Cdo command for NetCDF weather file
CDO DOES NOT HAVE a graphic environment! It only processes data. The commands run directly on the Linux terminal.
Note: For NetCDF data support, the installation of the NetCDF libraries are required;
Syntax: cdo input.nc output.nc or
cdo input.nc.
首先：export PATH=/public/software/anaconda2/bin:$PATH
1. File Information:
1 infon: information about the dataset listed by the variable name.
2 sinfon: information on the dataset in a reduced form.
3 griddes: shows information about the file domain.
4 zaxisdes: shows information about the vertical coordinate.
5 pardes: description of the variable or variables of the file.
cdo infon standard.nc | head
cdo infov ifile
cdo showtimestamp ~
cdo sellonlatbox ~
2. File Manipulation:
1 copy: duplicate the file and/or convert to another format.
2 cat: increases the number of file times.
3 merge: joins files with different variables.
4 split: divide the file (vertical level or time).
copy: duplicates the file and/or converts from one format to another.
Note:The files must have the same dimensions (latitude, longitude, vertical levels).
Example1: Converts from NetCDF to GRIB:
cdo -f grb copy pr.GPCP.1996.2005.nc pr.GPCP.1996.2005.grb
Example2: Converts from GRIB to NetCDF:
cdo -f nc copy pr.GPCP.1996.2005.grb pr.GPCP.1996.2005.nc
Example of concatenation with the copy operator:
cdo copy pr.GPCP.1996.1999.nc pr.GPCP.2000.2002.nc pr.1996.2002.nc
cat: increases the number of file times.
cdo cat pr.GPCP.1996.1999.nc pr.GPCP.2000.2002.nc pr.1996.2002.nc
merge: joins files with different variables.
cdo merge u.NCEP.R2.1996.2005.nc v.NCEP.R2.1996.2005.nc uv.nc
Note:Files do not necessarily have to have the same dimensions (latitude, longitude, vertical levels).
split: divide or separate the file.
Operator: splitlevel, splithour, splitday, splitmon, splitseas and splityear
Example: Separate file in months:
cdo splitmon pr.GPCP.1996.2005.nc month.
Where: month. represents a prefix or any name defined by user and pr.GPCP.1996.2005.nc is the input file.There will be 12 files created: month.01.nc, month.02.nc, ... ,month.12.nc.
3. Selection Operators:
1 sellevel: selects the vertical level.
2 selmon: selects the month of the file.
3 selseason: selects the season of the year.
4 selyear: selects the year file.
5 sellonlatbox: cut the data in the domain of interest
cdo selyear,2000/2002 tar.NCEP.R2.1996.2005.nc tar.2000.2002.nc
cdo selyear,1996/1998,2000 tar.NCEP.R2.1996.2005.nc tar.1996a1998.2000.nc
cdo seldate,2014-12-12T12:00:00, 2015-01-31T18:00:00 infile.nc outfile.nc
sellonlatbox: cut the data according to the area of interest.
cdo sellonlatbox,-90,-30,-60,10 pr.GPCP.1996.1999.nc pr.AS.nc
The convention will always be: longitude west, longitude east, latitude south and north latitude
Longitude values can be reported in either the 0 to 360 format and in the -180 to 180 format.
How to convert the global data to -180 to 180? Only works with global data.
cdo sellonlatbox,-180,180,-90,90 pr.GPCP.1996.1999.nc pr.global.nc
setcalendar: sets the calendar type of the file.
Possibilities: standard, proleptic_gregorian, 360_day, 365_day e366_day.
ncdump -h stan.nc
cdo setcalendar,365_day stan.nc st.nc
setmissval: sets a new missing value (UNDEF).
cdo setmissval,-999 tar.GFDL-ESM2M.1996.2005.nc tar.GFDL-ESM2M.nc
4. Operations with constants:
1 addc: sum to a constant.
2 subc: subtracts from a constant.
3 mulc: multiplies by a constant.
4 divc: divided by a constant.
乘法
Converting the precipitation of [kg/(m2 s)] to (mm/day).
cdo mulc,86400 pr.CANESM2.1996.2005.nc pr.mensal.nc
减法
Converting Kelvin temperature (K) to degrees Celsius (oC)
cdo subc,273.15 tar.NCEP.R2.1996.2005.nc tc.nc
5. Operations using two sets of data (uni or two-dimensional).
1 add: adds two fields.
2 sub: subtract two fields.
3 mul: multiplies two fields.
4 div: divide dois campos.
The files must have the same dimensions (latitude, longitude, vertical levels).
cdo sub pr.CANESM2.2005.nc pr.GPCP.2005.nc dif.nc
Annual statistical calculation. Converts the data to annual time scale
independent of the temporal resolution.
For example, if the data is hourly, daily or monthly, it will be converted
for an annual archive.
1 yearsum: annual sum.
2 yearmean: annual average.
3 yearavg: annual average.
Seasonal statistical calculation::
1 seassum: seasonal sum.
2 seasmean: seasonal average.
3 seasavg: seasonal average.
Statistical calculation - converts to monthly scale..
1 ymonsum: monthly sum.
2 ymonmean: monthly average.
3 ymonavg: monthly average.
Operators to calculate correlation.
fldcor ⇒ correlates all grid points for each time of the two
variables. The result is a time series of correlation between them.
timcor ⇒ performs the correlation at each grid point for all times
of the two variables. The result is a file spatial correlation with
only one time.
Example1: Correlation between precipitation and temperature. The
data are monthly referring to the year of 2004 (12 months or times).
cdo fldcor prec.mensal.nc temp.mensal.nc output.nc
Examples of other applications:
1 Calculate the anomaly.
2 Create masking.
3 Filling missing data.
4 Change NetCDF file values.
5 Text file conversion to NetCDF.
6 Conversion of a radiosonde to the NetCDF format.