Provide guidance on what is expected for certain outputs

[emilyanndavis]

In a future iteration, it could be helpful to provide specific guidance to help users identify potential problems.

For example:

• Stats for a masked LULC output should match stats for the input LULC

[davemfish]

It would be cool to have an attribute like range in the model spec, to describe the range of possible values. Beyond offering guidance, we could possibly do the check ourselves and only raise a flag in the report when there is a problem.