Link masked data by position

[Frederick Gent]

I have 3D MHD simulation data and want to track how the strength of the magnetic field varies with density or temperature across multiple snapshots. For any given snapshot how do I link the datasets by the location (pixel position, not common values) in order for example to compare the distribution of high magnetic energy across a temperature histogram? The data is in numpy 3D cubes. I can link the regions by selecting a block in a 2D map, but this is not what I require. I am looking to mask one dataset, based on the values of a different dataset.