I have a composite of SAR (28 images in a time series) data and I am trying to get the values of each pixel for the whole time series. I am looking at the coefficient of variation and trying to display pixel values as the time series progresses. I was able to get the coefficient of variation from the time series but it only shows me a static value. I was reading that possibly using multidimensional files would allow me to slice the data and drill down the pixels across the time series.

Can anyone give me any guidance?


You need to create a NetCDF file which can store multiple dimensions and parameters. ArcGIS and QGIS can display a NetCDF like a multi-band raster and there are plugins available to view using time.

You can generate NetCDF using command-line and GDAL: See here or the python library netcdf4-python

To interact with the NetCDF, use the python library Xarray. A useful tutorial is here: Multidimensional Arrays which shows you how to create a simple function to return data for a single location across multiple bands.

  • Hello James, thanks for the feedback. I've converted all of the GeoTiff's into netCDF format, but they only have lat/long dimensions. The time of acquisition is in the title of the GeoTiff. I downloaded anaconda and installed the above mentioned libraries, but I still do not know how to add a time dimension to an existing netCDF file. Do you know how to do this? Feb 19 '19 at 20:01
  • It is rather complicated but the answer to the below question is what you want. You will have to go through each step in the script and change based on your files e.g. how the date is formatted in the name, the projection etc: gis.stackexchange.com/questions/70458/… Feb 20 '19 at 9:59

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.