Sediment storage in alluvial valleys can strongly modulate the downstream migration of sediment and associated contaminants through landscapes. Traditional methods for routing contaminated sediment through valleys focus on in-channel sediment transport but ignore the influence of sediment exchanges with temporary sediment storage reservoirs outside the channel, such as floodplains. In theory, probabilistic analysis of particle trajectories through valleys offers a useful strategy for quantifying the influence of sediment storage on the downstream movement of contaminated sediment. This paper describes a field application and test of this theory, using 137Cs as a sediment tracer over 45 years (1952--1997), downstream of a historical effluent outfall at the Los Alamos National Laboratory (LANL), New Mexico. The theory is parameterized using a sediment budget based on field data and an estimate of the 137Cs release history at the upstream boundary. The uncalibrated model reasonably replicates the approximate magnitude and spatial distribution of channel- and floodplain-stored 137Cs measured in an independent field study. Model runs quantify the role of sediment storage in the long-term migration of a pulse of contaminated sediment, quantify the downstream impact of upstream mitigation, and mathematically decompose the future 137Cs flux near the LANL property boundary to evaluate the relative contributions of various upstream contaminant sources. The fate of many sediment-bound contaminants is determined by the relative timescales of contaminant degradation and particle residence time in different types of sedimentary environments. The theory provides a viable approach for quantifying the long-term movement of contaminated sediment through valleys.