TERRESTRIAL AND MARINE BIOMARKERS IN A SEASONALLY ICE-COVERED ARCTIC ESTUARY - INTEGRATION OF MULTIVARIATE AND BIOMARKER APPROACHES

Hydrocarbons, including alkanes, alkenes, hopane triterpenes and polyc yclic aromatic hydrocarbons (PAHs), plus sterols, n-alcohols and a num ber of higher plant triterpenoids have been determined for suspended p articulate, sediment trap and sediment samples taken from the Mackenzi e River and the adjacent Mackenzie shelf in the Beaufort Sea. These bi omarkers are valuable tracers of terrigenous and petrogenic inputs fro m the river to the shelf water column and are also useful for assessin g marine production in the Mackenzie River estuary. We use Principal C omponent Analysis (PCA) to provide a robust classification of biomarke rs according to their primary source (e.g. terrigenous, marine) and to identify which biomarkers covary. The Mackenzie River is the dominant source for n-alkanes, n-alcohols, sterols and triterpenoids from high er plants, diagenetic hopanes, petrogenic isoprenoids and parent (unsu bstituted) PAHs to the Mackenzie shelf. The riverine hydrocarbon signa ture of these markers is modified by preferential settling out of lith ic material relative to less dense higher plant debris. Seasonal marin e production of a suite of alkenes, sterols and alcohols from phytopla nkton and zooplankton is evident in water column and sediment trap sam ples, but these labile compounds tend not to be preserved in surficial sediments. Although few individual sterols provide unambiguous marker s of terrestrial or marine organic matter, PCA successfully classifies sterols as either principally marine or principally terrestrial. n-Al cohols are often overlooked as biomarkers, but we find that they too r eflect the relative contributions of zooplankton and terrigenous input s in the Mackenzie estuary. To quantify dispersal of riverine material on the adjacent shelf in the context of dilution by marine production , we develop a multivariate Partial Least Squares (PLS) path model. We find that the PLS results strengthen geochemical interpretations base d on individual biomarkers by providing a quantitative representation of the differences among samples. PLS successfully models the increase s in terrigenous particulate on the shelf with river flow and the effe ct of autochthonous production.