Engineered surface barriers (or covers) are used to isolate underlying contaminants from water, plants, animals, and humans. To understand the flow processes within a barrier and the barrier’s ability to store and release water, the field hydraulic properties of the barrier need to be known. A multiyear test of the evapotranspiration (ET) barrier at the US Department of Energy’s Hanford Site in the southeast of the state of Washington has yielded in situ soil water content and pressure data from multiple locations for a 9-yr period, offering the opportunity to estimate soil water retention properties at different locations and times. The upper 2-m layer of the ET barrier is a silt loam, and the top 1 m also contains 15% (w/w) pea gravel. Within this layer, valid monitoring data from 1995 to 2003 for four depths at 12 monitoring stations were used to determine the field water retention of the silt loam, with and without gravel. The data covered a wide range of wetness, from near saturation to the permanent wilting point, and each retention curve contained 51 to 96 data points. The data were described well with the commonly used van Genuchten water retention model. It was found that the spatial variation of the saturated and residual water content and the pore size distribution parameter was relatively small, while that of the van Genuchten α was relatively large. The effects of spatial variability of the retention properties appeared to be larger than the combined effects of added pea gravel and plant roots on the properties. Neither the primary wetting process in the winter season nor the drying process in the summer season nor time had a detectable effect on the water retention of the silt loam barrier.