The study was done twice, over 5 days in June '11 and 5 days in Aug. '11. Same conditions: contol, just glucose, just oil, just dispersant, oil + dispersant.
In many ways, the glucose is a better control, but they did well to have both. They looked at carbon flow over 5 days, temps are surface temps but the samples weren't free to freely mix with fresh water. They were also held at constant volume so a little new water was added each day to compensate for evaporation.
Prokaryotes liked the Corexit and mixed disp./oil in August. They were indifferent in June. They didn't much like the oil by itself, but they weren't zipped over the control. That makes their increase with the dispersant and dispersant/oil mixtures a bit more striking.
The diatoms were just pissed off. They dropped in numbers in pretty much every condition, both runs.
The dinoflagellates were where the action was. In all cases their numbers eventually plummetted. But in June the control, oil, and glucose conditions they spiked up just after the start of the test, while in the two conditions with dispersant they just dropped. I don't have an explanation for the rise in the control in June. In August, the dinoflagellate numbers just fell.
Ciliates were a big deal. The numbers look like noise. They were statistically significant, however, and are a large part of the findings. More often than not they increased in the control, glucose and oil-only samples. Small numbers, spotty. If you add the June and August numbers there's sort of a pattern that's absent in either. I think something else is going on, but don't know what it is and don't really care. I don't trust the stats, but there they are.
The "heterotrophic nanoflagellates" (think 'mini animals') in June did nothing. Small numbers, about as likely to increase with dispersant as decrease. In August they started high in the pre-test samples but then plummetted. Disp., glucose and control look roughly the same. They didn't like the disp. + oil mix, but liked the oil before dropping off to fairly high levels.
The prokaryotes liked the dispersant and oil, the paper says. But the authors are right--they might have been feeding on the other creatures' corpses. Hard to know for sure without isotope tagging them, but it's very likely.
One problem though, is the difference in just June and August. Their major findings deal with plankton that showed striking differences between the two datasets. Overall their numbers stand as an average, but they didn't have a good reason for the difference. I assume it's annual variation that everybody already knows about, but it matters, esp. since their starting samples were drawn from the water that their data is being used to make predictions about.
Another problem are the interplay of other nutrients. The dispersant and oil and varying concentrations of species altered the N and P levels. No clue what that would do, long term, esp. since in the real world that varies seasonally and also there's more mixing. The oil wasn't spread evenly, nor was the Corexit.
A third is the relative lack of light. They kept the water oxygenated and prevented statification. Good things. A lot of dinoflagellates photosynthesize. The tanks were shaded and to avoid introducing yet another variable bright lights were decided against. (Under the oil slick it's dark; light would penetrate other samples--best to exclude it.) The decline in dinoflagellates makes sense, but in the analysis this isn't a factor. This makes the effects of the other factors larger.
I wish ed research was 1/10th as rigorous (that would be orders of magnitude better than what we have). Still, a fourth problem is they can't say what the problem with the dispersant-containing samples is, on two grounds. One is an internal problem that I really don't much care about--their results don't match other results (different results but with both different dispersants and different temperatures, which is responsible). The serious issue is deciding whether the dispersant is a chemical or physical problem. Rerunning just the dispersant condition with two other samples, each containing droplets of something thought to be innocuous would clear it up.
They're dead to rights. If this pattern continues, it wouldn't lead to the oil's energy moving into the food chain. It would lead to a decrease in energy flow and the only way for the carbon to reach the surface would be through respiration. Or it could be sequestered on the Gulf floor. (That's not a bad outcome, to be honest.) But they were right in stopping at that. Did the pattern continue? (At this point it's not something that we need to set up an experiment about: It's just something we have to measure. They could have compared the microorganism level in their t = 0 samples with historical norms to answer that.)
Then again, the use that society makes of these kinds of studies is the really huge problem. Some scientists like having their studies wildly extrapolated because they like the use of the wild extarpolations without having to make them themselves. But this is essentially one drop of knowledge and we need at least a large lake: Is it the dispersion of the oil that's the problem? Chemicals in the dispersant? How could concentration of dispersant or droplets affect the results? Changes over weeks or months? Would added light matter? How does temperature matter? Starting conditions? Concentrations of other nutrients? Ignorance breeds fear.
= new reply since forum marked as read
and welcome and thank you for your informative post! 
