There is no succinct answer, that's the only issue in this "debate".

We can choose to consider the peak ground acceleration as G's applied uniformly over the height vertical vessel, however this does not correspond with vessel behavior during a seismic event.

We can enter deeply in details and conclude that a building code offers us more information about the actual behavior of vessel (and consequently give us realistic estimation about the horizontal seismic force and overturning seismic moment) but we need to face the issue of an appropriate R which is undocumented yet. That was my original warning in the thread.