Soil-water movement under natural-site and simulated waste-site conditions were compared by monitoring four experimental sites in the Mojave Desert, Nevada, during a 5-year period: one vegetated soil profile, one soil profile where vegetation was removed, and two nonvegetated test trenches. Precipitation ranged from 14 to 162 mm/yr. Temporal changes in water content measured by neutron probe were limited to the upper 0.5–1 m; values ranged from 0.01 to 0.19 m³/m³. Water potential and temperature were measured by thermocouple psychrometers; 77% remained operable for ≥4.5 years. For vegetated soil, precipitation that accumulated in the upper 0.75 m of soil was removed by evapotranspiration: water potentials decreased seasonally by 4 to >8 MPa. During 2 years with below-average precipitation, water potentials below the app arent root zone decreased by 2.3 (1.2-m depth) to 0.4 MPa (5-m depth), and the gradients became predominantly upward. Water potentials then rebounded during 2 years with near- and above-average precipitation, and seasonally variant water potential gradients were reestablished above the 4.2-m depth. Under nonvegetated waste-site conditions, data indicated the long-term accumulation and shallow, but continued, penetration of precipitation: water potentials showed moisture penetration to depths of 0.75−1.85 m. The method of simulated-waste drum placement (stacked versus random) and the associated differences in subsidence showed no measurable influence on the water balance of the trenches: subsidence totaled ≤13 mm during the study. Water potentials below the trenches and below the 2-m depth for the nonvegetated soil remained low (≈−5.5 to −7.5 MPa) and indicated the persistence of typically upward driving forces for isothermal water flow. Water fluxes estimated from water potential and temperature data suggested that isothermal liquid, isothermal vapor, and nonisothermal vapor flow need to be considered in the conceptualization of unsaturated flow at the field sites. Below the depth of temporal water content change, the estimated liquid fluxes ranged from 10⁻¹⁰ to 10⁻¹⁵ cm/s, isothermal vapor fluxes ranged from 10⁻¹⁰ to 10⁻¹³ cm/s, and the nonisothermal vapor fluxes ranged from 10⁻⁸ to 10⁻¹⁰cm/s.