exposing nash cascade storages to BMI for purposes of checkpointing/hotstarting

[fred-ogden]

Short description explaining the high-level reason for the new issue.

Current behavior

Nash cascade number and storages are not exposed through the BMI. Item identified by Dan Wright.

Expected behavior

For purposes of checkpointing/hotstarting, the number of Nash linear reservoirs in the cascade and the storage in each reservoir are required.

Steps to replicate behavior (include URLs)

Screenshots

[SnowHydrology]

@fred-ogden, @peckhams and @madMatchstick can give more details, but the full list of model variables is exposed via BMI in the owp-bmi-enhance branch. See the source code for bmi_cfe.c here, for example: https://github.com/NOAA-OWP/cfe/blob/owp-bmi-enhance/src/bmi_cfe.c

To perform checkpointing and hotstarting I believe we'll need more variables past the nash cascades that aren't currently available via BMI. We should start considering whether we'll want to use the serialization work that Scott and Jess have researched and performed, or find a way using BMI 2.0.