Do you need to quantify the amount of information lost due to satellite footprints or model grid boxed or unresolved inputs/processes in retrieval algorithms/model schemes with respect to a reference? Use my package called SpaTial Representation Error EstimaTor (STREET):

shortcut: pip install git+https://github.com/ahsouri/STREET

https://github.com/ahsouri/STREET

For many different applications including weather forecast modeling, or satellite retreivals, we are required to provide digital elevation maps at high resolution spatial resolution covering the whole globe. My toolbox will automatically download ALOS DEM at 30 m and save the subset of the data, depending on the boundary box defined by the user, into geotiff and netcdf formats.

shortcut: pip install demalos

https://github.com/ahsouri/demalos

It is sometimes useful to visualize the #GOES16/17 RGB imagery over your region/time of interest without having to pre-download the files. This simple toolbox will do the job.

To download the toolbox and see the examples, please visit: https://github.com/ahsouri/GOESVisualizer

shortcut: pip install GOESVisualizer

Can you guess what the left panel is showing? It's XCH4 retrieved from the synthetic MethaneSAT data over Barnett. The plume is completely buried by noise despite SNR~200 as XCH4 is extremely sensitive to the radiance accuracy. The right panel uses a wavelet to denoise it. The same method can be applied to other noisy images (such as CHOCHO, HCHO, and SO₂)

To download the toolbox and see the examples, please visit: https://github.com/ahsouri/MSATwtdenoiser

shortcut: pip install MSATwtdenoiser