Evaluating sediment yield estimations from large-scale hydrologic systems using the rating curve concept DEMETRIS ZARRIS, DEMETRIS KOUTSOYIANNIS Department of Water Resources, National Technical University of Athens, 5, Heroon Polutechneiou St., Athens, 15780, Greece; zarris@itia.ntua.gr Abstract: A new approach in studying sediment yield from large hydrologic systems is pre- sented that utilizes sediment rating curves in conjunction with reservoir sediment de- posits downstream of the measurement site. It is shown that the rating curves, even with inadequate measurements, can provide a good basis for the computation of sediment yield. Key words: sediment yield, rating curves, Acheloos river INTRODUCTION There are two widespread techniques for the estimating sediment yield of hydrologic ba- sins, namely sediment rating curves and measurement of reservoir deposits. The former is preferred when sediment yield es- timates in comprehensive time steps are re- quired, provided that a quite frequent sam- pling program was established and a substan- tial number of measurements in high dis- charges taken. The latter delivers a more ac- curate long-term quantitative estimate of the total sediment load, provided that the reten- tion efficiency of the reservoir is unified and the density of the deposits are well measured. Rating curves are drawing serious criticism from various researchers. For instance, FERGUSON (1986) argues that when the rat- ing parameters result from a log-log regres- sion between suspended sediment and river discharges, an underestimation of sediment yield is resulted. Another cause of the inac- curacies associated with the use of rating relationships is the fact that a large propor- tion of the total suspended sediment load is transported by a few major flood events, which represent only a very small propor- tion of the total time. In most cases, particu- larly in Mediterranean type catchments, most of the annual sediment load is transported over a few days around peak flow conditions. These observations indicate two important implications for the likely accuracy of rat- ing curve estimates of suspended sediment load. Firstly it means that a regular sampling programme is unlikely to provide samples representative of those periods when the majority of the load is transported. Secondly it means that because the rating plot is fitted by least squares to the whole range of dis- charge and sediment discharge of the avail- able data, its trend may be largely determined by the main mass of samples representative of low discharges and sediment discharges, and may therefore be unrepresentative of the conditions during which the majority of the Short scientific paper RMZ - Materials and Geoenvironment, Vol. 52, No. 1, 157-159, 2005 157