When two cells of fluid exchange mass, the temperature is occasionally applied to the wrong cell. If they're also out of thermal equilibrium this obviously creates or destroys heat. For example by rapid-tapping space I observed the following single-tick transition:
Left tile: 1013kg@1000K
Right tile: 1.013kg@500K
Left tile: firstname.lastname@example.orgK
Right tile: email@example.comK
The mass change is obviously 10% of the delta, and both temperatures are legitimate too; but they've been swapped. IMO this is the cause of
though I haven't been able to pin down the frequency (or 2d heat exchange theory) enough to be certain it accounts for exactly the right amount of missing heat.
Steps to Reproduce
1. Surround 2 adjacent tiles with neutronium to isolate them thermally. 2. Put 1000kg steam at 1000K in one, 1kg steam at 375K in the other. 3. Run for a tick. 4. Repeat 1-3 until the temperatures get swapped.
This issue has not been confirmed by a developer yet.