Create a non cold-start FATES test

[rgknox]

FATES currently has a method of initializing plant structure via reading in from a text based inventory file.

Here is documentation on how it is done:
https://github.com/NGEET/fates/wiki/Model-Initialization-Modes#Guide_for_a_Cold_Start_Simulation_with_Inventory_Case_3

We have mostly used this method to intiailize simulations at model testbed sites, like Barro Colorado Island Panama, and Sierra Nevada US.

However, this method could also be used to store stand structure data at the end of a simulation (gridded), which could be used to initialize another simulation, along with a finidat file (if desired, which it would in this case).

The proposal is that we create one of these inventory datasets, on a very coarse grid (f10 x f10), and use this to initialize FATE vegetation for some tests.

Pros: FATES tests need to be STRESS tested, and without several hundred years of simulation behind it, we are only testing saplings, not fully developed canopies. This would go towards rectifying that.

Cons: The files are TEXT. Not ideal, but a worthy stop-gap to be used until we save this stuff as a self-describing format.

[samsrabin]

Blocks #2722.