Water movement through a waste-trench cover under natural conditions at a low-level radioactive waste disposal site in northwestern Illinois was studied from July 1982 to June 1984, using tensiometers, a moisture probe, and meteorological instruments. Four methods were used to estimate seepage: the Darcy, zero-flux plane, surface-based water-budget, and groundwater-based water-budget methods. Annual seepage estimates ranged from 48 to 216 mm (5-23% of total precipitation), with most seepage occurring in spring. The Darcy method, although limited in accuracy by uncertainty in hydraulic conductivity, was capable of discretizing seepage in space and time and indicated that seepage varied by almost an order of magnitude across the width of the trench. Lowest seepage rates occurred near the center of the cover, where seepage was gradual. Highest rates occurred along the edge of the cover, where seepage was highly episodic, with 84% of the total there being traced to wetting fronts from 28 individual storms. Limitations of the zero-flux-plane method were severe enough for the method to be judged inappropriate for use in this study.Water movement through a waste-trench cover under natural conditions at a low-level radioactive waste disposal site in northwestern Illinois was studied from July 1982 to June 1984, using tensiometers, a moisture probe, and meteorological instruments. Four methods were used to estimate seepage: the Darcy, zero-flux plane, surface-based water-budget, and groundwater-based water-budget methods. Annual seepage estimates ranged from 48 to 216mm (5-23% of total precipitation), with most seepage occurring in spring. The Darcy method, although limited in accuracy by uncertainty in hydraulic conductivity, was capable of discretizing seepage in space and time and indicated that seepage varied by almost an order of magnitude across the width of the trench. Lowest seepage rates occurred near the center of the cover, where seepage was gradual. Highest rates occurred along the edge of the cover, where seepage was highly episodic, with 84% of the total there being traced to wetting fronts from 28 individual storms. Limitations of the zero-flux-plane method were severe enough for the method to be judged inappropriate for use in this study.