Just from the technical aspect of things, it's impossible to tell from this photo exactly where the horizon is. (The horizon is at one place, and that's exactly at eye level for a person who is looking straight ahead. Or at lens-level for a leveled camera. Since we don't know exactly how MR positioned the camera, we can't tell where the horizon is.) However, the horizon cannot be below the level of the water we see -- but it can be above it. All you have to do to know that is think about it. If you were at a high place, looking over water, then over an island, and then more water, which shows the true eye-level horizon, it's obvious that the far edge of the water in the foreground can't represent the horizon. If you lowered yourself a bit, so you could no longer see the actual eye-level horizon, the water below you would still (obviously) not constitute a true horizon. The true horizon is still at your eye-level, it's just out of sight.
It's equally true that the angle of the water and land makes a difference, which you can easily prove to yourself by taking an oblique photo of the line where a wall meets a floor, with a more or less level camera. The wall/floor line closer to you will be lower in the photo frame than the the wall-floor line further away. The same is true of land angled across an expanse of water, as it obviously is in this photo.
The problem with this photo is that the land angle is subtle enough that the line of the water could be mistaken for the true horizon. But it isn't. I think MR was either standing on shore, or standing up in a boat, and the true horizon is about at that line of portholes on the hull of the boat, or a hair lower.