It's very cool, (literally) that the study reveals this, but the cooling mechanism doesn't sound intuitive to me, it sounds arguable, and that the simulations might be just being tuned to suit measurements, but wouldn't accurately track changes. It's obviously a fiendishly difficult thing to try to simulate.
The photosynthesis process has to involve conversion of heat energy to things other than heat, and this has to be the essential mechanism of cooling. It's literally entropy reduction - creation.
(the opposite of what we do on the planet as a population, as stands, unpowered / unmonetised by solar)
The heat energy in the sun is known, but the efficiency of photosynthesis of the plankton, therefore reduction of temperature profile, isn't.
Take a volume of ocean with a known concentration of plankton, measured by a known volume and distribution of plankton sensors, and correlate with a known distribution of temperature sensors, to get a 3d volume distribution of heat vs plankton density, dependent on the known energy incident on the top surface of the volume, from the sun.
This might yield some useful models that could be used to replicate the effect under all conditions. It should also solve for a known value of plankton conversion efficiency.
But that might be an expensive study.
The irony is that it wouldn't be, if we were issuing the solar indexed stimulus necessary to convert our own net activity from temperature increase, to temperature decrease, there would be all the money and energy available needed to do these kinds of valuable studies.