The proportions don't make any difference. The only thing that you're calculating is the area described by the arc of a circle and the segment connecting its ends as a fraction of the total area of the circle. The depth of the cylinder is irrelevant in determining this.

Of course, you have to contend with the non-cylindrincal dissimilarities of both the tank and the jar. But we're not looking for precise measurement. Your method wouldn't provide that, either, as it's difficult to make the level of a tank the same after each filling. It also doesn't provide for arbitrary determination, which is certainly useful, if not actually required in this problem.