N-pools and forest age classes

[luyssaert]

When ORCHIDEE is run with age classes and a the forest in age class exceed the threshold such that they can promote to the next age class, all state variables (pools, age, stress, ...) Need to be promoted as well. In ORCHIDEE-CAN this has been implemented for all C related and stand structure variables. While merging ORCHIDEE-CN and ORCHIDEE-CAN into ORCHIDEE-CN-CAN, the following N-related variables are also moved from one age class to another: cn_leaf_min_season, nstress_season, soil_n_min, p_O2 and bact.
It needs to confirmed that this is indeed sufficient. Testing this is rather simple, run the model for a sufficiently long period with age classes (use the run.def for 37 PFTs). Sufficiently long means that all MTCs have moved at least from age class 1 (start of the simulation) to a higher age class. Then run the model with the same settings but now use the run.def for the 13 MTCs. In the absence of land cover change, the results should be identical. Peaks, drops, wiggles, etc following a change in age classes are suspicuous and should be confirmed or fixed.

[luyssaert]

The code that deals with promoting the variables from one age class to another is age_class_distr in sapiens_lcchange.f90.

The same variables that were added to age_class_distr were also added to mortality_clean in stomate_kill.f90. When a forest dies in, for example, age class 3 it will be replanted in age class 1 (by definition). This implies that all essential nitrogen variables will need to be moved from, in this example, age class 3 to age 1. This is done in mortality_clean and has been extensively tested for C. Now that N has been added in this merge, it needs to be tested for nitrogen as well. Testing this is rather simple, run the model for a sufficiently long period with age classes (use the run.def for 37 PFTs). Sufficiently long means that the forests start to die. This can be speed-up by decreasing the diameter (in the run.def) at which managed forest will be clear cut. Then run the model with the same settings but now use the run.def for the 13 MTCs. In the absence of land cover change, the results should be identical. Peaks, drops, wiggles, etc following a change in age classes are suspicious and should be confirmed or fixed.

[luyssaert]

cn_leaf_min_season, nstress_season, soil_n_min, p_O2 and bact were added to both age_class_distr and mortality_clean. This fixed the initial problem with having soil_n_min in PFTs without veget_max. The tests above should be performed to confirm that this is a sufficient fix.

[aresovsky]

The newly added N-related variables have been checked to ensure that their values are moved to the next age class correctly when promoted.