API 650 seismic "overturning" calculation assumes the development of two plastic hinges, one at the junction to the shell and the other at some distance inward from the shell. As a result of this model, a part of bottom and shell can be lifted off the ground and the weight of the content will resist overturning.
Details are given in a famous paper "Basis of seismic design provisions for welded steel oil storage tanks", authors Wozniak and Mitchell and In "Aboveground Storage Tanks" by Myers, chapter 6.1.5.
Based on this model, one can calculate the effective lift-off; see E.7.3.1 Method for Estimating Tank Uplift.
As you can see, API 650 does not consider a correlation of such value with Earthquake magnitude, vertical acceleration, etc, so I would consider E.7.3.1 gives the maximum tank response vertical displacement.
Returning to your question, the lift off is vertical (of course!) and theoretically this vertical deflection has nothing to do with nozzle's centerline elevation. I would add that the horizontal displacements should be increased to account for drift of the tank, and the drift depends on that elevation - see the case when the nozzle is top shell located.
As EC.7.3 PIPING FLEXIBILITY explains, the design displacements of Table E-8 "are a compromise of practical design considerations, economics and the probability that the piping connection will be at the point of maximum uplift. If one "estimated" the tank uplift using the simplified model in the appendix, the uplift will often exceed the values in Table E-8 unless the tank is in lower ground motion regions".