While meta-analyses are common in the health and some biological sciences, there is a lack of such analyses for petroleum-related marine research. Oil is a highly complex substance consisting of thousands of different compounds. Measurement limitations, different protocols and a lack of standards in recording and reporting various elements of laboratory experiments impede attempts to homogenize and compare data and identify trends. Nevertheless, oil toxicology research would benefit from meta-analyses, through which we could develop meaningful research questions and design robust experiments.
Here we report findings from an effort to quantitatively summarize results from oil toxicology studies on arctic and subarctic marine invertebrates. We discovered that the vast majority of studies was conducted on crustaceans, followed by molluscs. Analyzing the sensitivity of response measures across taxa we found that the most sensitive responses tend to rank low in ecological relevance, while less sensitive response measures tend to be more ecologically relevant. We further uncovered that crustaceans appear to be more sensitive to mechanically dispersed than chemically dispersed oil while the opposite seems true for molluscs, albeit not statistically significant. Both crustaceans and molluscs show a higher sensitivity to fresh than to weathered oil. No differences in the sensitivities of crustacean life stages were found. However, due to a lack of data, many questions remain unanswered. Our study revealed that while trends in responses can be elucidated, heterogeneous experimental protocols and reporting regimes prevent a proper meta-analysis.