At the beginning of each land timestep, the water in the unconfined aquifer is reset to 5000 [mm], if the water table is below the lowest hydrologically-active soil layer (see BalanceCheckMod.F90#L99). Though, this incorrect resetting of water in the unconfined aquifer does not impact instantaneous water balance, but it does impact water balance that is temporally integrated across multiple timesteps. Analysis of land-only simulations reveals that this error in the land model only impacts urban landunits within a land grid cell. The urban landunits in the ne30np4 surface dataset constitute 0.53% of the active land surface.
@golaz, Here is the description of the bug in the land model related to water balance. Let me know if you need additional information on it for the paper.
At the beginning of each land timestep, the water in the unconfined aquifer is reset to 5000 [mm], if the water table is below the lowest hydrologically-active soil layer (see BalanceCheckMod.F90#L99). Though, this incorrect resetting of water in the unconfined aquifer does not impact instantaneous water balance, but it does impact water balance that is temporally integrated across multiple timesteps. Analysis of land-only simulations reveals that this error in the land model only impacts urban landunits within a land grid cell. The urban landunits in the ne30np4 surface dataset constitute 0.53% of the active land surface.