Its a little hard to know for sure how the seat belts will cope with these sort of excessive forces. They are designed to serve their purpose - to reduce injuries in a survivable crash, not to restrain bodies exposed to forces far greater. But I doubt that there were a lot of belt failures.
A high proportion of unrestrained pax midflight is common. But probably not if there was a gradual increase in turbulence. Moderate turbulence will wake people up, and pax will be advised to put their belts on. In my limited experience of turbulence, most pax are pretty nervous and quick to comply.
No Mayday. No life jackets. Lots of possible explanations. But a common explanation (and one explanation that fits is more likely than lots of different explanations) is sudden and catastrophic loss of control. This situation would also be expected to result in lots of unrestrained passengers - which I believe is what we have here.
The seat cushions when removed from the seat do provide limited flotation. I think they are a legal option on non-overwater flights???? But I doubt that a row of seats with bodies attached would float. I would imagine this would sink fairly quickly. Something that can easily be tested however. Certainly there were not many seats recovered, and there were plenty of empty seats on that flight. A floating seat would be a much easier target to spot than a semi-submerged body, so I suspect they sank. Very limited positive buoyancy of some bodies will not keep the seats afloat.
Bodies themselves vary. Some float, some sink - depending mainly on fat content, whether the lungs filled with water or not, and clothing worn. A person dying at impact is more likely to float than a person who survived impact and subsequently drowned with water filled lungs. Sometimes a sunken body will refloat - but not always. Gas does form in the gastrointestinal tract, and this will create some positive buoyancy. But this gas formation is variable. Sometimes the body will belch (sorry) and this buoyancy will then be lost. The pressure of the gas will simply overcome the passive resistance of the esophageal sphincter. Sometimes the gas will be retained. It is actually a very fine balance - not surprising given we are mostly water, and the denser and less dense components pretty much cancel out. Try it swimming sometime - a few hundred ml more or less in your lungs affects whether you sink or float. Same with SCUBA - it doesn't take much extra weight belt or gas in your chamber to alter your buoyancy.
I can see wave action dislodging a body from a seat on the surface. Not after it has suck however - the seat and body will drift as one with the subsurface current.
On 22 October 1963 a BAC 1-11 G-ASHG entered a deep stall during a prototype test flight. The aircraft descended from FL180 to the ground over Cricklade, Wiltshire in 80 seconds - 13,500 fpm [133 knots]. Horizontal velocity at impact was effectively nil, as the aircraft belly-flopped and burned in-situ.
Thats 70m/sec. So if deformation allowed decelleration of cabin contents over 10 or so metres, figure a force around 25G. That sounds more likely than figures of 100G. Clearly still non-survivable, but perhaps more consistent with what we are seeing.