The 2008 dip is very real but pales in comparison to the annual rise and fall of CO2. That implies that vegetation is much more important than SST over the short run. That suggests fossil fuels are too (over the short run). Your sequester/release mechanism is viable over the long run but would require extremely rich-in-CO2 upwelling areas to be a factor in the short run. I do not recall reading anything about such areas in the present oceans, but I have not read a lot in that area.
Assume for a moment, the downwelling surface water temperature is 0 degrees C and the upwelling surface water temperature is 30 degrees C. By my rough calculations of the slopes of the absorption curve, if both SSTs are heated by a degree due to solar heating, then absorption drops by about 3.7%, while desorption increases by about 2.6%. If the downwelling sea surface temperature remains relatively constant because, as you say, it is a phenomena driven by the freezing of water and increased salinity, and if the geographical area where the downwelling resulting from this phenomena stays the same, then the amount of CO2 absorbed stays about the same, while the amount desorbed at the equator increases by about 2.6% if temperature there goes up by one degree C.
Downwelling might be caused by the fact that cold water is denser than warm water. I haven't compared salinity driven downwelling to temperature driven downwelling so I'm not positive which driving force is greater.
The point is that substantial changes in the difference between the amount of absorbed and desorbed CO2 can be caused by changes in the SST. Remember too that it takes hundreds of years, perhaps over a thousand years, for CO2 absorbed at the poles to be released at the equator. If so, temperature conditions at the poles a thousand years ago may determine how much CO2 is in the upwelling currents.
An article on this can be found here Link. It is not peer reviewed, so some caution may be needed. However, I have seen some poor papers in other areas of science get published because peer review on those papers was apparently poor or perhaps submitted to reviewers favorable to one side of a scientific controversy or perhaps to reviewers who do not have access to the data to rigorously check claims.
I suspect the annual cycles in CO2 correspond with the growing season or seasonal output of CO2. The great bulk of CO2 resides in the oceans, some 93% of it. The amount of difference between oceanic absorption and desorption is not insignificant, and the growth of CO2 in the atmosphere may well be controlled by it.
I'm going to bed. zzzz