Transportation of sediment as bed load and as suspended sediment

Transportation of sediment - the necessary link between erosion and deposition of sediment - reflects the redistribution of material within and the geomorphological evolution of a river basin. Quantitative analysis and understanding of the character of the sediment in motion usually requires a knowledge of the source areas, which in turn often are related to the land use pattern.

Obstacles, such as blocks of soil, that slide into the water on the erosion scarps, affect the flow conditions and especially the local bed configuration.

Bed-load particles move as a series of steps interrupted by periods of no motion, when the particles are a part of the bed material. Bed-load movement produces bed forms such as ripples, bars, dunes, etc. which in turn affect the flow conditions, the bank stability, and the conditions for navigation. Many attempts have been made to determine the bed-load discharge in rivers by direct measurements, with some type of apparatus, and by the use of empirical or theoretically deduced formulas. None of the proposed methods, however, have been universally accepted as completely adequate for the determination of bed-load disharge in natural river channels. I prefer to use a modification of Bagnold´s bed-load equation for computing the bed-load discharge.

Bed-load discharge rating curve for the Nam Theun river at station "Kengbit upstream" in Laos.

The rating curve is based on a modification of Bagnold´s bed-load equation. According to this modified equation

l_b=k₁(u²₁₀₀ - u²_100c)u₁₀₀

where l_b is the rate of bed-load transport per unit width of the stream bed, expressed as a weight per unit time (g . cm^-1 . sec^-1), k₁ a calibration coefficient related to the grain size of the bed material, u²₁₀₀ and u²_100c the actual and the critical flow velocity respectively 100 cm above the bed. The critical flow velocity is calculated by the equation worked out by Sundborg. According to my experience this rather simple bed-load equation, based on the excess stream power, often gives bed-load discharges in the proper order of magnitude.

An increase in stream flow is usually accompanied by an increase in suspended-sediment concentration. The sediment-concentration peak associated with a rise may precede, coincide with, or follow the water-discharge peak. The latter case is illustrated by the curve below.

The relationship between the sediment concentration, 5 cm below the water surface in vertical I of the outlet section of Lake Laitaure, and the water discharge in the section during the period July 9 - 24, 1957. (The outlet section is marked on the hydrographic map.)

Since the flood waves travel at a greater rate through the lake than the water and the suspended sediment, the peak in sediment concentration lags behind the peak in water discharge at the outlet of Lake Laitaure in northern Sweden. The sediment concentration may, therefore, be considerably lower during the rising than during the falling stage.

Suspended-sediment loads are often estimated by the flow-duration, rating curve (FDRC) method. The suspended-sediment rating curves show the empirical relation between suspended-sediment load, L (or y), and water discharge, Q (or x)). This relation, usally defined as a power function, L=aQ^b, can be formulated as either a linear or non-linear model to find the solution of the rating-curve parameters (a and b) These parameters are often estimated by the ordinary least squares regression of the log-transformed variables (L and Q). This usually leads to considerable underestimating of annual sediment yields, as examplified by curve C in the diagram below. The degree of underestimation increases with the degree of scatter about the rating curve.

Accumulated discharge of suspended sediment January - December 1971 at the mouth of the Swedish river Fyrisån, based on pentadmedia. A = values according to measured data, giving a yearly sum of 7402 tonnes. B = values according to calculations by a rating curve based on mean values of water discharge and suspended sediment for water discharge intervals of 5 m³/s, giving a yearly sum of 7355 tonnes = 99.37 % of A. (L = 0.66Q^1.45; r² = 0.93). C = values according to calculations by rating-curve parameters estimated by the ordinary least squares regression of the log-transformed variables (L and Q), giving a yearly sum of only 4897 tonnes = 66.16 % of A. (L = 1.22Q^1.09; r² = 0.80).

When calculating the relationship between the suspended-sediment load and the water discharge I prefer to group the values of water discharge and corresponding sediment concentration and sediment load into "water-discharge classes". The rating curves are then based on mean values of water discharge and suspended sediment for the selected water discharge intervals of equal range. The result of calculations by rating-curve parameters based on water-discharge classes is exemplified by curve B in the diagram above.

In order to test the reliability of the calculations and to correct the results, measured values, computed for days with known suspended-sediment concentration, should be compared with values for the same days, calculated by the rating curves.

For this river station the difference between measured values, computed for days with known suspended-sediment concentration and values for the same days, calculated by the equation of the rating curve, varied a lot but was less than 1% between the sum of measured and the sum of calculated loads for the 272 compared daily loads 1987 - 1990.

Monthly mean suspended-sediment load in the Nam Ngum river at Ban Na Luang.

The annual variation in suspended-sediment load is great at this river station. For the period 1987 - 1990 about 20% of the sediment load was discharged during a cumulated period of about 1% of the total. This means that it is very important to have a high sampling frequency during the flood periods.

Most ot the sediment transported by rivers, especially the coarser part of the sediment load, is primarily deposited close to the river mouth. Parts of the deposits will, however, later be redistributed, mainly by erosion, resuspension, and transportation to deeper areas during periods with high dynamic activity in the frontal lake or sea basin. The relationship between the rate of accumulation of river sediment in front of a river mouth and the sediment discharge in the feeding river may therefore sometimes be rather poor. This relationship may also be affected by slides taken place in front of the river mouth and by dredging and dumping of dredged material.

Sediments are capable of transporting loads of adsorbed nutrients, pesticides, heavy metals, and other toxins. Because sediment movement and sediment character is affected by the ever-changing environment, measurements should be done both for sediment quantity and sediment characteristics.

File in Swedish.

Back.