Grain nitrogen concentrogen

[byzheng]

working on #3605

Test method to calculate minimum nitrogen concentration of grain in #3606. Some improvement is observed for Australian wheat trials. Test for other crops.

[BrianCollinss]

This might be irrelevant to this issue. Currently accepted stats are indicators of model performance using a combination of parameters (some being genotypic, e.g., phenology) that were chosen and/or calibrated based on some observations. When there is a major change (like this) in the way the model handles biomass and N flow and deposition, don't we need to recalibrate the other parts of the models before we decide that any improvement observed is not just the model behaving randomly? In other words, if for example biomass was simulated reasonably well before a change, and now we are making this change, should we also make some changes in other parameters that govern biomass components?