At the opposite end of the spectrum from empirical studies based on large petrographic data bases are computer models created with the goal of predicting diagenesis in data-limited areas. These models fall broadly into three categories: (1) geochemical models that simulate chemical reactions between minerals, gases, and fluids, (2) fluid-flow models that simulate the hydrologic behavior of pore fluids through time, and (3) reaction-transport models that attempt to merge (couple) reactions and flow. Detailed description of the models is beyond the scope of this review, but can be found in publications by Chen et al. (submitted); Bryant et al. (1993), Hermanrud (1993), Plummer (1992), Steefel and Lasaga (1992), Harrison and Summa (1991), and Meshri and Ortoleva (1990), among others.
Over the past five to ten years, the increasing power of desktop computers has resulted in proliferation of computer models and applications. As these models have been more extensively applied, a series of significant challenges has arisen. There are computational limitations involved in solving complex geologic problems in real time, and researchers frequently lack well-constrained petrographic observations, as well as basic chemical and hydrologic data for the systems being considered. Given these limitations, the current models provide a useful laboratory in which to examine sensitivities and assess uncertainties, but they are a long way from being fully predictive, except in very simple settings. The next conceptual advances will likely depend on how well laboratory-derived kinetic parameters can be applied to complex geologic systems.