Evapotranspiration (ET) of irrigated urban plants is a large yet uncertain component of urban water budgets in semi-arid regions. A detailed understanding of plot-scale ET and its sensitivity to plant species composition is necessary to improve estimates of urban water vapour fluxes and water balance. We used portable enclosed chambers and empirical equations to quantify ET from (1) unshaded urban lawns covered exclusively by turfgrass and (2) urban lawns comprised of open-grown trees and turfgrass groundcover in the Los Angeles Metropolitan area. Turfgrass at all locations had a non-limiting supply of soil water because of regular sprinkler irrigation. ET of irrigated turfgrass reached a maximum of 10. 4 ± 1·3 mm d−1 and was always higher than plot-scale tree transpiration, which did not exceed 1 mm d−1. In summer, total plot ET of the lawns with trees was lower than lawns without trees by 0·9–3·9 mm d−1. Turfgrass ET was highly sensitive to solar radiation, and the ratio of ET of lawns with trees to ET of lawns without trees decreased with tree canopy cover. Hence, reductions in turfgrass ET caused by shading effects of open-grown trees were more important in influencing total landscape ET than the addition of tree transpiration. This suggests that low-density planting of trees that partially shade irrigated urban lawns may be a water-saving measure in semi-arid irrigated environments. Copyright © 2013 John Wiley & Sons, Ltd.