Assuming that the ice is floating at the beginning of the experiment, your analysis is incorrect: The water level will not change.
By Archimedes' Principle, a floating object displaces an amount of water equal in weight to itself. When ice melts, it becomes water. Therefore, the water into which the ice melts will exactly occupy the volume displaced by the unmelted ice.
(In practice, if we speak of ice floating on the sea, we should note that the melting ice will cause a slight reduction of the concentration of dissolved salts, and thus cause a small increase in total sea volume. But the effect in question would, as a practical matter, be insignificant).
I stand corrected, you are right. The only way to get a result in which the water level drops is to start with a container completely filled with ice. A trivial case and not at all what we were talking about. In any event, the level does not rise from the melting of floating ice.
Thanks for the correction, I posted in haste with out thinking it through.
Regards,
GtG