Discontinuous hindcast simulations of estuarine bathymetric change: A case study from Suisun Bay, California

Simulations of estuarine bathymetric change over decadal timescales require methods for idealization and reduction of forcing data and boundary conditions. Continuous simulations are hampered by computational and data limitations and results are rarely evaluated with observed bathymetric change data. Bathymetric change data for Suisun Bay, California span the 1867–1990 period with five bathymetric surveys during that period. The four periods of bathymetric change were modeled using a coupled hydrodynamic-sediment transport model operated at the tidal-timescale. The efficacy of idealization techniques was investigated by discontinuously simulating the four periods. The 1867–1887 period, used for calibration of wave energy and sediment parameters, was modeled with an average error of 37% while the remaining periods were modeled with error ranging from 23% to 121%. Variation in post-calibration performance is attributed to temporally variable sediment parameters and lack of bathymetric and configuration data for portions of Suisun Bay and the Delta. Modifying seaward sediment delivery and bed composition resulted in large performance increases for post-calibration periods suggesting that continuous simulation with constant parameters is unrealistic. Idealization techniques which accelerate morphological change should therefore be used with caution in estuaries where parameters may change on sub-decadal timescales. This study highlights the utility and shortcomings of estuarine geomorphic models for estimating past changes in forcing mechanisms such as sediment supply and bed composition. The results further stress the inherent difficulty of simulating estuarine changes over decadal timescales due to changes in configuration, benthic composition, and anthropogenic forcing such as dredging and channelization.

► We modeled estuarine bathymetric change over three past decadal-timescale intervals.
► We applied and tested input reduction and acceleration techniques.
► Performance was varied, and was sensitive to forcing functions.
► Abrupt changes in historical estuarine conditions may confound continuous simulation.