A sediment discharge computation procedure based on a time-integrated point sample.

Abstract

A procedure for computing total suspended sediment load using a single point-integrated sample is presented. A power velocity distribution and Laursen's concentration distribution equation (1980) are used: U/U = (x + 1) (y/D)ˣ and C/Cₐ = (a/y)ᶻ where x = 1/4 to 1/7 and z = w/(βκ√(τ₀/ρ); other symbols are as commonly used. The procedure was tested with USGS (1971) field data from the Rio Grande. Using nominal values of β, κ, and w results in estimates of total suspended concentration that agree sufficiently well with depth-integrated measurements corrected for unmeasured load; even better agreement was obtained when site-specific data are used to define the x and z exponents. The difference between total suspended load computed using this procedure (and a single measurement) and conventional computations based on depth-integrated measurements is well within sampling error. For fine sediment (small z value) errors in measurements do not have a large effect on the integrated sediment load. For coarse sediment, however, placement of the sampler, elevation of the bed, and other values must be known with more precision. The typical scatter in suspended sediment load versus discharge plots can be explained by considering possible changes in bed material over time. Laursen's (1958) relationship for total sediment load can be used to evaluate such changes and to calculate bed load. There are major advantages in estimating total suspended load using one time-integrated suspended-sediment point sample. Less field time is required; sampling costs are greatly reduced; and sampling can be more frequent and better timed to measure the changing sediment load. Automatic sampling procedures are more feasible.