Waves shouldn't have any significant influence, as all we see in the video are some small ripples and the light is mounted in fairly high above them. My money would be on, as you put it, "negative refraction". The listed temperature for the first clip 21 degrees, presumably Fahrenheit. Water freezes at 32 degrees, so because the pond isn't frozen, the water temperature must be greater than that of the air, and by a pretty significant amount, as corroborated by the strong inferior mirage seen at the opposing shoreline:
The red line in my insertion to mark the approximate location of where the inferior mirage starts. Later on, the second test is conducted, with the air temperature stated to be 38 degrees Fahrenheit. Note that there is very little obvious inferior mirage. It does look at time like there may be some, but I wonder if it's just from the light reflecting off the water:
Without actually knowing the water temperature, it's hard to make a definitive determination of what's going on. Based on the miraging that we see, and knowing that, generally, light deflects toward the colder medium, I would say it's just a bit of refraction in the upward direction.
As a final point, I'd juxtapose this with the following:

Source:
Source: https://www.youtube.com/watch?v=5ibkXsQoWd4
The fellow who took the video has done a lot of observations of the Turning Torso skyscraper in Sweden, and is very meticulous in specifying the details of his observations. For this one, he lists both the air and water temperature as being between 2-4 degrees Celsius, so in very close agreement. His observation point is also significantly above the water at nearly 4 meters, which reduces the possibility of significant thermal inversions directly over the water's surface.