Authors: Jeffrey Y. Jung, Thomas J. Mathis, Paul M. Craig, Bui Minh Hoa, Duy Huy Pham, Luis A. Bastidas, and Anurag Mishra
Dates: January 24-27, 2022
Location: Nashville, Tennessee
Background. Several potentially responsible parties at the Portland Harbor Superfund Site have cooperated to fund the development of a critically needed hydrodynamic and sediment transport model for the site. Given the spatial extent, expected costs, and morphologic complexity of the proposed site cleanup, the parties recognized the need for a robust and reliable model to predict long-term sediment transport processes in support of remedial design, recovery monitoring, and cost allocation.
Approach. To meet the requirements, DSI developed a modeling approach consisting of a hydrologic model of the City of Portland private and public drainages and multi-scale hydrodynamic and sediment transport models of the Lower Willamette and Columbia Rivers using the Environmental Fluid Dynamics Code Plus code (EFDC+). The multi-scale models include a large-scale (coarse grid) model for the entire domain, and a nested (fine grid) model at the scale to study individual sediment management areas. The SEDZLJ approach implemented in EFDC+ was used to simulate the various transport processes, including spatially varying sediment bed properties, bedload, suspended load, and settling of non-cohesive and cohesive particles. The model also features fully coupled morphologic feedback between changes in sediment bed elevations and the hydrodynamics.
Several lines of evidence were used to evaluate the model: qualitative patterns of erosion and deposition, net sedimentation rates, bed elevation changes, and TSS concentrations in the water column. Several bathymetric surveys were used to determine the observed bed elevation changes and support the calibration (2002-2004) and validation (2009-2018) simulations. Suspected areas of anthropogenic disturbance (propeller wash, dredging, capping, etc.) were identified to help understand discrepancies between observed and predicted patterns of erosion/deposition and to elucidate the role of natural forces in these areas.
Results. Deposition dominates the site, although many areas of natural erosion were also present. Both the large-scale and nested models predicted the observed net deposition volume at the site to within 11% during the calibration period and within 6% in the validation period. The sediment transport model predicted bed elevation changes at the site with a mean error of -0.5 cm/yr, and -0.1 cm/yr for the calibration and validation periods, respectively. The concentrations of TSS were accurately predicted in both the calibration and validation periods. These results demonstrate the EFDC+ based multi-scale sediment transport models developed for the Portland Harbor Superfund Site meet the project requirements and can be used as a general tool to support remedial design, recovery monitoring, and cost allocation.