Multivariate data analysis was used to assess the environmental fate, specifically the effect of biodegradation, of stranded oil after the Baltic Carrier oil spill on 3/29/2001. GC-MS data from samples (oil smudged stones) collected from the beach of Hestehoved, Denmark, an area affected by the spill, were analyzed by three different approaches, i.e., a univariate analysis using the diagnostic biodegradation ratios nC ₁₇ /(nC ₁₇ +Pr) and nC ₁₈ /(nC ₁₈ +Ph); multivariate data analysis of the isomeric pattern of tetramethylnaphthalenes; and a more comprehensive analysis of 21 diagnostic biodegradation ratios selected among a much larger number of variables by a selection method based on quantitative reliability criteria. The sampling strategy resulted in very heterogeneous analytical data, as biodegradation was very much dependent on the thickness of the oil layer as well as the local chemical and physical properties, e.g., exposure to water flushing, oxygen, etc. The univariate approach showed that the degree of biodegradation of alkanes increased from 0 to 8.5 mo while the samples collected at 12 and 16.5 mo indicated a lower degree of biodegradation. The third most comprehensive analysis showed that an initial biodegradation process occurred during 0-2.5 mo. Diagnostics ratios from samples taken at 8.5 mo after the spill indicated the highest degree of biodegradation. Several other biomarker ratios analyzed in the study, e.g., hopanes and steranes, were not yet affected by the biodegradation processes, as indicated by very consistent ratios with descriptive power-values close to unity. This is an abstract of a paper presented at the 25th Arctic and Marine Oilspill Program Technical Seminar (Calgary, Alberta, 6/11-13/2002).