http://nco.sf.net (Homepage, Mailing lists, Help)
http://github.com/nco/nco (Source Code, Issues, Releases)
What's new?
Version 5.2.8 mostly contains features and fixes for regridding
and climatology generation. Climatologies now allow the choice
between two statistics (sums and means) to output as regional/global
timeseries. The regridder fixes corner case behavior when the
sub-gridscale (SGS) fraction is a missing value (as in ELM/CLM).
Many improvements to the robustness and helpfulness of WARNINGs
and ERRORs are included. Upgrades are encouraged for everyone
who uses the regridder on ELM/CLM/CTSM datasets.
Enjoy,
Charlie
NEW FEATURES (full details always in ChangeLog):
A. ncclimo has two new options to control global/regional timeseries:
--rgn_stt=avg|sum and --sum_scl=sum_scl. These options allow the user
to choose which statistic, sums or averages, to output with
global/regional timeseries for all variables. The option --rgn_stt=avg
is identical to the current timeseries output by the --rgn_avg switch
(which will be deprecated eventually). When invoked with --rgn_stt=sum
the averaged field is multiplied by the sum of the area variable.
For fields that are area-intensive (e.g., fluxes per unit area) this
results in the total net flux over the area. However, the field must
employ the same area units as the Area variable for this to be true.
For example, fields given in m-2 would need to employ an Area variable
in m2. Unfortunately, many people love non-SI units so that is rarely
the case! For example, ELM and CLM archive Area in km2, so a scale
factor of one million is needed to correct the sum for many variables.
EAM and CAM use sr-1 for Area so it needs a different scale factor.
That is why ncclimo accepts a second new option --sum_scl=sum_scl.
This scale factor multiplies the integrated field value, and allows
the user to generate timeseries in the desired units for any field.
The whole procedure is model and variable-specific and we are open
to suggestions to make it more useful. Example usage is:
# Timeseries of global GPP in grams/s for ELM/CLM:
ncclimo -P elm --split --rgn_stt=sum --sum_scl=1.0e6 -v GPP ...
# Timeseries of global GPP in GT C/yr for ELM/CLM:
ncclimo -P elm --split --rgn_stt=sum --sum_scl=1.0e6*3600*24*365/1.0e12
-v GPP ...
# Timeseries of global column vapor in kg for EAM/CAM:
ncclimo -P eam --split --rgn_stt=sum --sum_scl=6.37122e6^2 -v TMQ ...
http://nco.sf.net/nco.html#ncclimo
http://nco.sf.net/nco.html#rgn_stt
http://nco.sf.net/nco.html#sum_scl
BUG FIXES:
A. ncks --chk_map fixes a mix-up introduced in 5.2.7 where the new
area_b diagnostics were mis-labeled and show the location of min/max
on the area_a grid instead of the area_b grid. The workaround is to
use 5.2.6 or earlier. The solution is to upgrade.
B. ncremap fixes a problem where portions of missing values (aka
_FillValue) could infect destination cells whose sub-gridscale (SGS)
fractions equal _FillValue. This problem can arise in ELM/CLM
datasets because their 'landfrac' variable sets non-land areas to the
missing value (a practice of questionable value). Thanks to Chloe
Whicker-Clarke for reporting this issue. There is no workaround. The
solution is to upgrade.
Full release statement at http://nco.sf.net/ANNOUNCE
--
Charlie Zender, Dept. of Earth System Science
University of California, Irvine 949-891-2429
