Deriving transport pathways in a sediment trend analysis (STA)

Sediment trend analysis (STA) is a technique that enables patterns of net sediment transport to be determined by relative changes in grain-size distributions of all naturally occurring sediments. In addition, STA can determine the dynamic behaviour of bottom sediments with respect to erosion, accretion or dynamic equilibrium. The data requirements for STA are the full grain-size distributions taken from sediment grab samples collected at a regular spacing over a particular area of interest. Two types of methods are presently used in the derivation of the transport pathways: a line-by-line approach, in which transport pathways are determined by searching for sample sequences in which the distributions change, in a statistically acceptable manner; and various vector approaches, in which each sample is compared to neighbouring sites from which a vector sum is calculated.The basic assumption for STA is that the processes that cause sediment transport will affect the statistics of sediment distributions in a predictable way. In reality, this type of analysis is complicated through the inclusion of a number of uncertainties, or noise. The goal of STA is to extract the information (the transport pathways) from the noisy signal (the grain-size distributions), an approach that is made difficult because neither the nature of the information nor the noise is known. Because of this, obtaining results by simply applying its theory in a “black-box” approach may provide poor solutions.The line-by-line approach to STA draws from communications theory to achieve a solution. In some communications systems, the information from many sources is combined into one signal. The resulting signal is, from a statistical viewpoint, nothing but noise. The extraction of specific information assumes that information is indeed present, and determines if that assumption is consistent with the received signal. For STA, the procedure is to assume a transport direction over an area comprising many sample sites. From this assumption, the predicted sediment trends are compared with the pathways determined by the actual samples. The assumed transport direction is repeatedly modified until the best fit is achieved.Vector techniques may be very helpful to provide insights and guidance to the line-by-line approach. However, the approach is ultimately necessary to finalize a credible solution as well as providing important further information on the dynamic behaviour of the sediments, and to delineate specific sediment transport environments.