Opened 7 years ago

Closed 6 years ago

#130 closed defect (wontfix)

Problem with interpolation of reference temperature in LMDZOR zoom mode

Reported by: maignan Owned by: jgipsl
Priority: major Milestone: IPSLCM6.v1
Component: Anthropogenic processes Version:
Keywords: Cc:

Description

Betreff: aggregate_p interpolation in stomate_io.f90: zoom works (but not elegant)
Datum: Freitag, 28. Februar 2014 18:45 CET
Von: "Juliane Otto" <Juliane.Otto@…>

Hi,

finally the interpolation (and the zoom) works on CURIE. I tested it with LMDZ-DOFOCO and LMDZ-trunk. I copied the changes to ORCHIDEE trunk so that the old interpolation in stomate_io.f90 is replaced by 'aggregate_p'. Here is the link to my directory in case you want to incorporate the changes into the svn of the trunk:

@CURIE:
/ccc/cont003/home/dsm/p529otto/COUP_ZOOM/modeles/ORCHIDEE

However, one little problem remains. Maybe you can find a solution for that, we would use it then in DOFOCO too. In the case that there aren't enough pixels to interpolate the temperature, these pixels get now the constant value of 13.9 °C. This is the mean earth surface temperature. However, this is a quick and dirty solution.

It would be more elegant to search for the nearest neighbour and use this value. I tried that and the code for that is in stomate_io.f90. I use the parameter 'neighbour' from grid.f90. However, the neighbours dimension doesn't correspond to the dimension of the variable trefe in get_reftemp, The error message is:

Subscript #1 of the array TREFE has value 469 which is greater than the upper bound of 288

This happens at the line:

                    IF (trefe(neighbours(ib,ic)) .NE. undef .AND. &
                        trefe(neighbours(ib,ic)) .NE. mstemp ) THEN

I guess this has something to do with the parallelisation of the code (when I tested it in on obelix it worked). This is out of my scope, I don't know much about it. In case you find a solution of it, could you led me know? We will use it in DOFOCO too.

Change History (7)

comment:1 Changed 7 years ago by maignan

Additional question by J. Otto (March, 7, 2014):
I would like to ask a question that came up during our DOFOCO meeting: using the mean global average temperature for the pixels without interpolation - do you think that is feasible? Or do you think that is too simple?

comment:2 Changed 7 years ago by nvuilsce

  • Owner changed from somebody to jgipsl
  • Status changed from new to assigned

comment:3 Changed 7 years ago by nvuilsce

Comment from Ticket Meeting (13/05/2014): One could consider to remove the reftemp file as N. Viovy did in his code version. To discuss. At least, one should only read its content when starting from scratch and then from a restart file.

comment:4 follow-up: Changed 7 years ago by nvuilsce

Note from Ticket meeting : To harmonize with the read of reftemp in sechiba in the MICT version.

comment:5 Changed 6 years ago by jgipsl

  • Milestone set to IPSLCM6.v1

comment:6 in reply to: ↑ 4 Changed 6 years ago by jgipsl

Replying to nvuilsce:

Note from Ticket meeting : To harmonize with the read of reftemp in sechiba in the MICT version.

In thermosoil, if read_reftemp=true, the reftemp.nc file is read to initialize the vertical soil temperature (ptn). The variable REFTEMP_FILE in thermosoil is now changed to SOIL_REFTEMP_FILE to distinguish from the REFTEMP_FILE variable in stomate_io used to initialize the t2m longterm temperature. [2441]

comment:7 Changed 6 years ago by jgipsl

  • Resolution set to wontfix
  • Status changed from assigned to closed

In rev [2441], the default is now to calculate the reference 2m temperature instead of reading if from file. See ticket #158

The reading of reftemp.nc is still possible if setting REFTEMP_FILE=filename (default:NONE). The interpolation method will not be revised as this is only an optional non recommended method. Ticket will be closed without changes in the code.

Note: See TracTickets for help on using tickets.