Read numerical value of variable in NetCDF4 using Python

Question

I need to undestand how I can get the numerical value of one channel IR_108 ( variable) in NetCDF4 and some other data:

import numpy as np
from netCDF4 import Dataset
 path = '/home/data/Mmultic3kmNC4_msg04_201905080200.
 ...: nc'
In [78]: f=netCDF4.Dataset(path,'r')
 f.variables.keys()
Out[80]: odict_keys(['time', 'dtime', 'IR_016', 'commentaires', 'satellite', 'geos', 'ImageNavigation', 'GeosCoordinateSystem', 'Y', 'X', 'Albedo_to_Native_count_IR_016', 'IR_039', 'Temp_to_Native_count_IR_039', 'IR_087', 'Temp_to_Native_count_IR_087', 'IR_097', 'Temp_to_Native_count_IR_097', 'IR_108', 'Temp_to_Native_count_IR_108', 'IR_120', 'Temp_to_Native_count_IR_120', 'IR_134', 'Temp_to_Native_count_IR_134', 'VIS006', 'Albedo_to_Native_count_VIS006', 'VIS008', 'Albedo_to_Native_count_VIS008', 'WV_062', 'Temp_to_Native_count_WV_062', 'WV_073', 'Temp_to_Native_count_WV_073'])
In [81]: IR=f.variables['IR_108']
In [82]: IR
Out[82]: 
<class 'netCDF4._netCDF4.Variable'>
int16 IR_108(ny, nx)
 standard_name: toa_brightness_temperature
 long_name: 10.8 microns infra-red channel MSG Imager
 units: K
 scale_factor: 0.01
 add_offset: 273.15
 valid_min: -27315
 valid_max: 32767
 _FillValue: -32768
 available: yes
 comment: channel 9
 nuc: 931.122
 alpha: 0.9983
 beta: 0.6256
 CalibrationParametersOrigin: rttov
 _CoordinateSystems: GeosCoordinateSystem
 grid_mapping: GeosCoordinateSystem
 resolution: 3.0
 Offset : radiance = offset + CN x slope: -10.456819486590666
 Slope : radiance = offset + CN x slope: 0.2050356762076601
unlimited dimensions: 
current shape = (3712, 3712)
filling on
print(IR.dimensions)
('ny', 'nx')

fisrt question The dimensions (nx,ny) is the row and column? or the numerical value corresponds to coordinates x;y ?

What I would like to understand from the information on the variable IR_108 how I can have the numerical values in this channel and also the geographical coordinates X Y or lat and lon.

The idea that I have points x, y ( 2150,525) which corresponds to the line and column and I will have to extract from these x y the numerical value of the variable in this pixel. For example, the pixel 2150 row 525 corresponds to the value IR_108 = 4.7.

Answer 1

All the X,Y coordinates are stored in arrays and can be extracted by:

x=f.variables['X'][:]
y=f.variables['Y'][:]

Let's say we need to extract the value of IR_108 at a specific location (X=500000, Y=250000). Get the array indices for the data point nearest this position:

idx=(np.abs(x-500000)).argmin()
idy=(np.abs(y-250000)).argmin()

The grid coords for these indices may be double-checked by getting the values of X by x[idx] and Y by y[idy]. Then the value of IR_108 at this location is:

val_IR_108 = f.variables['IR_108'][idy,idx]

Which is 281.56 on y,x indexing.

• Sarah Tohami: thank you for answers the value 281.56 is the value of IR but is a InfraRed value in K i need to convert value in C i think , for the lat and lon how i can get the value in the same position?
• JimT: For a lat/lon you could use pyproj to convert from lat/lon into the grid coordinates and then do the same procedure. You will need to setup pyproj with the projection the data is in.