It isn't guaranteed. The internal variability will be a result of the physics, dynamics and chemistry within the model which differs between compsets. fwhist and fwschist differ in their representation of chemistry (interactive of specified), so if interactive chemistry e.g., ozone variability, affects the variability of your field of interest, you shouldn't expect them to be the same. So, it depends on what compsets you're considering and what region you're interested in. To leading order, I think you should expect the variability to be similar in most regards, but if you're looking in a region where the differences between the compsets may matter, they may not be.
