It's usually modeled as a fairly clean break. Above it the water is less dense because of thermal expansion and (in some places) salinity differences. Below it the water is more dense. Mixing is relatively rare because winds and tides and the like seldom get down the 100 m or so necessary. Even hurricanes don't necessary cause much mixing.
You get mixing in the polar regions, where the temperature difference makes mixing between the layers easy and where ice formation tends to give the surface water higher salinity.
You get it along the equator, esp. in the Pacific, where the winds usually push a lump of water up into the SE Asian basin and the Equatorial Countercurrent has it flow east. The result is that the S. Pacific gyre typically causes the deep, cold water to get closer to the surface, close enough for wind to create some mixing. There's also and Eq. CC in the Indian ocean and in the Atlantic, but they're weaker.
The El nino usually results when winds allow the Eq. CC to bring a lot of warm water to the S. Am. west coast.
I wish the pop sci article had actually said where the mixing was occuring, the mechanisms, and where the biggest delta Ts were. But that would be information.
Egads. I actually learned something when I read Thurman's book? Argh.
= new reply since forum marked as read