Not really. Unless you care to elaborate based on your own professional experience, this is not what I've seen in the scientific sector. When a lab or organization uses Macs it tends to be localized to a particular group and almost never runs OS X natively. But admittedly my experience is limited to the US National Labs, JPL, and CERN.
Again, just because there are Macs in science doesn't mean creation of a "Macs in Science" forum will lead to any productive discussion about it. Look at how dead the Distributed Computing forum is now.
Whether or not the use of a Mac is localized to a particular group isn't really here or there - a lone graphic artist in a Windows-based office is equally localized. If a single user, or lab group, uses a Mac and the rest of the university doesn't, what does that actually matter for the poster?
In fact, that localization *is* a topic of discussion - the need for a Mac to "play well with others" has a much more expansive definition of "others" for most university settings.
My experience is contrary to yours - the Mac users in academia I know are keen to use OS X natively as much as possible, though admittedly many of them have a terminal window up as well. But 90% of what I do can run natively in OS X, and that number would be 100% if SAS would just come out with a Mac version, or I could be bothered to go learn STATA.
In terms of differences between most power users - Grids and other "server" type infrastructure becomes important to people whose training is frankly, not in that area (I for example, was busy getting degrees in biology), multithreading is something you *can* do, rather than just griping that Adobe/Apple/Whoever hasn't implemented it well yet, and the UNIX side of the machine will probably be seeing heavy use.
"Macs in the Workplace" threads pop up every now and then, but they're in the vast minority.
True, but that's kind of why sub-forums exist in the first place, no? So that low traffic threads don't simply vanish off the main pages?
That being said, the question was whether or not it would generate enough traffic to be justified. The answer to that may very well be no.