Traditionally, poplars (Populus) have been planted to control erosion on New Zealand's hill-slopes, because of their capacity to dry out and bind together the soil, by reducing effective rainfall and increasing evapotranspiration and soil strength. However, the effect of widely spaced poplars on the partitioning of soil water and rainfall has not been reported. This study determined rainfall partitioning for 18 mid-spring days in a mature P. deltoides (Bart. ex Marsh, Clone I78)-pasture association (37 stems per hectare, unevenly spaced at 16.4 +/- 0.4 m) and compared it with a traditional open pasture system in grazed areas of a hill environment. Tree transpiration was measured by the heat pulse technique. A time-driven mathematical model was used to set a zero offset, adjust anomalous values and describe simultaneous sap velocity time courses of trees. The model showed that daylight sap flow velocities can be represented with a nonlinear Beta function (R(2) > 0.98), and differences in the parameters representing the initiation, duration and conformation of the sap velocity can be tested statistically to discern tree transpiration differences during the day. Evapotranspiration was greater for the poplar-pasture association than for the open pasture (2.7-3.0 versus 2.2 mm day(-1)). The tree canopy alone contributed 0.92 mm day(-1) as transpiration and 1.37 mm day(-1) as interception, whereas evapotranspiration of the pasture understory was only 0.4-0.6 mm day(-1). Despite the higher water use of the poplar-pasture association, soil water in the 0-300 mm soil stratum was higher than, or similar to, that of the open pasture. Tree shading decreased evapotranspiration and pasture accumulation under the trees.