^__ 13_ 



velocities were slightly lower), but varied significantly in direction. As will be discussed 

 later, the choice of velocity input, near bottom current meter or ADCIRC vertically 

 averaged velocity results, had little impact on the results for total erosion from the site. 

 Tidal elevation inputs were derived from the ADCIRC simulations. Sediment boundary 

 condition inputs are TSS concentrations at the inflow boundaries. Since no boundary 

 condition data were available for these simulations, all inflow boundary TSS concentrations 

 were assumed to be zero. This may result m. artificially low amounts of re-deposition of 

 sediment at the site. The resulting bathymetry change can therefore be considered more of 

 a gross erosion flux rather than a net flux of sediment at the sediment/water interface. The 

 zero boundary conditions may seem unrealistic, especially during storms, but the nature of 

 the PDS, widi many rocky ridges, may block most sediment from being advected to the 

 site from other locations if sediments are not high in the water column. Therefore the 

 assimiption of no inflow boundary concentrations can be considered a conservative 

 assumption, i.e. assuming the worst case condition of all sediment being blocked fi-om 

 entering the PDS. 



As a first step in choosing the appropriate sediment characteristics for the site, the 

 field data collected during the February-May 1996 survey by SAIC at the PDS were 

 analyzed. The sediments collected by grab sample and visually analyzed were described as 

 a mixture of rock, gravel, sand, silt and clay (McDowell and Pace, 1996). The 

 predominant sediment types are sand, silt and clay. The LTFATE model can not, at 

 present, model both cohesionless (sand and coarse silt) and cohesive (silt and clay) 

 sediments. Therefore a single sediment type needed to be chosen. To select a sediment 

 type, events (storms) within the survey period were simulated for both cohesive and 

 cohesionless sediments and the results compared to the TSS data collected during those 

 periods. These experiments indicated that, if classified as cohesionless coarse silt (0.06 

 mm), insufficient sediment was transported to accurately reflect the SAIC measured TSS 

 concentrations for the April 1996 events. Two relevant data sets were available to assist in 

 determining the classification of PDS sediments for LTFATE simulations. The furst was 

 sediment size analysis of 16 cores extracted from the PDS in July, 1992. Twelve of the 

 sixteen sediment samples were greater than 50% (by dry weight) silt and clay (U.S. Army 

 Corps of Engineers, 1996). The second data set, sediment core samples collected during 

 summer 1995 from 23 locations in Portland Harbor (which can be assumed to be similar to 

 sediments placed at the PDS) and analyzed by the Army Corps (Report Number 446- 

 50274-1), indicate that 17 of 23 core samples were greater than 50% silt and clay material. 

 Therefore for all modeling purposes the sediments were described as fme-grain cohesive. 

 This would seem reasonable for mixed sand/silt/clay sediments because research has 

 indicated that sediments with even a relatively moderate fraction of silt/clay will behave 

 more like cohesive sediments than like pure sands (for example, the erosion potential will 

 decrease significantly with depth below the sedunent/water interface). Research has 



A Predictive Model for Sediment Transport at the Portland Disposal Site, Maine 



