“…Our analysis is, to our knowledge, more rigorous and thorough than previous works about coarsening in space. In [6,8,12,19], relatively accurate convergence estimates are provided, but they are based on local Fourier analysis (LFA), and, therefore, heuristic for parabolic problems, as this approach is by nature unable to track rigorously nonnormality effects. Moreover, because of the many parameters involved, analytical studies are intractable.…”