[en] Reactive transport modelling can be time consuming and memory-intensive, especially for large-scale, long-term simulations with a large number of chemical components and interactions. The objective of this research was to develop a parallel version of MIN3P-THCm, a general purpose multicomponent reactive transport code for variably saturated porous media. The resulting program, entitled ParMIN3P-THCm, is able to deal with the significant computational burden of reactive transport simulations involving large spatial scales and long time frames and can be run efficiently on machines ranging from desktop PCs, shared-memory workstations, to distributed-memory supercomputers. Parallelization of MIN3P-THCm (ParMIN3P-THCm) was achieved through the domain decomposition method based on PETSc (Portable Extensible Toolkit for Scientific Computation) libraries. PETSc is also used as the parallel solver package, and for data structure and message communication. A hybrid MPI and OpenMP parallel programming approach is implemented in the code to take advantage of leadership-class supercomputers that combine both shared memory and distributed memory architectures. Features of the code include a modular input file, parallel configuration file, and parallel I/O, with potential expansibility to incorporate additional features in the near feature such as high-performance I/O using parallel HDF5, as well as parallel multigrid and unstructured grid methods. ParMIN3P-THCm has been developed from the ground up for parallel scalability and has been run using up to 768 processors with problem sizes up to 100 million unknowns. The code has demonstrated excellent speedup for reactive transport simulation problems using 8 processors on a local shared-memory workstation, 128 processors on the WestGrid supercomputer using MPI parallelization and 768 processors on the WestGrid supercomputer using hybrid MPI-OpenMP parallelization. The code has shown strong scalability in modelling large-scale reactive transport problems. (author)