Spatial and modal superconvergence of the discontinuous Galerkin method for linear equations (Q2399155)

The authors employ the discontinuous Galerkin (DG) finite element method with a \(p\)-degree polynomial basis to the linear advection equation in order to derive a partial differential equation (PDE) which the numerical solution solves exactly. Polynomial solutions to this PDE are obtained which are losely related to the \(\frac{p}{p+1}\) Padé approximant of the exponential function. For a uniform mesh of \(N\) elements it is obtained \( (p+1)N\) such independent polynomial solutions of which \(N\) can be viewed as physical and \(Np\) as non-physical. It is further obtained that the accumulation error of the physical mode is of order \(2p+1\). These results are then used to investigate the superconvergence of the DG method on uniform grids and to establish a connection between spatial superconvergence and the superaccuracies in dissipation and dispersion errors of the scheme. Finally, it is obtained that for a class of initial projections on a uniform mesh, the superconvergent points of the numerical error tend exponentially quickly towards the downwind based Radau points.