I think you're right in that encounter design is "the" big way to make PC flight a non-issue, but I think that might be part of the problem: forcing DMs to design encounters with flight in mind makes flight problematic, since it puts a limit on the DM's ability to do whatever they want.
DM's wanting to do whatever they want is the big problem, and not flight. This is easily demonstrated by the fact that so many of the problems encountered at the gaming table are DMs who believe the role entails them to do what they want, have the story they want, and to have as a payout the actualization of the thing they formerly imagined. One of the greatest temptations as a DM is to rehearse the encounter in your mind how you think it will go, always of course enlarging the excitement of the encounter and the glory that is due you from the players. IMO, this is the root of most of the table disfunction that results from the DM side of the screen.
As a DM, I must be on gaurd against wanting the T-Rex to be a big fight. I must be on gaurd against thinking that a T-Rex is the only thing that makes sense in this context. In short, I must be on gaurd against thinking that I'm the only source of material in the story and that I am the sole author and contributer and therefore must have my way.
No normal T-Rex (or terrasque) is going to worry a flying character too much.
Bah. You are not thinking this through.
First of all, no encounter which is optional - that is where the player has no stake in continuing or winning it - worries a character with the ability to flee. Flying gives you in many cases the ability to flee if for no other reason than it usually about doubles your movement rate. However, assuming that the T-Rex must be defeated, there are numerous ways to make the T-Rex relevant even against a party where everyone has wings.
a) Thick Foilage: If the T-Rex's lair is a dense jungle where a canopy of leaves and branches provides a partial ceiling and underbrush limits vision to 10-15', the flying character will have limited or no ability to take advantage of its mobility. If it flies above the T-Rex the jungle canopy provides 100% concealment and significant cover. It cannot fly among the canopy without colliding with the many obstacles. If it flies below the canopy, then either its within range of the snapping jaws or else neither can be seen nor can see its foe to utilize ranged attacks. And if it lacks less than excellent manueverability, the foilage will make engaging while flying very difficult. Close range against a feathered tiger striped T-Rex (very likely the 'normal' sort given what we know now) in a gaudy dense jungle is I think appropriate enough to not be considered pure metagaming.
b) Cave or Dungeon: The usual combat environment is enclosed, so that the ceiling might not be very far above the T-Rex's massive head. There is no where for a flier to hide or to evade to. Note that if you want to do this in a wildnerness setting, simply put the lair in the ruins of a megalithic temple. Sure, there may be some gaps allowing in sunlight, but they are likely to be too small to fly through. If the PC's flee the temple complex, then have it surrounded by terrain as #1 above.
c) Protect the innocent: One way to force a player to go into close combat is force the character to choose between evading and loosing something that they may prize. If the T-Rex is about to scarf down a maiden sacrifice, the PC's may have no choice but to interpose themselves between the potential victim and the threat, and this means quite certainly not flying up above the T-Rex's reach but rather going into base to base and putting some threat zones on the monster. It need not only be an innocent in distress. A treasure which is about to be devoured by a fire elemental/rust monster/disenchanter/black pudding/etc. should work out as well.
Again, encounter design. Good encounter design not only makes the encounter more tactical in nature, it also makes it less redundant and more emmersive. This increases the emotional payout for winning the scenario. A T-Rex in an effective 60'x80' flat square space (whether explicitly dressed as such or not) is not very interesting whether or not the PC's can fly. So don't design that encounter even if you've taken flying out of the mix, and indeed expecially if you've taken out the the 'this is boring, I win' button.
And I want to point out that this is limiting, but its not restraining. If you are shooting a movie and you wish to have the T-Rex be a threat to the good guys, you are going to do this same sort of encounter design for similar reasons. Sure, you could in theory put the T-Rex into a shot on a broad flat plain where he can be seen at a great distance quite clearly, but if you do this then it won't be this T-Rex at this time which is going to be a threat to your movies protagonists. Rather, this will be an establishing shot where we hang Chekov's T-Rex on he wall and say to the audience, "Keep this in mind." The good movie maker is going to have that T-Rex jump out and attack the party when its in a claustrophobic situation and when the protagonist is unready and unable or unwilling to escape. He does all this for the same reasons that the good game master does.
In designing a game for a wide audience, it would seem prudent to avoid anything that forced the DM's hand. Flight would certainly fit in that category.
It doesn't work that way. There is no substitute for good DMing. Period. The necessities of good DMing are the only things that force the DM's hand. Everything else is negotiable. To say that you need a game that avoids the necessity of good encounter design is like saying you want a language that avoids the necessity of good writing, or a film medium that avoids the necessity of well shot visuals. The game is the encounter design; it is the heart and soul of the game. This is most obvious in games of the normal sort that only provide a single scenario, rather than in RPG's which are more like scenario toolkits, but its just as true of RPGs. You can make the toolkit as high of quality as you like, and that's great, but without craftsmanship the game will still suck. Taking away flight doesn't reduce the necessity of craftsmanship; it merely reduces the number of tools in the toolbox.