The lower formation may even be as old as 170 or so million years ago, whereas the latest part of the formation may have just barely squeeked into the Oxfordian. The lowest part may have even squeeked into the older Bajocian. The Mamenchisaurus Assemblage therefore probably is more like 160 million years ago, vs. 156-146 million years ago for the Morrison, depending on how deeply into it you go. Of course, none of this means that you can pinpoint precisely in what year a dinosaur fossil died and was buried, so sometimes you have to take the best guess and average in the range. If the Shunosaurus-Omeisaurus Assemblage is thus about 170 million years ago, and the Mamenchisaurus Assemblage is 160 million years ago, then the "Camarasaurus-Diplodocus Assemblage" of the Morrison is about 150 million years ago; although we can speak at least a little bit intelligently about the fact that that's the midpoint of a ten million year spread, and that we can say a few provisional things about how things changed over that ten million years. But still; the point is that the Mamenchisaurus Assemblage represents an earlier stage of the "Jurassic Aspect" fauna.
Now, we have to be a little bit careful. The map below of the Late Jurassic gives some indication of why the Tendaguru, Lourinha and Morrison would be similar; they hadn't been separated for very long or by very far. As sea levels surged throughout the Jurassic, they would have allowed faunal crossover of large animals to a wide degree, so it's not surprising to see that we see either the same genera, or at least very closely related genera within the same families at all three locations. China, on the other hand, you can see is quite far away. In the preceding Middle Jurassic periods, this connectedness would be even more marked. Also; Mamenchisaurus Assemblage animals are found in more recent formations than the Shanshaximiao, so some of the same animals may have actually been coterminous with the Morrison after all (at least the early part of it); but not from this formation specifically.
So when we see faunal differences, we have to remember that time and space both play a factor. The paper says this as well, but the only variable that it really analyzes or postulates as having a significant impact is time. That said, I think that they're on to something. Let me summarize their main lines of evidence.
- Freshwater fishes and bivalves found are consistent with a later Middle Jurassic age.
- Mamenchisaurus is abundant in the upper member, but completely absent in the lower.
- Short-necked cetiosaur grade sauropods like Shunosaurus are completely absent from the upper member, but abundant in the lower.
- Extreme rarity of any coelurosaurs in the lower member
- Relatively primitive megalosaur-grade therapods, getting progressively more advanced and replaced by carnosaurs as we move from lower to upper.
- Calibrated phylogenies show an earlier grade to the Mamenchisaurus assemblage than to the Morrison.
- Presence of primitive ornithischians like fabrosaurs in the lower member.
- Overall more primitive sauropod genera, lack of shared families with the three formations of Europe-North America-Africa.
- Overall lack of relative diversity (Mamenchisaurus in several species, vs. at least half a dozen genera of sauropod for the Morrison)
- Fewer mammals and small-bodied therapods
- No iguanodontid-grade ornithopods in China yet.
They note that overall there is less collection done, compared with the Morrison in particular, so more work can yet be put together to detail the differences. They also note that the Mamenchisaurus Assemblage seems to be coterminous in time with the fossil asssemblages associated with Chuanjiesaurus Assemblage from the more southerly Yunnan province, which raises interesting questions about localization between the Yunnan and Sichuan basins. Which, after all, aren't that far apart from each other.
Curiously, they also note that the Mamenchisaurus Assemblage carries forward into the late Jurassic in the Suining and Penglaizhen Formations, although those are much more poorly known and understood than the Upper Shazimiao. They include this nifty chart:
And that chart doesn't even have the Shishugou Formation (of Guanlong and Yinlong fame) in the Junggar basin which also has Mamenchisaurus fossils, the carnosaur Sinraptor and megalosaurs like Monolophosaurus. And what paleontological post would be complete with just charts and maps? Here's Yangchuanosaurus. Rawr!