No one is "perfectly" adapted. Life is a constant struggle for survival. The fish are also getting faster and better at evading the seals, a drop in temperature ushers in an ice age.
The females would go right on selecting against any male which tried to adapt outside the species boundaries. Or is some females velociraptor supposed to have said to herself:
"Say, you know, that Alvin over there sure looks weird with those wings and that beak, but that's sure gonna be useful for flying some day...
That was funny, thanks for the laugh medved. Seriously, evolution doesnt argue that beaks and wings just spontaneously appear out of nowhere in one generation. We are talking about "simple" adaptations and improvements thereupon - longer claws, faster arms, thicker skin etc.
As for flight... this page offers a simple hypothesis for how flight may have evolved in birds.
For starters, every one of these things would be antifunctional until the day on which the whole thing came together, so that the chances of evolving any of these things by any process resembling evolution (mutations plus selection) would amount to an infinitessimal, i.e. one divided by some gigantic number.
In probability theory, to compute the probability of two things happening at once, you multiply the probabilities together. That says that the likelihood of all these things ever happening, best case, is ten or twelve such infinitessimals multiplied together, i.e. a tenth or twelth-order infinitessimal. The whole history of the universe isn't long enough for that to happen once.
All of that was the best case. In real life, it's even worse than that. In real life, natural selection could not plausibly select for hoped-for functionality, which is what would be required in order to evolve flight feathers on something which could not fly apriori. In real life, all you'd ever get would some sort of a random walk around some starting point, rather than the unidircetional march towards a future requirement which evolution requires.
And the real killer, i.e. the thing which simply kills evolutionism dead, is the following consideration: In real life, assuming you were to somehow miraculously evolve the first feature you'd need to become a flying bird, then by the time another 10,000 generations rolled around and you evolved the second such reature, the first, having been disfunctional/antifunctional all the while, would have DE-EVOLVED and either disappeared altogether or become vestigial.
An idea of how hard it would truly be for "proto-bird" (TM) to make it to flying-bird status can be gotten from the case of the escaped chicken.
Consider that man raises chickens in gigantic abundance, and that on many farms, these are not even caged. Consider the numbers of such chickens which must have escaped in all of recorded history; look in the sky overhead: where are all of their wild-living descendants??
Why are there no wild chickens in the skies above us???
A flying bird requires a baker's dozen highly specialized systems, including flight feathers, wings, a special light bone structure, specialized flow-through design hearts and lungs vastly more efficient than ours, specialized tails and balance parameters, and a number of other things. Now, you can imagine the difficulty involved for something like a dinosaur which did not have any of these things to evolve them all, but the feral chicken
In other words, if there's any chance whatsoever of a non-flying creature evolving into a flying bird, then surely, surely the feral chicken, close as it is, could RE-EVOLVE back into being a flying bird. They're only missing the tiniest fraction of whatever is involved.
They've got wings, tails, and flight feathers, and the whold nine yards. In their domestic state, they can fly albeit badly; they are entirely similar to what you might expect of an evolutionist's proto-bird, in the final stage of evolving into a flight-worthy condition.
According to evolutionist dogma, at least a few of these should very quickly finish evolving back into something like a normal flying bird, once having escaped, and then the progeny of those few should very quickly fill the skies.
But the sky holds no wild chickens. In real life, against real settings, real predators, real conditions, the imperfect flight features do not suffice to save them.
In real life, if you ever lose the tiniest part of some complex trait or capability, you will never get it back. In the real world, if you lack the tiniest part of some complex trait or capability, then, other than possibly via some genetic engineering process, you will never get it.
Thus we see that "proto-bird" (TM) not only couldn't make it the entire journey which he is supposed to have, he couldn't even make it the last yard if we spotted him the thousand miles minus the yard.
The basic question is: How in hell is some velociraptor supposed to make it the thousand miles, if history proves that a creature which amounts to the final stage of such a development cannot make it the final yard of such a process?
And that is exactly why natural selection does not work as an agent for evolution. It takes too long. Even evolutionists admit that transformation of one species into another takes millions of years. As FDR said 'people do not live in the long run, they have to eat every day' or something close to that. A species, an organism cannot wait millions of years to eat. Environmental changes can occur in the twinkling of an eye. Even the one you mention, took place in a comparatively short time. The freezing of Greenland took only a few centuries, the drying of North Africa at most a millenia. So no, there was no time for evolution, natural selection or whatever, to induce new faculties in a species to make it able to better adapt to the changed conditions.