Currently, lusconv files are static (there is only one).
The lusmatrix files can be dynamic (so, accounting for efficiency improvement, loss of fertility, policy implementation), which can result in changes to the competitive advantage of the land system types. For example: only particular land systems receive efficiency improvements (e.g. irrigated systems, or only traditional systems due to certain policy), which can result that these systems can provide more compared to other systems in later years, than in the initial years.
This can lead to issues in the model run, as the model still promotes systems that were defined as most suitable in the initial years, but are not most suitable in the later years.
Assigning different values to the lusconv file in the beginning does not do the same trick, as it promotes wrong systems at the beginning (when they haven't yet received efficiency improvements).
Originally from: Ziga Malek