Search

Peters et al. (2015) criticize the one-run kinetics method that my colleagues and I developed and promoted (Waples et al., 2002, 2010; Waples and Nowaczewski, 2013), and further claim that source-rock kinetics derived by the standard multirun methods are superior. Specifically, they state that Arrhenius ( A ) factors determined mathematically from multirun kinetics are significantly more accurate than those obtained using other methods. They also claim that those A factors and the associated activation-energy ( E a) distributions are sufficiently accurate in an absolute sense to be used, with confidence and without any further quality control, in modeling hydrocarbon generation. I disagree strongly with those various statements. Any discussion of the kinetics of hydrocarbon generation must begin with an understanding that the parallel first-order kinetic description used by almost everyone is simply a convenient construct, and it is clearly not a mechanistic description of what happens in nature (e.g., Stainforth, 2009). Most researchers in this field believe that the reactions involved in hydrocarbon generation are mainly chain reactions rather than simple decompositions, that reaction schemes do not involve a group of discrete parallel processes, and that generation yields various intermediate products on the way to final products (e.g., Stainforth, 2009; Tang and Ma, 2009). We therefore seek a useful kinetic system, rather than a correct one, and we must be prepared to make compromises and simplifications to achieve this goal. That said, not all compromises and simplifications are equally valid or acceptable. The parallel first-order kinetic model has been remarkably successful for three decades in modeling the evolution of the S2 peak during Rock-Eval–type pyrolysis of the great majority of kerogens, although exceptions have been noted where the S2 peak is narrower than first-order kinetics can explain (e.g., Burnham et al., 1996; Stainforth, 2009). Because generation of the …