We constructed nitrogen (N) budgets for the lawns of three simulated residences built to test the environmental impacts of three different residential landscape designs in southern California. The three designs included: a ''Typical'' lawn planted with cool season tall fescue (Schedonorus phoenix), fertilized at the recommended rate for this species (192 kg -1 ha -1 year -1 ) and irrigated with an automatic timer; a design intended to lower N and water requirements (''Low Input'') with the warm season seashore paspalum (Paspalum vaginatum) fertilized at 123 kg -1 ha -1 year -1 and irrigated with a soil moisture-based system; and a design incorporating local best practices (''Low Impact'' lawn) that included the native sedge species Carex, fertilized at 48 kg -1 ha -1 year -1 and irrigated by a weather station-based system. Plant N uptake accounted for 33.2 ± 0.5 (tall fescue), 53.7 ± 0.7 (seashore paspalum), and 12.2 ± 1.3 % (Carex) of annual N inputs, while estimated N retention in soil was relatively large and similar in the three lawns (41-46 %). At lower N and water inputs than Typical, Low Input showed the highest annual clipping yield and N uptake, although it also had higher denitrification rates. Leaching inorganic N losses remained low even from the Typical lawn (2 %), while gaseous N losses were highly variable. The Low Input lawn was most efficient in retaining N with relatively low water and N costs, although its fertilization rates could be further reduced to lower gaseous N losses. Our results suggest that the choice of a warm-season, C 4 turf species with reduced rates of irrigation and fertilization is effective in this semi-arid region to maintain high productivity and N retention in plants and soils at low N and water inputs.