Hot nighttime temperatures in the urban core of Phoenix, AZ create daunting challenges for city officials as they seek to attract new residents to a revitalized downtown. We investigated the tradeoffs between water use and nighttime cooling for three residential urban-design scenarios of 10 inner-city census tracts of interest to the City. We used the LUMPS model (Local-Scale Urban Meteorological Parameterization Scheme, after S. Grimmond and T. Oke, 2002) to examine the variation in evaporation. We then estimated possible differences in cooling based on output of 2000-2400 LST sensible heat fluxes calculated by LUMPS. The LUMPS model compares well with estimates of outdoor water use from City records (r2=.89). Our estimates of cooling were commensurate with spatial variations of ASTER nighttime satellite temperatures at 2035 LST (r2=.81). We ran LUMPS to simulate consequences of implementing: (1) a more compact core with an increase in residential building density of 10%, (2) an oasis core with a 20% gain in irrigated vegetation, and (3) a desert core, with irrigated landscaping decreased by 10%. Overall, Scenario 2 with increased vegetation produced significant increases in nighttime cooling, but with accompanying large increases in water use. Increases in vegetative density produced larger cooling effects in sparsely vegetated core residential locales than in densely vegetated neighborhoods. Reducing vegetation as a water saving strategy slightly reduces water consumption across the 10 tracts, but may cause even warmer nights. LUMPS provides credible land cover-based estimates of implementing different UHI-mitigation methods and urban design strategies.