A 2-year study was conducted to quantify the actual evapotranspiration (ETa) of three woody ornamental trees placed under three different leaching fractions (LFs). Argentine mesquite (Prosopis alba Grisebach), desert willow [Chilopsis linearis (Cav.) Sweet var. linearis], and southern live oak (Quercus virginiana Mill.) (nursery seedling selection) were planted as 3.8-, 18.9-, or 56.8-liter container nursery stock outdoors in 190-liter plastic lysimeters in which weekly hydrologic balances were maintained. Weekly storage changes were measured with a portable hoist-load cell apparatus. Irrigations were applied to maintain LFs of +0.25, 0.00, or -0.25 (theoretical) based on the equation irrigation (I) = ETa/(1 - LF). Tree height, trunk diameter, canopy volume, leaf area index, total leaf area (oak only) and dry weight were monitored during the experiment or measured at final harvest. Average yearly ETa was significantly influenced by planting size (oak and willow, P ≤ 0.001) and leaching fraction imposed (P ≤ 0.001). Multiple regressions accounting for the variability in average yearly ETa were comprised of different growth and water management variables depending on the species. LF, trunk diameter, and canopy volume accounted for 92% (P ≤ 0.001) of the variability in the average yearly ETa of oak. Monthly ETa data were also evaluated, with multiple regressions based on data from nonwater-deficit trees, such that LF could be ignored. In the case of desert willow, monthly potential ET and trunk diameter accounted for 88% (P ≤ 0.001) of the variability in the monthly ETa. Results suggest that irrigators could apply water to arid urban landscapes more efficiently if irrigations were scheduled based on such information.