Spatial complexity in channel geometry indicates that accurate prediction of sediment transport requires modeling in at least two dimensions. However, a one-dimensional model may be the only practical or possible alternative, especially for longer river reaches of practical concern in river management or landscape modeling. We have developed a one-dimensional model of the Colorado River through upper Grand Canyon that addresses this problem by reach averaging the channel properties and predicting changes in sand storage using separate source and sink functions coupled to the sand routing model. The model incorporates results from the application of a two-dimensional model of flow, sand transport, and bed evolution, and a new algorithm for setting the near-bed sand boundary condition for sand transported over an exposed bouldery bed. Model predictions were compared to measurements of sand discharge during intermittent tributary inputs and varying discharges controlled by dam releases. The model predictions generally agree well with the timing and magnitude of measured sand discharges but tend to overpredict sand discharge during the early stages of a high release designed to redistribute sand to higher-elevation deposits.