WebPEP Domain related error in INT2LM read_nc_gdefs_ext_lm – in #11: EXTPAR / WebPEP

in #11: EXTPAR / WebPEP

<p> Hi there, </p> <p> currently I’m trying to run <span class="caps"> COSMO </span> with a slightly different domain compared to <span class="caps"> EURO </span> -Cordex. So I used WebPEP to create a domain which is just one grid point larger at the western boundary. But then INT2LM – fed with the <span class="caps"> EXTPAR </span> -File – returns an error: </p> <span class="caps"> OPEN </span> : ncdf-file: …/DOMAINS/europe-022_w+1.nc external data file namelist input ie_ext 233 233 je_ext 226 226 lonll -30.7399997711182 -30.7400000000000 latll -25.5200004577637 -25.5200000000000 lonur 20.2999992370605 20.3000000000000 latur 23.9799995422363 23.9800000000000 dlon 0.219999995724908 0.220000000000000 dlat 0.220000000000000 0.220000000000000 pollat 39.2500000000000 39.2500000000000 pollon -162.000000000000 -162.000000000000 polgam 0.000000000000000E+000 0.000000000000000E+000 wrong grid for LM external parameters <p> This error I do not understand. It looks like an rounding issue. Or INT2LM seems to be especially picky about rounding error. <br/> Here is the link to my WebPEP session: http://www.clm-community.eu/web_pep/domains/domain2017031409012/ </p> <p> Cheers, Klaus </p>

  @redc_migration in #3c19259

<p> Hi there, </p> <p> currently I’m trying to run <span class="caps"> COSMO </span> with a slightly different domain compared to <span class="caps"> EURO </span> -Cordex. So I used WebPEP to create a domain which is just one grid point larger at the western boundary. But then INT2LM – fed with the <span class="caps"> EXTPAR </span> -File – returns an error: </p> <span class="caps"> OPEN </span> : ncdf-file: …/DOMAINS/europe-022_w+1.nc external data file namelist input ie_ext 233 233 je_ext 226 226 lonll -30.7399997711182 -30.7400000000000 latll -25.5200004577637 -25.5200000000000 lonur 20.2999992370605 20.3000000000000 latur 23.9799995422363 23.9800000000000 dlon 0.219999995724908 0.220000000000000 dlat 0.220000000000000 0.220000000000000 pollat 39.2500000000000 39.2500000000000 pollon -162.000000000000 -162.000000000000 polgam 0.000000000000000E+000 0.000000000000000E+000 wrong grid for LM external parameters <p> This error I do not understand. It looks like an rounding issue. Or INT2LM seems to be especially picky about rounding error. <br/> Here is the link to my WebPEP session: http://www.clm-community.eu/web_pep/domains/domain2017031409012/ </p> <p> Cheers, Klaus </p>

WebPEP Domain related error in INT2LM read_nc_gdefs_ext_lm

re wrote

Hi there,

currently I’m trying to run COSMO with a slightly different domain compared to EURO -Cordex. So I used WebPEP to create a domain which is just one grid point larger at the western boundary. But then INT2LM – fed with the EXTPAR -File – returns an error:

OPEN : ncdf-file: …/DOMAINS/europe-022_w+1.nc external data file namelist input ie_ext 233 233 je_ext 226 226 lonll -30.7399997711182 -30.7400000000000 latll -25.5200004577637 -25.5200000000000 lonur 20.2999992370605 20.3000000000000 latur 23.9799995422363 23.9800000000000 dlon 0.219999995724908 0.220000000000000 dlat 0.220000000000000 0.220000000000000 pollat 39.2500000000000 39.2500000000000 pollon -162.000000000000 -162.000000000000 polgam 0.000000000000000E+000 0.000000000000000E+000 wrong grid for LM external parameters

This error I do not understand. It looks like an rounding issue. Or INT2LM seems to be especially picky about rounding error.
Here is the link to my WebPEP session: http://www.clm-community.eu/web_pep/domains/domain2017031409012/

Cheers, Klaus

View in channel
<p> Hi Klaus, </p> <p> I am not quite sure whether I understand correctly what you did. <br/> Your “target” domain size for the <span class="caps"> CCLM </span> run is 233*226 (ie_ext*je_ext),isn’t it? <br/> And you have chosen the identical domain and size for the external data set, haven’t you? <br/> To see the file <span class="caps"> INPUT </span> of INT2LM would be helpful. </p> <p> If you did it in this way, then INT2LM would never run. <br/> The size of the domain of external data has to be larger then the size of the actual <span class="caps"> CCLM </span> domain. <br/> I am not quite sure, but at least by 2 further grid points at each of the four sides. </p> <p> I always recommend to chose the domain of the external data set rather large. Then you can “play” with <br/> the size of the <span class="caps"> CCLM </span> domain, with shifts of start_lon and start_lat, etc. without creating new external data sets. </p> <p> Hans-Juergen </p>

  @hans-jürgenpanitz in #1f87894

<p> Hi Klaus, </p> <p> I am not quite sure whether I understand correctly what you did. <br/> Your “target” domain size for the <span class="caps"> CCLM </span> run is 233*226 (ie_ext*je_ext),isn’t it? <br/> And you have chosen the identical domain and size for the external data set, haven’t you? <br/> To see the file <span class="caps"> INPUT </span> of INT2LM would be helpful. </p> <p> If you did it in this way, then INT2LM would never run. <br/> The size of the domain of external data has to be larger then the size of the actual <span class="caps"> CCLM </span> domain. <br/> I am not quite sure, but at least by 2 further grid points at each of the four sides. </p> <p> I always recommend to chose the domain of the external data set rather large. Then you can “play” with <br/> the size of the <span class="caps"> CCLM </span> domain, with shifts of start_lon and start_lat, etc. without creating new external data sets. </p> <p> Hans-Juergen </p>
HP replied

Hi Klaus,

I am not quite sure whether I understand correctly what you did.
Your “target” domain size for the CCLM run is 233*226 (ie_ext*je_ext),isn’t it?
And you have chosen the identical domain and size for the external data set, haven’t you?
To see the file INPUT of INT2LM would be helpful.

If you did it in this way, then INT2LM would never run.
The size of the domain of external data has to be larger then the size of the actual CCLM domain.
I am not quite sure, but at least by 2 further grid points at each of the four sides.

I always recommend to chose the domain of the external data set rather large. Then you can “play” with
the size of the CCLM domain, with shifts of start_lon and start_lat, etc. without creating new external data sets.

Hans-Juergen

View in channel
<p> Hi Hans-Jürgen, </p> <p> thanks for the information. I didn’t have in mind, that the <span class="caps"> EXTPAR </span> dataset has to be on a enlarged grid compared to the <span class="caps"> COSMO </span> domain. <br/> Now the error message makes sense. </p> <p> Klaus </p>

  @redc_migration in #040a8a0

<p> Hi Hans-Jürgen, </p> <p> thanks for the information. I didn’t have in mind, that the <span class="caps"> EXTPAR </span> dataset has to be on a enlarged grid compared to the <span class="caps"> COSMO </span> domain. <br/> Now the error message makes sense. </p> <p> Klaus </p>
re replied

Hi Hans-Jürgen,

thanks for the information. I didn’t have in mind, that the EXTPAR dataset has to be on a enlarged grid compared to the COSMO domain.
Now the error message makes sense.

Klaus

View in channel