Reconstructing fire regimes of the past can provide a valuable frame of reference for understanding the impact of human land uses on contemporary fire and forest structure, but methods for reconstructing past fire regimes are under re-evaluation. In the present article, a common method of characterizing surface fire regimes, using composite fire intervals from fire scars, is shown to significantly underestimate the length of the fire rotation and population mean fire interval in Grand Canyon landscapes where these parameters are known. Also, the evidence and interpretation that past high-severity fire was uncommon in ponderosa pine landscapes in Grand Canyon National Park are challenged. Together, these two concerns mean that an alternative characterization of the fire regime, which has very different implications, cannot be excluded. Management aimed at lowering fire risk, as a means of restoration, does not presently have a sound scientific basis, if it uses the composite fire interval as a measure of the fire regime or is based on fire history research that lacks adequate analysis of past high-severity fire.