Models of chemical reaction systems can be complex as they need to include information regarding the reactions and the mass and heat transfers. The commonly used state variables, namely, concentrations and temperatures, express the interplay between many phenomena. As a consequence, each state variable is affected by several rate processes. On the other hand, it is well known that it is possible to partition the state space into a reaction invariant subspace and its orthogonal complement using a linear transformation involving the reaction stoichiometry. This paper uses a more sophisticated linear transformation to partition the state space into various subspaces, each one linked to a single rate process such as a particular reaction, mass transfer or heat transfer. The implications of this partitioning are discussed with respect to several applications related to data reconciliation, state and rate estimation, modeling, identification, control and optimization of reaction systems. (C) 2017 Elsevier Ltd. All rights reserved.