1: \begin{abstract}

2: We study the limiting behavior of a random dynamic system driven by 

3: a stochastic chain. Our main interest is in the chains that are not necessarily

4: ergodic but rather decomposable into ergodic classes. To investigate 

5: the conditions under which the ergodic classes of a model can be identified, 

6: we introduce and study an $\ell_1$-approximation and infinite flow graph of 

7: the model. We show that the $\ell_1$-approximations of random chains preserve 

8: certain limiting behavior. Using the $\ell_1$-approximations, we show how 

9: the connectivity of the infinite flow graph is related to the structure of 

10: the ergodic groups of the model. Our main result of this paper provides 

11: conditions under which the ergodicity groups of the model can be identified

12: by considering the connected components in the infinite flow graph. We provide 

13: two applications of our main result to random networks, namely broadcast 

14: over time-varying networks and networks with random link failure. 

15: \end{abstract}

16: