•Biopore- and matrix-dominated samples were isolated to estimate macropore flow.•Biopores and percolating biopores were separated for biopore-dominated samples.•Characteristics of percolating biopores were better at predicting water or air flow.•The mean diameter of the limiting layer (MDLL) was the best predictor. The different types of soil macropores (e.g., biopores and non-biopores) vary in the conductivity of water or air due to the difference in the 3D pore characteristics. The objectives of this study were to reveal which types of macropores and which macropore characteristics played the most important roles in regulating water or air flow. Intact soil columns sampled from the subsoil of a long-term fertilization experiment were scanned by medical X-ray computed tomography (CT), and subsequently, saturated hydraulic conductivity (Ks) and air permeability at −12 cm water potential (Ka12) were measured. The 3D characteristics of macropores were then analyzed with image analysis. The biopores and the percolating biopores that connected the top and the bottom of a soil column were separated for the biopore-dominated samples (with percolating biopores) and their 3D characteristics were also quantified. Our results showed that the mean macropore diameter of the limiting layer (MDLL) presented the best relationships with Ks and Ka12 compared with the other macropore characteristics for all the samples. The biopores and percolating biopores contributed 27.8–74.5% and 3.26–64.4% of the volume of total macropores, respectively, for the biopore-dominated samples. The hydraulic radius, mean diameter, compactness, global and local connectivities, and MDLL of biopores, especially those of percolating biopores, were generally larger than those of total macropores. The global connectivity (Γ) of biopores performed very well for estimating Ks and Ka12. The MDLL of percolating biopores could predict Ks much better than the MDLL of biopores and total macropores. Moreover, the performance of MDLL for estimating Ka12 was as good as the MDLL of biopores but was much better than the MDLL of total macropores. The findings of this study reveal that the MDLL is a more useful parameter in estimating saturated hydraulic conductivity and air permeability at low water potential than the other CT-derived pore characteristics.