The netCDF Operators NCO version 4.9.1 are woke.
http://nco.sf.net (Homepage, Mailing lists, Help)
http://github.com/nco (Source Code, Issues, Releases, Developers)
What's new?
Version 4.9.1 contains an important bugfix for vertical interpolation,
and two small new features for ncclimo and ncpdq. It also fixes some
autoconf-build issues that affected some users with Intel compilers,
and those attempting parallel builds of NCO.
Work on NCO 4.9.2 has commenced and will improve NCO weight-generator
accuracy, reduce vertical interpolation memory use, and supply more
accurate weight generation options for rectangular lat-lon grids.
Work on NCO 5.0.0 has commenced "under the hood". A key leap in that
release will be support for netCDF4 user-defined types. Printing of
netCDF4 user-defined types ENUM and VLEN is ready now (though
unsupported) with the --udt flag. 5.0.0 will contain the finished
version of that, and include options for invoking mbtempest in place
of tempest.
Enjoy,
Charlie
NEW FEATURES (full details always in ChangeLog):
A. ncremap changes the default regridder algorithm from ESMF
"bilinear" to NCO "nco_con", a first-order conservative algorithm
(similar to ESMF "conserve") that requires no external
dependencies. Because the default algorithm comes with NCO, users
without access to ESMF and/or TempestRemap can perform a full
regridding workflow. Of course many/most users will still wish to
employ ESMF or TR, and their algorithms are easily accessed:
ncremap -a nco -s grd_src.nc -d grd_dst.nc -m map.nc # New default
ncremap -a bilin -s grd_src.nc -d grd_dst.nc -m map.nc # Old default
ncremap -a fv2fv -s grd_src.nc -d grd_dst.nc -m map.nc # Tempest
http://nco.sf.net/nco.html#alg_typ
http://nco.sf.net/nco.html#ncremap
B. ncclimo can now output spatial-average timeseries.
The new --glb_avg (or --global_average) option causes ncclimo to
perform a spatial mean over the timeseries constructed from the
input data. The main limitation is that it only works for
timeseries whose horizontal spatial dimension is "ncol" and
whose area weight is named "area", as in CAM SE/EAM.
Let us know if this function would be useful to you for other
models and we can easily add options to make these hardwired values
specifiable from the command-line.
ls *mdl*.nc | ncclimo --glb_avg -s 1979 -e 1981 --drc_out=drc_out
http://nco.sf.net/nco.html#ncclimo
http://nco.sf.net/nco.html#glb_avg
C. ncpdq has a new map, flt_dbl, to convert single-precision to
double-precision floating point variables. The resulting variables
are not packed. All single-precision variables, including
coordinates, are promoted. Some software requires geospatial data
in double precision, and this feature simplifies the dataset
conversion.
ncpdq -M flt_dbl in.nc out.nc
ncpdq -M sgl_dbl in.nc out.nc
http://nco.sf.net/nco.html#flt_dbl
http://nco.sf.net/nco.html#ncpdq
BUG FIXES:
A. Fix a difficulty that caused Intel compilers to fail to build NCO.
Now the netcdf.h header, and thus the EXTERNL token, are always
included and defined during autoconf header checks.
Thanks to Matthew Thompson of NASA for reporting the issue, and to
Dave Allured of NOAA for providing the fix.
B. Print WARNING instead of dying when requested to copy the
_NCProperties attribute. That attribute should not but can be
exposed through the netCDF API in some circumstances, and this new
treatment prevents that leakage from breaking workflows. Thanks to
Rotislav for reporting and Dave Allured for helpful discussions.
This is really a feature yet listed as a bug-fix because it is
an incomprehensible technical workaround, the kind of thing users
should never need to know about.
C. Fix an error when interpolating to/from a pure pressure grid.
The minimum pressure was not always being initialized, which caused
(highly) erroneous values when the interpolation type was mss_val.
This bug only affects NCO installations built by compilers that do
not automatically initialize allocated memory to preset values.
Affected output datasets would be obviously corrupted, so nothing
subtle about this bug. There is no workaround, the only fix is to
upgrade. Thanks to Sterling Baldwin for reporting this.
D. Fix misuse of autoconf library references. This foiled parallel
builds, and caused difficulty compiling ncap2. Thanks to Daniel
Macks for identifying and solving the issue.
Full release statement at http://nco.sf.net/ANNOUNCE
--
Charlie Zender, Earth System Sci. & Computer Sci.
University of California, Irvine 949-891-2429 )'(
