Scientists today conduct new research via software-based experimentation and validation in a host of disciplines, including materials science, life sciences, astronomy, and physics. Scientific software represents a significant investment due to its complexity and longevity, its life spanning sometimes decades. Unfortunately, there is little reuse of scientific software beyond small libraries, increasing development and maintenance costs. Modern workflow and grid technologies offer a promising medium for reuse in this domain, however reuse at the workflow level is quite different than that of software libraries. To alleviate this disconnect, we have developed KADRE, a domain-specific architecture recovery approach and toolset to aid automatic and accurate identification of workflow components in existing scientific software. KADRE improves upon traditional state of the art general cluster techniques, helping to promote componentbased reuse of scientific kernels within the domain. Categ...