| Abstract: | In a 350 m wide reach of the braided, gravel-bed Ashley River, the surface layer of the bed material was sampled in 141 areas of homogeneous graded sediment along seven cross-sections, and 30 kg bulk samples were collected at 86 randomly selected locations along the cross-sections. At one location, a single 854 kg sample composed of 28 subsamples was also collected. Analysis of the single large sample indicate that accurate determination of mean grain size D at that site requires, desirably, a sample of ～ 100 kg, but that samples in which the weight of the largest stone is less than 5 per cent of the total weight have unbiased estimates of D. Spatial variability of bulk material is such that 228 and 50 samples are needed to estimate D to ± 10 and ± 20 per cent respectively of the true value; requirements for estimating inclusive graphic standard duration are only 11 and 3 respectively. The grain-size distribution of the surface layer is only weakly related to the bulk material beneath. The results of ‘Wolman sampling’ along 12 cross-sections at two pace intervals (average 120 stones per cross-section) indicate that estimation of overall surface D to ±10 and ±20 per cent would require sampling along 64 and 14 cross-sections respectively. It is concluded that accurate characterization of bed sediment in gravel-bed rivers is very demanding of labour and resources, and that careful planning is needed to ensure efficiency and meaningful results. |
