metab.bookkeep currently doesn't use both midnights, and that some might interpret that as a better representation. But to do that, we may want to change metab.bookkeep, the metab code, and update documentation to make it clear.
There are concerns about NEP != GPP + R; should we revisit midnight to midnight input data and / or calculating GPP differently than it is calculated now?
metab.bookkeep currently doesn't use both midnights, and that some might interpret that as a better representation. But to do that, we may want to change metab.bookkeep, the metab code, and update documentation to make it clear.
Extensive discussion on the issue on PR #121