Inter-comparison of black carbon measurement methods for simulated open biomass burning emissions

Biomass burning (BB) is a major source of black carbon (BC), but comparing BC content of different smoke-impacted air masses may be uncertain if different measurement techniques are used to quantify the BC, or if non-BC fractions influence a given measurement. To investigate these potential issues, five instruments reporting BC were compared in well-mixed smoke during the FIREX laboratory campaign in 2016, including two filter-based absorption instruments; one in situ absorption instrument; a laser-induced incandescence instrument; and a thermal-optical instrument. BB aerosols were generated using fuels common to wildfires in the Western US in a relatively controlled environment, with BC concentrations ranging from roughly 10–100 μg m⁻³ (55 total fires). Applying the Bland-Altman graphical approach, systematic biases and proportional biases were identified between the selected reference instrument (in situ absorption) and the other four instruments. BC emission factors (EF_BC) derived from the thermal-optical instrument, laser-induced incandescence instrument, and filter-based absorption instruments were, on average, 83%, 39% and 66%, greater than the in situ absorption instrument, respectively. To understand why these differences exist, principal component analysis combined with a K-means clustering algorithm was implemented to group different fires into three clusters based on several co-dependent fire-related parameters (modified combustion efficiency (MCE), single scattering albedo (SSA) at 870 nm, organic carbon/elemental carbon ratio (OC/EC ratio), and absorption Ångström exponents (AAE)); clusters are nominally referred to as “Black”, “Mixed”, and “Brown” based on the mean SSA and AAE values for each. The best agreement among all instruments was observed for the “Black” cluster (mean EF_BC ratio = 1.89, for the fires with mean SSA = 0.31 and AAE = 1.44); this agreement worsened for the “Mixed” (mean EF_BC ratio = 2.94, for the fires with mean SSA = 0.80 and AAE = 1.92) and “Brown” clusters (mean EF_BC ratio = 3.12, for the fires with mean SSA = 0.96 and AAE = 2.50), likely due to the increased presence of externally (or internally) mixed aerosols that altered the chemical and optical properties of the aerosols. In general, the discrepancies observed among the BC instruments from this work agree with or slightly exceed the ones from previous ambient and laboratory studies. Care should be taken when interpreting different BC measurements in BB smoke because large artifacts can occur due to co-emitted materials.