Missing update in the previous commit.
109 lines of code changed in 1 file:
* Bugs corrections:
- sequential gcm fixed
- parallel gcm compilation fixed ; to be tested
* Some generic operations moved from infotrac to readTracFile
* Fixed algebrical reduction routine, used in the isotopes parameters file.
* Additional component "comp" in the tracers descriptor derived type "tra",
specifying the model component name(s) (cf. tracers sections) it belongs.
* isotopes class selection tool fixed.
338 lines of code changed in 13 files:
Extension of the tracers management.
The tracers files can be:
1) "traceur.def": old format, with:
- the number of tracers on the first line
- one line for each tracer: <tracer name> <hadv> <vadv> [<parent name>]
2) "tracer.def": new format with one section each model component.
3) "tracer_<name>.def": new format with a single section.
The formats 2 and 3 reading is driven by the "type_trac" key, which can be a
coma-separated list of components.
* Format 2: read the sections from the "tracer.def" file.
* format 3: read one section each "tracer_<section name>.def" file.
* the first line of a section is "&<section name>
* the other lines start with a tracer name followed by <key>=<val> pairs.
* the "default" tracer name is reserved ; the other tracers of the section
inherit its <key>=<val>, except for the keys that are redefined locally.
This format helps keeping the tracers files compact, thanks to the "default"
special tracer and the three levels of factorization:
* on the tracers names: a tracer name can be a coma-separated list of tracers
=> all the tracers of the list have the same <key>=<val> properties
* on the parents names: the value of the "parent" property can be a
coma-separated list of tracers
=> only possible for geographic tagging tracers
* on the phases: the property "phases" is [g](l][s] (gas/liquid/solid)
Read information is stored in the vector "tracers(:)", of derived type "tra".
"isotopes_params.def" is a similar file, with one section each isotopes family.
It contains a database of isotopes properties ; if there are second generation
tracers (isotopes), the corresponding sections are read.
Read information is stored in the vector "isotopes(:)", of derived type "iso".
The "getKey" function helps to get the values of the parameters stored in
"tracers" or "isotopes".
4749 lines of code changed in 85 files:
Update the branch to the current trunk.
3280 lines of code changed in 164 files:
Fix for daily input data usage in limit_netcdf: remove 1 day shift between champan and champtime.
4 lines of code changed in 1 file:
Fix of the rev. 3231 for ozone horizontal interpolation.
Fields pre-r3231 were one point shifted in longitude in the easter part of the globe due to an index offset.
Difference in small, thanks to the quasi-zonal symetry of the 3D ozone fields.
2 lines of code changed in 3 files:
Fix in regr_horiz_time_climoz_m: <= instead of < in an index search loop to avoid
a possible error in the 3D ozone interpolation when input and output longitudes
numbers are equal.
Few typos in the comments corrected.
0 lines of code changed in 4 files:
Fix to restore compilation with gfortran + single length for strings.
17 lines of code changed in 1 file:
In limit_netcdf.f90:
* add a richer dictionary of recognized SST "units" attributes (taken from udunits).
* in case the unit is not specified, try to determine it from the values in the file:
mean zonal temperature at the equator for SST and maximum value for SIC.
0 lines of code changed in 2 files:
* Fix: Sig_bot was used before being initialized in "regr_pr_time_av".
* Add an optional output argument to the "dyn_tropopause" function:
the table of the last tropospheric layer in each column.
* A filter for irrealistic tropopause pressure (lower than 80hPa or
greater than 400hPa) is included in "dyn_tropopause". This is safer
if this function is not used together with "regr_pr_time_av" (which
contains already this filter), which is the case for the LMDZ version
with coupled REPROBUS and INCA chemistry.
13 lines of code changed in 3 files:
Desactivate additional debug outputs.
1 lines of code changed in 1 file:
Modification of the ozone interpolation routine to ensure more realistic
profiles, in particular when the tropopause adjustment is activated, as
well as an improved accuracy on the chemical tropopause of streched ozone
field:
* stretching function has no longer a trapezoidal shape, but is a hat
function, which top is at the target (ldmz) tropopause, with lower /
upper bounds respectively lower / higher than the lowest / highest
tropopause (lmdz and file). This helps to avoid stretching the profile
too far from the strictly required zone.
* revert to a classical linear vertical interpolation on log-pressure
because the stairs-shape effect of the conservative interpolation can not
be avoided, even at second order method. Note that the second order
conservative method is still used for Cariolle linearized ozone
coefficients interpolation: the regr_pr_time routine now switches from
one method to the other depending on flag "Ploc" (pressures must be given
at layers interfaces - Ploc=='I' - for conservative method and layers
centers - Ploc=='C' - for the regular one).
* force the stretching function to be discretized on at least three
levels to avoid the effect of the stretching to be reduced to a single
point shifting of the ozone profile when discretized.
249 lines of code changed in 3 files:
Fix the ozone vertical interpolation routines. Main features:
* lower initial pressure for chemical tropopause detection routine,
to avoid ozone holes, in 1998 in particular.
* min and max tropopauses pressure ensured for both field (chemical)
and lmdz (dynamical) tropopauses pressures for more safety.
* fix in the second order term of interpolation.
0 lines of code changed in 4 files: