Mass budget accounting requires new calculations and modifications

[matthewhoffman]

Currently it is not possible to close the mass budget for grounded or floating ice. It is possible to get close with extensive post-processing calculations, but this accounting should happen online and be a standard output from all runs that closes BFB.

Some known issues are:

• GL mask isn’t really GL mask
• subglacial lakes need to be excluded from ice shelf
• non-dynamic floating fringe should be part of grounded mask
• GtoF migrations flux is necessary to close budget and GLflux needs to make use of appropriate GL position when the GL moves
• calculations may need to occur online to capture changes to mask during the timestep
• possibly we need special masks for budgets

cc: @trhille

[trhille]

Using https://github.com/trhille/Humboldt-scripts/blob/debug_budgets/recalculate_GL_flux_with_edgeVelocities.py at commit trhille/Humboldt-scripts@845637c we are able to get reasonable results for plotting, but fractional differences early in the run are still too high. Here the bottom panel compares the magenta and dashed blue curves in the top panel.

[trhille]

* GL mask isn’t really GL mask

Addressed in PR #37

[trhille]

• subglacial lakes need to be excluded from ice shelf

Adressed in PR #54