Wrong coordinates for U-velocity – in #10: INT2LM

in #10: INT2LM

<p> Dear colleagues, <br/> I have a problem with a nesting simulation cclm-to-cclm (13 km to 6.6 km). I have run the int2lm and become an error (see attached slurm-1071576.out). The error is ‘Wrong coordinates for U-velocity’. I didn’t see this error earlier. I have suggested thaint the problem is somewhere in coordinate system of coarse grid output, but I don’t know how to fix it. I have tried to add ‘lushift_in=.TRUE., lvshift_in=.TRUE.,’ to the <span class="caps"> INPUT </span> , but result was the same (see attached slurm-1074010.out). Maybe, I didn’t prescribe something in coarse model output (attached <span class="caps"> INPUT </span> _IO from the <span class="caps"> COSMO </span> - <span class="caps"> CLM </span> 13 km run). <br/> Could you suggest something about this problem? Thank you! </p>

  @vladimirplatonov in #ab33693

<p> Dear colleagues, <br/> I have a problem with a nesting simulation cclm-to-cclm (13 km to 6.6 km). I have run the int2lm and become an error (see attached slurm-1071576.out). The error is ‘Wrong coordinates for U-velocity’. I didn’t see this error earlier. I have suggested thaint the problem is somewhere in coordinate system of coarse grid output, but I don’t know how to fix it. I have tried to add ‘lushift_in=.TRUE., lvshift_in=.TRUE.,’ to the <span class="caps"> INPUT </span> , but result was the same (see attached slurm-1074010.out). Maybe, I didn’t prescribe something in coarse model output (attached <span class="caps"> INPUT </span> _IO from the <span class="caps"> COSMO </span> - <span class="caps"> CLM </span> 13 km run). <br/> Could you suggest something about this problem? Thank you! </p>

Wrong coordinates for U-velocity

VP wrote

Dear colleagues,
I have a problem with a nesting simulation cclm-to-cclm (13 km to 6.6 km). I have run the int2lm and become an error (see attached slurm-1071576.out). The error is ‘Wrong coordinates for U-velocity’. I didn’t see this error earlier. I have suggested thaint the problem is somewhere in coordinate system of coarse grid output, but I don’t know how to fix it. I have tried to add ‘lushift_in=.TRUE., lvshift_in=.TRUE.,’ to the INPUT , but result was the same (see attached slurm-1074010.out). Maybe, I didn’t prescribe something in coarse model output (attached INPUT _IO from the COSMO - CLM 13 km run).
Could you suggest something about this problem? Thank you!

View in channel
<p> In the <span class="caps"> GRIBOUT </span> namelist of <span class="caps"> CCLM </span> for output files in directory <code> out01 </code> you set <br/> <code> luvmasspoint=.TRUE., </code> <br/> This interpolates the U and V to mass points. For nesting a higher resolution <span class="caps"> CCLM </span> U and V have to stay on the flux points. <br/> Either re-run the 13km <span class="caps"> CCLM </span> with <br/> <code> luvmasspoint=.FALSE., </code> <br/> or just delete the <code> luvmasspoint=.TRUE., </code> line, since <code> luvmasspoint=.FALSE., </code> is the default. </p>

  @burkhardtrockel in #e861a0c

<p> In the <span class="caps"> GRIBOUT </span> namelist of <span class="caps"> CCLM </span> for output files in directory <code> out01 </code> you set <br/> <code> luvmasspoint=.TRUE., </code> <br/> This interpolates the U and V to mass points. For nesting a higher resolution <span class="caps"> CCLM </span> U and V have to stay on the flux points. <br/> Either re-run the 13km <span class="caps"> CCLM </span> with <br/> <code> luvmasspoint=.FALSE., </code> <br/> or just delete the <code> luvmasspoint=.TRUE., </code> line, since <code> luvmasspoint=.FALSE., </code> is the default. </p>
BR replied

In the GRIBOUT namelist of CCLM for output files in directory out01 you set
luvmasspoint=.TRUE.,
This interpolates the U and V to mass points. For nesting a higher resolution CCLM U and V have to stay on the flux points.
Either re-run the 13km CCLM with
luvmasspoint=.FALSE.,
or just delete the luvmasspoint=.TRUE., line, since luvmasspoint=.FALSE., is the default.

View in channel
<p> So, can I make nesting simulation without re-run? Which parameters should I change in int2lm <span class="caps"> INPUT </span> ? Sorry for the foolish question… </p>

  @vladimirplatonov in #a844883

<p> So, can I make nesting simulation without re-run? Which parameters should I change in int2lm <span class="caps"> INPUT </span> ? Sorry for the foolish question… </p>
VP replied

So, can I make nesting simulation without re-run? Which parameters should I change in int2lm INPUT ? Sorry for the foolish question…

View in channel
<p> Hallo Vladimir, <br/> I’m sorry, but I don’t know any int2lm namelist parameter which allows using u/v at masspoints in cosmo output for preparation of <span class="caps"> IBC </span> s at higher resolution. <br/> Burkhardt, did I oversee something? <br/> Shall we contact Uli Schaettler? </p> <p> If you do not have the computing time to resimulate everything, you have to interpolate u/v to the points at the boundaries of the grid boxes. <br/> Additionally the dependencies and grid coordinates have to be corrected in the files. <br/> Burkhardt, do you have a script for that? Me not. <br/> The coordinates can be generated by producing one correct output file. <br/> In this file you can see, which attributes need to be changed as well. </p> <p> Regards, Andreas </p>

  @andreaswill in #a30e63e

<p> Hallo Vladimir, <br/> I’m sorry, but I don’t know any int2lm namelist parameter which allows using u/v at masspoints in cosmo output for preparation of <span class="caps"> IBC </span> s at higher resolution. <br/> Burkhardt, did I oversee something? <br/> Shall we contact Uli Schaettler? </p> <p> If you do not have the computing time to resimulate everything, you have to interpolate u/v to the points at the boundaries of the grid boxes. <br/> Additionally the dependencies and grid coordinates have to be corrected in the files. <br/> Burkhardt, do you have a script for that? Me not. <br/> The coordinates can be generated by producing one correct output file. <br/> In this file you can see, which attributes need to be changed as well. </p> <p> Regards, Andreas </p>
AW replied

Hallo Vladimir,
I’m sorry, but I don’t know any int2lm namelist parameter which allows using u/v at masspoints in cosmo output for preparation of IBC s at higher resolution.
Burkhardt, did I oversee something?
Shall we contact Uli Schaettler?

If you do not have the computing time to resimulate everything, you have to interpolate u/v to the points at the boundaries of the grid boxes.
Additionally the dependencies and grid coordinates have to be corrected in the files.
Burkhardt, do you have a script for that? Me not.
The coordinates can be generated by producing one correct output file.
In this file you can see, which attributes need to be changed as well.

Regards, Andreas

View in channel