On gradient calculation in flux correction method

Flux Correction method is a family of edge-based schemes for solving hyperbolic systems on unstructured meshes. The cruical operation there is a nodal gradient calculation of physical variables with at least second order of accuracy. There are two well-known procedures meeting this condition. One is based on Least Squares method and the other one is based on spectral elements. In this paper we compare resulting schemes and discuss their problems.