|
Table of contents
Chapter 1. Technical overview
Chapter 2. FULL-POS post-processing
and interpolation
Chapter 3. Parallel implementation
REFERENCES
|
|
Next
Section
Previous Section
Contains variables and pointers for grid point fields used for horizontal
post-processing. Variables are initialised in setup routine setup/SUFPSC2B.
No variable in namelist. All variables are DM-global.
List of variables:
| |
• NFPVT0: number of auxiliary
surface grid point fields. |
| |
• MATS0 : pointer for
output surface temperature. |
| |
• MATSI : pointer for
interpolated surface temperature. |
| |
• MAHS0 : pointer for
output surface relative moisture. |
| |
• MASDO : pointer for
(standard deviation of orography)*g. |
| |
• MADOU : pointer for
zonal component of topography principal axis. |
| |
• MADOV : pointer for
meridian component of topography principal axis. |
| |
• MAACT : pointer for
anisotropy coefficient of topography. |
PTRFPB2 has to be split later into YOMFPB2
containing NFPVT0 and a pointer (directory pointer)
PTRFPB2 containing pointers.
Contains variables relative to FULL-POS working arrays (level 4). Variables
are initialised in setup routine setup/SU4FPOS.
No variable in namelist. All variables are DM-global.
Post-processing on 3D dynamical variables:
| |
• NTFPP : total number
of fields in pressure levels. |
| |
• NTFPH : total number
of fields in height levels. |
| |
• NTFPTH: total number
of fields in  -levels. |
| |
• NTFPPV: total number
of fields in potential vorticity levels. |
| |
• NTFPS : total number
of fields in  -levels. |
| |
• NOFPP : total number
of pressure levels for each field. Array of dimension NFP3DF. |
| |
• NOFPH : total number
of height levels for each field. Array of dimension NFP3DF. |
| |
• NOFPTH: total number
of  -levels for each field. Array of dimension NFP3DF. |
| |
• NOFPPV: total number
of potential vorticity levels for each field. Array of dimension NFP3DF. |
| |
• NOFPS : total number
of  -levels for each field. Array of dimension NFP3DF. |
| |
• NAFPP : total number
of subdomains for each pressure level of each field. Array of dimension
(NFP3P,NFP3DF). |
| |
• NAFPH : total number
of subdomains for each height level of each field. Array of dimension
(NFP3H,NFP3DF). |
| |
• NAFPT : total number
of subdomains for each  -level of each field. Array of dimension
(NFP3TH,NFP3DF). |
| |
• NAFPV : total number
of subdomains for each potential vorticity level of each field. Array
of dimension (NFP3PV,NFP3DF). |
| |
• NAFPS : total number
of subdomains for each  -level of each field. Array of dimension
(NFP3S,NFP3DF). |
| |
• NCPFP : field pointers
in pressure levels. Array of dimension NFP3DF. |
| |
• NCHFP : field pointers
in height levels. Array of dimension NFP3DF. |
| |
• NCTFP : field pointers
in  -levels. Array of dimension NFP3DF. |
| |
• NCVFP : field pointers
in potential vorticity levels. Array of dimension NFP3DF. |
| |
• NCSFP : field pointers
in  -levels. Array of dimension NFP3DF. |
| |
• NPFP : pressure level
pointers for each field. Array of dimension (NFP3P,NFP3DF). |
| |
• NHFP : height level
pointers for each field. Array of dimension (NFP3H,NFP3DF). |
| |
• NTFP :  -level pointers
for each field. Array of dimension (NFP3TH,NFP3DF). |
| |
• NVFP : potential vorticity
level pointers for each field. Array of dimension (NFP3PV,NFP3DF). |
| |
• NSFP :  -level pointers
for each field. Array of dimension (NFP3S,NFP3DF). |
| |
• NIPFP : subdomain index
for each pressure level of each field. Array of dimension (NFPDOM,NFP3P,NFP3DF). |
| |
• NIHFP : subdomain index
for each height level of each field. Array of dimension (NFPDOM,NFP3H,NFP3DF). |
| |
• NITFP : subdomain index
for each  -level of each field. Array of dimension (NFPDOM,NFP3TH,NFP3DF). |
| |
• NIVFP : subdomain index
for each potential vorticity level of each field. Array of dimension
(NFPDOM,NFP3PV,NFP3DF). |
| |
• NISFP : subdomain index
for each  -level of each field. Array of dimension (NFPDOM,NFP3S,NFP3DF). |
Post-processing on 2D dynamical variables:
| |
• NOFP2F: total number
of 2D fields. |
| |
• NOFP2A: total number
of subdomains for each 2D field. Array of dimension NFP2DF. |
| |
• NC2FP : 2D field pointers.
Array of dimension NFP2DF. |
| |
• NI2FP : subdomain index
for each 2D field. Array of dimension (NFPDOM,NFP2DF). |
Post-processing on lagged physical and instantaneous fluxes variables:
| |
• NPHYFPB9: number of
physical fields. |
| |
• NXFUFPB9: number of
instantaneous fluxes. |
| |
• NAPHYFP9: number of
subdomains for each physical field. Array of dimension NFPPHY. |
| |
• NAXFUFP9: number of
subdomains for each instantaneous flux. Array of dimension
NFPXFU. |
| |
• NCPHYFP9: field pointers
of physics. Array of dimension NFPPHY. |
| |
• NCXFUFP9: field pointers
of instantaneous fluxes. Array of dimension NFPXFU. |
| |
• NIPHYFP9: subdomain
index for each physical field. Array of dimension (NFPDOM,NFPPHY). |
| |
• NIXFUFP9: subdomain
index for each instantaneous flux. Array of dimension (NFPDOM,NFPXFU). |
Contains ARPEGE fields descriptors. Variables are initialised in setup routine
setup/SUAFN. All variables are DM-global.
ARPEGE fields descriptors:
| |
• CNAM3DS, C1NAM3:
3D dynamical fields. Global variable in equivalence with the 43 variables
CNZ to CNUA16 described in paragraph
"3D dynamical fields". |
| |
• CNAM2DS, C1NAM2:
2D dynamical fields. Global variable in equivalence with the 11 variables
CNSP to CNLNSP described in paragraph
"2D dynamical fields". |
| |
• CNAMPDS, C1NAMP:
physical surface fields. Global variable in equivalence with the 109
variables CNLSM to CNPSU30 described
in paragraph "Physical surface fields". Fields CNSTL1
to CNPSU30 are used only in the ECMWF physics. |
| |
• CNAMCDS, C1NAMC:
surface cumulated fluxes (CFU). Global variable in equivalence with
the 53 variables CNCLSP to CNCTP
described in paragraph "Surface cumulated fluxes". |
| |
• CNAMXDS, C1NAMX:
surface instantaneous fluxes (XFU). Global variable in equivalence
with the 41 variables CNXTCC to CNXSNS
described in paragraph "Surface instantaneous fluxes". |
3D dynamical fields (DYN3D):
Remark:
There are 13 additional fields "free upper air field nr 4" (DYN3D number
31) to "free upper air field nr 16" (DYN3D number 43) which are used only
at ECMWF, the corresponding quantities CNUA4 to CNUA16
and NBUA4 to NBUA16 are not in namelist
NAMAFN.
2D dynamical fields (DYN2D):
Remarks:
| |
• there is one additional field "logarithm
of surface pressure" (DYN2D number 11) which is used only at ECMWF,
the corresponding quantities CNLNSP and NBLNSP
are not in namelist NAMAFN. |
| |
• fields DYN2D number 06 to 10 have a different
definition at ECMWF (optional surface fields), what is written in
the previous table is valid only for METEO-FRANCE. |
Physical surface fields (PHYSOL):
Remarks:
| |
• fields PHYSOL numbers 21 to 36 are used
only at ECMWF, the corresponding quantities CNSTL1
to CNSRC and NBSTL1 to NBSRC
are not in namelist NAMAFN. |
| |
• there are additional fields PHYSOL numbers
37 to 79 (not detailed here) which are used only at ECMWF, the corresponding
quantities CN... and NB... are
not in namelist NAMAFN. These fields are diagnostic
ones and some of them are present in the list of CFU or XFU fields
available at METEO-FRANCE. |
| |
• there are additional optional fields
PHYSOL numbers 80 to 109 (not detailed here) which are used only at
ECMWF, the corresponding quantities CNPSU1 to CNPSU30
and NBPSU1 to NBPSU30 are not in
namelist NAMAFN. |
Surface cumulated fluxes:
Surface instantaneous fluxes:
Number of bits for coding in FULL-POS:
| |
• NBIT3DS: 3D dynamical
fields. Global variable in equivalence with the 43 variables NBZ
to NBUA16 described in paragraph "3D dynamical fields". |
| |
• NBIT2DS: 2D dynamical
fields. Global variable in equivalence with the 11 variables NBSP
to NBLNSP described in paragraph "2D dynamical fields". |
| |
• NBITPDS: physical surface
fields. Global variable in equivalence with the 20 variables NBLSM
to NBDPAT described in paragraph "Physical surface
fields". Fields NBSTL1 to NBSRC
are used only in the ECMWF physics and are not yet implemented in
the cycles 17 and 18 of ARPEGE/IFS. |
| |
• NBITCDS: surface cumulated
fluxes (CFU). Global variable in equivalence with the 53 variables
NBCLSP to NBCTP described in paragraph
"Surface cumulated fluxes". |
| |
• NBITXDS: surface instantaneous
fluxes (XFU). Global variable in equivalence with the 41 variables
NBXTCC to NBXSNS described in paragraph
"Surface instantaneous fluxes". |
| |
• Default value is 24 for all these quantities. |
Scalar/vector descriptor (0 for scalar, 1 for vector):
| |
• NVECPDS: physical surface
fields. Global variable in equivalence with the 20 variables NVLSM
to NVDPAT (same suffix as variables, the name of
which starts by NB...). Fields NVSTL1
to NVSRC are used only in the ECMWF physics and
are not yet implemented in the cycles 17 and 18 of ARPEGE/IFS. |
| |
• NVECCDS: surface cumulated
fluxes (CFU). Global variable in equivalence with the 53 variables
NVCLSP to NVCTP (same suffix as
variables, the name of which starts by NB...). |
| |
• NVECXDS: surface instantaneous
fluxes (XFU). Global variable in equivalence with the 41 variables
NVXTCC to NVXSNS (same suffix as
variables, the name of which starts by NB...). |
| |
• Default value is 1 for NVCUSS,
NVCVSS, NVCUGW, NVCVGW,
NVX10U, NVX10V, 0 for the other
quantities. |
Use of land-sea mask for horizontal interpolations (0=no; 1=yes):
| |
• NITMPDS: physical surface
fields. Global variable in equivalence with the 20 variables NILSM
to NIDPAT (same suffix as variables, the name of
which starts by NB...). Fields NISTL1
to NISRC are used only in the ECMWF physics and
are not yet implemented in the cycles 17 and 18 of ARPEGE/IFS. |
| |
• NITMCDS: surface cumulated
fluxes (CFU). Global variable in equivalence with the 53 variables
NICLSP to NICTP (same suffix as
variables, the name of which starts by NB...). |
| |
• NITMXDS: surface instantaneous
fluxes (XFU). Global variable in equivalence with the 41 variables
NIXTCC to NIXSNS (same suffix as
variables, the name of which starts by NB...). |
| |
• Default value is 0 for NILSM,
NIGFIS, NISDOG, 1 for the other
physical surface fields quantities, 1 for NICUSS,
NICVSS, NICUGW, NICVGW,
NIX10U, NIX10V, NIX2T,
NIX2SH, NIX2RH, NIXX2T,
NIXN2T, 0 for the other CFU or XFU quantities. |
Namelist NAMAFN:
The following variables of YOMAFN are in namelist NAMAFN:
CNZ, CNT, CNU, CNV,
CNQ, CNR, CNS,
CNVV, CNVOR, CNDIV, CNTH,
CNPSI, CNKHI, CNW, CNTPW,
CNCLF, CNWND, CNETH,
CNABS, CNSTD, CNSHD,
CNPV, CNWPV, CNSCVA,
CNP, CNUA1, CNUA2,
CNUA3, CNSP, CNMSL, CNFIS,
CNGM, CNFOL, CNSU1,
CNSU2, CNSU3, CNSU4,
CNSU5, CNLSM, CNGFIS,
CNST, CNDST, CNRDST,
CNSSW, CNDSW, CNRDSW,
CNCSSW, CNCDSW, CNSD,
CNSR, CNBSR, CNAL,
CNEMIS, CNSDOG, CNVEG, CNLAN,
CNACOT, CNDPAT, CNCLSP,
CNCCP, CNCLSS, CNCCSF,
CNCUSS, CNCVSS, CNCSSH,
CNCSLH, CNCTSP, CNCTCC,
CNCBLD, CNCSSR, CNCSTR,
CNCTSR, CNCTTR, CNCCCC,
CNCHCC, CNCMCC, CNCLCC,
CNCUGW, CNCVGW, CNCE,
CNCS, CNCLHE, CNCLHS,
CNCC, CNCWS, CNCSNS,
CNCQTO, CNCTO3, CNCTME,
CNCICE, CNCLI, CNCCVU,
CNCCVV, CNCCVQ, CNCCVS,
CNCTUQ, CNCTUS, CNCSOC,
CNCTHC, CNCSOP, CNCTOP,
CNCSOD, CNCTHD, CNCFON,
CNCCHS, CNCEAS, CNCSRU,
CNCDRU, CNCIRU, CNCETP,
CNCTP, CNXTCC, CNX10U,
CNX10V, CNX2T, CNX2SH,
CNX2RH, CNXCCC, CNXHCC,
CNXMCC, CNXLCC, CNXX2T,
CNXN2T, CNXC, CNXCVU,
CNXCVV, CNXCVQ, CNXCVS,
CNXTUU, CNXTUV, CNXTUQ,
CNXTUS, CNXGDU, CNXGDV,
CNXLSP, CNXCP, CNXLSS,
CNXCSF, CNXSSR, CNXSTR,
CNXTSR, CNXTTR, CNXUBL,
CNXVBL, CNXTBL, CNXQBL,
CNXGBL, CNXST, CNXDT,
CNXSW, CNXDW, CNXSNS,
NBZ, NBT, NBU,
NBV, NBQ, NBR, NBS,
NBVV, NBVOR, NBDIV, NBTH,
NBPSI, NBKHI, NBW, NBTPW,
NBCLF, NBWND, NBETH,
NBABS, NBSTD, NBSHD,
NBPV, NBWPV, NBSCVA,
NBP, NBUA1, NBUA2, NBUA3,
NBSP, NBMSL, NBFIS,
NBGM, NBFOL, NBSU1, NBSU2,
NBSU3, NBSU4, NBSU5,
NBLSM, NBGFIS, NBST,
NBDST, NBRDST, NBSSW,
NBDSW, NBRDSW, NBCSSW,
NBCDSW, NBSD, NBSR, NBBSR,
NBAL, NBEMIS, NBSDOG,
NBVEG, NBLAN, NBACOT,
NBDPAT, NBCLSP, NBCCP,
NBCLSS, NBCCSF, NBCUSS,
NBCVSS, NBCSSH, NBCSLH,
NBCTSP, NBCTCC, NBCBLD,
NBCSSR, NBCSTR, NBCTSR,
NBCTTR, NBCCCC, NBCHCC,
NBCMCC, NBCLCC, NBCUGW,
NBCVGW, NBCE, NBCS,
NBCLHE, NBCLHS, NBCC, NBCWS,
NBCSNS, NBCQTO, NBCTO3,
NBCTME, NBCICE, NBCLI,
NBCCVU, NBCCVV, NBCCVQ,
NBCCVS, NBCTUQ, NBCTUS,
NBCSOC, NBCTHC, NBCSOP,
NBCTOP, NBCSOD, NBCTHD,
NBCFON, NBCCHS, NBCEAS,
NBCSRU, NBCDRU, NBCIRU,
NBCETP, NBCTP, NBXTCC,
NBX10U, NBX10V, NBX2T,
NBX2SH, NBX2RH, NBXCCC,
NBXHCC, NBXMCC, NBXLCC,
NBXX2T, NBXN2T, NBXC,
NBXCVU, NBXCVV, NBXCVQ,
NBXCVS, NBXTUU, NBXTUV,
NBXTUQ, NBXTUS, NBXGDU,
NBXGDV, NBXLSP, NBXCP,
NBXLSS, NBXCSF, NBXSSR,
NBXSTR, NBXTSR, NBXTTR,
NBXUBL, NBXVBL, NBXTBL,
NBXQBL, NBXGBL, NBXST,
NBXDT, NBXSW, NBXDW,
NBXSNS, NILSM, NIGFIS,
NIST, NIDST, NIRDST,
NISSW, NIDSW, NIRDSW,
NICSSW, NICDSW, NISD,
NISR, NIBSR, NIAL,
NIEMIS, NISDOG, NIVEG, NILAN,
NIACOT, NIDPAT.
Contains buffer for basic fields resulting from APACHE in FULL-POS. Variables
are initialised in setup routine setup/SUFPSC2B.
No variable in namelist. All variables are DM-local.
List of variables:
| |
• NLENAFPBL: length of
buffer. |
| |
• NSTAAFPB : start adresses
for post-processed packets of points within buffer. |
Absolute pointers of the basic fields which have been horizontally then
vertically post-processed. Variables are initialised in setup routine setup/SUAFPDS.
No variable in namelist. All variables are DM-global.
Field pointers:
| |
• MFPUPT2: for upper air
pressure. |
| |
• MFPAPT2: for cloud fraction. |
| |
• MFPUT2 : for zonal component
of wind. |
| |
• MFPVT2 : for meridian
component of wind. |
| |
• MFPTT2 : for temperature. |
| |
• MFPQT2 : for specific
humidity. |
| |
• MFPWT2 : for liquid
water. |
| |
• MFPSVT2: for passive
scalar variables. |
| |
• MFPSPT2: for surface
pressure. |
Other variables:
| |
• NAFPOS : number of fields. |
| |
• NFPALEV: maximum number
of fields. |
Contains control variables. The following variables also present in namelist
NAMCT0 can be useful for FULL-POS (these variables are
DM-global).
| |
• NCONF: configuration
(default value is 1). |
| |
• LFPOS: FULL-POS main
switch (default value is .FALSE.). |
| |
• NFRPOS: post-processing
frequency (default value depends on configuration). |
| |
• NPOSTS: post-processing
control array (default value depends on configuration). |
| |
• LIOGAUX: I/O on unfitted
vertically post-processed fields (default value depends on configuration). |
| |
• LIOFOU1: I/O on Fourier
data (default value depends on configuration). |
| |
• LIOLPOL: I/O on Legendre
polynomial (default value depends on configuration). |
| |
• LIOSPEC: I/O on saved
spectral data (default value depends on configuration). |
| |
• LIOSUA: I/O on saved
upper air grid-point data (default value depends on configuration). |
| |
• LIOSSU: I/O on saved
surface data (default value depends on configuration). |
| |
• LIOSCF: I/O on saved
cumulated fluxes data (default value depends on configuration). |
| |
• LIOSXF: I/O on saved
instantaneous fluxes data (default value depends on configuration). |
| |
• LMLTSK: switch for multitasking
(default value depends on configuration). |
| |
• NTASKS: number of processors
(default value depends on configuration). |
| |
• CNPPATH: path name for
selection files (default value is ' '). |
| |
• CFPNCF: file name for
FULL-POS control file (default value is 'ncf927'). |
The following variables also present in namelist NAMPAR0
can be useful for FULL-POS when running distributed memory jobs (these variables
are DM-global).
| |
• NPRGPNS: number of processors
used in A-direction during grid-point phase in North-South direction. |
| |
• NPRGPEW: number of processors
used in B-direction during grid-point phase in East-West direction. |
| |
• NPRTRW: number of processors
used in A-direction during transform phase in wave space. Default
value is the same as NPRGPNS. |
| |
• NPRTRNS: number of processors
used in A-direction during transform phase in North-South direction
( used to dimension some arrays used in Fourier transforms). Default
value is the same as NPRGPNS. |
| |
• NPRTRV: number of processors
used in B-direction during transform phase in vertical direction.
Default value is the same as NPRTREW. |
| |
• NPROCK: not currently
used. |
| |
• NPROC: total number
of processors requested (DM-global). NPROC has to
be equal to the products NPROCA*NPROCB,
NPRGPNS*NPRGPEW, NPRTRW*NPRTRV.
Default value is 1 for distributed memory, 0 for shared memory. |
| |
• NOUTPUT: type of output
(DM-global). |
| |
• 0 = no diagnostic output; |
| |
• 1 = only diagnostic output from processor
number 1. |
| |
• 2 = diagnostic output from all processors
into separate files. |
Default value is 1.
| |
• LMESSP: .TRUE./.FALSE.:
distributed memory/ shared memory run (DM-global). Default value is
.FALSE. . |
| |
• LMPDIAG: extensive message
passing diagnostic output requested if .TRUE. (DM-global). Default
value is .FALSE. . |
The following variables also present in namelist NAMPAR1
can be useful for FULL-POS when running distributed memory jobs (these variables
are DM-global). They are read and initialised in setup/SUMP0.
| |
• NINSTR1: number of time
steps instrumented (DM-global). Default value is 0. |
| |
• NINSTR2: number of time
steps instrumented (DM-global). Default value is 0. |
Contains buffer for fully post-processed dynamical fields. Variables are
initialised in setup routine setup/SUFPSC2B.
No variable in namelist. All variables are DM-local.
List of variables:
| |
• NLENDFPBL: length of
buffer. |
| |
• NSTADFPB : start adresses
for post-processed packets of points within buffer. |
Contains dimensioning variables. The following variables are related with
FULL-POS.
Variables computed in SUDIM:
| |
• NDLON: maximum number
of Gaussian/Lobatto longitudes (DM-global). |
| |
• NFPXLEV: maximum number
of post-processing levels (DM-global). |
| |
• NPMAX: post-processing
truncation (DM-global, default value is generally NSMAX). |
| |
• NFPPHYB: maximum number
of post-processed physical fields (DM-global). |
| |
• NPROMA: working dimension
of (vertical) post-processing rows (DM-global, default value depends
on configuration). |
| |
• NGPBLKS: number of NPROMA-subpackets
in a processor (DM-local). |
| |
• NDGSAFPH: modified lower
bound for latitude (DM-local). |
| |
• NDGENFPH: modified upper
bound for latitude (DM-local). |
Variables computed under SUFPDIM:
| |
• NFPGT1: maximum number
of fields to be vertically post-processed (DM-global). |
| |
• NFPAUXB: maximum number
of vertically post-processed fields to remain unfitted (DM-global). |
| |
• NFPSPA: maximum number
of underived vertically post-processed fields to be fitted (DM-global). |
| |
• NFPSPD: maximum number
of derived vertically post-processed fields to be fitted, given one
horizontal subdomain (DM-global). |
| |
• NFPSPB: maximum number
of derived vertically post-processed fields to be fitted, given the
maximum number of horizontal subdomains (DM-global). |
| |
• NFPGT0B: maximum number
of fields to be horizontally post-processed (DM-global). |
| |
• NFPDYNB: maximum number
of post-processed dynamical fields (DM-global). |
| |
• NFPAVEC: maximum number
of vectorial underived vertically post-processed fields to be fitted
(DM-global). |
| |
• NFPBVEC: maximum number
of vectorial derived vertically post-processed fields to be fitted,
given the maximum number of horizontal subdomains (DM-global). |
| |
• NFPASCA: maximum number
of scalar underived vertically post-processed fields to be fitted
(DM-global). |
| |
• NFPBSCA: maximum number
of scalar derived vertically post-processed fields to be fitted, given
the maximum number of horizontal subdomains (DM-global). |
| |
• NFPLEI: maximum number
of useful fields in the working arrays for the inverse Legendre transforms
(DM-global). |
| |
• NFPLED: maximum number
of useful fields in the working arrays for the direct Legendre transforms
on wind (DM-global). |
Namelist NAMDIM:
The following variables of YOMDIM are in namelist NAMDIM:
NPROMA, NPMAX.
Contains dimensions for FULL-POS on one-level-type dynamical fields. Variables
are initialised in setup routine setup/SU4FPOS.
No variable in namelist. All variables are DM-global.
Number of fields in each array/buffer for dynamics:
| |
• NGT1FP : number of fields
in GT1 (fields to be fitted). |
| |
• NAUXFPB: number of fields
in GAUXBUF (fields to remain unfitted). |
| |
• NAUX3FP: number of 3D
fields in GAUXBUF (fields to remain unfitted). |
| |
• NSPAFP : number of useful
fields in SPAFP. |
| |
• NSPBFP : number of useful
fields in SPBFP. |
| |
• NSPDFP : number of useful
fields in SPDFP. |
| |
• NGT0FPB: number of fields
in GT0BUF during an horizontal post-processing time
step. |
| |
• NGT03FP: number of 3D
fields in GT0BUF during an horizontal post-processing
time step. |
| |
• NDYNFPB: number of fields
in GDFPBUF. |
Dimensions for dynamical fields post-processing:
| |
• NOFP3F : total number
of 3D fields. |
| |
• NOFP3A : total number
of subdomains for each level of each field. Array of dimension (NFPXLEV,NFP3DF). |
| |
• NOFPLEV: total number
of levels for each field. Array of dimension NFP3DF. |
| |
• NAVECFP: number of vectorial
underived vertically post-processed fields to be fitted. |
| |
• NBVECFP: number of vectorial
derived vertically post-processed fields to be fitted. |
| |
• NASCAFP: number of scalar
underived vertically post-processed fields to be fitted. |
| |
• NBSCAFP: number of scalar
derived vertically post-processed fields to be fitted. |
| |
• NLEIFP : number of fields
in the working arrays for the inverse Legendre transforms. |
| |
• NLEDFP : number of fields
in the working arrays for the direct Legendre transforms. |
Dimensions for physical fields and fluxes post-processing:
| |
• NPHYFPB: number of physical
fields. |
| |
• NCFUFPB: number of cumulated
fluxes. |
| |
• NXFUFPB: number of instantaneous
fluxes. |
Contains information relative to lagged variables needed to write out post-processed
fields. Variables are initialised in the sequence pp_obs/DYNFPOS
 setup/SUVFPOSL.
No variable in namelist. All variables are DM-global.
Horizontally post-processed dynamical variables:
| |
• NDYNFPL : number of
fields in GDFPBUF. |
| |
• NFLDFPL : field pointer.
Array of dimension NFPDYNB. |
| |
• RLEVFPL : level value.
Array of dimension NFPDYNB. |
| |
• NBITFPL : number of
bits when writing output file. Array of dimension NFPDYNB. |
| |
• NDOMFPL : number of
subdomains for each field. Array of dimension NFPDYNB. |
| |
• NINDFPL : indices of
subdomains for each field. Array of dimension (NFPDOM,NFPDYNB). |
| |
• NAUXFPL : number of
fields in GAUXBUF (fields to remain unfitted). |
| |
• NAUX3FPL: number of
3D fields in GAUXBUF (fields to remain unfitted). |
| |
• NGT0FPL : number of
fields in GT0BUF during an horizontal post-processing
time step. |
| |
• NGT03FPL: number of
3D fields in GT0BUF during an horizontal post-processing
time step. |
Spectral variables:
| |
• NSPAFPL : number of
fields in SPAFP. |
| |
• NSPDFPL : number of
fields in SPDFP. |
| |
• NFLDAFPL: field pointers
in SPAFP. Array of dimension NFPSPA. |
| |
• NFLDDFPL: field pointers
in SPDFP. Array of dimension NFPSPD. |
| |
• RLEVAFPL: level values
in SPAFP. Array of dimension NFPSPA. |
| |
• RLEVDFPL: level values
in SPDFP. Array of dimension NFPSPD. |
Grid-point variables:
| |
• NFLDFPXL: field pointers
in GAUXBUF. Array of dimension NFPAUXB. |
| |
• RLEVFPXL: level values
in GAUXBUF. Array of dimension NFPAUXB. |
| |
• NBITFPXL: number of
bit when writing out GAUXBUF. Array of dimension
NFPAUXB. |
Contains scientific and technical variables for post-processing. Variables
are initialised in setup routine setup/SUFPC.
All variables are DM-global.
Technical variables.
| |
• CFPDIR: prefix (path)
for the output files (default value is 'PF'). |
| |
• CFPIDEN: identificator
of the output files. |
| |
• CFPFMT: format of the
output files, can take the following values: |
| |
• 'MODEL' for output in spherical harmonics. |
| |
• 'GAUSS' for output in grid-point space
on Gaussian grid (covering the global sphere). |
| |
• 'LELAM' for output on a grid of kind
'ALADIN' (spectral or grid-point coefficients). |
| |
• 'LALON' for a grid of kind "latitudes
* longitudes'. |
Default is 'GAUSS' in ARPEGE/IFS, 'LELAM' in ALADIN.
| |
• CFPDOM, C1FPDOM:
names of the subdomains. Names have at maximum 7 characters. If CFPFMT
= 'GAUSS' or 'LELAM' only one output domain is allowed. If CFPFMT
= 'LALON' the maximum of output subdomains allowed is 10. By default,
one output domain is requested, CFPDOM(1)='000' and
CFPDOM(i)=" for i>1. |
| |
• CFP3DF, C1FP3DF:
names of the 3D dynamics fields to be post-processed. Names have at
maximum 12 characters. By default CFP3DF contains
blanks (no 3D dynamical field to post- process). |
| |
• CFP2DF, C1FP2DF:
names of the 2D dynamics fields to be post-processed. Names have at
maximum 16 characters. By default CFP2DF contains
blanks (no 2D dynamical field to post- process). |
| |
• CFPPHY, C1FPPHY:
names of the physical fields to be post-processed. Names have at maximum
16 characters. By default CFPPHY contains blanks
(no physical field to post-process). |
| |
• CFPCFU, C1FPCFU:
names of the cumulated fluxes fields to be post-processed. Names have
at maximum 16 characters. By default CFPCFU contains
blanks (no cumulated fluxes field to post- process). |
| |
• CFPXFU, C1FPXFU:
names of the instantaneous fluxes fields to be post-processed. Names
have at maximum 16 characters. By default CFPXFU
contains blanks (no instantaneous fluxes field to post-process). |
| |
• remark: variables CFP(XXX)
and C1FP(XXX) are in equivalence (same quantity)
but differently dimensioned. |
| |
• MFP3DF: gribcodes of
the 3D dynamics fields to be post-processed (variable used at ECMWF
only). |
| |
• MFP2DF: gribcodes of
the 2D dynamics fields to be post-processed (variable used at ECMWF
only). |
| |
• MFPPHY: gribcodes of
the physical fields to be post-processed (variable used at ECMWF only). |
| |
• MFPCFU: gribcodes of
the cumulated fluxes fields to be post-processed (variable used at
ECMWF only). |
| |
• MFPXFU: gribcodes of
the instantaneous fluxes fields to be post-processed (variable used
at ECMWF only). |
| |
• RFP3P: list of post-processing
pressure levels in Pa (no default value). |
| |
• RFP3H: list of post-processing
height (above orography) levels in meters (no default value). |
| |
• RFP3TH: list of post-processing
potential temperature levels in Kelvins (no default value). |
| |
• RFP3PV: list of post-processing
potential vorticity levels in S.I. units (no default value). |
| |
• NRFP3S: list of post-processing
 -levels (pseudo-configuration 927 only). No default value. |
| |
• LFPCNT: control varying
output variables according to time step. |
| |
• LFPNORM: print out the
mean of each post-processed field for each subdomain (default value
is .TRUE.). |
| |
• LTRACEFP: trace for
FULL-POS (additional prints on listing, default value is .FALSE.). |
| |
• NFPGRIB: level of GRIB
coding in output file ARPEGE/ALADIN for grid-point arrays (default
value is 2). |
| |
• NFPGRSP: level of GRIB
coding in output file ARPEGE/ALADIN for spectral arrays (default value
is 2). |
| |
• LFPBACK: .TRUE. to return
to initial geometry (case LFPSPEC) (default value
is .FALSE.). |
| |
• NFPDOM: useful dimension
of CFPDOM. |
| |
• NFP3DF: useful dimension
of CFP3DF (default value is the number of 3D wanted
dynamical fields). |
| |
• NFP2DF: useful dimension
of CFP2DF (default value is the number of 2D wanted
dynamical fields). |
| |
• NFPPHY: useful dimension
of CFPPHY (default value is the number of wanted
surface physical fields). |
| |
• NFPCFU: useful dimension
of CFPCFU. |
| |
• NFPXFU: useful dimension
of CFPXFU. |
| |
• NFP3P: useful dimension
of RFP3P. |
| |
• NFP3H: useful dimension
of RFP3H. |
| |
• NFP3TH: useful dimension
of RFP3TH. |
| |
• NFP3PV: useful dimension
of RFP3PV. |
| |
• NFP3S: useful dimension
of NRFP3S. |
Scientific variables.
| |
• LFPSPEC: .T. if post-processed
dynamical fields are written out as spectral coefficients, .F. if
post- processed dynamical fields are written out as grid point values
(default value is .FALSE., except if second part of a pseudo-configuration
927 or E927). |
| |
• LFIT2D: 1 if 2D dynamical
post-processed fields should be fitted, 0 otherwise. Default value
is .T. . |
| |
• LFITP: .T. if post-processed
fields on pressure levels should be fitted, .F. otherwise. Default
value is .T. . |
| |
• LFITH: .T. if post-processed
fields on height levels should be fitted, .F. otherwise. |
| |
• LFITT: .T. if post-processed
fields on  -levels should be fitted, .F. otherwise. Default
value is .F. . |
| |
• LFITV: .T. if post-processed
fields on PV-levels should be fitted, .F. otherwise. Default
value is .F. . |
| |
• LFITS: .T. if post-processed
fields on  -levels should be fitted, .F. otherwise. |
| |
• LFPUVDZ: .T. if divergence
and vorticity should be post-processed with respect to the wind components
(default value is .TRUE.). |
| |
• LFPUVKP: .T. if stream
function and potential velocity should be post-processed with respect
to the wind components, otherwise they are post-processed with respect
to the divergence and the vorticity (default value is .TRUE.). |
| |
• NFPCLI: usage level
for climatology: |
| |
• 0: no climatology (default value). |
| |
• 1: orography and land-sea mask of output
only. In your script, the local name of this external climatology
file should be "const.clim.DDDDDDD", where DDDDDDD is the variable
CFPDOM of the namelist NAMFPC.
This file should be of the same geometry as the post- processing file
you want to make. |
| |
• 2: all available climatological fields
of the current month. Use of auxiliary land-sea mask, orography, surface
temperature, relative surface wetness, deep soil temperature, relative
deep soil wetness, albedo, emissivity, standard deviation of orography,
percentage of vegetation, roughness length, anisotropy coefficient
of topography, direction of the main axis of topography and snow depth.
For that you need to make 2 climatology files: the first one, named
"Const.Clim", should be in the input model geometry, the second one,
named "const.clim.DDDDDDD", where DDDDDDD is the variable
CFPDOM of the namelist NAMFPC, should be
in the output (post-processing) geometry. |
| |
• 3: shifting mean from the climatological
fields of the current month to the ones of the closest month. |
| |
• NDOMFP: domain definition
of fields in files; NDOMFP = 1: physical fields are
grid-points stored on (C+I)+E and dynamical fields are spectral (option
available in ALADIN only); NDOMFP = -1: physical
and dynamical fields are grid-points stored on (C+I)+E; NDOMFP
= 0: physical and dynamical fields are grid-points stored on (C+I).
C, I, E mean respectively conservation domain, intermediate zone and
extension zone. Default value depends on configuration: |
| |
• if CFPFMT='MODEL', LFPSPEC=.F.,
default value is -1. |
| |
• if CFPFMT='GAUSS','LELAM'
or 'LALON' and LFPSPEC=.F., default value is 0. |
| |
• if CFPFMT='GAUSS' or
'LELAM' and LFPSPEC=.T., default value is -1 for
the first part, 1 for the second part. |
| |
• LFPQ: .T. if specific
humidity is interpolated (then deduct relative humidity), .F. if relative
humidity is interpolated (then deduct specific humidity). Default
value is .FALSE. . It is better to use the default value (.FALSE.):
in this case, relative humidity is conserved within displacement of
the planetary boundary layer and it is interpolated vertically. Relative
humidity is considered to have better properties for an interpolation
than mixing ratio even if it is not a conservative quantity. |
| |
• LASQ: .T. if apparent
surface humidity is set to 80% saturation (default value is .FALSE.). |
| |
• WSXI: maximum surface
moisture in input. |
| |
• WDXI: maximum deep soil
moisture in input. |
| |
• WSXO: maximum surface
moisture in output. Default value is -999. . |
| |
• WDXO: maximum deep soil
moisture in output. Default value is -999. . |
| |
• WCXO: maximum "climatological"
moisture in output. Default value is -999. . |
| |
• LSNOWI: snow in the
input file. Default value is equal to LNEIGE. |
| |
• LSNOWO: snow in the
output file. Default value is equal to LNEIGE. |
| |
• FPBL: critical thickness
of PBL. Default value is 17500 Pa. |
| |
• RFPCORR: critical orography
difference for correcting surface temperature through standard profile,
in J/kg. Default value is 300*g. |
| |
• RFPMXZ: non-critical
maximum difference between interpolated orography and output orography,
in J/kg. Default value is 3000*g. |
| |
• LFPSPLIT: .TRUE. if
split of FULL-POS post-processing of dynamical fields. |
| |
• NFPSPLIT: number of
dynamical fields to be post-processed at a time. |
| |
• MFP3DYN : maximum number
of 3D-dynamical fields needed for LFPART2. |
| |
• MFP2DYN : maximum number
of 2D-dynamical fields needed for LFPART2. |
| |
• NFPTRSPLIT: pointer
to actual post-processing package of dynamical fields. |
Remarks.
| |
• Note that if you ask for post-processing
of dynamical fields which are not spectral in the model, then these
post-processed fields will not be spectrally fitted, even if the corresponding
switch LFIT... is .TRUE. . |
| |
• Note that if you wish to post-process
upper air dynamical fields on height levels or hybrid levels, it is
not possible to apply such spectral fit because the horizontal interpolations
are performed before the vertical interpolation in order to respect
the displacement of the planetary boundary layer. |
Namelist NAMFPC:
The following variables of YOMFPC are in namelist NAMFPC:
CFPDIR, CFPFMT, CFPDOM,
CFP3DF, CFP2DF, CFPPHY,
CFPCFU, CFPXFU, LFPSPEC,
LFPBACK, RFP3P, RFP3H,
RFP3TH, RFP3PV, NRFP3S,
NFPGRIB, NFPGRSP, NDOMFP,
LFIT2D, LFITP, LFITT,
LFITV, LFITS, RFPCORR,
LFPUVDZ, LFPUVKP, NFPCLI,
LFPQ, WSXO, WDXO, WCXO,
LSNOWI, LSNOWO, LFPNORM,
LTRACEFP, FPBL, RFPMXZ,
LASQ, MFP3DF, MFP2DF,
MFPPHY, NFP3DF, NFP2DF,
NFPPHY, LFPSPLIT, NFPSPLIT.
Variables concerning the boundaries and the horizontal dimensions of each
output subdomain. Variables are initialised in setup routine setup/SUFPD.
For all kinds of output (sub)domains:
| |
• NLAT: number of latitudes
for each output (sub)domain (poles not included if a Gaussian grid
is required); array of integers; corresponds to the variable NDGLG
of the output grid(s) (DM-global). |
| |
• NLON: number of longitudes
for each output (sub)domain; array of integers; corresponds to the
variable NDLON of the output grid(s) (DM-global). |
| |
• RLATN: northern latitude
in degrees for each output (sub)domain; array of reals (DM-global). |
| |
• RLATS: southern latitude
in degrees for each output (sub)domain; array of reals (DM-global). |
| |
• RLONW: western longitude
in degrees for each output (sub)domain; array of reals (DM-global). |
| |
• RLONE: eastern longitude
in degrees for each output (sub)domain; array of reals (DM-global). |
| |
• NFPSIZEG: number of
points in each subdomain (DM-global). |
| |
• NFPSIZEL: DM-local version
of variable NFPSIZEG, its value is computed in SUFPG. |
| |
• NFPGP: maximum number
of output points (DM-global). This is the sum on all the subdomains
of the NFPSIZEG. Reduction of the grid near the poles
for global output grids is not taken in account, so NFPSIZEG
can be above the actual total number of points where interpolations
will have to be done. |
| |
• NFPD1: first dimension
of the array to be written on output file (DM-global). NFPD1
can be above NFPGP for global domains (NFPD1
counts polar latitudes contrary to NFPGP) and ALADIN
domains (NFPD1 can count extension zone points according
to the value of NDOMFP in NAMFPC
but NFPGP does not count any point of the extension
zone). |
These variables have many possible default values, according to the content
of the namelist NAMFPC.
ARPEGE default values (LELAM=.FALSE.) for NLAT, NLON, RLATN, RLATS, RLONW,
RLONE:
| |
• if CFPFMT='GAUSS' or
CFPFMT='MODEL', the default values are: NLAT=NDGL,
NLON=NDLON, while the other variables
are not useful. |
| |
• if CFPFMT='LALON', there
are many possible default values, according to the value of CFPDOM
in NAMFPC: |
If CFPDOM(j) is uninitialised or unknown from FULL-POS,
then the default values are all null, and you must at least specify in the
namelist the values of NLAT and NLON,
otherwise the job will abort.
| |
• If CFPFMT='LELAM', there
are many possible default values, according to the value of CFPDOM
in NAMFPC (see SUFPD, for real
quantities there are often more digits than written in the following
table): |
ALADIN default values (LELAM=.FALSE.) for NLAT, NLON, RLATN, RLATS, RLONW,
RLONE:
They are not detailed here and computation is sometimes complicated, see
routine SUFPD for more details.
Additional variables required for a grid ALADIN:
If you ask for a grid ALADIN (CFPFMT='LELAM'), there are
additionnal variables that you can specify:
| |
• NFPLUN: actual first
row of longitude (DM-global). |
| |
• NFPLUX: actual last
row of longitude (DM-global). |
| |
• NFPGUN: actual first
row of latitude (DM-global). |
| |
• NFPGUX: actual last
row of latitude (DM-global). |
| |
• NFPZONL: half the size
of area I in abscissa x (DM-global). |
| |
• NFPZONG: half the size
of area I in ordinate y (DM-global). |
There are many possible default values, according to the value of CFPDOM
in NAMFPC; for NFPLUX and NFPGUX
there are two possible default values according to the value of NFPDOM,
the first one is valid when NFPDOM=0, the second one is
valid when NFPDOM=1 or -1:
For CFPFMT='MODEL','GAUSS' or 'LALON' these variables are
not useful, they have default values equal to 0 excepted in the case ALADIN
when CFPFMT='MODEL' or 'LALON': defaults are NFPLUN=NDLUN,
NFPLUX=NDLUX, NFPGUN=NDGUN,
NFPGUX=NDGUX, NFPZONL=NBZONL,
NFPZONG=NBZONG.
Namelist NAMFPD:
The following variables of YOMFPD are in namelist NAMFPD:
NLAT, NLON, RLATN, RLATS,
RLONW, RLONE, NFPLUN,
NFPLUX, NFPGUN, NFPGUX,
NFPZONL, NFPZONG.
Variables defining the FULL-POS filter. Variables are initialised in setup
routine setup/SUFPF. All variables are
DM-global.
List of variables:
| |
• NFMAX : maximum truncation
of the output subdomains (default value depends on configuration). |
| |
• LFPFIL: .TRUE. if "THX"
filter for derived fields is active (default value is .TRUE. in ARPEGE
excepted in the first part of a 927-type configuration, .FALSE. in
ALADIN). |
| |
• RFPEPS: maximum relative
distance between the asymptot and the filtering function outside the
window. Default value is 0.1 . |
| |
• NFPWID: width of the
window where the filtering function is strongly decreasing; defined
as a deviation. Default value is 12 if \Nc £ 24, 24 if \Nc >
24 and \Nc £ 48, 48 if \Nc > 48. |
| |
• RFPFIL: value of the
filter for each zonal wavenumber and each subdomain. |
| |
• RFPMAT: filtering matrixes. |
| |
• LFPBEL: .TRUE. if "bell-shape"
filter is active for other fields than geopotential, temperature or
mean sea level pressure. |
| |
• RFPBEL: coefficient
of the exponential function in the "bell-shape" filter for other fields
than geopotential, temperature or mean sea level pressure. |
| |
• LFPBEG: .TRUE. if "bell-shape"
filter is active for geopotential. |
| |
• RFPBEG: coefficient
of the exponential function in the "bell-shape" filter for geopotential. |
| |
• LFPBET: .TRUE. if "bell-shape"
filter is active for temperature. |
| |
• RFPBET: coefficient
of the exponential function in the "bell-shape" filter for temperature. |
| |
• LFPBEP: .TRUE. if "bell-shape"
filter is active for mean sea level pressure. |
| |
• RFPBEP: coefficient
of the exponential function in the "bell-shape" filter for mean sea
level pressure. |
| |
• LFPBEH: .TRUE. if "bell-shape"
filter is active for relative moisture. |
| |
• RFPBEH: coefficient
of the exponential function in the "bell-shape" filter for relative
moisture. |
| |
• LFPBED: .TRUE. if "bell-shape"
filter is active for "derivatives". |
| |
• RFPBED: coefficient
of the exponential function in the "bell-shape" filter for "derivatives". |
| |
• LFPBE : .TRUE. if any
of LFPBE... is .TRUE. . |
The array below summarizes the existing "Bell-shaped" filters (default values
are parenthesis):
Variables LFPBED to LFPBEL are .FALSE.
in the first part of a 927-type configuration LFPSPEC=.TRUE.
.
Furthermore, in ARPEGE/IFS, it is also possible to perform at the end
an overtruncation of each post-processed field. This overtruncation is equivalent
to the so-called "THX" filter, but it is applied in the stretched spectral
space and then n0 (see formula ( [Ref: deffilterTHXARPEGE] ) ) is
equal to the variable NPMAX in the namelist NAMDIM.
Note that this overtruncation is active only if NPMAX
is less than NSMAX.
Namelist NAMFPF:
The following variables of YOMFPF are in namelist NAMFPF:
NFMAX, LFPFIL, RFPEPS,
NFPWID, LFPBEL, RFPBEL,
LFPBEG, RFPBEG, LFPBET,
RFPBET, LFPBEP, RFPBEP,
LFPBEH, RFPBEH, LFPBED,
RFPBED.
Variables defining the characteristics of the (transformed) output geometry.
Variables are initialised in setup routine setup/SUFPG.
Variables concerning all kinds of output subdomains:
| |
• NFPMAX: truncation order
(DM-global). |
| |
• If CFPFMT='GAUSS', NFPMAX
is the truncation corresponding to the output grid; default value
is NSMAX. |
| |
• If CFPFMT='LELAM', NFPMAX
is the meridian truncation; default value satisfies to the identity
3* NFPMAX +1 = NLAT. |
| |
• If CFPFMT='LALON', NFPMAX
is the truncation of the Gaussian grid which would have been defined
by NLAT latitudes and NLON longitudes;
default value satisfies to the identity 3* NFPMAX
+1 = min( NLAT,NLON). |
| |
• RFPLA: latitudes of
the output points (DM-local). |
| |
• RFPLA1: latitude of
the first output point in a DM-global sense (DM-global). |
| |
• RFPLANLAT1: latitude
of the output point number NLAT(1) in a DM-global
sense (DM-global). |
| |
• RFPLO: longitudes of
the output points (DM-local). |
| |
• NFPRGPG: effective number
of output points (DM-global). |
| |
• NFPRGPL: DM-local effective
number of output points for one processor (DM-local). |
| |
• NFPRGPLX: maximum of
the NFPRGPL of all the processors (DM-global). |
| |
• LFPOSHOR: .TRUE. for
actual horizontal post-processing (DM-global). |
Variables concerning Gaussian grid or ALADIN subdomain as output subdomain:
| |
• NFPLEV : number of levels;
default is NFLEVG (DM-global). |
| |
• FPMUCEN: sine of the
latitude of the pole of interest (DM-global); default value is RMUCEN
for global outputs, 1 for limited area outputs. |
| |
• FPLOCEN: longitude of
the pole of interest (DM-global); default value is RLOCEN
for global outputs, 0 for limited area outputs. |
| |
• FPVALH : A coefficients
of vertical system (DM-global); default value is VAH. |
| |
• FPVBH : B coefficients
of vertical system (DM-global); default value is VBH. |
| |
• FPVP00 : reference pressure
(DM-global); default value is VP00. |
| |
• RFPGM : mapping factor
(DM-local). |
| |
• RFPNORX: zonal component
of rotation matrix for vector outputs (DM-local). |
| |
• RFPNORY: meridian component
of rotation matrix for vector outputs (DM-local). |
Variables concerning Gaussian grid as output subdomain:
| |
• 1: number of points proportional to  . |
| |
• 2: number of points read on namelist
NAMFPG. |
| |
• default value is NHTYP
if NLAT = NDGLG, zero in the other
cases. |
| |
• FPNLGINC: increment
to get non-linear grid (DM-global); default value is 1. |
| |
• NFPRGRI : number of
active points on a parallel (DM-global); default value is
NLOENG if NLAT = NDGLG,
NLON in the other cases. |
| |
• NFPMEN : maximum zonal
wave number for a given parallel (DM-global). |
| |
• FPSTRET : stretching
factor (DM-global); default value is RSTRET for
global outputs, 1 for limited area outputs. |
| |
• 1: no tilting; pole of high resolution
at the northern pole of the real Earth. |
| |
• 2: tilting; pole of high resolution at
a different location than the northern pole of the real Earth. |
Default value is NSTTYP for global outputs, 1 for limited
area outputs.
| |
• NFPQUAD : quadrature
( 1 : Gauss ; 2 : Lobatto); (DM-global); default value is NQUAD
for global outputs, 1 for limited area outputs. |
| |
• RFPMU : array containing
the sines of the output Gaussian latitudes (DM-global). |
| |
• LFPOLE : .TRUE. if points
at the poles are required; default value is .TRUE if CFPFMT='MODEL'
or pseudo-configurations 927 and E927, .FALSE. elsewhere (DM-global). |
Variables concerning ALADIN grid as output subdomain:
| |
• FPBETA: angle of rotation
in degrees (DM-global); default value is zero in ARPEGE/IFS, zero
in ALADIN if CFPFMT='MODEL' and EBETA
in ALADIN in the other cases. |
| |
• NMFPMAX: meridian truncation
order (DM-global); default value satisfies to the identity 3* NFPMAX
+1 = NLAT if CFPFMT='LELAM' and
default value is 1 in the other cases. |
| |
• LFPMAP: .T./.F. if the
domain is defined through its coordinates/wavelengths (DM-global). |
| |
• FPLX: wavelength in
longitude (DM-global); default is 5448300. in ARPEGE/IFS if CFPFMT='LELAM',
0 in ARPEGE/IFS if CFPFMT='MODEL', 'LALON' or 'GAUSS',
0 in ALADIN if CFPFMT='GAUSS' and ELX
in ALADIN if CFPFMT='MODEL', 'LALON' or 'LELAM'. |
| |
• FPLY: wavelength in
latitude (DM-global); default is 5448300. in ARPEGE/IFS if
CFPFMT='LELAM', 0 in ARPEGE/IFS if CFPFMT='MODEL',
'LALON' or 'GAUSS', 0 in ALADIN if CFPFMT='GAUSS'
and ELY in ALADIN if CFPFMT='MODEL',
'LALON' or 'LELAM'. |
| |
• NFPROTEQ: parameter
for rotation (DM-global). |
| |
• 1: the reference point of rotation (FIR,LAR)
is rotated to equator, the north pole is on the new Greenwich meridian. |
Default is 0 in ARPEGE/IFS, 0 in ALADIN if CFPFMT='GAUSS'
and NROTEQ in ALADIN if CFPFMT='MODEL',
'LALON' or 'LELAM'.
| |
• FPLONR: LAR
geographical longitude of reference point of rotation, in degrees
(DM-global); default value is zero in ARPEGE/IFS, 0 in ALADIN if CFPFMT='GAUSS'
and ELONR in ALADIN if CFPFMT='MODEL',
'LALON' or 'LELAM'. |
| |
• FPLATR: FIR
geographical latitude of reference point of rotation, in degrees (DM-global);
default value is zero in ARPEGE/IFS, 0 in ALADIN if CFPFMT='GAUSS'
and ELATR in ALADIN if CFPFMT='MODEL',
'LALON' or 'LELAM'. |
| |
• FPLON0: LA0
geographical longitude of reference for the projection, in degrees
(DM-global); default value is 0. in ARPEGE/IFS excepted in the case
CFPFMT='LELAM' (various default values according
to CFPDOM, see routine SUFPG1),
0. in ALADIN if CFPFMT='GAUSS' and ELON0
in ALADIN if CFPFMT='MODEL', 'LALON' or 'LELAM'. |
| |
• FPLAT0: FI0
geographical latitude of reference for the projection, in degrees
(DM-global); default value is 0. in ARPEGE/IFS excepted in the case
CFPFMT='LELAM' (various default values according
to CFPDOM, see routine SUFPG1),
0. in ALADIN if CFPFMT='GAUSS' and ELAT0
in ALADIN if CFPFMT='MODEL', 'LALON' or 'LELAM'. |
| |
• FPRPK: K projection
parameter and definition (DM-global); default value is -9. in ARPEGE/IFS
excepted in the case CFPFMT='LELAM' (various default
values according to CFPDOM, see routine SUFPG1),
-9. in ALADIN if CFPFMT='GAUSS' and ERPK
in ALADIN if CFPFMT='MODEL', 'LALON' or 'LELAM'. |
| |
• NFPSOTRP: isotropy parameter
under projection (DM-global); default value is 0. in ARPEGE/ IFS excepted
in the case CFPFMT='LELAM' (generally 1.), 0. in
ALADIN if CFPFMT='GAUSS' and NSOTRP
in ALADIN if CFPFMT='MODEL', 'LALON' or 'LELAM'. |
| |
• NFPGIV0: choice of reference
point for projection (DM-global); default value is zero in ARPEGE/
IFS, 0. in ALADIN if CFPFMT='GAUSS' and NGIV0
in ALADIN if CFPFMT='MODEL', 'LALON' or 'LELAM'. |
Variables concerning ALADIN or 'LALON' grid as output subdomain:
| |
• RFPDX: grid size in
meters along x if projection (ALADIN grid); longitude increment
in radians if spherical geometry (DM-global). |
| |
• RFPDY: grid size in
meters along y if projection (ALADIN grid); latitude increment
in radians if spherical geometry (DM-global). |
Namelist NAMFPG:
The following variables of YOMFPG are in namelist NAMFPG:
NFPMAX, NFPLEV, FPMUCEN,
FPLOCEN, FPVALH, FPVBH,
NFPHTYP, NFPRGRI, FPSTRET,
NFPTTYP, NFPQUAD, FPBETA,
LFPOLE, FPVP00, NMFPMAX,
FPLX, FPLY, NFPROTEQ,
FPLONR, FPLATR, FPLON0,
FPLAT0, FPRPK, NFPSOTRP,
NFPGIV0, FPNLGINC.
Contains control variables for FULL-POS IO scheme. Variables are initialised
in setup routine setup/SUFPIOS.
Writing to output files: variable NFPXFLD.
To write post-processed fields on an output file, you first extract them
from an array (or a work file). Rather than extracting the fields one after
the other, the fields are grouped in packets, and these packets of fields
are extracted one after the other. You can specify the maximum number of
fields in a packet by modifying the variable NFPXFLD in
the namelist NAMFPIOS. The default value is the maximum
possible depending on the number of fields to produce. Raising the value
of NFPXFLD should save CPU time to the detriment of the
memory cost, and vice versa.
Variables for auxiliary fields:
| |
• LIOFPR : switch for
IO scheme (DM-global). |
| |
• CIOFPR : pathname (DM-global). |
| |
• NIOFPR : MIO package
file reference (DM-global). |
| |
• NBLIOFPR: buffer length
for IO (DM-local). |
| |
• NIOBFFPR: number of
buffers for IO (DM-global). |
| |
• NPCKFFPR: packing factor
(DM-global). |
| |
• NEXPBFPR: number of
bits used for exponent when packing (DM-global). |
Variables for working fields:
| |
• LIOFPW : switch for
IO scheme (DM-global). |
| |
• CIOFPW : pathname (DM-global). |
| |
• NIOFPW : MIO package
file reference (DM-global). |
| |
• NBLIOFPW: buffer length
for IO (DM-local). |
| |
• NIOBFFPW: number of
buffers for IO (DM-global). |
| |
• NPCKFFPW: packing factor
(DM-global). |
| |
• NEXPBFPW: number of
bits used for exponent when packing (DM-global). |
Variables for dynamics fields:
| |
• LIOFPD : switch for
IO scheme (DM-global). |
| |
• CIOFPD : pathname (DM-global). |
| |
• NIOFPD : MIO package
file reference (DM-global). |
| |
• NBLIOFPD: buffer length
for IO (DM-local). |
| |
• NIOBFFPD: number of
buffers for IO (DM-global). |
| |
• NPCKFFPD: packing factor
(DM-global). |
| |
• NEXPBFPD: number of
bits used for exponent when packing (DM-global). |
Variables for physics fields:
| |
• LIOFPP : switch for
IO scheme (DM-global). |
| |
• CIOFPP : pathname (DM-global). |
| |
• NIOFPP : MIO package
file reference (DM-global). |
| |
• NBLIOFPP: buffer length
for IO (DM-local). |
| |
• NIOBFFPP: number of
buffers for IO (DM-global). |
| |
• NPCKFFPP: packing factor
(DM-global). |
| |
• NEXPBFPP: number of
bits used for exponent when packing (DM-global). |
Variables for fields from APACHE:
| |
• LIOFPA : switch for
IO scheme (DM-global). |
| |
• CIOFPA : pathname (DM-global). |
| |
• NIOFPA : MIO package
file reference (DM-global). |
| |
• NBLIOFPA: buffer length
for IO (DM-local). |
| |
• NIOBFFPA: number of
buffers for IO (DM-global). |
| |
• NPCKFFPA: packing factor
(DM-global). |
| |
• NEXPBFPA: number of
bits used for exponent when packing (DM-global). |
Additional remarks and some default values:
For the horizontal post-processing, the data flows have been separated in
4 parts:
| |
• post-processed dynamical data. |
| |
• post-processed physical and fluxes data. |
| |
• weights for horizontal interpolations. |
| |
• output geometry and climatology. |
| |
• horizontally pre-processed dynamical
data (for post-processing on height levels or hybrid levels). |
To spare memory space, each of these data flows can be controlled by a specific
I/O scheme. By default, no I/O scheme is activated. If you wish to activate
I/O schemes, then use the variables which are in the namelist NAMFPIOS
as the array below describes it (defaults values are the ones in parenthesis).
Namelist NAMFPIOS:
The following variables of YOMFPIOS are in namelist NAMFPIOS:
NFPXFLD, CIOFPR, CIOFPW,
CIOFPD, CIOFPP, CIOFPA,
LIOFPR, NPCKFFPR, NEXPBFPR,
LIOFPW, NPCKFFPW, NEXPBFPW,
LIOFPD, NPCKFFPD, NEXPBFPD,
LIOFPP, NPCKFFPP, NEXPBFPP,
LIOFPA, NPCKFFPA, NEXPBFPA.
Useless, has to be removed.
Contains post-processing file-handling variables. Variables are initialised
in setup routine setup/SUFPOPH. No variable
in namelist.
List of variables:
| |
• CFPFN: path file name
for output files (DM-global). |
| |
• CFPCA: names of output
frames (DM-global). |
Contains control variables for SCAN2MSM or SCAN2MDM
FULL-POS rows. Variables are initialised in setup routine setup/SUFPSC2.
List of variables:
| |
• NFPLENR: number of active
points in each post-processing row (DM-local). |
| |
• NFPROMAG: working dimension
for horizontal post-processing (DM-global if its value is specified
in the namelist NAMFPSC2, DM-local in the contrary,
default value depends on the processor if distributed memory). |
| |
• NFPROMAL: working dimension
for horizontal post-processing, DM-local version of NFPROMAG. |
| |
• NFPBSR: number of sub-rows
in each post-processing row (DM-local). |
| |
• NFPTBSRL: total number
of sub-rows in post-processing buffer (DM-local). |
Additional remarks about variables NFPROMAG and NFPROMAL:
Variable NFPROMAG is very useful to find a compromise between
the vectorization and the memory cost in the post-processing for the lagged
computations (horizontal interpolations in HPOSLAG and
additional calculations in ENDVPOS), and it is also very
useful to make efficient I/O schemes. The variable NFPROMAL
represents the horizontal depth of work in the lagged computations of post-processing;
it is obvious that this variable has some similarity with the famous variable
NPROMA (in namelist NAMDIM) or the less
famous variable NSEGM (in namelist NAM926),
but its use is more flexible. For shared memory NFPROMAL=NFPROMAG
if NFPROMAG is an odd integer, NFPROMAL=NFPROMAG+1
if NFPROMAG is an even integer; this is also the case
if NFPROMAG is taken to its (DM-local) default value for
distributed memory, this is not necessary to define a unic DM-global value
(but desirable in the future). For distributed memory when a DM-global value
of NFPROMAG is given in the namelist NAMFPSC2,
NFPROMAL depends on the processor, its maximum value is
equal to NFPROMAG if NFPROMAG is an odd
integer, NFPROMAG+1 if NFPROMAG is an
even integer. Variable NFPROMAG is used only in SUFPSC2,
the remainder of the code uses only NFPROMAL. The way how
the post-processing data flow is managed in the post-processing arrays/work
files is explained in detail in the sections 'Some shared memory features'
and 'Some distributed memory features'. For summary:
| |
• The model data is split in packets of
points of fixed size NPROMA; each model packet is
treated by one task, and is independent from the other packets. For
distributed memory one processor treats NGPTOT model
grid-points, with a subdivision into NPROMA-packets,
but this subdivision is active for not lagged computations only. |
| |
• Each output point (i.e. where post-processing
is performed) is related to one model packet: the one which contains
the model point which is the nearest from it. That way, the output
points are sorted out, so that each post-processing packet (i.e. the
group of all the output points which are related to the same model
packet) has a specific size. For shared memory (resp. distributed
memory) the model-packets used to compute NFPROMAG
and NFPROMAL are the NPROMA-packets
(resp. the NGPTOT-packets). |
| |
• In order to use the I/O subroutines already
available in ARPEGE/IFS, each post-processing packet has been split
in sub-rows of fixed size NFPROMAL. Each post-processing
packet is treated by one task which performs a loop on the number
of sub-rows in the current post-processing packet. |
| |
• Note that an empty post-processing packet
is characterized by a number of sub-rows equal to zero; as a consequence,
empty post-processing packets do not lead to memory spill. On the
opposite, each non-empty post-processing packet contains a specific
number of unused words. |
Searching for a compromise between vectorization and memory cost, the value
of NFPROMAG can be any integer from 1 to the size of the
biggest post-processing packet. The bigger NFPROMAG will
be, the more the vectorization will be efficient, and the more memory will
be needed. The default value of NFPROMAG is the mean size
of the post-processing packets if shared memory, the total number of post-processing
points treated by the current processor if distributed memory, as this value
appeared to be the best compromise. However, if the I/O schemes are activated,
the size of the biggest post-processing packet may be a more efficient value.
In the listing, the user will find more information to find out the value
to adopt: just before "COMMON YOMFPSC2" is printed out, the following information
is written (information depends on the processor for distributed memory
but is not very useful in this case because there is only one value per
processor which is used to compute the following statistics: the standard
deviation is zero):
| |
• size of the smallest post-processing
packet. |
| |
• size of the biggest post-processing packet. |
| |
• mean size of the post-processing packets. |
| |
• standard deviation of the size of the
post-processing packets. |
You can also know the memory you have lost in the post-processing arrays:
just locate "TOTAL MEMORY LOST IN WORDS =" .
Namelist NAMFPSC2:
The following variable of YOMFPSC2 is in namelist NAMFPSC2:
NFPROMAG.
Contains control variables for SCAN2MSM or SCAN2MDM
FULL-POS sub-rows. Variables are initialised in setup routine setup/SUFPSC2B.
No variable in namelist.
List of variables:
| |
• NFPROF: number of active
points in each post-processing sub-row (DM-local). |
| |
• NFPINT: number of points
to be computed through 12-points interpolations (NFPINT(1,.))
and through averaging of a box (NFPINT(2,.)). DM-local
variable. |
| |
• NFPSORT: pointer of
the interlaced post-processing points in the sorted array (DM-local). |
| |
• NFPSTAP: sub-row start
pointer for each row (DM-local). |
Contains FULL-POS spectral arrays. Variables are initialised after direct
spectral transforms. All variables are DM-global. No variable in namelist.
List of variables:
| |
• SPAFP: underived vertically
post-processed fields. |
| |
• SPBFP: derived vertically
post-processed fields for all required subdomains. |
| |
• SPDFP: derived vertically
post-processed fields for one subdomain. |
Contains buffer FPTS0: Auxiliary surface grid point array
for post-processing on height levels above an output orography and for surface
horizontal post-processing above an output orography (DM-local). Variable
is computed in routine pp_obs/GPOS. No
variable in namelist.
Some variables in YOMIOS (which does not contain only FULL-POS
variables) can be useful for FULL-POS. These variables are DM-global. For
the vertical post-processing, the data flows have been separated in 2 parts:
| |
• fitted post-processed fields. |
| |
• unfitted post-processed fields. |
To spare memory space, the data flow of unfitted post-processed fields can
be controlled by a specific I/O scheme. By default, this I/O scheme is activated
only at ECMWF (LECMWF=.TRUE.). If you wish to activate/deactivate
it, just set LIOGAUX=.TRUE./.FALSE. in the namelist
NAMCT0. Then you can change other related variables in the namelist
NAMIOS:
| |
• CIOGAUX: work file name
of unfitted vertically post-processed fields; default is 'miogaux'. |
| |
• NPCKFGX: packing factor
for I/O on vertically post-processed fields; default is 1. |
| |
• NEXPBGX: number of bits
used for exponent when packing I/O on vertically post-processed fields;
default is 6. |
Other useful variables in YOMIOS and NAMIOS:
| |
• CIOFOU1: work file name
for Fourier data. |
| |
• NBLFOU : segment length
for complex I/O on Fourier data. |
| |
• CIOLPOL: work file name
for Legendre polynomials. |
| |
• CIOTMDT: work file name
of grid-point fields at t-dt. |
| |
• NPCKFT9: packing factor
for I/O on grid-point fields at t-dt. |
| |
• NEXPBT9: number of bits
for exp. when packing grid-point fields at t-dt. |
| |
• CIOGPUA: work file name
of upper air grid-point fields. |
| |
• NPCKFGU: packing factor
for I/O on upper air grid-point fields. |
| |
• NEXPBGU: number of bits
for exp. when packing upper air grid-point fields. |
| |
• CIOGPP : work file name
of surface fields. |
| |
• NPCKFGP: packing factor
for I/O on surface fields. |
| |
• NEXPBGP: number of bits
for exp. when packing surface fields. |
| |
• CIOCFU : work file name
of cumulated fluxes. |
| |
• NPCKFCF: packing factor
for I/O on cumulated fluxes. |
| |
• NEXPBCF: number of bits
for exp. when packing cumulated fluxes. |
| |
• CIOXFU : work file name
of instantaneous fluxes. |
| |
• NPCKFXF: packing factor
for I/O on instantaneous fluxes. |
| |
• NEXPBXF: number of bits
for exp. when packing instantaneous fluxes. |
| |
• CIOSPEC: work file name
of saved spectral data. |
| |
• CIOSUA : work file name
of saved upper air grid-point data. |
| |
• CIOSSU : work file name
of saved surface data. |
| |
• CIOSCF : work file name
of saved cumulated fluxes data. |
| |
• CIOSXF : work file name
of saved instantaneous fluxes data. |
Default value of a variable of the type CIONAME (for ex.
CIOGPP) is 'mioname' (for ex. 'miogpp'). Default value
of NBLFOU depends on configuration. Default value of NPCK...
variables is 1. Default value of NEXP... variables is 6.
See routine SUIOS for more details.
Variables specific to the distributed memory environment. Some of these
variables can be used also for shared memory. The following variables of
YOMMP (which does not contain only FULL-POS variables)
can be useful for FULL-POS. Variables are generally computed in setup/SUMP0
(called by setup/SU0Y0MA), setup/SUMP,
sometimes setup/SUCT0.
Variables describing distributed memory parallelization:
| |
• MYPOS(1): own position
in x-direction (DM-global). |
| |
• MYPOS(2): own position
in y-direction (DM-global). |
| |
• MYPROC: own processor,
quasi 1D position in torus (DM-local). |
| |
• MYSETA: own processor
set A (set A DM-local, set B DM-global). |
| |
• MYSETB: own processor
set B (set A DM-global, set B DM-local). |
| |
• MYPID: own process identifier
(DM-local). |
| |
• MYFATHER: father process
(DM-local). |
| |
• NSONS: number of son
processes (DM-local). |
| |
• NPHASE: number of phases
in the recursive transposition (DM-global). Default value is 1. |
| |
• NLAGA: lagging factor
for tranpositions in A-direction (DM-global). Default value is 0. |
| |
• NLAGB: lagging factor
for tranpositions in B-direction (DM-global). Default value is 0. |
| |
• NSLPAD: number of pad
words initialised to a huge number at either of side of the interpolation
buffer halo, used to trap halo problem (DM-global). Default value
is 0. |
| |
• LSPLITIN: input data
provided in sequential files if .TRUE. or in directories if .FALSE.
(DM- global). Default value is .FALSE. . |
| |
• LSPLITOUT: output data
provided in sequential files if .TRUE. or in directories if .FALSE.
(DM- global). Default value is .TRUE. . |
| |
• NSTRIN: number of processors
required to perform input processing (DM-global). Default value is
1. |
| |
• NSTROUT: number of processors
required to perform output processing (DM-global). Default value is
0. |
| |
• NFLDIN: number of input
fields to be buffered during distribution (DM-global). Default value
is 0. |
| |
• NFLDOUT: number of output
fields to be buffered during gathering (DM-global). Default value
is 0. |
| |
• NPRCIDS: array containing
the process identificators (DM-local). |
| |
• LAPPLE: defines apple
or orange grid point decomposition (DM-global). Default value is .FALSE.
. |
| |
• LSPLIT: .TRUE./.FALSE.:
latitudes are shared/not shared between sets (DM-global). Default
value is .TRUE. for configurations between 100 and 199, .FALSE. for
other configurations. |
| |
• LSLSYNC: .TRUE. if communication
between processors for interpolation buffers reads/writes in synchronisation
(DM-global). Default value is .TRUE. . |
Variables describing the partitioning of data:
| |
• NDWIDE: width of interpolation
buffer halo region (DM-global). Default value is 32. |
| |
• NPROCM: gives process
which is responsible for the zonal wave number m (DM-global). |
| |
• NUMPROCGP: gives processor
which is responsible for model grid point (DM-global). |
| |
• NUMPROCFP: gives processor
which is responsible for FULL-POS horizontal interpolation point jfp,
where jfp = 1 to NFPRGPG (DM-global). |
| |
• NUMPP: tells for how
many wave numbers the processes are responsible for the truncation
NSMAX (DM-local). |
| |
• NALLMS: global wave
numbers for all processes (DM-global). |
| |
• NPTRMS: pointer to first
wave number of given process in NALLMS (DM-global). |
| |
• NLATLS: first latitude
for which current process calculates Legendre polynomials (DM-global). |
| |
• NLATLE: last latitude
for which current process calculates Legendre polynomials (DM-global). |
| |
• NPROCL: gives process
which is responsible for latitude in grid-point space (DM-global). |
| |
• NPROCLF: gives process
which is responsible for latitude in Fourier space (DM-global). |
| |
• NULTPP: number of latitudes
in given process in grid-point space (DM-global). |
| |
• NULTPPF: number of latitudes
in given process in Fourier space (DM-global). |
| |
• MYLATS: latitude numbers
mapped to current process in grid-point space (DM-global). |
| |
• NPTRLS: pointer to first
global latitude of given process in grid-point space (DM-global). |
| |
• NPTRLSF: pointer to
first global latitude of given process in Fourier space (DM-global). |
| |
• NFRSTLAT: first latitude
of each set in grid-point space (DM-global). |
| |
• NFRSTLATF: first latitude
of each set in Fourier space (DM-global). |
| |
• NFRSTLOFF: offset for
first lat of own set in grid-point space (DM-local). |
| |
• NFRSTLOFFF: offset for
first lat of own set in Fourier space (DM-local). |
| |
• NLSTLAT: last latitude
of each set in grid-point space (DM-global). |
| |
• NLSTLATF: last latitude
of each set in Fourier space (DM-global). |
| |
• NPTRFRSTLAT: pointer
to first latitude in grid-point space (DM-global). |
| |
• NPTRFRSTLATF: pointer
to first latitude in Fourier space (DM-global). |
| |
• NPTRLSTLAT: pointer
to last latitude in grid-point space (DM-global). |
| |
• NPTRLSTLATF: pointer
to last latitude in Fourier space (DM-global). |
| |
• NPTRFLOFF: offset for
pointer to first latitude of own set in grid-point space (DM-local). |
| |
• NPTRFLOFFF: offset for
pointer to first latitude of own set in Fourier space (DM-local). |
| |
• NPTRLAT: pointer to
start of latitude in grid-point space (DM-global). |
| |
• NPTRLATF: pointer to
start of latitude in Fourier space (DM-global). |
| |
• LSPLITLAT: .TRUE. if
latitude is split over two A sets (DM-global). |
| |
• MYFRSTACTLAT: first
actual latitude of each processor in grid-point space (DM-global). |
| |
• MYLSTACTLAT: last actual
latitude of each processor in grid-point space (DM-global). |
| |
• MYSENDA: processors
to which current processor will send messages during the (recursive)
A- transposition (set A DM-local, set B DM-global). |
| |
• MYRECVA: processors
from which current processor will receive messages during the (recursive)
A-transposition (set A DM-local, set B DM-global). |
| |
• MYSENDB: processors
to which current processor will send messages during the (recursive)
B- transposition (DM-local). |
| |
• MYRECVB: processors
from which current processor will receive messages during the (recursive)
B-transposition (DM-local). |
| |
• MYSENDSETA: process-set
to which current processor will send messages during the (recursive)
A-transposition (set A DM-local, set B DM-global). |
| |
• MYRECVSETA: process-set
to which current processor will receive messages during the (recursive)
A-transposition (set A DM-local, set B DM-global). |
| |
• MYSENDSETB: process-set
to which current processor will send messages during the (recursive)
B-transposition (set A DM-global, set B DM-local). |
| |
• MYRECVSETB: process-set
to which current processor will receive messages during the (recursive)
B-transposition (set A DM-global, set B DM-local). |
| |
• MYSENDG: processors
to which current processor will send messages during global communication
(DM-local). |
| |
• MYRECVG: processors
from which current processor will receive messages during global communication
(DM-local). |
| |
• MYMSGS: identification
of the origin of the messages during recursive transposition (DM-local). |
| |
• NPTRSV: pointer to first
spectral wave column on given process for NSMAX
arrays (DM-global). |
| |
• NPTRSVV: as NPTRSV
but for full m-columns (DM-global). |
| |
• NPTRMF: distribution
of m-columns among B sets used for semi-implicit calculations
on this processor (full m-columns case, set A DM-global, set
B DM-local). |
| |
• NSPSTAF: pointer to
where each m-column starts; used for semi-implicit (full
m-columns case, set A DM-global, set B DM-local). |
| |
• NUMLL: distribution
of levels on B-processor set (DM-global). |
| |
• NPTRLL: pointer to first
level on this B-processor set (set A DM-global, set B DM-local). |
| |
• NPSP: equals 1 if pressure
is handled by this B-processor set (set A DM-global, set B DM-local). |
| |
• NPSURF: equals 1 if
surface pressure is handled by this B-processor set (set A DM-global,
set B DM-local). |
| |
• NBSETLEV: on which B-set
do the level belongs (set A DM-global, set B DM-local). |
| |
• NBSETSP: min(NFLEVG+1,NPROCB)
(set A DM-global, set B DM-local). |
| |
• MYLEVS: level numbers
mapped to current process (set A DM-global, set B DM-local). |
| |
• NVMODIST: normal modes
mapped to the different B-sets (set A DM-global, set B DM-local). |
| |
• NSPEC2V: number of spectral
columns computed by this process for NSMAX arrays
(DM-local). |
| |
• NSPEC2VF: number of
spectral columns computed by this process if complete columns are
required (DM-local). |
| |
• NSTA: start position
of grid columns on latitudes (set A DM-global, set B DM-local). |
| |
• NONL: number of grid
columns on latitudes (set A DM-global, set B DM-local). |
| |
• NGPSET: number of grid
points of this processor in grid point space which also belong to
it in Fourier space (set A DM-global, set B DM-local). |
| |
• NAPLAT: number of apple
latitudes at the poles (DM-global). Default value is 0. |
| |
• NGPTRSEND: defines grid
columns to be sent to each B-set during TRGTOL (DM-local). |
| |
• NGPTRRECV: defines B-sets
to receive from grid columns during TRGTOL (DM-local). |
SUFPCSET and SUFPRSET variables (based on NFPWIDE) used for FULL-POS
interpolations:
| |
• NAFPB1: local inner
dimension of interpolation buffer used for FULL-POS interpolations
(DM- local). |
| |
• NFPSTA: interpolation
buffer start position of grid columns on latitudes (set A DM-global,
set B DM-local). |
| |
• NFPONL: interpolation
buffer number of grid column on latitudes (set A DM-global, set B
DM- local). |
| |
• NFPOFF: interpolation
buffer offset to start of each row in FPBUF1 and
FPBUF2 (set A DM- global, set B DM-local). |
| |
• NFPEXT: interpolation
buffer extend latitudes over poles support (set A DM-global, set B
DM- local). |
| |
• NFPSTLAT: latitudinal
position in the interpolation buffer during the data sending (DM-local). |
| |
• NFPSTLON: longitudinal
position in the interpolation buffer during the data sending (DM-local). |
| |
• NFPRSTLAT: latitudinal
position in the interpolation buffer during the data reception (DM-local). |
| |
• NFPRSTLON: longitudinal
position in the interpolation buffer during the data reception (DM-
local). |
| |
• NFPSENDNUM: length of
the sent interpolation buffer (DM-local). |
| |
• NFPRECVNUM: length of
the received interpolation buffer (DM-local). |
| |
• NFPCORE: offsets to
core region points in interpolation buffer (set A DM-global, set B
DM- local). |
Some other variables:
| |
• NCOMBFLEN: size of communication
buffer. This is the maximum per processor buffer space (in words)
that the ARPEGE/IFS should use for one or more sends before receives
are issued from destination processors (DM-global). Default value
is 1800000. |
| |
• LBIDIR: .TRUE. if bi-directional
transpositions are preferred (DM-global). Only implemented in the
A-direction and require even number of processors. Default value is
.FALSE. . |
| |
• LSNDS: logical indicating
whether to send (resp. receive) data to southern set in TRLTOG
(resp. TRGTOL) (DM-local). |
| |
• LSNDN: logical indicating
whether to send (resp. receive) data to northern set in TRLTOG
(resp. TRGTOL) (DM-local). |
| |
• LRCVS: logical indicating
whether to receive (resp. send) data from southern set in TRLTOG
(resp. TRGTOL) (DM-local). |
| |
• LRCVN: logical indicating
whether to receive (resp. send) data from northern set in TRLTOG
(resp. TRGTOL) (DM-local). |
| |
• NSNDS: number of messages
to send (resp. receive) to southern set in TRLTOG
(resp. TRGTOL) (DM-local). |
| |
• NSNDN: number of messages
to send (resp. receive) to northern set in TRLTOG
(resp. TRGTOL) (DM-local). |
| |
• Other distributed memory environment
variables which are currently in another place, and for which YOMMP
would be a better place: |
| |
• NUMP (in YOMDIM):
number of spectral waves handled by this processor (DM-local).
NUMP=NSMAX+1 for shared memory. |
| |
• NUMCP (in YOMDIM):
same as NUMP, but related to NCMAX
(DM-local). NUMCP=NCMAX+1 for shared
memory. |
| |
• NUMTP (in YOMDIM):
same as NUMP, but related to NTMAX
(DM-local). NUMTP=NTMAX+1 for shared
memory. |
| |
• MYMS (in YOMLAP):
actual wave numbers handled by each processor (DM-global). |
Namelist NAMPAR1:
Initialize variables that control layout of distribution. The following
variables of YOMMP are in namelist NAMPAR1,
namelist reading and default values calculations are done in SUMP0:
LSPLIT, LSLSYNC, LAPPLE,
LBIDIR, NSLPAD, NAPLAT,
NPHASE, NLAGA, NLAGB,
NSTRIN, NSTROUT, NFLDIN,
NFLDOUT, NVALAG, NCOSTLAG,
NLAGBDY, LSPLITIN, LSPLITOUT,
NCOMBFLEN
Variables specific to the distributed memory environment. DM-global variables
mainly used to handle reading and writing of grib data. Some of these variables
can be used also for shared memory. Some variables in YOMMPG
(which does not contain only FULL-POS variables) can be useful for FULL-POS.
No variable in namelist.
| |
• NPOSSP: defines partitioning
of global spectral fields among processors (DM-global). |
| |
• NPOSCP: as
NPOSSP for NCMAX arrays (DM-global). |
| |
• NPOSTP: as
NPOSSP for NTMAX arrays (DM-global). |
| |
• NDIM0G: defines partitioning
of global spectral fields among processors (DM-global). |
Some variables in YOMOP (which does not contain only FULL-POS
variables) and namelist NAMOPH can be useful for FULL-POS.
| |
• LINC: .T. to get time
stamp in hours rather than in time steps (default value is .FALSE.).
DM- global variable. |
Contains buffer for fully post-processed physical fields. Variables are
initialised in setup routine setup/SUFPSC2B.
No variable in namelist.
List of variables:
| |
• NLENPFPBL: length of
buffer (DM-local). |
| |
• GPFPBUF : buffer (DM-local). |
| |
• NSTAPFPB : start adresses
for post-processed packets of points within buffer (DM-local). |
Contains buffer of auxiliary fields (geometry, climatology) for horizontal
post-processing. Variables are initialised in setup routines setup/SURFPBUF
for buffer RFPBUF and setup/SUFPSC2B
for other variables. No variable in namelist.
List of variables:
| |
• NLENRFPBL: length of
buffer (DM-local). |
| |
• RFPBUF : buffer containing
output climatology and geometry (DM-local). |
| |
• NSTARFPB : start adresses
for post-processed packets of points within buffer (DM-local). |
Description of auxiliary fields for horizontal post-processing. Variables
are initialised in setup routine setup/SURFPDS.
Variables are DM-global. No variable in namelist.
Field pointers:
| |
• MFPGMO : for output
mapping factor. |
| |
• MFPGNXO: for output
x-component of rotation matrix. |
| |
• MFPGNYO: for output
y-component of rotation matrix. |
| |
• MFPLSMO: for output
land-sea mask. |
| |
• MFPFISO: for output
surface geopotential. |
| |
• MFPFISI: for interpolated
input surface geopotential. |
| |
• MFPCSTO: for output
climatological surface temperature. |
| |
• MFPCSWO: for output
climatological relative surface moisture. |
| |
• MFPSDO : for output
climatological snow cover. |
| |
• MFPRSTO: for output
relaxation deep soil temperature. |
| |
• MFPRSWO: for output
relaxation deep soil moisture. |
| |
• MFPDSTO: for output
climatological deep soil temperature. |
| |
• MFPDSWO: for output
climatological relative deep soil moisture. |
| |
• MFPALBO: for output
albedo. |
| |
• MFPEMIO: for output
emissivity. |
| |
• MFPDEVO: for output
(standard deviation) . |
| |
• MFPVEGO: for output
vegetation. |
| |
• MFPGZ0O: for output
 . |
| |
• MFPANIO: for output
anisotropy. |
| |
• MFPDIRO: for output
direction. |
Other variables:
| |
• NFPVSO: number of climatological
fields. |
| |
• CFPVSO: ARPEGE field
names of the climatological fields. |
| |
• NRFPOS: total number
of fields in RFPBUF. |
Contains parameters used to control vectorisation and memory space. Some
of these parameters are used in FULL-POS. No variable in namelist.
| |
• NSLBCT: packet control
array for SCAN2MSM or SCAN2MDM
(DM-local). Packets are NPROMA-packets for shared
memory, NGPTOT-packets for distributed memory. |
| |
• NSLBCT(1,.): start row
of packet. |
| |
• NSLBCT(2,.): end row
of packet. |
| |
• NSLBCT(3,.): 1 if non-lagged
grid-point computations on packet, 0 if not. |
| |
• NSLBCT(4,.): 1 if lagged
grid-point computations on packet, 0 if not. |
| |
• NSLBCT(5,.): usage of
SLBUF1 for packet (semi-Lagrangian scheme). |
| |
• NSLBCT(6,.): usage of
SLBUF2 for packet (semi-Lagrangian and Eulerian schemes). |
| |
• NSLBCT(7,.): usage of
FPBUF1 for packet (FULL-POS interpolations for fields
other than surface physics). |
| |
• NSLBCT(8,.): usage of
FPBUF2 for packet (FULL-POS interpolations for surface
physics). |
| |
• NSLBCT(5,.) to NSLBCT(8,.)
are useless for distributed memory. For distributed memory
NSLBCT(1,1)=NSLBCT(2,1)=NSLBCT(3,1)=
NSLBCT(4,1)=1, the other values are useless. |
| |
• NSLBR: number of packets
of rows (DM-local). Packets are NPROMA-packets for
shared memory, NGPTOT-packets for distributed memory.
NSLBR=1 for distributed memory. |
| |
• NLOCKFP1: multitasking
lock for interpolation buffer FPBUF1 used in FULL-POS.
Useless for distributed memory. |
| |
• NLOCKFP2: multitasking
lock for interpolation buffer FPBUF2 used in FULL-POS.
Useless for distributed memory. |
| |
• NLOCKOBL: multitasking
lock for COBSLAGAD and COBSLAG.
Useless for distributed memory. |
| |
• NFPB1EV: event for interpolation
buffer FPBUF1 used in FULL-POS. Useless for distributed
memory. |
| |
• NFPB2EV: event for interpolation
buffer FPBUF2 used in FULL-POS. Useless for distributed
memory. |
| |
• NWEVFP1: number of tasks
waiting for interpolation buffer FPBUF1 event used
in FULL-POS. Useless for distributed memory. |
| |
• NWEVFP2: number of tasks
waiting for interpolation buffer FPBUF2 event used
in FULL-POS. Useless for distributed memory. |
| |
• NSLWIDE: number of rows
the model lagged part runs behind (DM-global). |
| |
• NFPWIDE: number of rows
the FULL-POS lagged part runs behind (DM-global). |
| |
• for shared memory: number of buffers
(last dimension) for interpolation buffer FPBUF1
used in FULL-POS. |
| |
• for distributed memory: used to dimension
an interpolation buffer but only in the non lagged part of calculations;
NFPB1=NGPBLKS (DM-local). |
| |
• for shared memory: number of buffers
(last dimension) for interpolation buffer FPBUF2
used in FULL-POS. |
| |
• for distributed memory: used to dimension
an interpolation buffer but only in the non lagged part of calculations;
NFPB2=NGPBLKS (DM-local). |
| |
• NSTABUF: start address
for each row in grid-point buffers (DM-local). |
| |
• NLATPBF: packet pointer
for each row in interpolation buffers (DM-local). |
| |
• NDIST: start address
for each row in grid-point calculations (DM-local). |
| |
• NDIEND: end address
for each row in grid-point calculations (DM-local). |
| |
• NCIST: start address
for each row of grid-point field in buffer (DM-local). |
| |
• NCIEND: end address
for each row of grid-point field in buffer (DM-local). |
| |
• NBFPFP1 (resp. NBFPFP2): |
| |
• elements (1,.): pointer to packet contained
in each buffer (FPBUF1, resp. FPBUF2). |
| |
• elements (2,.): remaining users of buffer
(FPBUF1, resp. FPBUF2). |
DM-global variables.
| |
• NSC2EVH: event for each
row indicating non-lagged part complete for FULL-POS. Useless for
distributed memory. |
Contains switches to use FULL-POS in an incremental variational application.
Variables are initialised in setup routine var/SUVAR
(move to directory setup will be desirable in the future).
Variables are DM-global.
List of variables:
| |
• LARCHFP: .T. => FULL-POS
will be used for communication between the high and low resolution
run. Default is .F. |
| |
• LREFFP : .T. => the
reference field is post-processed at high resolution. Default is .F. |
| |
• LRFTLFP: .T. => the
reference field is post-processed at low resolution. Default is .F. |
| |
• LANAFP : .T. => the
final (analysis) field is post-processed. Default is .F. |
Namelist NAMVFP:
The following variable of YOMVFP is in namelist NAMVFP:
LARCHFP.
Contains control variables for vertical post-processing. Variables are initialised
in setup routine setup/SUVFPOS and can
be modified by routine transform/UPDVPOS.
All variables are DM-global. No variable in namelist.
Control variables for post-processing of 3D variables.
| |
• LFIT3 : .T./.F. if 3D
post-processed fields are fitted/not fitted. |
| |
• Scalar variables NADD...:
address in array GT1. |
| |
• NADDWIND : address of
wind. |
| |
• NADDPSI : address of
potential velocity. |
| |
• NADDDIV : address of
divergence. |
| |
• NADDVOR : address of
relative vorticity. |
| |
• NADDABS : address of
absolute vorticity. |
| |
• NXFPLEV : number of
post-processing levels. |
| |
• XFPLEV : values of post-processing
levels. Array dimensioned to NXFPLEV . |
| |
• Scalar variables NLEV...:
number of post-processing (pp) levels for different variables. |
| |
• NLEVGEOP : number of
pp levels for geopotential. |
| |
• NLEVTEMP : number of
pp levels for temperature. |
| |
• NLEVUMOM : number of
pp levels for U-wind component. |
| |
• NLEVVMOM : number of
pp levels for V-wind component. |
| |
• NLEVSHUM : number of
pp levels for specific humidity. |
| |
• NLEVRHUM : number of
pp levels for relative humidity. |
| |
• NLEVVVEL : number of
pp levels for vertical velocity. |
| |
• NLEVLWAT : number of
pp levels for liquid water. |
| |
• NLEVSNOW : number of
pp levels for solid water. |
| |
• NLEVCLFR : number of
pp levels for cloud fraction. |
| |
• NLEVSCVA : number of
pp levels for passive scalars. |
| |
• NLEVDIVE : number of
pp levels for divergence. |
| |
• NLEVVORT : number of
pp levels for relative vorticity. |
| |
• NLEVPSI : number of
pp levels for velocity potential. |
| |
• NLEVKHI : number of
pp levels for stream function. |
| |
• NLEVTETA : number of
pp levels for potential temperature. |
| |
• NLEVTHPW : number of
pp levels for  . |
| |
• NLEVWIND : number of
pp levels for wind velocity. |
| |
• NLEVEPTH : number of
pp levels for equivalent potential temperature. |
| |
• NLEVABSV : number of
pp levels for absolute vorticity. |
| |
• NLEVSTDF : number of
pp levels for stretching deformation. |
| |
• NLEVSHDF : number of
pp levels for shearing deformation. |
| |
• NLEVPOTV : number of
pp levels for potential vorticity. |
| |
• NLEVWEPV : number of
pp levels for wet potential vorticity. |
| |
• NLEVPRES : number of
pp levels for pressure. |
| |
• NLEVUA01 to NLEVUA16
: number of pp levels for upper air fields nr 01 to 16. |
| |
• Arrays NLVP... (dimensioned
with NFPXLEV): level pointers for different variables. |
| |
• NLVPGEOP : level pointers
for geopotential. |
| |
• NLVPTEMP : level pointers
for temperature. |
| |
• NLVPUMOM : level pointers
for U-wind component. |
| |
• NLVPVMOM : level pointers
for V-wind component. |
| |
• NLVPSHUM : level pointers
for specific humidity. |
| |
• NLVPRHUM : level pointers
for relative humidity. |
| |
• NLVPVVEL : level pointers
for vertical velocity. |
| |
• NLVPLWAT : level pointers
for liquid water. |
| |
• NLVPSNOW : level pointers
for solid water. |
| |
• NLVPCLFR : level pointers
for cloud fraction. |
| |
• NLVPSCVA : level pointers
for passive scalars. |
| |
• NLVPDIVE : level pointers
for divergence. |
| |
• NLVPVORT : level pointers
for relative vorticity. |
| |
• NLVPPSI : level pointers
for velocity potential. |
| |
• NLVPKHI : level pointers
for stream function. |
| |
• NLVPTETA : level pointers
for potential temperature. |
| |
• NLVPTHPW : level pointers
for  . |
| |
• NLVPWIND : level pointers
for wind velocity. |
| |
• NLVPEPTH : level pointers
for equivalent potential temperature. |
| |
• NLVPABSV : level pointers
for absolute vorticity. |
| |
• NLVPSTDF : level pointers
for stretching deformation. |
| |
• NLVPSHDF : level pointers
for shearing deformation. |
| |
• NLVPPOTV : level pointers
for potential vorticity. |
| |
• NLVPWEPV : level pointers
for wet potential vorticity. |
| |
• NLVPPRES : level pointers
for pressure. |
| |
• NLVPUA01 to NLVPUA16
: level pointers for upper air fields nr 01 to 16. |
| |
• Arrays NSDO... (dimensioned
with NFPXLEV): number of subdomains for each level
of different variables. |
| |
• NSDOGEOP : number of
subdomains for each level of geopotential. |
| |
• NSDOTEMP : number of
subdomains for each level of temperature. |
| |
• NSDOUMOM : number of
subdomains for each level of U-wind component. |
| |
• NSDOVMOM : number of
subdomains for each level of V-wind component. |
| |
• NSDOSHUM : number of
subdomains for each level of specific humidity. |
| |
• NSDORHUM : number of
subdomains for each level of relative humidity. |
| |
• NSDOVVEL : number of
subdomains for each level of vertical velocity. |
| |
• NSDOLWAT : number of
subdomains for each level of liquid water. |
| |
• NSDOSNOW : number of
subdomains for each level of solid water. |
| |
• NSDOCLFR : number of
subdomains for each level of cloud fraction. |
| |
• NSDOSCVA : number of
subdomains for each level of passive scalars. |
| |
• NSDODIVE : number of
subdomains for each level of divergence. |
| |
• NSDOVORT : number of
subdomains for each level of relative vorticity. |
| |
• NSDOPSI : number of
subdomains for each level of velocity potential. |
| |
• NSDOKHI : number of
subdomains for each level of stream function. |
| |
• NSDOTETA : number of
subdomains for each level of potential temperature. |
| |
• NSDOTHPW : number of
subdomains for each level of  . |
| |
• NSDOWIND : number of
subdomains for each level of wind velocity. |
| |
• NSDOEPTH : number of
subdomains for each level of equivalent potential temperature. |
| |
• NSDOABSV : number of
subdomains for each level of absolute vorticity. |
| |
• NSDOSTDF : number of
subdomains for each level of stretching deformation. |
| |
• NSDOSHDF : number of
subdomains for each level of shearing deformation. |
| |
• NSDOPOTV : number of
subdomains for each level of potential vorticity. |
| |
• NSDOWEPV : number of
subdomains for each level of wet potential vorticity. |
| |
• NSDOPRES : number of
subdomains for each level of pressure. |
| |
• NSDOUA01 to NSDOUA16
: number of subdomains for each level of upper air fields nr 01 to
16. |
| |
• Arrays NSDP... (dimensioned
with (NFPDOM,NFPXLEV)): subdomains pointers for different
variables. |
| |
• NSDPGEOP : subdomains
pointers for geopotential. |
| |
• NSDPTEMP : subdomains
pointers for temperature. |
| |
• NSDPUMOM : subdomains
pointers for U-wind component. |
| |
• NSDPVMOM : subdomains
pointers for V-wind component. |
| |
• NSDPSHUM : subdomains
pointers for specific humidity. |
| |
• NSDPRHUM : subdomains
pointers for relative humidity. |
| |
• NSDPVVEL : subdomains
pointers for vertical velocity. |
| |
• NSDPLWAT : subdomains
pointers for liquid water. |
| |
• NSDPSNOW : subdomains
pointers for solid water. |
| |
• NSDPCLFR : subdomains
pointers for cloud fraction. |
| |
• NSDPSCVA : subdomains
pointers for passive scalars. |
| |
• NSDPDIVE : subdomains
pointers for divergence. |
| |
• NSDPVORT : subdomains
pointers for relative vorticity. |
| |
• NSDPPSI : subdomains
pointers for velocity potential. |
| |
• NSDPKHI : subdomains
pointers for stream function. |
| |
• NSDPTETA : subdomains
pointers for potential temperature. |
| |
• NSDPTHPW : subdomains
pointers for  . |
| |
• NSDPWIND : subdomains
pointers for wind velocity. |
| |
• NSDPEPTH : subdomains
pointers for equivalent potential temperature. |
| |
• NSDPABSV : subdomains
pointers for absolute vorticity. |
| |
• NSDPSTDF : subdomains
pointers for stretching deformation. |
| |
• NSDPSHDF : subdomains
pointers for shearing deformation. |
| |
• NSDPPOTV : subdomains
pointers for potential vorticity. |
| |
• NSDPWEPV : subdomains
pointers for wet potential vorticity. |
| |
• NSDPPRES : subdomains
pointers for pressure. |
| |
• NSDPUA01 to NSDPUA16
: subdomains pointers for upper air fields nr 01 to 16. |
Control variables for post-processing of 2D variables.
| |
• LFIT2 : .T./.F. if 2D
post-processed fields are fitted/not fitted. |
| |
• Logical variables LFP(XXX):
.T./.F. if post-processing/no post-processing (pp/no pp) of variable
(XXX) |
| |
• LFPMSLP : .T./.F. if
pp/no pp of mean sea level pressure. |
| |
• LFPSP : .T./.F. if pp/no
pp of surface pressure. |
| |
• LFPORO : .T./.F. if
pp/no pp of orography in the model. |
| |
• LFPGM : .T./.F. if pp/no
pp of mapping factor (grid-point only). |
| |
• LFPFOL : .T./.F. if
pp/np pp of tropopause folding indicator (grid-point only). |
| |
• LFPSU1 : .T./.F. if
pp/no pp of surface field nr 1. |
| |
• LFPSU2 : .T./.F. if
pp/no pp of surface field nr 2. |
| |
• LFPSU3 : .T./.F. if
pp/no pp of surface field nr 3. |
| |
• LFPSU4 : .T./.F. if
pp/no pp of surface field nr 4. |
| |
• LFPSU5 : .T./.F. if
pp/no pp of surface field nr 5. |
| |
• Scalars NSDO...: number
of subdomains for different variables. |
| |
• NSDOMSLP : number of
subdomains for mean sea level pressure. |
| |
• NSDOSP : number of subdomains
for surface pressure. |
| |
• NSDOORO : number of
subdomains for orography in the model. |
| |
• NSDOGM : number of subdomains
for mapping factor. |
| |
• NSDOFOL : number of
subdomains for tropopause folding indicator. |
| |
• NSDOSU1 : number of
subdomains for surface field nr 1. |
| |
• NSDOSU2 : number of
subdomains for surface field nr 2. |
| |
• NSDOSU3 : number of
subdomains for surface field nr 3. |
| |
• NSDOSU4 : number of
subdomains for surface field nr 4. |
| |
• NSDOSU5 : number of
subdomains for surface field nr 5. |
| |
• Arrays NSDP... (dimensioned
with NFPDOM): subdomain pointers for different variables. |
| |
• NSDPMSLP : subdomain
pointers for mean sea level pressure. |
| |
• NSDPSP : subdomain pointers
for surface pressure. |
| |
• NSDPORO : subdomain
pointers for orography in the model. |
| |
• NSDPGM : subdomain pointers
for mapping factor. |
| |
• NSDPFOL : subdomain
pointers for tropopause folding indicator. |
| |
• NSDPSU1 : subdomain
pointers for surface field nr 1. |
| |
• NSDPSU2 : subdomain
pointers for surface field nr 2. |
| |
• NSDPSU3 : subdomain
pointers for surface field nr 3. |
| |
• NSDPSU4 : subdomain
pointers for surface field nr 4. |
| |
• NSDPSU5 : subdomain
pointers for surface field nr 5. |
Contains buffer of working fields (indices, weights) for horizontal post-processing.
Variables are initialised in setup routines setup/SUWFPBUF
for buffer WFPBUF and setup/SUFPSC2B
for other variables. No variable in namelist.
List of variables:
| |
• NLENWFPBL: length of
buffer (DM-local). |
| |
• WFPBUF : buffer containing
working fields for horizontal interpolations (size of the box, indices,
weights, land/sea mask) (DM-local). |
| |
• NSTAWFPB : start adresses
for post-processed packets of points within buffer (DM-local). |
Description of working fields for horizontal post-processing. Variables
are initialised in setup routine setup/SUWFPDS.
Variables are DM-global. No variable in namelist.
Field pointers:
| |
• MBOX : for the size
of the box. |
| |
• MILA : pointer for the
array containing the nearest northern latitude (lat 1 of figure
5.2). |
| |
• MILON : pointer for
the array containing the longitude index of the point A1 of
figure 5.2 . |
| |
• MILOS : pointer for
the array containing the longitude index of the point A2 of
figure 5.2 . |
| |
• MILONN: pointer for
the array containing the longitude index of the point A0 of
figure 5.2 . |
| |
• MILOSS: pointer for
the array containing the longitude index of the point A3 of
figure 5.2 . |
| |
• MWSTD : for the first
weight without land-sea mask. |
| |
• MWLSM : for the first
weight with land-sea mask for scalars. |
Other variables:
| |
• NWFPOS: total number
of fields in WFPBUF. |
These namelists are NAMFPPHY (read in routine setup/SUFPPHY),
NAMFPDY2, NAMFPDYP, NAMFPDYH,
NAMFPDYV, NAMFPDYT and NAMFPDYS
(read in routine setup/SUFPDYN). These
namelists can be used to make an accurate list of post-processed fields.
In ordinary case, at each post-processing time step, all the fields that
are written in the namelist NAMFPC are post-processed at
all the levels and for all the output domains written in the namelist NAMFPC.
However, it is possible to get, at each post-processing time step, exactly
the fields you wish, and nothing more: in that case, you have to make other
namelists file which will contain the selection of the fields you wish to
get. First, you have to set in NAMCT0 the variable CNPPATH
as the directory where the selection files will be. Under this directory,
the name of a selection file must be xxtDDDDHHMM, where DDDDHHMM specifies
the date/time of the post-processing time step. Then a selection file must
contain 7 new namelists (see above).
Namelist NAMFPPHY:
Namelist NAMFPPHY is for physical fields and fluxes; it
contains the following DM-global parameters:
| |
• CLPHY: names of selected
physical fields; array of 16 characters. |
| |
• CLDPHY: names of selected
subdomains, separated by ":", for each selected physical field; characters
array (example: 'DOM1:DOM2:DOM3'). |
| |
• CLCFU: names of selected
cumulated fluxes fields; array of 16 characters. |
| |
• CLDCFU: names of selected
subdomains, separated by ":", for each selected cumulated flux; characters
array (example : 'DOM1:DOM2:DOM3'). |
| |
• CLXFU: names of selected
instantaneous fluxes fields; array of 16 characters. |
| |
• CLDXFU: names of selected
subdomains, separated by ":", for each selected instantaneous flux;
characters array (example : 'DOM1:DOM2:DOM3'). |
Namelist NAMFPDY2:
Namelist NAMFPDY2 is for 2D dynamical fields; it contains
the following DM-global parameters:
| |
• CL2DF: names of selected
fields; array of 16 characters. |
| |
• CLD2DF: names of selected
subdomains, separated by ":", for each selected field; characters
array (example : 'DOM1:DOM2:DOM3'). |
Namelists NAMFPDYP, NAMFPDYH, NAMFPDYV, NAMFPDYT and NAMFPDYS:
These namelists are for fields post-processed respectively on pressure levels,
height levels, potential vorticity levels, isentropic levels and hybrid
levels; they contain the same following DM-global parameters:
| |
• CL3DF: names of selected
fields; array of 16 characters. |
| |
• IL3DF: list of selected
indexes of post-processing levels, for each selected field; integer
array of 2 dimensions; first subscript is for selected post-processing
level; second one is the local field index. |
| |
• CLD3DF: names of selected
subdomains, separated by ":", for each selected level and each selected
field; characters array of 2 dimensions; first subscript is the local
index of the selected level; second one is the local field index (example
: 'DOM1:DOM2:DOM3'). |
Note that all the fields or levels or domains written in a selection file
should be included in the main namelist NAMFPC, otherwise
the job will abort. Default value of namelist quantities is 0 for IL3DF
and ' ' for character quantities.
Pointer PTRFP4.
Contains pointers for FULL-POS on one-level-type dynamical fields. All variables
are DM-global. Variables are initialised in setup routine setup/SU4FPOS.
Horizontally post-processed dynamical array:
| |
• NFLDFPD: field pointer.
Array of dimension NFPDYNB. |
| |
• RLEVFPD: level value.
Array of dimension NFPDYNB. |
| |
• NBITFPD: number of bits
when writing output file. Array of dimension NFPDYNB. |
| |
• NDOMFPD: number of subdomains
for each field. Array of dimension NFPDYNB. |
| |
• NINDFPD: indexes of
subdomains for each field. Array of dimension (NFPDOM,NFPDYNB). |
Physics and fluxes:
| |
• NFLDFPP: field pointer.
Array of dimension NFPPHYB. |
| |
• FPARITP: parity array
(+1 for scalar; -1 for vector). Array of dimension NFPPHYB. |
| |
• NITMFPP: land-sea mask
array for interpolation (0 if without; 1 if with). Array of dimension
NFPPHYB. |
| |
• NBITFPP: number of bits
when writing output file. Array of dimension NFPPHYB. |
| |
• NDOMFPP: number of subdomains
for each field. Array of dimension NFPPHYB. |
| |
• NINDFPP: indexes of
subdomains for each field. Array of dimension (NFPDOM,NFPPHYB). |
Spectral arrays:
| |
• NFLDAFP: field pointers
in SPAFP. Array of dimension NFPSPA. |
| |
• NFLDDFP: field pointers
in SPDFP. Array of dimension NFPSPD. |
| |
• RLEVAFP: level values
in SPAFP. Array of dimension NFPSPA. |
| |
• RLEVDFP: level values
in SPDFP. Array of dimension NFPSPD. |
Grid point array:
| |
• NFLDFPX: field pointers
in GAUXBUF. Array of dimension NFPAUXB. |
| |
• RLEVFPX: level values
in GAUXBUF. Array of dimension NFPAUXB. |
| |
• NBITFPX: number of bit
when writing out GAUXBUF. Array of dimension NFPAUXB. |
Other variables:
| |
• NC3FP: field pointers.
Array of dimension NFP3DF. |
| |
• X3FP : level values
per field. Array of dimension (NFPXLEV,NFP3DF). |
| |
• NI3FP: subdomain index
for each level of each field. Array of dimension (NFPDOM,NFPXLEV,NFP3DF). |
Next Section
Previous Section
|
