It is nice that you guys are including as much information as possible, however your overthinking the process and getting off track.
Your thermal scan is inconclusive at this point.
As it appears you discovered this thermal exception after leaving the site and doing your report, it is what it is. It is now required that you go back and specifically address what you have found utilizing the proper techniques to analyze the situation that you suspect. The Preparation process for taking thermal imaging is more important than anything else in this application.
You have less than a 2° temperature differential in your scan and this must be increased. Also time is a factor. You need to shut down the system and let it stabilize to ambient conditions. Take a baseline scan which will indicate moisture during the cooling process of the floor. This is the same as doing a flat roof inspection. Temperatures remain for a longer period of time when water is involved versus normal conduction of the materials.
Then turn on the system and watch the thermal pattern and its rate of conductivity across the floor. If this is a water leak, the thermal pattern be more defined compared to what we are looking at in your scan right now.
Due to the density of these floors, conductive temperature rise will be slow and blurry regardless. This is where a real thermal camera is required.
Looking at what you have right now, I would be more inclined to call this a depth of piping/density changes in the flooring system than a fluid leak.
It is also recommended that you utilize isotherms to bring out temperature differentials that are smaller than the palette differential you are using. This pulls out temperature differentials of less than 1/10 of 1°C and provides a pattern of flow over time.