I am trying to predict land use to the future based on the past changes and factors influencing the changes. Now, I have 2 transition periods: 1989-2001 and 2001-2011 (I will use only one for modeling). Both transition matrices are quite different since the rate of change of the classes were different (also the underlying reasons). In this case how should I decide which transition period should I choose for the modeling? I think even if I choose 1989-2011 (as a whole) I need a reason why I chose that. Any kind of help is really appreciated.