Reactive flow models for explosives are usually developed by choosing an empirical form for the reaction rate and calibrating parameters against initiation experiments. In Shock wave initiation, experimental data almost always comprise mechanical measurements such as shock speed, material speed, compression, and pressure. However, we know from chemistry that reaction rates depend on temperature as well as on the mechanical state. This is one person why mechanically based reaction rates do not extrapolate well outwith the range of states used to normalize them. For instance, mechanical reaction models which match single-shock initiation generally fail to reproduce multiple-shock initiation phenomena.