I’ve written before about the probability and plausibility of various artistic reconstructions but this can extend beyond simple soft tissues and to the heart of behaviour and lifestyles. Dinosaurs were, in general, highly diverse and lived in many varied environments and different ecosystems over a long period of time. We are still digging into the diversity of dinosaurs and while ‘oddities’ like Pelicanimimus, Limusaurus and the alvarezsaurs as a whole have turned up, there are very likely more interesting and unusual things to come. Even taxa that are already described might turn out to be rather more odd than realised with further discoveries.
Now this is all a bit ephemeral and no you can’t really put numbers on these kinds of assessments but as an example, I think it far more likely than not that *one* of the known big theropods was a dedicated scavenger (or rather perhaps, much more of a scavenger than any other large theropod). I does make ecological sense that the odd carnivore was rather more jackal or hyena-like than its relatives and exploited a different nice. We might not ever get any good evidence for this from the fossil record. The putative anatomical specialisations might be confounded by preservation or evolutionary history. But I think it’ll be there.
From a scientific perspective therefore, I don’t think you can say much more than “based on what we know of living ecosystems, diversity and evolution, it’s likely that something out of the allosauroids, tyrannosaurs and abelisaurs was a behaviourally specialised scavenger”. You certainly can’t name one species over another as a likely candidate, and it’s extremely unlikely that all of a group had gone down this route. However, if you want to draw this kind of thing then I’d have no problem with that and I don’t think many people necessarily would, or should.
No, based on the available you can’t say “this is Carnotaurus the scavenger” but what’s wrong with “Carnotaurus may have been a scavenger”? In terms of producing interesting and thought provoking artwork, I think this should be encouraged. After all, if you stick to what is absolutely known or is very likely true for most dinosaurs there’s quite a limited set of things. The only real trick is to stick to probabilities – nothing wrong with say a largely featherless ornithomimid say, but if you only ever draw bald ones then you’ve probably gone too far. Exceptions to the rule occur, but exceptions shouldn’t be the rule.