I just read your paper about the accuracy of various gradient reconstruction methods. Actually, we've written a similar paper which you can find in the link below (it's on ResearchGate). It focuses on a variant of the Green-Gauss gradient that is very popular, where the face centre values are obtained with linear interpolation from the values at the cell centres on either side of the face. Despite its popularity (it is the default scheme in OpenFOAM, for example, and one of three choices in Fluent) our study shows that it is zeroth-order accurate on unstructured grids.

Unfortunately, when we tried to publish it in J. Comp. Phys. and Computers & Fluids it was rejected by both because the reviewers claimed that the method is second-order accurate. However, our study includes a mathematical analysis and numerical tests with two different codes, one of which is OpenFOAM, so the possibility of a false conclusion is quite small. So, I wanted to ask the following: In your paper (which we were not aware of) you discuss other variants of the Green-Gauss gradient, which are at least first-order accurate on unstructured grids, but you do not discuss the variant that we refer to in our paper, despite its popularity. Did you deliberately omit it because you knew or suspected of its inconsistency?

Thanks a lot!

Alexandros

Article A critical analysis of some popular methods for the discreti...