In PR #1217 we added diagnostics that track if the time integrated net mass flux through live vegetation and litter (respectively) matches the quantity in the states themselves. This bias between the integrated net flux and the state is written to history and can be tracked. In this PR @JessicaNeedham suggests that we consider having a graceful failure if this divergence gets too large.
This user switch would probably be in the fates parameter file, as we reserve switches in the namelist for turning on/off major modules.
The failure is in place and needs only be un-commented, see ChecksBalancesMod.F90:CheckIntegratedMassPools()