Suppose in Python, I have read in a netCDF4 file, as follows:

import netCDF4

ds = netCDF4.Dataset('fire_weather_index_2018.nc')

The aspects of the netCDF4 is as follows:

<class 'netCDF4._netCDF4.Dataset'>
root group (NETCDF4 data model, file format HDF5):
    Conventions: CF-1.4
    created_by: R, packages ncdf4 and raster (version 3.0-7)
    date: 2019-10-16 00:07:39
    dimensions(sizes): Longitude(1440), Latitude(721), Time(365)
    variables(dimensions): int32 crs(), float64 Longitude(Longitude), float64 Latitude(Latitude), float64 Time(Time), float32 FWI(Time, Latitude, Longitude)

When trying to access a particular datapoint in the file (for example, for day 200, at the coordinates of (27, 47), though this error occurs for every datapoint), as follows:

ds['FWI'][200, 27, 47]

It says the data is masked, but when I query directly ds['FWI'], there are plenty of datapoints. In fact, I can map it accordingly on a world map and all the data would show up.

How do I access / extract a particular data point for the 'FWI' variable in this case?

  • 1
    When you say coordinate, do you mean latitude/longitude?
    – snowman2
    Nov 12 at 14:36
  • @snowman2 yes, that's right
    – NicTam
    Nov 12 at 14:55
  • This may help: gis.stackexchange.com/questions/358036/…
    – snowman2
    Nov 12 at 15:46
  • You might look in your Longitude(1440), Latitude(721) variables to see which index matches your coordinates. You may be getting a masked value at whatever coordinate happens to be stored at Latitude(27) Longitude(47) in those arrays.
    – Dave X
    Nov 15 at 19:09
  • Based on the 721 x 1440, the data array is probably 0.25° resolution and you are indexing 27,47 in it. ds['FWI'][200, 27, 47] could well be at 27/720*90-90=-86.625° 47/1440*180-180=-174.125°, signs depending on the ordering of the Latitude, longitude variables.
    – Dave X
    Nov 16 at 22:25

You are getting a masked value because you are ignoring a level of abstraction and treating your coordinates as indices. The dataset probably holds a masked result for [27,47] at 83.25°N, 11.75°E, which is well north of Svalbard.

You need to use the indices that match your coordinates:

import netCDF4
import numpy as np

ds = netCDF4.Dataset('fire_weather_index_2018.nc')


jj = np.argwhere(ds['Latitude'][:]==27)[0][0]  # first matching lat
ii = np.argwhere(ds['Longitude'][:]==47)[0][0] # first matching lon


ds['FWI'][200, jj, ii]

If the coordinates don't have an exact match (e.g. 26.3°N,47.3°E), you can look for the nearest ones:

jj = np.argmin((ds['Latitude'][:]-27.3)**2)
ii = np.argmin((ds['Longitude'][:]-47.3)**2)

print(ds['FWI'][200, jj, ii])

(I found some comparable data at https://zenodo.org/record/3539654#.YajbsvHMLMI )

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.