Stokes drift in momentum equation

[amrapallig]

Estimation of mean velocities (uvel and vvel) needs the Stokes drift term as per the Craik-Leibovich momentum equation.

@BrodiePearson and I discussed that until MPAS is coupled with the wave, the following changes could serve as a temporary fix.

             uvel(k,iCell) = uvel(k,iCell) - dt_small*(uw(i1,k,iCell) - uw(i1,k+1,iCell)) /  &
               (ze(k,iCell) - ze(k+1,iCell)) + dt_small*fCell(iCell)*vStokes(k,iCell)
            vvel(k,iCell) = vvel(k,iCell) - dt_small*(vw(i1,k,iCell) - vw(i1,k+1,iCell)) /  &
               (ze(k,iCell) - ze(k+1,iCell)) - dt_small*fCell(iCell)*uStokes(k,iCell)

@vanroekel, do you suggest adding Stokes term in mpas_ocn_adcReconstruct.F or outside of the module?

With the above correction (red dashed), ADC profiles for V and vw seem qualitatively correct compared to LES (black solid). Blue dashed line is the one discussed in Issue #26 (comment) with vwtend6. config parameters are the same as in issue #26 (comment). Following plot corresponds to 96th-hour output with an average of 24 hr. ADC is still showing overmixing.