Review and update data product levels

[dlebauer]

Some of the level 1 data products are really Level 2.

Examples:

• the *_ind.nc files generated by the hyperspectral pipeline are currently found next to their parent *.nc files, but should be moved to the Level 2 data products.
• Level 1 fullfields have a lot of artifacts
  • thumbnails and csv files should be level 3 because they are spatially aggregated and not at the same resolution as the raw data.
  • on the other hand, as csv files that are otherwise in BETYdb shouldn't be exposed to end users. If we want to provide season level summaries of one or more traits in a csv file (like what we are providing to D3M), we should provide these as a single file by date under Level 3.

Creating this issue, and will add more examples as they come up.

[dlebauer]

@max-zilla bumping this as we move toward publication it will be helpful to get this straight, and at least make sure that the data products are easily discoverable