The water resources of the Upper Colorado River Basin have been extensively developed for water supply, irrigation, and power generation through water storage in upstream reservoirs during spring runoff and subsequent releases during the remainder of the year. The net effect of water-resource development has been to substantially modify the predevelopment annual hydrograph as well as the timing and amount of sediment delivery from the upper Green River and the Yampa River Basins tributaries to the main-stem reaches where endangered native fish populations have been observed. The U.S. Geological Survey, in cooperation with the Colorado Division of Wildlife and the U.S. Fish and Wildlife Service, began a study to identify sediment source reaches in the Green River main stem and the lower Yampa and Little Snake Rivers and to identify sediment-transport relations that would be useful in assessing the potential effects of hydrograph modification by reservoir operation on sedimentation at identified razorback spawning bars in the Green River. The need for additional data collection is evaluated at each sampling site. Sediment loads were calculated at five key areas within the watershed by using instantaneous measurements of streamflow, suspended-sediment concentration, and bedload. Sediment loads were computed at each site for two modes of transport (suspended load and bedload), as well as for the total-sediment load (suspended load plus bedload) where both modes were sampled. Sediment loads also were calculated for sediment particle-size range (silt-and-clay, and sand-and-gravel sizes) if laboratory size analysis had been performed on the sample, and by hydrograph season. Sediment-transport curves were developed for each type of sediment load by a least-squares regression of logarithmic-transformed data. Transport equations for suspended load and total load had coefficients of determination of at least 0.72 at all of the sampling sites except Little Snake River near Lily, Colorado. Bedload transport equations at the five sites had coefficients of determination that ranged from 0.40 (Yampa River at Deerlodge Park, Colorado) to 0.80 (Yampa River above Little Snake River near Maybell, Colorado). Transport equations for silt and clay-size material had coefficients of determination that ranged from 0.46 to 0.82. Where particle-size data were available (Yampa River at Deerlodge Park, Colorado, and Green River near Jensen, Utah), transport equations for the smaller particle sizes (fine sand) tended to have higher coefficients of determination than the equations for coarser sizes (medium and coarse sand, and very coarse sand and gravel). Because the data had to be subdivided into at least two subsets (rising-limb, falling-limb and, occasionally, base-flow periods), the seasonal transport equations generally were based on relatively few samples. All transport equations probably could be improved by additional data collected at strategically timed periods.