When I say situational, I mean how often a good spot for the wonder is available. Desert areas are usually not that good.
Yes, I know what you mean. My point is that focusing on that particular situation is misguided, because there's nothing that makes a situational terrain type any more or less constraining than other situational game conditions.
Take the recently-departed Machu Picchu. You can think of it as being situational because it needs a mountain - easy right? But it's also situational because it needs a certain number of cities with profitable connections. Okay, getting harder, but surely that's just a matter of choice whether you play tall or wide? Except that now you're looking not just for the MP city spot, but also for suitable spots for every city you want to connect it to - whether you play tall or wide is itself going to be constrained by the landscape. Oh, and you need the city you actually build the thing in to have enough production to do so, as well as its mountain, or you need to be relying on another city to provide you with an engineer through choices you've made earlier in the game.
The same hidden situational concerns apply to Neuschwanstein or Statue of Liberty - sure you can build extra castles, but the Wonder's effects are tightly constrained by the same sorts of considerations about how many decent city spots you have access to, as well as the tech strategy adopted and already discussed. Statue of Liberty gets you +8 production per city, a figure suggested earlier on this thread? Sure, if you've got the food to sustain that, the money to maintain factories and windmills in non-production cities (allowing that everyone wants a workshop), no cities on hills (since you need the windmills), and enough coal for a factory in every city. etc. etc.
Just in terms of geography, look at the image I posted a couple of days ago again:
Then I was asking people to look at York. Now look at Nottingham, or just the landscape as a whole. Which is a less desirable location - York for Petra, or Nottingham or a hypothetical city a tile or two east of York or west of Nottingham, for closer access to the mountains (and hence Neuschwanstein or Machu Picchu)? There's very little useful production land anywhere near those mountains around Nottingham. If I sited Nottingham one tile west I'd lose Mt. Sinai, while if I sited York closer to the mountains I'd have lost the city's wheat. Neuschwanstein comes much too late to justify settling a suboptimal city I'd be keeping for the whole game, and a city placed there is never going to have the food to sustain a population high enough to make an observatory worthwhile.
Sure, that's a single anecdotal situation, but deserts really aren't as bad as often claimed and good sites are consistent enough that people base strategies around Petra. A desert hill is identical to any other hill, improved desert iron or stone with the appropriate supporting buildings have output equal to a mined hill, strategic resources and incense are somewhat common in desert, oases and floodplains can provide food - as, eventually (with Fertilizer), can wheat. You might well get one or two bare desert tiles that are no good for anything, but people actively settle near mountains which also have no tile output, and which quite often come in chains (finding a good production city spot next to an isolated counts-as-mountain NW is rather rarer than finding a good desert city spot).