RESUMEN
The residual air saturation plays a crucial role in modeling hydrological processes of groundwater and the migration and distribution of contaminants in subsurface environments. However, the influence of factors such as media properties, displacement history, and hydrodynamic conditions on the residual air saturation is not consistent across different displacement scenarios. We conducted consecutive drainage-imbibition cycles in sand-packed columns under hydraulic conditions resembling natural subsurface environments, to investigate the impact of wetting flow rate, initial fluid state, and number of imbibition rounds (NIR) on residual air saturation. The results indicate that residual air saturation changes throughout the imbibition process, with variations separated into three distinct stages, namely, unstable residual air saturation (Sgr-u), momentary residual air saturation (Sgr-m), and stable residual air saturation (Sgr). The results also suggest that the transition from Sgr-u to Sgr is driven by changes in hydraulic pressure and gradient; the calculated values followed the following trend: Sgr > Sgr-u > Sgr-m. An increase in capillary number, which ranged from 1.46 × 10-7 to 3.07 × 10-6, increased Sgr-u and Sgr-m in some columns. The increase in Sgr ranged from 0.034 to 0.117 across all the experimental columns; this consistent increase can be explained by water film expansion at the primary wetting front along with a strengthening of the hydraulic gradient during water injection. Both the pre-covered water film on the sand grain surface and a pore-to-throat aspect ratio of up to 4.42 were identified as important factors for the increased residual air saturation observed during the imbibition process. Initial air saturation (Sai) positively influenced all three types of residual air saturation, while initial capillary pressure (Pci) exhibited a more pronounced inhibitory effect on residual air saturation, as it can partly characterized the initial connectivity of the air phase generated under different drying flow rates. Under identical wetting flow rate conditions, Sgr was higher during the second imbibition than during the first imbibition due to variations in initial fluid state, involving both fluid distribution and the concentration of dissolved air in the pore water. In contrast, NIR did not have an obvious effect on the three types of residual air saturation. This work aims to provide empirical evidences and offer further insights into the capture of non-wetting phases in groundwater environments, as well as to put forward some potential suggestion for future investigations on the retention and migration of contaminants that involves multiphase interface interactions in subsurface environments.
