Dear Laurent and others,
I just tested the SOSIE3 executable (compiled with ifort) with the provided examples. It works well with almost all examples.
The execution using namelist.example3 (regridding from an ORCA1 grid, I didn't change any entry there) generates a segmentation fault:
forrtl: severe (174): SIGSEGV, segmentation fault occurred
I attached the logfiles from the console and from the debugger, where it occurs.

Do you have a hint, what to do to make ot work?
Many thanks in advance,
Janna

Hi there. I did some digging in the docs, and I couldn't figure out whether or not SOSIE supports ORCA025 conversion into a .nc file in the ORCA1 or ORCA2 format? It's not my area of expertise but I need to convert it for someone else.

Thanks!

Good afternoon
I tried to compile sosie but something went wrong. netcdf library version 4.4.1.1 ifort version 14.0.1
[zhaonemo@mgt02 sosie-master]$ make
ifort -I/public/app/nemolib/netcdf/include -c src/io_ezcdf.f90 -o obj/io_ezcdf.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_conf.f90 -o obj/mod_conf.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_scoord.f90 -o obj/mod_scoord.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_init.f90 -o obj/mod_init.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_manip.f90 -o obj/mod_manip.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_grids.f90 -o obj/mod_grids.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_bdrown.f90 -o obj/mod_bdrown.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_drown.f90 -o obj/mod_drown.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_akima_2d.f90 -o obj/mod_akima_2d.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_poly.f90 -o obj/mod_poly.o
ifort -xHOST -O3 -i4 -module mod/ -I/public/app/nemolib/netcdf/include -c src/mod_bilin_2d.f90 -o obj/mod_bilin_2d.o
ifort -xHOST -O3 -i4 -module mod/ -c src/mod_akima_1d.f90 -o obj/mod_akima_1d.o
src/mod_akima_1d.f90(85): error #6404: This name does not have a type, and must have an explicit type. [FINDLOC]
k1_1 = FINDLOC(imiss1, 0, 1)
-------------------^
compilation aborted for src/mod_akima_1d.f90 (code 1)
make: *** [obj/mod_akima_1d.o] Error 1

Hello.

I am new to programming and I just tried to interpolate an ERA5 file towards a CMEMS grid file using sosie3.x interpolation tool. (with grid types are regular). However, I get an error message like this:
Namelist_Final.pdf

./sosie3.x: symbol lookup error: ./sosie3.x: undefined symbol: __netcdf_MOD_nf90_open

Namelistm seems to work fine but as soon as the interpolation is about to start this error message appears. I attach also the namelist file that I use and I copy and paste the output when I try to run sosie3.x:

==========================================================
S O S I E version 3.0

• namelist = namelist

Opening namelist "namelist"

The current field is NOT a component of a vector!

Source grid is declared as regular

Target grid is declared as regular

Source file: /home/sofiad/SWR_MARCH_2D.nc

Method for 2D interoplation: bilin

Time record name in source file: time

File containing the source grid: /home/sofiad/SWR_MARCH_2D.nc

Longitude name: longitude

Latitude name: latitude

Source grid mask: missing_value

Source mask variable:

Variable to be treated:msnswrf

Level to treat: 1

Name of variable on the output file:temp

Units of variable to be treated:

Long name of variable to be treated:

Target grid: /home/sofiad/CMEMS_thetao_depth_0_2017-03.nc

Longitude on target grid: longitude

Latitude on target grid: latitude

Longitude name in target file: longitude

Latitude name in target file: latitude

Target land-sea mask file: /home/sofiad/mask_CMEMS_2D_FIN.nc

Name of land-sea mask variable on target grid: thetao

Output file: ./temp_regular-regular_monthly_1998.nc

Shall we use DROWN on source fields: T

=> about DROWN action:
-> continental penetration in # of pixels: 100
-> # of time to apply a smoothing on continents after DROWN: 50
-> is the mask on the source field is changing with time : F

East west periodicity of source grid: 0

Masking output file: T

Value for masked points in output file: -9999.00000

East west periodicity of target grid: 2

Time is not controlled and will be the same than in source file!

             ---------------------

Starting interpolation for config "regular-regular"...

./sosie3.x: symbol lookup error: ./sosie3.x: undefined symbol: __netcdf_MOD_nf90_open

Could you please help me with this?
Thank you in advance for your reply,
Kind regards,
Sofi

I am trying to interpolate ORCA05 fields onto the ORCA025 grid and found that the deallocation of i1dum here: https://github.com/brodeau/sosie/blob/master/src/mod_manip.f90#L1135 breaks the loop. Commenting this line results in success.

I have been using sosie to "drown" some atmospheric forcing files. In order to use these files with MOM6, I need to clean up after sosie:

1. Sosie sets a blank "calendar" attribute on the time variable - it was there in the parent file as "gregorian".
2. Sosie lies about the "valid_range", setting it smaller than it actually is in the file. Easiest is to delete this attribute (it wasn't there in the parent file), but one could also set it to correctly match the range used in the file.

I can clean up with ncatted, but why should I have to?

Hello,

I wanted to ask whether sosie is able to interpolated from C-grid to A-grid data. I have vertical velocities w in a native NEMO C-grid and I am planning to use an A-grid file to interpolate the vertical velocities in A grid as well.

Will that work? How will I know if my file is interpolated in the A-grid.
Thank you very much for your answer.
Kind regards,
Sofi

Hello.

I am new to programming and I just tried to interpolate 3D so data from NEMO-ORCA2 to EN.4 analiysis grid file using sosie3.x interpolation tool. I tried to imitate namelist.example2, but it always prompts that the variable cannot be found. And I can’t find which variable can't be found. All file was provided by ./data , and i will attach my namelist file and err file.
Could you please help me with this?
Thank you in advance for your reply,
Kind regards,
Sunhh
files.zip

Dear sirs,

Good afternoon. My name is Sofia and I am a student currently using sosie.x for interpolation in NEMOMED8 (drown method). I recently encountered a problem, during which I could not seem to apply the sosie interpolation to another model file (Not NEMO), because the output was not interpolated properly for some reason. Perhaps the grid of the model (curvilinear) file has a particuliarity that sosie does not recognise (?). Could you please give me your insights on that?

I am attaching here a .docx file called "sosie.out" where it shows where exactly is the error encountered during the .map creation file and I am also attaching images taken with ncview of the original .nc file I wanted to interpolate and its variables. (lon,lat,sst). It is showing the Mediterranean region only that the grid is tilted or rotated somehow. (It is the Mediterranean Sea region I am interested in)

Finally I am attaching a file with two images (from ncview), to show you exactly what is the difference between the original file and the interpolated. In this file the interpolated plot is named : sst_tr04_nas-583_dm.nc and the original file is the plot named : tr04_nas_583_dm_2005.nc.

I hope this information will be helpful. Thank you in advance for your time and help,
Kind regards,
Sofia Darmaraki

sosie_out.docx

hello,

I have compiled the SOSIE successfully. while running the test case I am getting the fellowing errors.

-bash: sosie3.x: command not found

I used this command to run it
sosie3.x -f namelist.example1

I also download and untag the fellowing libraries
data.tar.gz and data_rims.tar.gz

Please help me to solve this errors ASAP.

Best Regards
Neelam

Hello,

I cloned the master branch and compiled it using make.macro_ifort_NESH on the respective machine (including debugging flags -traceback -fpe0 -CB).
When running the examples that use bilinear interpolation (e.g. example1_bilin, example3, example5_bilin) I get the following error message:

 Error occured in procedure P2D_MAPPING_AB !
 
 Treated file     = sosie_mapping_360x180-ORCA1.nc
 Treated variable = metrics (masking)
 
 --> aborting program
 
 Netcdf message was :
 NetCDF: Not a valid data type or _FillValue type mismatch

Examples that use akima interpolation work fine.
A similar issues has been reported here and the issue was solved by falling back to an older intel and netcdf environment. I tested all the environments that are available on the machine, without any success.

Any suggestions what might be going wrong here?

Cheers,
Patrick

The inverse vector correction does not seem to work anymore in the most recent commits... I see, that the -x and -y switches have been remove in commit 4de7074 aready (I was using an earlier version before), however when eliminating -x and -y from my command

old:
corr_vect.x -I -i gird_U.nc grid_V.nc -t time_counter -x vozocrtx -y vomecrty -m mask.nc
new:
corr_vect.x -I -i gird_U.nc grid_V.nc -t time_counter -m mask.nc

I just get the "Usage"-message as if I typed corr_vect.x -h. Is the inverse correction not supported anymore? Or do I need to give specific names to the variables for this to work?
Any help would be much appreciated!

Hi,

The last google drive link of the user's guide does not work for me. Please, check.

Thank you,
Francesco

I cloned your repository (commit 710176c)
compiling works well (see my make.macro below), during run time this call in io_ezcdf.F90 line 1031

CALL prepare_nc(idx_f, cdt, lx, ly, cv_lo, cv_la, cv_t, vextrema, &
            &            id_x, id_y, id_t, id_lo, id_la, id_tim, crtn,cf_in,cv_in, &
            &            attr_lon=attr_lon, attr_lat=attr_lat, attr_time=attr_time) 

leads to the following error message:
sosie.x: string.c:177: NC_check_name: Assertion `name != ((void *)0)' failed.
(no output is generated, while the interpolation has been successful).

Instead I now use the following work around NOT providing all these attributes (since attr_lon, attr_lat and attr_time are optional - this is just fine for me).

CALL prepare_nc(idx_f, cdt, lx, ly, cv_lo, cv_la, cv_t, vextrema, &
            &            id_x, id_y, id_t, id_lo, id_la, id_tim,crtn,cf_in,cv_in)

To compile I used the following make.macro

# ============================================
# Fortran compiler:
FC = ifort
# Root directory for Netcdf:
NETCDF_DIR = /sfs/fs3/sw/netcdf/netcdf-4.1.3/netcdf-4.1.3_intel
# Linking argument: usually -lnetcdf or -lnetcdff (or both):
L_NCDF = -lnetcdf -lnetcdff
# Fortran compilation flags:
# -- Production
FF = -O3 -i4 -xHost -module mod/
# -- Debugging
#FF = -O0 -i4 -traceback -CB -module mod/
# Directory to install binaries:
INSTALL_DIR = ./bin

Good evening

I'm trying to compile SOSIE on the HLRN machine in Berlin, Germany.
I have done the following:

module load git
git clone -b '3.0' https://github.com/brodeau/sosie
cd sosie/

module load netcdf/intel/4.7.3 hdf5/intel/1.10.5
module load intel/18.0.6
cp macro/make.macro_intel_HLRN make.macro
export NETCDF_DIR=$(nf-config --prefix)

So I've got the Intel Fortran, netCDF/HDF5 and set the NETCDF_DIR variable.

The compilation works, and the lib/libsosie.a is created. But then compilation fails with

a - obj/mod_interp.o
a - obj/mod_nemotools.o
a - obj/mod_poly.o
ranlib lib/libsosie.a

ifort  src/sosie.f90 -o bin/sosie3.x -L./lib -lsosie -L/sw/dataformats/netcdf/intel.18/4.7.3/skl/lib -lnetcdf -lnetcdff ; 
src/sosie.f90(41): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [MOD_CONF]
   USE mod_conf     !* important parameters, namelist, arrays, etc...
-------^
src/sosie.f90(42): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [MOD_INIT]
   USE mod_init     !* important parameters, namelist, arrays, etc...
-------^
src/sosie.f90(43): error #7002: Error in opening the compiled module file.  Check INCLUDE paths.   [MOD_GRIDS]
   USE mod_grids    !* grided-domain module

(full error message is attached).

Does anyone know what to do? As I understand it, the mod_conf should be in the libsosie.a which is clearly being linked here.
Also tried compiling on my Macbook (using Gfortran etc) and got the same error...

Best wishes
Joakim
sosie_compile_error.txt

Hi,
tried to compile trunk version of sosie on Gustafson at BSC with GCC4.9.2 & netcdf4.2
Had to modify 3 routines to make it work :

• io_ezcdf.f90 : in the READ statements, had to change '(i)' by '(i4)'
• interp_to_ground_track.f90 : in the READ statment, had to change '(f)' to '(f8.0)'. Also, had to change "xdum_r8 = SIGN(1.,180.-xlon_gt)" by "xdum_r8 = SIGN(1.,180.-REAL(xlon_gt,4))"
• mod_grids.f90 : Had to change " SIGN(1.,180.-xdum)" by " SIGN(1.,180.-REAL(xdum,4))"

Thomas

Hello,
I need to interpolate a dataset from native ORCA1 grid (namely CESM tn1v3 grid) to ORCA025 grid (namely CESM tn0.25v3 grid).

This is the input namelist:

!! -------------------
!! Namelist for SOSIE
!! -------------------
!!
!! Interpolation of SST from the ORCA1 to the ORCA025 grid
!!
!!
&ndom_src
csource    = 'tn1v3'
ivect      = 0
l_reg_src  = .false.
cf_src     = 'sst_tn1v3.nc'
cv_src     = 'SST_cpl'
cv_t_src   = 'time'
cf_x_src   = 'sst_tn1v3.nc'
cv_lon_src = 'xc'
cv_lat_src = 'yc'
cf_lsm_src = 'missing_value'
cv_lsm_src = ''
ewper_src  = 0
/
!!
!!
&ndom_trg
ctarget    = 'tn025v3'
l_reg_trg  = .false.
cf_x_trg   = tn0.25v3_L50_mesh_mask.nc'
cv_lon_trg = 'nav_lon'
cv_lat_trg = 'nav_lat'
cf_lsm_trg = 'missing_value'
cv_lsm_trg = ''
ewper_trg  = 0
/
!!
!!
&ninterp
cmethod     = 'bilin'
!!
idrown      = 50, 10, .false.
ixtrpl_bot  = 0
l_save_drwn = .false.
ismooth     = 0
jt1         = 0
jt2         = 0
jplev       = 1
vmax        = 1.E6
vmin        = -1.E6
ismooth_out = 0
/
!!
!!
&noutput
cv_out    = 'SST_cpl'
cu_out    = ''
cln_out   = ''
cv_t_out  = 'time'
cd_out    = 'output'
cextra    = 'Monthly_SST_amip'
lmout     = .false.
rmiss_val = -9999.
lct       = .false.
t0        = 0.
t_stp     = 0.
cv_z_out  = ''
/
!!

Variable attributes from input file read:

float SST_cpl(time, y, x) ;
		SST_cpl:standard_name = "BCS Pseudo sea_surface_temperature" ;
		SST_cpl:long_name = "Constructed mid-month Sea Surface Temperature" ;
		SST_cpl:units = "degC" ;
		SST_cpl:coordinates = "yc xc" ;
		SST_cpl:_FillValue = 1.e+20f ;
		SST_cpl:missing_value = 1.e+20f ;
		SST_cpl:cell_methods = "time: point" ;

I get the following error:

Error occured in procedure P2D_MAPPING_AB !
Treated file     = sosie_mapping_tn1v3-tn025v3.nc
Treated variable = metrics (masking)
 --> aborting program
Netcdf message was :
NetCDF: Not a valid data type or _FillValue type mismatch

I am using ifort 19.0.5.281 20190815 and netCDF-cxx4 4.3.1.
Any suggestion?

UPDATE
Also running the provided examples namelist.example1_bilin and namelist.example3 leads to the same issue.

Hi,
I've generated a 2D-mapping file from ORCA12 to ORCA025.
Both configurations got the same number of vertical levels.
Is there a tricky way to generate the 3D mapping file based on this existing 2D mapping file, without running the (costly) mapping file generation process of sosie on all levels ?
As 3D interpolation is a level-by-level interpolation, in this case I guess it's just a mater of duplicating in a smart way the existing weights...
Thanks
Thomas.

Hi there,

I am executing sosie.x (version 2.6 as that is the only version my computing server can use), and I am getting this error:
$ sosie.x -f namelist.example1
$ sosie.x: error while loading shared libraries: libnetcdf.so.18: cannot open shared object file: No such file or directory

I have downloaded and unzipped the files too. I'm not too sure what has gone wrong?

Thanks,
Yusuf

Hello.
I tried to interpolate 3D so data from NEMO-ORCA2 to EN.4 analiysis grid file using sosie3.x interpolation tool. I tried to imitate namelist.example2, but it always prompts that the variable cannot be found. Then i found that running namelist.example2 directly does not succeed, it will prompt Invalid argument.Could you please help me with this?

Thank you in advance for your reply,
Kind regards,
Sunhh

Hi there. Compiling SOSIE on Debian 9 will yield this result. Note that the exact same install and compilation process under Ubuntu 18.04 works fine.

Version information:

gcc --version
gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516

gfortran --version
GNU Fortran (Debian 6.3.0-18+deb9u1) 6.3.0 20170516

uname -a 
Linux 4.9.0-6-amd64 #1 SMP Debian 4.9.88-1+deb9u1 (2018-05-07) x86_64 GNU/Linux

nc-config --all
This netCDF 4.6.1 has been built with the following features:

  --cc        -> gcc
  --cflags    -> -I/usr/local/include -I/usr/local/include
  --libs      -> -L/usr/local/lib -lnetcdf

  --has-c++   -> no
  --cxx       ->

  --has-c++4  -> no
  --cxx4      ->

  --has-fortran-> yes
  --fc        -> gfortran
  --fflags    -> -I/usr/local/include
  --flibs     -> -L/usr/local/lib -lnetcdff -L/usr/local/lib -lnetcdf -lnetcdf
  --has-f90   -> no
  --has-f03   -> yes

  --has-dap   -> no
  --has-dap4  -> no
  --has-nc2   -> yes
  --has-nc4   -> no
  --has-hdf5  -> no
  --has-hdf4  -> no
  --has-logging-> no
  --has-pnetcdf-> no
  --has-szlib ->
  --has-parallel -> no
  --has-cdf5 -> yes

  --prefix    -> /usr/local
  --includedir-> /usr/local/include
  --libdir    -> /usr/local/lib
  --version   -> netCDF 4.6.1

Installation / Compilation:

sudo apt install git gcc gfortran make build-essential \
     netcdf-bin ncview libnetcdff-dev libnetcdff-dbg  \
     libnetcdf-dev 

git clone https://github.com/brodeau/sosie
cd sosie
git fetch && git checkout 3.0
cp ./macro/make.macro_gfortran_Debian ./make.macro
make && sudo make install 

Behaviour:

Error: There is no specific function for the generic ‘nf90_def_var’ at (1)
src/io_ezcdf.f90:959:39:

       IF (PRESENT(cv_dt7)) CALL sherr( NF90_DEF_VAR(idf, TRIM(cv_dt7), NF90_FLOAT, idtd, idv7, deflate_level=9), crtn,cf_in,cv_dt7 )
                                       1
Error: There is no specific function for the generic ‘nf90_def_var’ at (1)
src/io_ezcdf.f90:960:39:

       IF (PRESENT(cv_dt8)) CALL sherr( NF90_DEF_VAR(idf, TRIM(cv_dt8), NF90_FLOAT, idtd, idv8, deflate_level=9), crtn,cf_in,cv_dt8 )
                                       1
Error: There is no specific function for the generic ‘nf90_def_var’ at (1)
make: *** [obj/io_ezcdf.o] Error 1

Hi completely new to Fortran and tried to build the sosie for evaluation.
On WSL ubuntu, installed gfortran, libnetcdf-dev, libnetcdff-dev.
Can you provide general instruction how to create 'make.macro' and compile ?

Hello!
After installing official version Sosie 3.0 I have met some error message when trying to interpolate from ORCA onto 360x180 grid:
../bin/sosie3.x -f namelist.example3
...

Error occured in procedure P2D_MAPPING_AB !

Treated file = sosie_mapping_ORCA1-360x180.nc
Treated variable = metrics (masking)

--> aborting program

Netcdf message was :
NetCDF: Not a valid data type or _FillValue type mismatch

It seems that a file “sosie_mapping_ORCA1-360x180.nc” cannot be generated.
Could you please help me with this problem?

Kind wishes,
Vyacheslav Khon

Dear sirs,

I have installed Sosie version 3.0, using Ubuntu and gfortan v.10.
I am trying to insert latitude and longitude values to a netcdf file with ORCA75 grid by running the following:

QUOTE

PATH TO SOSIE BIN (interpolation software executable)

sosie_bin=~/sosie/bin/sosie3.x

PATH TO GEOMETRY FILE (input grid file)

grid=./GRID.nc

WHAT TO INTERPOLATE (directory, name, year, parameter)

expdir=.
expname=CHOR_AS
year=1901
variable=somxl010

TARGET RESOLUTION: 1 for 1x1 regular grid (WOA grid) ; 2 for 0.5x0.5 regular grid

tar_res=2

##################### RUN

if [ $tar_res -eq 1 ]; then
fog=woa_grid.nc
olo="lon"
ola="lat"
onm="WOA"
elif [ $tar_res -eq 2 ]; then
fog=spheric
olo="0.5"
ola="0.5"
onm="0.5x0.5"
else
echo "Target resolution not recognized"
exit -9
fi

fin=${expdir}/${expname}${year}.nc
fun=unp${expname}_${year}.nc

NetCDF Unpacking

ncpdq -h -U $fin -O $fun || exit -2

cat > sosie_namelist << EOF
&ninput
ivect = 0
lregin = F
cf_in = '$fun'
cv_in = '$variable'
cv_t_in = 'time_counter'
jt1 = 0
jt2 = 0
jplev = 1
cf_x_in = '$grid'
cv_lon_in = 'lon'
cv_lat_in = 'lat'
cf_lsm_in = '$grid'
cv_lsm_in = 'tmsk'
ldrown = T
ewper = 2
vmax = 1.E28
vmin = -1.E28
/
&n3d
cf_z_in = '$grid'
cv_z_in = 'dep'
cf_z_out = '/path/to/depth/file.nc'
cv_z_out = 'dep'
/
!!
!!
!!
!! *****************************
!! Output Grid characteristics :
!! *****************************
!!
&noutput
lregout = T
cf_x_out = '$fog'
cv_lon_out = '$olo'
cv_lat_out = '$ola'
cf_lsm_out = ''
cv_lsm_out = ''
lmout = F
rmaskvalue = -9999.
lct = F
t0 = 0.
t_stp = 0.25
/
!!
!!
!!
!!
!! *******************************
!! Netcdf output characteristics :
!! *******************************
!!
&nnetcdf
cmethod = 'bilin'
cv_l_out = 'lon'
cv_p_out = 'lat'
cv_t_out = 'time'
cv_out = '$variable'
cu_out = 'missing'
cu_t = 's'
cln_out = '$variable'
cd_out = '.'
csource = 'ORCA05'
ctarget = '$onm'
cextra = '$year'
/
EOF

$sosie_bin -f sosie_namelist || exit -2
rm -f sosie_namelist
exit 0
UNQUOTE

The files that I am using are GRID.nc that has the required lat and lon, and CHOR_AS_1901.nc in which I want to insert the lat and lon values. Below you can find the files information:

QUOTE
netcdf GRID {
dimensions:
y = 511 ;
x = 722 ;
z = 75 ;
variables:
double kdcoast2d(y, x) ;
kdcoast2d:missing_value = 1.e+20f ;
int topo(y, x) ;
int mbathy(y, x) ;
float vmsk(z, y, x) ;
float umsk(z, y, x) ;
float dz(z) ;
float ff(y, x) ;
float dy(y, x) ;
float dx(y, x) ;
float tmsk(z, y, x) ;
float lat(y, x) ;
float dep(z) ;
float lon(y, x) ;
float lsm(y, x) ;
double dcoast2d(y, x) ;
dcoast2d:missing_value = 1.e+20f ;

// global attributes:
:DOMAIN_number_total = 1 ;
:DOMAIN_number = 0 ;
:DOMAIN_dimensions_ids = 1, 2 ;
:DOMAIN_size_global = 722, 511 ;
:DOMAIN_size_local = 722, 511 ;
:DOMAIN_position_first = 1, 1 ;
:DOMAIN_position_last = 722, 511 ;
:DOMAIN_halo_size_start = 0, 0 ;
:DOMAIN_halo_size_end = 0, 0 ;
:DOMAIN_type = "BOX" ;
}
UNQUOTE

QUOTE
netcdf CHOR_AS_1901 {
dimensions:
time_counter = UNLIMITED ; // (0 currently)
y = 511 ;
x = 722 ;
variables:
double somxl010(y, x) ;
somxl010:_FillValue = NaN ;
}
UNQUOTE

I would really appreciate your assistance on this error.
Thank you very much in advance for your assistance.

With kind regards,
Dionysia Rigatou

Dear @brodeau and all,
Happy new year!
When I try to use the 3.0 Sosie
Sosie.log
(branch master) on a ECMWF netcdf file at 1/10deg to interpolate it on a 1/48deg grid, I receive the error below:

.......
=> going to disregard points of target domain with lon < -75.00 and lon > 42.00

Program received signal SIGSEGV, Segmentation fault.
0x000000000047e292 in mod_grids::trg_domain () at src/mod_grids.f90:211
211 xdum = degE_to_degWE(xdum)

I attached the complete log obtained with gdb.
Do you have any idea on the possible causes?
bests,
Sergio

Do i understand correctly that i have to do the inverse steps to get a correct interpolation from a Nemo output to a 2 polar e.g etopo grid than explained in example 4 on: https://brodeau.github.io/sosie/ -
this means first correction of the vectors and then interpolation?
corr_vect.x ...
and then
sosie.x -f try.U
sosie.x -f try.V

What is the correction actually doing? - Is it already accounting for the projection of the given curvilinear vectors along the model grid to a real-mercator N-S E-W grid?
so is the output already aligned along real n-s and e-w lines?

Thank you for your time.
Cheers Patricia

I have ran into an issue from converting from MODIS gridded data to ORCA1 using sosie. There is a noticible change in the geometric shape of many of the landmasses, particularly Antarctica and Africa. I'm just wondering if this is normal or maybe I have gone wrong in the configurations?

You can see in the attached images below.