What do correlations tell us about anthropogenic–biogenic interactions and SOA formation in the Sacramento plume during CARES?

Abstract. During the Carbonaceous Aerosols and Radiative Effects Study (CARES) the US Department of Energy (DOE) G-1 aircraft was used to sample aerosol and gas phase compounds in the Sacramento, CA, plume and surrounding region. We present data from 66 plume transects obtained during 13 flights in which southwesterly winds transported the plume towards the foothills of the Sierra Nevada. Plume transport occurred partly over land with high isoprene emission rates. Our objective is to empirically determine whether organic aerosol (OA) can be attributed to anthropogenic or biogenic sources, and to determine whether there is a synergistic effect whereby OA concentrations are enhanced by the simultaneous presence of high concentrations of carbon monoxide (CO) and either isoprene, MVK + MACR (sum of methyl vinyl ketone and methacrolein), or methanol, which are taken as tracers of anthropogenic and biogenic emissions, respectively. Linear and bilinear correlations between OA, CO, and each of three biogenic tracers, “Bio”, for individual plume transects indicate that most of the variance in OA over short timescales and distance scales can be explained by CO. For each transect and species a plume perturbation, (i.e., ΔOA, defined as the difference between 90th and 10th percentiles) was defined and regressions done amongst Δ values in order to probe day-to-day and location-dependent variability. Species that predicted the largest fraction of the variance in ΔOA were ΔO₃ and ΔCO. Background OA was highly correlated with background methanol and poorly correlated with other tracers. Because background OA was  ∼  60 % of peak OA in the urban plume, peak OA should be primarily biogenic and therefore non-fossil, even though the day-to-day and spatial variability of plume OA is best described by an anthropogenic tracer, CO. Transects were split into subsets according to the percentile rankings of ΔCO and ΔBio, similar to an approach used by Setyan et al. (2012) and Shilling et al. (2013) to determine if anthropogenic–biogenic (A–B) interactions enhance OA production. As found earlier, ΔOA in the data subset having high ΔCO and high ΔBio was several-fold greater than in other subsets. Part of this difference is consistent with a synergistic interaction between anthropogenic and biogenic precursors and part to an independent linear dependence of ΔOA on precursors. The highest values of ΔO₃, along with high temperatures, clear skies, and poor ventilation, also occurred in the high ΔCO–high ΔBio data set. A complicated mix of A–B interactions can result. After taking into account linear effects as predicted from low concentration data, an A–B enhancement of OA by a factor of 1.2 to 1.5 is estimated.