The proper design of a bioremediation strategy for petroleum-contaminated sites requires a reasonable estimate of the biodegradation rate constant, which is not easy due to spatial heterogeneity. Accordingly, predictive models were developed by completing a bioventing study at the meso-scale. Reactors holding 4 kg of disturbed soil were tested using five different types of soils. Using statistical analysis, a two stage process was observed, with a fast rate of hydrocarbon degradation in the first 8 d and a slower rate in the remaining 22 d. Review of the correlations showed that the initial population of petroleum degrading bacteria and increasing silt content had a positive effect on the degradation. A negative impact on the degradation rate was seen by increasing the fraction of organic matter and clay content. Comparison of previously completed micro-scale and meso-scale degradation rates gave a scale-up factor (SF) of 1.8±0.5. Soils with an increased sand fraction had slightly higher SF values, whereas soils high in organic matter content had lower SF values. The measured SF values and developed correlations will help practitioners with site closure decisions, and indicate the need for additional SF work to allow better transfer of meso-scale data to the field.