allow confident identification of dredged material layers from two areas of the same source 

 area on the seafloor at the PDS. The fine fraction data were most promising as a 

 diagnostic tool to allow differentiation of the layers at the Royal River Project capped 

 mound. Because of the importance of this goal, fine fraction results from the cores 

 collected during the postcap survey were statistically analyzed to evaluate the visual 

 distinctions in the sediment layers of the Portland Capping Project disposal mound. Two 

 overall methods of analysis were conducted on the sample dataset (Section 3.10.1). The 

 first approach analyzed the entire dataset with no a priori classification of the material 

 (Section 3.10.2). The second approach was to measure the strength of the classifications 

 of postcap core samples into CDM, pseudo-UDM, and ambient units based on visual 

 descriptions and microscopic analysis (Sections 3.10.3 and 3.10.4). 



The techniques described below provided statistical data using a variety of 

 approaches, and results were considered as a whole in this report. The tests were designed 

 to address several questions/hypotheses. First, the analysis was necessary to compare the 

 sample groupings assessed by the core visual descriptions in relation to the detailed 

 microscopic analyses. Although the visual and sample data provide qualitative evidence 

 for identifiable layers on the seafloor, these statistical analyses were able to address the 

 significance of the differences between the CDM, pseudo-UDM, and ambient layers. 

 Second, the relative correlation between the multiple variables, including biological, 

 physical, and environmental factors, was assessed. The results are discussed further in the 

 Discussion (Section 5.0). 



3.10.1 Sample Selection and Database Description 



Forty samples from seven of nine cores taken during the postcap coring survey were 

 used for the analyses (Section 3.7.2). Data from cores H2 and F2, which were 

 waterlogged and had unreliable stratigraphies, were excluded, as well as sample CI (53-57 

 cm) which contained only three microfossil individuals. For the mineralogy data, the 

 abundances that were quantified prior to analysis (absent=0, rare=l, common=2, and 

 abundant =3) were used for the statistical analyses described below. For the microfossil 

 data analysis, the number of individuals of each species of foraminifera and thecamoebian 

 were entered for each sample. 



3.10.2 Clustering and Multi-Dimensional Scaling Analysis 



Using PRIMER software (Plymouth Marine Laboratory, UK), two complimentary 

 multivariate techniques were employed to assess the mineralogy and microfossil data: 

 clustering and non-metric multi-dimensional scaling (MDS; Clark and Warwick 1994). 

 Both of these tests were conducted on the sample database (Section 3.10.1) with no a 

 priori classification of the samples from the visual core descriptions. These tests are 

 described briefly below, with references provided for further information. Following the 

 The Portland Disposal Site Capping Demonstration Project, 1995-1997 



