Accurate modeling of air flow in porous media is significantly important for various branches of geosciences and engineering that include, but are not limited to, hydrology, hydrogeophysics, petroleum engineering, and environmental engineering (Assouline et al., 2016;Honarpour et al., 1986;Springer et al., 1995). The simulation of air movement in unsaturated porous media needs the accurate estimate of relative air permeability (RAP) as a function of the water content (Springer et al., 1995). The RAP, in fact, is a required parameter for the numerical simulation of subsurface two-phase flow under isothermal (e.g., Kueper & Frind, 1991;Zang et al., 2019) and thermal conditions (e.g., Mohanty & Yang, 2013;Vanderborght et al., 2017), which has significant implications for many geophysical issues such as the investigation of soil aeration and evapotranspiration (e.g., Ben-Noah & Friedman, 2018), the determination of transport of organic contaminants in the underground (e.g., Essaid et al., 2015), the simulation of reservoir CO 2 injection Abstract Appropriate simulation of air flow in porous media needs an accurate estimate of relative air permeability (RAP). Combining 2 traditional and 2 fractal water retention curves (WRCs) respectively with 6 permeability equations, we obtained 12 statistical and 12 fractal RAP models. All these models were subsequently tested with 31 experimental datasets of disturbed soils to examine their predictive ability for RAP. Results showed that irrespective of using traditional or fractal WRCs, the six permeability equations exhibit consistent performances for RAP prediction. The modified Burdine (Yang & Mohanty,
