large differences in version INT2LM_2.00_clm4 und INT2LM_2.05_clm2 – in #10: INT2LM

in #10: INT2LM

<p> Dear all, </p> <p> I want to test ERA5 as forcing data. As I was at first experiencing problems with the INT2LM version 2.00_clm4 in doing this, I tried it also with the newer version INT2LM_2.05_clm2. </p> <p> I was now interested in the differences of the versions, as I want maybe to switch between ERA40+CCLM-0.11° and ERA5 forcing during the evaluation run (19790101) and therefore maybe also to switch the INT2LM version. </p> <p> I did a small test: for 1977_12 I ran with the ERA40+CCLM-0.11° forcing with INT2LM_2.05_clm2 and INT2LM_2.00_clm4. I compared the output with the two different versions and found quite some large differences, especially in the Alps, but also else-where in the domain (which is a <span class="caps"> COSMO </span> -DE domain with 2.8km grid spacing). I attach some plots of the differences at some arbitrary point in time (it does not change much during the month) of T_S (INT2LM output) and PP (level 45, INT2LM output), T_2M ( <span class="caps"> CCLM </span> output) and <span class="caps"> CLCT </span> ( <span class="caps"> CCLM </span> output). </p> <p> Question: did anyone already compare the 2 versions? is this realistic to have these large differences? What could be the reason? </p> <p> Thanks! Best wishes, Susanne </p>

  @susannebrienen in #15b28d8

<p> Dear all, </p> <p> I want to test ERA5 as forcing data. As I was at first experiencing problems with the INT2LM version 2.00_clm4 in doing this, I tried it also with the newer version INT2LM_2.05_clm2. </p> <p> I was now interested in the differences of the versions, as I want maybe to switch between ERA40+CCLM-0.11° and ERA5 forcing during the evaluation run (19790101) and therefore maybe also to switch the INT2LM version. </p> <p> I did a small test: for 1977_12 I ran with the ERA40+CCLM-0.11° forcing with INT2LM_2.05_clm2 and INT2LM_2.00_clm4. I compared the output with the two different versions and found quite some large differences, especially in the Alps, but also else-where in the domain (which is a <span class="caps"> COSMO </span> -DE domain with 2.8km grid spacing). I attach some plots of the differences at some arbitrary point in time (it does not change much during the month) of T_S (INT2LM output) and PP (level 45, INT2LM output), T_2M ( <span class="caps"> CCLM </span> output) and <span class="caps"> CLCT </span> ( <span class="caps"> CCLM </span> output). </p> <p> Question: did anyone already compare the 2 versions? is this realistic to have these large differences? What could be the reason? </p> <p> Thanks! Best wishes, Susanne </p>

large differences in version INT2LM_2.00_clm4 und INT2LM_2.05_clm2

SB wrote

Dear all,

I want to test ERA5 as forcing data. As I was at first experiencing problems with the INT2LM version 2.00_clm4 in doing this, I tried it also with the newer version INT2LM_2.05_clm2.

I was now interested in the differences of the versions, as I want maybe to switch between ERA40+CCLM-0.11° and ERA5 forcing during the evaluation run (19790101) and therefore maybe also to switch the INT2LM version.

I did a small test: for 1977_12 I ran with the ERA40+CCLM-0.11° forcing with INT2LM_2.05_clm2 and INT2LM_2.00_clm4. I compared the output with the two different versions and found quite some large differences, especially in the Alps, but also else-where in the domain (which is a COSMO -DE domain with 2.8km grid spacing). I attach some plots of the differences at some arbitrary point in time (it does not change much during the month) of T_S (INT2LM output) and PP (level 45, INT2LM output), T_2M ( CCLM output) and CLCT ( CCLM output).

Question: did anyone already compare the 2 versions? is this realistic to have these large differences? What could be the reason?

Thanks! Best wishes, Susanne

View in channel
<p> Dear Susanne, </p> <p> My first idea when I saw your diffs was the following from 2.03 (short version): </p> <p> <pre><code class="text"><br/> - New option for vertical adaptation of boundary layer profiles (T, u, v, w, rel. hum., hydrometeors) from the input orography to the <span class="caps">COSMO</span> orography in case of non-hydrostatic input models (<span class="caps">COSMO</span>, <span class="caps">ICON</span>, UM, CM). The new option can be selected by setting the new namelist parameter</code></pre> </p> itype_profiles_vert_interp = 2 (“1” is the previous method, default is 2!) <p> </p> <p> But this change is only for non-hydrostatic models. So if the description is correct and there is no bug in the implentation, these changes should not cause your diff (you could still test, as this is the most disruptive change that I can remember in this context). </p> <p> Could it be the following change? </p> <pre>&lt;code class="text"&gt; New option to ensure consistency of hydrostatic balanced pressure with the COSMO RK-core if one has chosen "lbalance_pp = .TRUE.". Uses hydrostatic pressure integration routines from COSMO src_artifdata.f90, which correctly take into account the type of the dynamical core. &lt;/code&gt;</pre> <p> Best regards <br/> Daniel </p>

  @danielrieger in #b348e5b

<p> Dear Susanne, </p> <p> My first idea when I saw your diffs was the following from 2.03 (short version): </p> <p> <pre><code class="text"><br/> - New option for vertical adaptation of boundary layer profiles (T, u, v, w, rel. hum., hydrometeors) from the input orography to the <span class="caps">COSMO</span> orography in case of non-hydrostatic input models (<span class="caps">COSMO</span>, <span class="caps">ICON</span>, UM, CM). The new option can be selected by setting the new namelist parameter</code></pre> </p> itype_profiles_vert_interp = 2 (“1” is the previous method, default is 2!) <p> </p> <p> But this change is only for non-hydrostatic models. So if the description is correct and there is no bug in the implentation, these changes should not cause your diff (you could still test, as this is the most disruptive change that I can remember in this context). </p> <p> Could it be the following change? </p> <pre>&lt;code class="text"&gt; New option to ensure consistency of hydrostatic balanced pressure with the COSMO RK-core if one has chosen "lbalance_pp = .TRUE.". Uses hydrostatic pressure integration routines from COSMO src_artifdata.f90, which correctly take into account the type of the dynamical core. &lt;/code&gt;</pre> <p> Best regards <br/> Daniel </p>
DR replied

Dear Susanne,

My first idea when I saw your diffs was the following from 2.03 (short version): 



-  New option for vertical adaptation of boundary layer profiles (T, u, v, w, rel. hum., hydrometeors)
   from the input orography to the COSMO orography in case of non-hydrostatic input models
   (COSMO, ICON, UM, CM).
   The new option can be selected by setting the new namelist parameter

itype_profiles_vert_interp = 2 (“1” is the previous method, default is 2!)

But this change is only for non-hydrostatic models. So if the description is correct and there is no bug in the implentation, these changes should not cause your diff (you could still test, as this is the most disruptive change that I can remember in this context).

Could it be the following change? 

<code class="text">
New option to ensure consistency of hydrostatic balanced pressure with the
    COSMO RK-core if one has chosen "lbalance_pp = .TRUE.". Uses hydrostatic pressure integration
    routines from COSMO src_artifdata.f90, which correctly take
    into account the type of the dynamical core.
</code>

Best regards
Daniel

View in channel
<p> Hi Daniel, </p> <p> I have seen these comments, and already did an additional test setting: <br/> &gt; itype_balance_pp=1, itype_profiles_vert_interp=1, lmultlay_deepsoil_clim_hcorr=.FALSE., <br/> which means using the old defaults methods. (in the old version I had lbalance_pp=T, which is also hard-coded in the new version). <br/> But the differences are in the same order of magnitude, even between the two INT2LM_2.05_clm2 runs with the two different namelist settings. </p> <p> I started with the configuration of the <span class="caps"> CORDEX </span> - <span class="caps"> FPS </span> -Convection simulations, which used INT2LM_2.00_clm4. I attach the settings. For INT2LM_2.05_clm2 I just deleted lbalance_pp (and then changed the 3 namelist parameters in the sensitivity test mentioned above), but maybe other settings would be necessary to adapt? </p> <p> Best wishes, <br/> Susanne </p>

  @susannebrienen in #7285e35

<p> Hi Daniel, </p> <p> I have seen these comments, and already did an additional test setting: <br/> &gt; itype_balance_pp=1, itype_profiles_vert_interp=1, lmultlay_deepsoil_clim_hcorr=.FALSE., <br/> which means using the old defaults methods. (in the old version I had lbalance_pp=T, which is also hard-coded in the new version). <br/> But the differences are in the same order of magnitude, even between the two INT2LM_2.05_clm2 runs with the two different namelist settings. </p> <p> I started with the configuration of the <span class="caps"> CORDEX </span> - <span class="caps"> FPS </span> -Convection simulations, which used INT2LM_2.00_clm4. I attach the settings. For INT2LM_2.05_clm2 I just deleted lbalance_pp (and then changed the 3 namelist parameters in the sensitivity test mentioned above), but maybe other settings would be necessary to adapt? </p> <p> Best wishes, <br/> Susanne </p>
SB replied

Hi Daniel,

I have seen these comments, and already did an additional test setting:
> itype_balance_pp=1, itype_profiles_vert_interp=1, lmultlay_deepsoil_clim_hcorr=.FALSE.,
which means using the old defaults methods. (in the old version I had lbalance_pp=T, which is also hard-coded in the new version).
But the differences are in the same order of magnitude, even between the two INT2LM_2.05_clm2 runs with the two different namelist settings.

I started with the configuration of the CORDEX - FPS -Convection simulations, which used INT2LM_2.00_clm4. I attach the settings. For INT2LM_2.05_clm2 I just deleted lbalance_pp (and then changed the 3 namelist parameters in the sensitivity test mentioned above), but maybe other settings would be necessary to adapt?

Best wishes,
Susanne

View in channel