Extracting time series at a large number of grid points from a global grid
Added by David Heineke almost 2 years ago
Good day,
I am trying to extract time series from a global grid (see attached) at a large number of grid points (~10,000, see attached indices).
I am working in Python using CDO version 1.9.9rc1 on a Windows Subsystem for Linux (WSL).
As looping over the grid points takes too much time, I have tried to first select the required grid cells and then to distribute the horizontal grid as follows:
reduced_fields = cdo.selgridcell(*indices, input = ERA5_19800101_19800102.nc)
cdo.distgrid(10543, input = reduced_fields, output = 'test_')
However, the distgrid function seems to return only one time step per grid cell. Can you please advise? Are there any alternative methods to do this?
Thanks in advance.
Kind regards,
David
Replies (3)
RE: Extracting time series at a large number of grid points from a global grid - Added by Karin Meier-Fleischer almost 2 years ago
Hi David,
for me it works and each single file has 16 time steps. Maybe it is a problem caused by the Windows WSL installation?
print(cdo.__version__()) 1.5.4 print(cdo.version()) 1.9.9
I used the following Python script to do your calculations:
indices_file = 'Coverage_coast_ind2.txt'
infile = 'ERA5_19800101_19800102.nc'
indices = [ e.rstrip('\n') for e in open(indices_file, 'r')]
reduced_fields = cdo.selgridcell(*indices, input=infile)
!mkdir -p tmp_files
cdo.distgrid(10543, input=reduced_fields, output='./tmp_files/test_')
E.g. the first file:
cdo sinfon ./tmp_files/test_00000.nc
File format : NetCDF2
-1 : Institut Source T Steptype Levels Num Points Num Dtype : Parameter name
1 : unknown unknown v instant 1 1 1 1 I16 : u10
2 : unknown unknown v instant 1 1 1 1 I16 : v10
3 : unknown unknown v instant 1 1 1 1 I16 : mdts
4 : unknown unknown v instant 1 1 1 1 I16 : mdww
5 : unknown unknown v instant 1 1 1 1 I16 : mpts
6 : unknown unknown v instant 1 1 1 1 I16 : mpww
7 : unknown unknown v instant 1 1 1 1 I16 : mwd
8 : unknown unknown v instant 1 1 1 1 I16 : mwp
9 : unknown unknown v instant 1 1 1 1 I16 : wmb
10 : unknown unknown v instant 1 1 1 1 I16 : pp1d
11 : unknown unknown v instant 1 1 1 1 I16 : swh
12 : unknown unknown v instant 1 1 1 1 I16 : shts
13 : unknown unknown v instant 1 1 1 1 I16 : shww
Grid coordinates :
1 : unstructured : points=1
lon : 71 degrees_east
lat : 77 degrees_north
Vertical coordinates :
1 : surface : levels=1
Time coordinate : 16 steps
RefTime = 1900-01-01 00:00:00 Units = hours Calendar = gregorian
YYYY-MM-DD hh:mm:ss YYYY-MM-DD hh:mm:ss YYYY-MM-DD hh:mm:ss YYYY-MM-DD hh:mm:ss
1980-01-01 00:00:00 1980-01-01 03:00:00 1980-01-01 06:00:00 1980-01-01 09:00:00
1980-01-01 12:00:00 1980-01-01 15:00:00 1980-01-01 18:00:00 1980-01-01 21:00:00
1980-01-02 00:00:00 1980-01-02 03:00:00 1980-01-02 06:00:00 1980-01-02 09:00:00
1980-01-02 12:00:00 1980-01-02 15:00:00 1980-01-02 18:00:00 1980-01-02 21:00:00
-Karin
RE: Extracting time series at a large number of grid points from a global grid - Added by David Heineke almost 2 years ago
Thank you very much for your quick response Karin.
I do think it a problem caused by the Windows WSL. Is there any way to update to a more recent version, or should I use a different platform?
print(cdo.__version__()) 1.5.3rc1 print(cdo.version()) 1.9.9rc1
Kind regards,
David
RE: Extracting time series at a large number of grid points from a global grid - Added by Karin Meier-Fleischer almost 2 years ago
Most of our Windows users are using Cygwin to run CDO. Maybe it is worth a try to use conda (from miniconda) to install cdo and python-cdo.