The single-assignment property of concurrent logic programming languages results in a large memory bandwidth requirement and low spatial locality in heap-based implementations. As large amounts of garbage are generated, it becomes necessary to salvage used memory frequently and efficiently, with a garbage collector. Another approach is to detect when a data structure becomes garbage and reuse it. In concurrent languages it is particularly difficult to determine when a data structure is garbage and suitable for destructive update. Dynamic schemes, such as reference counting, incur space and time overheads that can be unacceptable. In contrast, static-analysis techniques can be used to identify data objects whose storage may be reused. Information from static analysis can be used by the compiler in generating appropriate instructions for reuse, incurring little or no runtime overhead. In this paper we present a new method of reuse detection based on abstract interpretation. We present empirical performance measurements comparing the new scheme against binary reference counting (MRB). It is shown that our proposed static analysis can achieve most of the benefits of MRB and improve the execution time.