1: \begin{abstract}

2: \noindent 

3: We develop a general polynomial chaos (gPC) based stochastic Galerkin (SG) for

4: hyperbolic equations with random and singular coefficients. Due to the singular nature of the solution, the standard gPC-SG methods may suffer from a poor or even

5: non convergence. Taking advantage of the fact that the discrete solution,

6: by the central type  finite difference or finite volume approximations

7: in space and time for example, 

8: is smoother, we first discretize the equation by a smooth finite difference or

9: finite volume scheme, and then use the gPC-SG approximation to the discrete

10: system. The jump condition at the interface is treated using the immersed

11: upwind methods introduced in \cite{Jin:2009pro, Wen:2005ueba}.

12: This yields a method that converges with the spectral accuracy

13: for finite mesh size and time step.  We use a linear hyperbolic equation

14: with discontinuous and random coefficient, and the Liouville equation with

15: discontinuous and random potential, to illustrate our idea, with both one

16: and second order spatial discretizations. Spectral convergence is established 

17: for the first equation, and numerical examples for both equations show the

18: desired accuracy of the method.

19: 

20: \bigskip

21: 

22: \noindent{\bf Key words.} hyperbolic equation, random coefficient, potential barrier, stochastic Galerkin, polynomial chaos\\

23: {\bf AMS subject classifications.} 35L02, 65M06, 65M60, 65C30\\

24: \end{abstract}

25: