116 



As with the microfossil plots, the pseudo-UDM samples were divided into different 

 shades of blue according to the clustering on the dendrogram. Samples that shared 68% or 

 greater similarities were circled. The MDS plot does show general groupings of the 

 samples with some overlap between the CDM and pseudo-UDM samples. 



4.7.2 Analysis of Similarities 



After analyzing the microfossil and mineralogical results of the core samples and 

 conducting clustering analysis, there was an overlap of characteristics between pseudo- 

 UDM and CDM. Therefore, a null hypothesis was tested to determine the statistical 

 significance of the difference between CDM and pseudo-UDM samples. Again, the null 

 hypothesis was that no differences existed between the CDM and pseudo-UDM samples. 



An ANOSIM randomization test was conducted on microfossil data. ANOSIM is 

 based on a non-parametric test analogous to standard parametric analysis of variance 

 (ANOVA). For this test, the classification of the samples occurred prior to analysis and 

 ambient samples were not included. The test compared the difference between the CDM 

 and pseudo-UDM samples with the differences in the samples within each group displayed 

 on the MDS plot. The program calculated a global R value of 0.297. The null hypothesis, 

 that no differences exist between the CDM and pseudo-UDM samples, was rejected with a 

 significance level of p < 0.001. 



4.7.3 Discriminant Statistics 



Using SPSS® Professional Statistics 6.1, we performed a discriminant statistical 

 analysis on the mineralogy and microfossil results from the core samples. Discriminant 

 statistics is a multi-variable technique to measure the degree of association between groups 

 of data. Because the groups were pre-determined based on the visual descriptions, the 

 success of discriminant classification allowed an estimate of the acmal differences or 

 similarities between groups. 



The microfossil data were grouped into five categories for this analysis: freshwater 

 thecamoebians, marsh foraminifera, mudflat foraminifera, shelf agglutinated foraminifera, 

 and shelf calcareous foraminifera. The relative abundance of the five groups of species 

 were calculated for individual samples. Mineralogical parameter abundances were used as 

 described above. Each sample was then grouped with the AMB, pseudo-UDM, or CDM 

 classifications based on the visual core descriptions and on microfossil analyses. 



Following separation into groups, the group means, standard deviation, and 

 discriminant scores were calculated, and the scores were graphed according to two 

 canonical discriminant functions. The canonical functions represented the ordination axes 

 that best separated the pre-determined groups. The SPSS program then determined the 

 The Portland Disposal Site Capping Demonstration Project, 1995-1997 



