[MITgcm-support] KPP parameters not read?

[Donald Slater]

Hi Patrick and Jean-Michel,

Thank you - KPP is definitely now reading the values I define. As Jean-Michel suggests, if you put a value in which does not make sense, KPP appears to use the default value without producing an error.

Donald


> On Oct 18, 2017, at 11:21 AM, Jean-Michel Campin <jmc at mit.edu> wrote:
> 
> Hi Donald,
> 
> I think Patrick is right, in the sense that the "." is misplaced
> (the tricky thing is that it might be compiler dependent like many
> subtle features of namelist reading).
> 
> And the reason why it does not produce any error when reading data.kpp (again, 
> this could be compiler dependent), as opposed to a wrong line in an other package 
> parameter file, is likely due to the "IOSTAT=errIO)" part in the namelist read
> code (in kpp_readparms.F); most other packages (including  pkg/icefront)
> use the simplest form to read a namelist, without IOSTAT).
> In this regard, this is similar to the previous message you pointed at:
>> This reminds me a little of another message on the support list (http://mailman.mitgcm.org/pipermail/mitgcm-support/2005-July/003245.html).
> 
> Going to change kpp_readparms.F very soon so that it will stop if a wrong line is
> found in "data.kpp", like the other package parameter files.
> 
> The other thing that does not help here with "data.kpp" is that the parameters
> are not reported to STDOUT (unlike most other packages).
> 
> Cheers,
> Jean-Michel
> 
> On Wed, Oct 18, 2017 at 09:29:04AM -0700, Donald Slater wrote:
>> Hi Jean-Michel,
>> 
>> Here it is.
>> As I said before, the simulation runs fine even if I change to e.g. ???difmo = Donald.,???
>> 
>> Thanks,
>> Donald
> 
> 
>> 
>>> On Oct 18, 2017, at 9:14 AM, Jean-Michel Campin <jmc at mit.edu> wrote:
>>> 
>>> Hi Donald,
>>> 
>>> Could you attach the parameter file "data.kpp" that you are using
>>> with specified value for difm0, difs0 and dift0 that are ignored ?
>>> 
>>> Cheers,
>>> Jean-Michel
>>> 
>>> On Tue, Oct 17, 2017 at 04:45:11PM -0700, Donald Slater wrote:
>>>> Just to clarify, as in retrospect my previous email was a little confusing.
>>>> 
>>>> I do have "useKPP=.TRUE." in data.pkg. And I know KPP is running because I get the expected variable vertical viscosity when I print out the relevant diagnostic.
>>>> 
>>>> But I still have the issue with the simulation appearing to ignore the parameters in data.kpp
>>>> 
>>>> Thanks,
>>>> Donald
>>>> 
>>>> 
>>>>> On Oct 17, 2017, at 4:13 PM, Dimitris Menemenlis <dmenemenlis at gmail.com> wrote:
>>>>> 
>>>>> Did you specify "useKPP=.TRUE." in data.pkg?
>>>>> 
>>>>>> On Oct 17, 2017, at 4:06 PM, Donald Slater <daslater at ucsd.edu> wrote:
>>>>>> 
>>>>>> Hi,
>>>>>> 
>>>>>> I am running some simulations with the KPP package enabled, but I think I have a strange problem reading parameters from data.kpp. I noticed this because I changed some parameters (difm0, difs0 and dift0) without having any effect on my simulation.
>>>>>> 
>>>>>> The output stream from my simulation suggests that data.kpp is being read (i.e. I get the lines "KPP_INIT: opening data.kpp ??? OPEN_COPY_DATA_FILE: opening file data.kpp ... KPP_INIT: finished reading data.kpp???).
>>>>>> 
>>>>>> However if I try to trip the simulation up by inserting things which don???t make sense into data.kpp, for example setting "difm0 = Donald???, the simulation carries on as usual with the default value for difm0 and no error messages. In contrast, if I insert something silly into data.icefront (e.g. icefrontthetasurface = Donald), the simulation immediately stops. This suggests to me that somehow difm0 (and perhaps the other parameters in data.kpp) are not actually being read.
>>>>>> 
>>>>>> This reminds me a little of another message on the support list (http://mailman.mitgcm.org/pipermail/mitgcm-support/2005-July/003245.html). I???ve looked at kpp_readparms and compared it to similar files for other packages but can???t see anything amiss.
>>>>>> 
>>>>>> Has anyone experienced this or can offer any pointers as to what might be going on?
>>>>>> 
>>>>>> Many thanks,
>>>>>> Donald
>>>>>> 
>>>>>> _______________________________________________
>>>>>> MITgcm-support mailing list
>>>>>> MITgcm-support at mitgcm.org
>>>>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
>>>>> 
>>>>> _______________________________________________
>>>>> MITgcm-support mailing list
>>>>> MITgcm-support at mitgcm.org
>>>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
>>>> 
>>>> _______________________________________________
>>>> MITgcm-support mailing list
>>>> MITgcm-support at mitgcm.org
>>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
>>> _______________________________________________
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
>> 
> 
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support