The often counter-intuitive behavior of dynamic systems requires profound training and experience to predict. Simulations can provide a practice arena in which users can experiment with different policies, and thus learn why certain behavior arises. However, common tools for creating such simulations provide no explanation of the exhibited behavior, and it is up to the user to reason about causes and effects. Without insight into the system dynamics field, such an analysis can be highly challenging to perform. This paper looks into how an explanation of system dynamic models can be dynamically generated in order to scaffold the users? comprehension of the system. The objective of the generated explanation is to enhance the users? understanding of the particular dynamic system, but also to stimulate their dynamical thinking in general.