Irregular clipping of NC data with CDO

CDO Processing climate data , In especial NetCDF The data format is very convenient , It can be said to be the Swiss Army knife for climate data processing .

This article takes CMIP Take the data , Give an example to illustrate how to use CDO Yes NetCDF Data clipping .

Rule tailoring

Rule tailoring is generally two methods ：

1. Cut according to longitude and latitude

Cutting according to longitude and latitude requires a certain longitude and latitude box, That is to say lonmin,lonmax,latmin,latmax, This selection operation is in CDO What we use sellonlatbox command , For example, we should select the scope of China from the global data , Then you can use the following command ：

cdo sellonlatbox,73,136,3,54 global.nc china.nc

• global.nc It's the input file
• china.nc It's the output file

The above four numbers correspond to your target area Minimum longitude 、 Maximum longitude 、 Minimum latitude 、 Maximum latitude .

2. According to the grid ID To cut

This kind is generally used for special grids , Like Gauss N16 Grid this , Example usage is as follows :

cdo selindexbox,60,11,3,11 infile outfile

( Not very familiar with , Not much , Because... Is rarely used )

Irregular clipping

The results of the above rules are generally like this ：

Rule tailoring

But many times what we want is the basis shp The result of irregular clipping of the file , Like this ：

Irregular clipping

First we need to know , Irregular clipping is to cut unnecessary meshes mask（ Set to nodata）, Not to delete those grids ！

There are many ways to get the above results , For example, use R、Python、MATLAB, Other number owners have also written a lot , I won't repeat . Here is a non open programming language , Methods completed directly on the command line .

The tools you want to use are CDO and GDAL（ To make GIS It's usually installed GDAL）, The idea is ：

• Gridding shp File get grid file ;
• Take the above file as a mask and do conditional operations with the global data .

1. Gridding shp file

We can use it conveniently on the command line gdal_rasterize Command to shp Rasterize the file ：

gdal_rasterize -of netCDF -burn 1 -tr 0.01 0.01 china.sha china.nc

• china.shp： The vector boundary of your study area
• china.nc： Output nc file （ Used as a mask）

By the command above , We got china.nc：

mask

2. Conduct mask

CDO Provides ifthen Command to help us complete this step , Its use for ：

cdo ifthen china.nc global.nc china_masked.nc

The final result is as follows ：

mask result

• ifthen The command has three arguments , The first is mask file , The second is to carry out mask The file of , The third is the output ;
• It should be noted that ,mask file （ This example is china.nc） The resolution of needs to match global.nc Consistent resolution , That's up there gdal_rasterize Of tr The following parameters need to be consistent with global.nc The grid size of is consistent .

How to reverse mask Well ？

Put the above ifthen The order was changed to ifnotthen that will do .

Reference resources

【1】cdo Official user documentation