This may not explain the issue entirely. The Polynesians did not have big ships, or very tall ships, at most 20-odd feet above the water, and they had no optical aid to view distant objects. They could have used mountain tops to detect earth curvature, but most Pacific islands are very flat. My guess is they wouldn't have noticed the "hull up/hull down" phenomenon unless they were already looking for it. I know I've never been able to see it without binoculars.

But they must have noticed the roll of the celestial sphere as they traveled north and south. We know they made detailed star charts for navigational purposes, and that they logged which stars passed over different islands in order to determine their latitude, so they certainly noticed how a star's max angular distance above the horizon changed as they went north and south--they used that effect to navigate.

Their most ambitious known voyage was from New Zealand to Hawaii, and we know they did the round trip repeatedly, they must have crossed the equator numerous times, so we know they understood the concept of "celestial equator". They would see stars in one hemisphere that were totally invisible from the other, and they would have noticed there were two celestial poles, one in the north and one in the southern hemisphere, 180 degrees apart.

The only earth/sky geometry which would give this effect would be a round planet surrounded by a spherical sky. The only observation which could contradict this is the erroneous "common sense" idea that the "down" direction is the same everywhere, a reasonable but incorrect assumption. It would take a great mental leap for them to arrive at this insight, but I think these guys could have handled it.

And like I said, steering a midwatch under a starry sky, the Milky Way and a million stars above in a totally black sky, and the geometry of your star chart memorized after a lifetime at sea--yeah. I'm convinced at least some of them figured it out. They were very heavy dudes.