Oil type and temperature dependent biodegradation dynamics. Combining chemical and microbial community data through multivariate analysis.

Abstract

Background This study investigates a comparative multivariate approach for studying the biodegradation of chemically dispersed oil. The rationale for this approach lies in the inherent complexity of the data and challenges associated with comparing multiple experiments with inconsistent sampling points, with respect to inferring correlations and visualizing multiple datasets with numerous variables. We aim to identify novel correlations among microbial community composition, the chemical change of individual petroleum hydrocarbons, oil type and temperature by creating modelled datasets from inconsistent sampling time points. Four different incubation experiments were conducted with freshly collected Norwegian seawater and either Grane and Troll oil dispersed with Corexit 9500. Incubations were conducted at two different temperatures (5 °C and 13 °C) over a period of 64 days. Results PCA analysis of modelled chemical datasets and calculated half-lives revealed differences in the biodegradation of individual hydrocarbons among temperatures and oil types. At 5 °C, most n-alkanes biodegraded faster in heavy Grane oil compared to light Troll oil. PCA analysis of modelled microbial community datasets reveal differences between temperature and oil type, especially at low temperature. For both oils, Colwelliaceae and Oceanospirillaceae were more prominent in the colder incubation (5 °C) than the warmer (13 °C). Overall, Colwelliaceae, Oceanospirillaceae, Flavobacteriaceae, Rhodobacteraceae, Alteromonadaceae and Piscirickettsiaceae consistently dominated the microbial community at both temperatures and in both oil types. Other families known to include oil-degrading bacteria were also identified, such as Alcanivoracaceae, Methylophilaceae, Sphingomonadaceae and Erythrobacteraceae, but they were all present in dispersed oil incubations at a low abundance (< 1%). Conclusions In the current study, our goal was to introduce a comparative multivariate approach for studying the biodegradation of dispersed oil, including curve-fitted models of datasets for a greater data resolution and comparability. By applying these approaches, we have shown how different temperatures and oil types influence the biodegradation of oil in incubations with inconsistent sampling points. Clustering analysis revealed further how temperature and oil type influence single compound depletion and microbial community composition. Finally, correlation analysis of degraders community, with single compound data, revealed complexity beneath usual abundance cut-offs used for microbial community data in biodegradation studies.

Oil type and temperature dependent biodegradation dynamics. Combining chemical and microbial community data through multivariate analysis.