Negative N pools (BiomeE)

[marcadella]

On master using either p-model or gs leuning, some N pools are negative. It was true before the big master update from last week.

• N_yrMin < 0from day one, because:
  • vegn%annualN < 0 in SOMdecomposition() because:
  • N_loss is not compensated by the other fluxes and we start at 0.
• soilN < 0 after a couple of days because:
  • vegn%psoil_sl%n%n14 < 0 because of the correction done in Recover_N_balance().
• annualfixedN = 0 since it is only updated by vegn%annualfixedN = vegn%annualfixedN + cc%annualfixedN * cc%nindivs, but since it is initialized to 0, it stays at 0.
• Same for n_deadtrees, c_deadtrees, and m_turnover are all zero because their update happens after annual_diagnostics(). N_P2S_yr is probably only partially filled too.
• The daily and annual N pools and fluxes are scaled by a factor 1000, which is not consistent with the units (and not consistent with the C pools and fluxes).
• With do_closedN_run = True, the C:N ratio in soils is > 300...

[marcadella]

@stineb Negative psoil_sl%n%n14 comes from Recover_N_balance() where you correct if vegn%totN != vegn%initialN0. You seemed to assumed that the discrepancy came from vegn%psoil_sl%n%n14 exclusively but it might be that other pools have issues too, leading to an over-correction of psoil_sl. I am afraid the only way to fix the issue is therefore to fix the underlying model so that the N pools are correct in the first place.

[marcadella]

A simple alternative quick fix would be to spread evenly the deltaN over all the pools (proportionally to the pool sizes). Still a hack, but at least it would preserve positive values in the pools.

[marcadella]

Solved by using the solution proposed above.