He's still going to need to prove that he can start winning African-american support away from Hillary if he's going to win the whole thing even if he does win Nevada. He needs to be able to put a dent in that block and win parts over.
He needs to dent it (i.e., do better in South Carolina than expected), but he doesn't need to overcome that deficit entirely. South Carolina is demographically Clinton's third strongest state - it has the fourth highest proportion of minorities in the Dem caucus, but has enough older people to put it over Texas. If Clinton were to lose South Carolina, the race is over that day and everyone would know, literally no questions asked. She'd probably even concede the day after.
Demographically speaking, the bellwether primary states for the Democrats are Michigan and Florida (which was why they wanted to push forward in '08, famously). Heck, even winning Nevada would be a massive blow to Clinton, it's her 14th most favourable state demographically - hence why Sanders wouldn't even need to win there and still be considered threatening (although I think he can because Nevada doesn't actually match its demographics very well, it's surprisingly liberal and high-information).
A "genuinely threatening" Sanders looks something like this:
a) win Iowa (even narrowly)
b) double-digit win in New Hampshire
c) push Clinton very close in Nevada - not necessarily a win, but keep the margin below 6%
d) keep the gap in South Carolina inside 12% or so.
He'd be behind after Super Tuesday because generally speaking those states are more favourable than the average to Clinton, but not so far behind given those numbers. It would be competitive. If he actually swept Iowa-NH-Nevada, he'd be the mild favourite at that point, IMO.
Obviously very unlikely, I'm just saying what it would look like.
Superdelegates is irrelevant. If Sanders won the ordinary delegates, I cannot see the superdelegates possibly overruling it. That'd destroy the party, you'd have a mass sit-out come the presidential.