Deep percolation (DP) from irrigation may be important for groundwater recharge in irrigated agricultural river corridors of arid regions, yet few studies of this physiographic setting have characterized both percolation and its direct effects on groundwater levels. The objectives of our study in a sandy loam, flood-irrigated, alfalfa-grass field in northern New Mexico were to (1) compare DP below the 1 m effective root zone based on water balance method (WBM) and Root Zone Water Quality Model (RZWQM) simulations, and (2) characterize effects of DP on shallow groundwater levels. Irrigation water applications were metered, and automated instrumentation measured soil water content and climate data for WBM calculations and RZWQM simulations. Groundwater response was characterized by recorded below-field water levels in four experimental wells. DP varied with initial soil water content and water application amount, ranging from 5 to 18 cm (mean 11.2 4.1 SD) with the WBM and from 6 to 17 cm (10.6 3.8 SD) with RZWQM (using 0.0005 cm3 cm-3 macroporosity). Across irrigation events, there was high correlation (r = 0.90) between WBM and RZWQM DP. Peak water level response (up to 38 cm) varied from 8 to 16 h after irrigation onset depending on well location and water application amount. Study results show that flood irrigation is a significant source of shallow groundwater recharge. The high correlation between calculated and simulated deep percolation without iterative model calibration indicates that RZWQM can be a useful tool to estimate DP and extend localized field studies to larger spatial scales.