Derivation of sediment resuspension rates from acoustic backscatter time-series in tidal waters

Abstract

Extending oceanographic forecasting models beyond dynamics to ecological parameters involves simulation of concentrations of suspended particulate matter (SPM). The latter will require assimilation of both in situ and remote sensing observations. Assimilation will need to reconcile both types of observations with modelling responses for a variety of both resuspension and settling velocity parameters. This study develops a systematic approach to this problem. Time series of suspended sediment particulate matter (SPM) concentration are routinely obtained via indirect optical, acoustic and satellite instrumentation. However, translating such measurements into components contributed by localised sediment resuspension and horizontal advection is severely complicated by uncertainties concerning the specific SPM characteristics which cannot easily be measured in situ. Since some estimate of synoptic tidal currents is generally available, resuspension-transport-deposition models can be used to interpret these SPM concentration time series. Here, a novel methodology, incorporating an optimisation procedure and a 1-D Lagrangian particle tracking model, is developed to automate this interpretation and indicate the nature of the associated SPM. Utilising calibrated acoustic backscatter measurements from Acoustic Doppler Current Profilers, a downhill simplex optimisation method minimises the least squares coefficient of determination ( R 2) between model and observed SPM concentration time series. Advection of a linear “background” concentration gradient is incorporated into the SPM model, and the optimisation procedure decouples observed SPM concentration time series into background and resuspension components. The model has been validated in three independent ways and good agreement between derived model parameters and independent observations has been found for settling velocity, background concentration gradients and erosion rates. Using data from two contrasting sites in the Mersey estuary and Dover Straits, agreement for concentrations involved 0.61 < R 2 < 0.83. A modular design provides scope for more complex formulations and improvements of 20% in R 2 occurred when a time varying eddy diffusivity was employed.