Magnitude–frequency analysis of coarse suspended sediment discharges in northwestern Greece

Abstract

Sediment rating curves (SRCs) constitute a satisfactory deterministic approach to represent the sediment regime in catchments with limited or poor-quality records. Conversely, stochastic approaches like magnitude–frequency analysis have only lately been used directly on sediment records. This study aims to provide insight on the impact of SRC methodologies on the assessment of sediment discharge (Qs) variability. Three theoretical probability distributions were fitted to simulated mean monthly coarse (>100 μm) suspended Qs time series. The latter were developed by applying four SRC methodologies to the discharge records of 11 cross-sections of four major rivers in northwestern Greece. The analysis indicates that the SRC method used can influence the shape of the marginal distribution of the resulting time series. Its improper selection may lead to a dramatic underestimation of the sediment load maxima. The effect of different return periods per SRC method and theoretical distribution on Qs estimation are also assessed.