1: \begin{abstract}

2:   We present a new perspective on the use of weighted essentially nonoscillatory (WENO) reconstructions in high-order methods for scalar hyperbolic conservation laws.

3:   The main focus of this work is on nonlinear stabilization of

4:   continuous Galerkin (CG) approximations. The

5:   proposed methodology also provides an interesting alternative to

6:   WENO-based limiters for  discontinuous Galerkin (DG) methods.

7:   Unlike Runge--Kutta DG schemes that

8: overwrite finite element solutions with WENO reconstructions, our approach uses a reconstruction-based smoothness

9: sensor to blend the numerical viscosity

10: operators of high- and low-order stabilization terms.

11: The so-defined  WENO approximation introduces low-order nonlinear

12: diffusion in the vicinity of shocks, while preserving the high-order accuracy

13: of a linearly stable baseline discretization in regions

14: where the exact solution is sufficiently smooth. The underlying reconstruction

15: procedure performs Hermite interpolation on stencils consisting of a

16: mesh cell and its neighbors. The

17: amount of numerical dissipation depends on the relative differences between partial

18: derivatives of reconstructed 

19: candidate polynomials and those of the underlying finite element approximation.

20: All derivatives are taken

21: into account by the employed smoothness sensor. To assess the accuracy

22: of our CG-WENO scheme, we derive error estimates and perform

23: numerical experiments. In particular, we prove that the consistency

24: error of the nonlinear stabilization is of the order $p+1/2$,

25: where $p$ is the polynomial degree. This estimate is optimal for general

26: meshes. For uniform meshes and smooth

27: exact solutions, the experimentally observed rate of convergence is as high

28: as $p+1$.

29: 

30: \end{abstract}

31: