Open pedrocol opened 1 year ago
There might be a problem with the existing formulation if we want to use it in ALE coordinates.
Adcroft et al., 2008: 'In general, the Montgomery form will tend to manifest a pressure gradient error when compressibility or the choice of vertical coordinate leads to along surface variations of alpha.' Being alpha = 1/rho - 1/rho_0.
The Montgomery potential is used as a correction of the pressure gradient formulation. As we can see, if the vertical coordinate is not isopycnal, the density difference in the same layer will lead to a variation of the pressure gradient. Or at least this is my understanding at the moment.
The idea is to evaluate potential problems with the existing pressure gradient formulation and propose ways of action