provide.

Dashed black line: observed
Orange line: Default or unchanged
Blue line: modified

 

[Ming Chen]

t2 in wrfinput is only used at the first time step. After that, it is diagnosed based on similarity theory. This is why the model result is not sensitive to initial T2 value.

What sensitivity tests are you trying to do?

 

[Liliputian_123]

Hi Dr. Ming Chen,

I am trying to improve the bias of simulated Heat Index (HI). Here's what we have done:

1. modified temperature by adding 1 and 3 deg C at all time steps in the input file-----no changes in the output as shown in my previous post.
2. Modified the parameter p in the ACM2 PBL scheme------we see changes in the RH and temp but no significant changes in the computed HI.
I think I am missing somethng fundamental here. Any suggestions to solve this is appreciated.


3. Now I am planning to multiply 0.5 to the input wind speed. Shall I do that too to the first time step in the input like temperature?

 

[haiqingsong]

Hi Dr. Ming Chen,

I am trying to improve the bias of simulated Heat Index (HI). Here's what we have done:

1. modified temperature by adding 1 and 3 deg C at all time steps in the input file-----no changes in the output as shown in my previous post.
2. Modified the parameter p in the ACM2 PBL scheme------we see changes in the RH and temp but no significant changes in the computed HI.
I think I am missing somethng fundamental here. Any suggestions to solve this is appreciated.
View attachment 18649

3. Now I am planning to multiply 0.5 to the input wind speed. Shall I do that too to the first time step in the input like temperature?

T2 is the diagnostic variable, so changing the initial value does not affect subsequent T2.

 

[Ming Chen]

@ Liliputian_123
The bias of HI exhibits a clear diurnal cycle, which is definitely related to the ACM physics. I don't think modification of initial data can help solving the problem, since the model physics will evolve quickly and dominate the result.

I would suggest that you read the papers below and figure out what factors affect HI calculation. This will be a good starting point for you to move on.

Pleim, J. E., 2006: A simple, efficient solution of flux-profile relationships in the atmospheric surface layer, J. Appl. Meteor. and Clim., 45, 341–347.
doi:10.1175/JAM2339.1

Pleim, J. E., and A. Xiu, 1995: Development and testing of a surface flux and planetary boundary layer model for application in mesoscale models. J. Appl. Meteor., 34, 16-32.
doi:10.1175/1520-0450-34.1.16

 

[Liliputian_123]

Thank you for this! I'll read the suggested papers.