The traditional application domain for C has been system software and other programs where performance is so critical that it's worth the tradeoff in higher development and maintenance costs. If you were writing a production-quality virtual machine, you might consider C.
Some still argue that it has a place in environments constrained in other ways, such as limited-resource embedded systems.
If by "computational science" you mean computer science / computer engineering education, you may still find C in environments which want to expose students to "the bare metal". If you mean computing in support of scientific research, you might be surprised at how much FORTRAN code is around.
You might also be amazed at the amount of cycles still spent on COBOL.
I tend to think of the above as "taxicab" languages. They aren't glamorous by any stretch of the imagination, but are still found all over the place due to their workaday utility.
As for other languages, one often finds Java in the big-server corporate/enterprise environment, as well as in many open-source projects. The JVM-as-platform is host for a rapidly-growing variety of alternate languages, many of which leverage the wide range of available Java libraries to avoid reinventing the wheel. These alternative languages may be open-source projects, academic efforts, or individual efforts; examples include Scala, Fan, JRuby, Jython, Clojure, Groovy, and hundreds more.
All of Python, Ruby, Perl, Lisp, Squeak, etc. have active, enthusiastic user communities with applications all over the landscape, especially including web development.
Functional languages, such as Haskell, OCaml, Scala, Erlang, F#, etc. were primarily academic, research, or niche efforts in the past, but are seeing much more interest in a wide range of application areas due to the increasing interest in high reliability, high code re-use, and multi-core/parallel programming.
The list goes ever on...