Yeah, it's quite a stretch to call it water, an ocean, a reservoir...
But they sound good as a headline.
I was just going to say that. Idiot journalists keep running with that line, but there's no ocean. The "oceans" of "water" are stored dissolved within the molecular matrix of the ringwoodite in the form of hydroxide. It's also 410-660km below the surface, while the deepest borehole ever dug was 12.3km.
The article title is definitely misleading (it's an "ocean" as much as oil locked up in oil sands is a "sea") but the volume of water they suspect is down there is what intrigues me. Not necessarily for us to harvest anytime soon, since even digging down oil wells more than 8 miles hasn't been done, let alone going below the crust.
What interests me more is "dehydration melting" - a phase used in the paper I had not heard until now. If I'm understanding the phenomenon correctly, when magma in the mantle comes into contact with water trapped within rocks like ringwoodite it heats it up, turning it into steam and melting mantle rock, and potentially causing significant seismic activity (and when it occurs closer to the surface, volcanic eruptions).
Even if that water isn't accessible to us it's amazing that most of the H20 on the planet is basically trapped at that depth below the crust.