I made a top view. At this angle, the water mirror changes to a transparent medium with optical fractures of the bottom. I did a lot of trial experiments and found a more or less good option. I took a model of a gypsum wall panel in the form of waves.
Panel
I also gave it in the MTL d 0.3. Ni 2 file. I assigned a minimum thickness of 0.1 cm. If the relief is directed upwards, then there is a strong reflection of the sky, the bottom of the pool is almost invisible. I turned the model upside down. And everything has changed!! Now there is only optical refraction and there are no reflections of the sky at all. It looks very much like clear water with waves. Although, of course, in reality the water will be more difficult, with glare. This mode loads down relief for a very long time, much longer than the relief up with mirror reflections of the sky. Interesting.. But as a result, there is an illusion of water movement, the rosette circles do not lose their shapes, because they do not intersect with the shape of horizontal water. The Ni coefficient greatly changes the tile pattern under water, I gave a 3D view for comparison. All circles are 1mm above the top of the horizontal water relief, so the shapes of the circles are clean.
If the relief is turned up, it will reflect the sky with clouds mostly, the bottom is almost invisible., mainly reflections of the walls of the pool above the water. So for a view with clear water, you need to turn the model upside down.
