Spatial and temporal variability in aggregated grain-size distributions, with implications for sediment dynamics

The grain-size distribution of bottom sediments has important implications for diverse aspects of sediment dynamics, including prediction of the critical boundary shear stress and calibration of suspended sediment sensors. Past sampling strategies to obtain estimates of the seabed grain-size distribution typically have not considered spatial and temporal variability, and have been insufficient to resolve potential millimeter-scale vertical variations in grain size. Moreover, laboratory analyses have been predicated on chemically and/or ultrasonically disaggregating the sediments before resolving particle diameter, therefore the more dynamically relevant in situ grain-size spectrum is not measured. To test for such effects, three sites on the northern California continental shelf comprising a cross-shelf transect from a sandy, inner shelf (60 m) site, to a muddy, mid-shelf (90 m) site and a relict, outer-shelf (130 m) site Were studied, Replicate box cores were collected over two winter field seasons, and multiple subcores from each box core were vertically sectioned at 2-mm intervals. Gentle wet sieving techniques were used to determine the mass fraction in the <20, 20-44, 45-62, 63-89, 90-179, 180-300 and >300 mu m size classes. In addition, a lesser number of standard disaggregated grain-size analyses were performed using a Coulter Counter. Results from the sandy, inner-shelf site indicate the presence of an ephemeral fine-grained (<20 mu m) surface layer (0-2 mm) that substantively alters the grain-size distribution ''seen'' by the flow. In addition, there is evidence for a progressive and substantial winnowing of fine-grained sediment from the surface layer over the course of a winter storm season. At the deeper sites, the upper 2 mm of the bed contained 5-20% more material <20 mu m than deeper portions of the sediment column. At all three sites, the level of small- (10s of centimeters) and intermediate- (10s of meters) scale spatial variability is modest. In addition, the disaggregated grain-size distribution al the two muddy sites is, in all cases, markedly finer than the in situ grain-size distribution. Therefore, calibrations and predictions based on a knowledge of the size distribution of the primary (i.e. disaggregated) particles could be in serious error. Copyright (C) 1996 Elsevier Science Ltd