Date of Award
Master of Science in Experimental Statistics
Computer Science and Experimental Statistics
Statistical relationships between a set of physical and chemical variates and a vector of species abundances are explored with the objective of examining the influence of waste disposal on the benthic macrofaunal assemblages of the New York Bight Apex area. The problem of microenvironmental variability is handled by post collection stratification of samples based on the physical characteristics of each grab sample, rather than classical spatial strata definitions. Three physical variables used to define strata are: mean particle size, percent organic matter, and total concentration of heavy metals. Variations in species densities within strata over time are investigated. Multivariate discriminant techniques are used to search for major statistical differences in species abundance between strata.
A sequential sampling plan is considered as an approach for detecting important shifts in species densities. Upper bounds on the average number of samples required to reach a decision are estimated for a number of different situations: varying σ2χ1- χ2 and the acceptable error. Applications to certain types of environmental monitoring programs are considered.
Walker, Henry A., "Environmental Impact Assessment Exploratory Techniques and Sampling Design" (1979). Open Access Master's Theses. Paper 2116.