Abstract
Odorous emissions are increasingly recognized as a critical environmental issue due to their pronounced influence on public perception and quality of life. Among the various sources, crude oil storage facilities are frequently associated with odour nuisance, while their emission characterization remains inherently complex. External floating-roof storage tanks (EFRT), in particular, are characterized by exclusively diffuse emissions, which limit direct on-site sampling and require the application of indirect methodologies for the estimation of the odour emission flux. Additional complexity arises from the operational variability of these systems, as different crude oils—each exhibiting distinct and highly variable odour characteristics—may be stored within the same tank over the course of a year. Moreover, Volatile Organic Compounds (VOC) losses from storage tanks are commonly estimated using the US EPA AP-42 methodology, which provides annual average emission rates. Such long-term averaged estimates may be poorly representative for odour impact assessments, which are typically based on short-term peak concentrations, and may therefore underestimate the occurrence and intensity of odour annoyance episodes. This paper presents a case study referring to a crude oil storage site equipped with EFRTs. The objective is to provide an in-depth characterization of this highly complex emission scenario through refined emission estimation and atmospheric dispersion simulations with the CALPUFF model. Experimental approaches are proposed for the determination of the odour potential of the crude oil mixtures stored in the tanks, explicitly accounting for the temporal variability of stored products. In parallel, a practical methodology for the estimation of VOC wall losses, associated with liquid withdrawal operations, is developed, avoiding long-term averaging assumptions. Overall, the conclusions of this study highlight that the odour potential of the different crude oils is highly heterogeneous, emphasizing the importance of accounting for the quantity of each crude type stored within the tanks. Furthermore, wall losses associated with withdrawal operations are shown to be highly intermittent, characterized by periods of negligible emissions alternated with short-term emission peaks. Finally, the odour impact assessment shows that the comparison between the 98th percentile and the maximum concentrations leads to substantially different impact patterns, mainly driven by the adoption of time-resolved odour emission fluxes.
