Empirical model-building and response surface
Empirical model-building and response surface
Applied multivariate statistics for the social sciences
Applied multivariate statistics for the social sciences
| Hi-index | 0.48 |
The United States Environmental Protection Agency (USEPA) initiated a study of bioremediation to further reduce oil on contaminated beaches after the March 1989 Exxon Valdez oil spill in Prince William Sound, Alaska. Effects of two methods for delivering mixtures of nitrogen and phosphorus fertilizers to oiled beaches were compared to a control. Several forms of hydrocarbon variables were selected as indicators of fertilizer effects in reducing residual oil amounts over that which would disappear by natural causes. The resulting data set was analyzed by multivariate statistical procedures to illustrate a protocol for selection of an optimum set of variables to assess reduction of residual oil through nutrient enrichment applications and to determine statistically significant differences among the treatment application methods. One subset of variables, nC18, phytane, and the sum of amounts of hydrocarbons nC18-nC27, was effective for prediction of amount of residual oil while another subset of variables, phytane, nC22, ratio of nC18/phytane, and residual dry oil, was successful for distinguishing differences in treatments by multivariate analysis of variance and its' associated canonical analysis. The ranked order of differences from small to large in variable responses to fertilization application methods were control, liquid-solid application, and sprinkler application of nitrogen and phosphorus fertilizers.