GEOENVIRONMENTAL APPLICATION OF STATISTICAL TECHNIQUES USED IN MARINEMINERAL EXPLORATION

Recent research efforts have involved the development of computer prog rams and quantitative techniques designed specifically to facilitate t he reduction and interpretation of large geochemical data sets acquire d during marine mineral exploration surveys. The purpose of this study is to examine the applicability of these same techniques in developin g meaningful interpretations of geoenvironmental data. A suite of sedi ment samples from the Mobile-Tombigbee River system in southern Alabam a were subjected to bulk geochemical and mineralogical analyses. A Q-m ode factor analysis of the geochemical data suggests reveals that > 98 % of the total variance in data is explained by two factors: (1) a rel atively labile component of predominantly detrital aluminosilicates an d iron and manganese oxyhydroxides, and (2) a mixture of quartz, felds par, and ilmentite. The geochemical associations in both of these fact ors primarily reflect differences in sediment grain size. A linear mod eling procedure was used to determine specific associations between ce rtain trace metals and different mineralogical host phases, including composite clay (illite and illite-smectite) quartz + albite, kaolinite , goethite, and ilmenite. This information facilitates an understandin g of the distribution of trace metals in the sediment, thereby giving insight to the potential bioavailability and environmental threat of v arious trace metal contaminants. Elements (or that portion of elements ) associated with the more labile sedimentary components, such as oxyh ydroxide and organic carbon phases, represent the greatest environment al hazard, due to their potential release during change of sedimentary conditions.