Recent work has emphasized development of full-range water-retention functions that are applicable under both wet and dry soil conditions, but evaluation of such functions in numerical modeling has been limited. Here we show that simulations using the Rossi-Nimmo (RN) full-range function compared favorably with those using the common Brooks-Corey function and that the RN function can improve prediction of water potentials in near-surface soil, particularly under dry conditions. Simulations using the RN function also improved prediction of temperatures throughout the soil profile. Such improvements could be important for calculations of liquid and vapor flow in near-surface soils and in deep unsaturated zones of arid and semiarid regions. ¿ 2000 American Geophysical Union