Characteristics of Atmospheric Ice Nucleating Particles Associated with Biomass Burning in the US: Prescribed Burning and Wildfires
July 08, 2013
Christina McCluskey
Committee: Sonia Kreidenweis (advisor), Jeff Pierce, Paul DeMott, John Volckens (Environmental and Radiological Health Sciences)
Abstract
Insufficient knowledge about the sources and number concentrations of atmospheric ice nucleating particles (INP) leads to large uncertainties in understanding the interaction of aerosols with cloud processes, such as cloud life time and precipitation rates. An increasingly important source of aerosol in the United States is biomass burning, particularly in the form of prescribed burning and wildfires in the southeastern and western U.S., respectively. Prior field and laboratory observations have suggested that biomass burning can be a source of INP. However, emissions from biomass burning are complex, varying with combustion efficiency, fuel type and plume age and dilution. Thus, this potentially important source of INP is poorly characterized. This study utilizes measurements of INP from a diverse set of biomass burning events to better understand INP associated with biomass burning in the U.S.Prescribed burns in Georgia and Colorado, two Colorado wildfires and a laboratory burn were monitored for INP number concentrations (nINP) using the Colorado State University continuous-flow diffusion chamber (CFDC) to activate INP in the condensation/immersion freezing nucleation mode. Additional measurements included total particle number concentrations, number concentrations of particles with diameters larger than 500nm (n500nm), aerosol mass concentrations, carbon monoxide concentrations and chemically-speciated bulk aerosol filter samples. Additionally, activated INP were collected onto TEM grids downstream of the CFDC, isolating INP for single particle chemical and morphological analyses. These fires varied by fuel type, including wiregrass, longleaf pine and ponderosa pine, and also in combustion efficiency, ranging from highly flaming to a mixture of flaming and smoldering. Additionally, plume histories were different between the fires including aged plumes from the wildfires and young smoke from the prescribed burns.
Concentrations at standardized temperature and pressure of INP measured at -30°C were averaged over the sampling periods each day and ranged from 16.2 (±34.83) to 122.64 (±100.62) L-1 during the Georgia prescribed burns and from 44.66 (±14.10) to 148.53 (±12.60) L-1 during the Colorado wildfires.
The relationship between nINP and total particle number concentrations, evident within prescribed burning plumes, was degraded within aged smoke plumes from the wildfires, limiting the utility of this relationship for comparing laboratory and field data. Larger particles, represented by n500nm, are less vulnerable to plume processing. Our measurements indicated that for a given n500nm, nINP emitted from the wildfires were nearly an order of magnitude higher than found in prescribed fire emissions. That is, nINP represented a much larger fraction of n500nm in wildfires as compared with prescribed fires. Further, an existing parameterization for “global” nINP that relates INP abundance to n500nm under-predicted and over-predicted INP emitted from wildfires and prescribed burns, respectively. Reasons for the differences between INP characteristics in these emissions were explored, including variations in combustion efficiency, fuel type, transport time and environmental conditions.
Combustion efficiency and fuel type were eliminated as controlling factors by comparing samples with contrasting combustion efficiencies and fuel types. Transport time was eliminated because the expected impact would be to reduce n500nm, thus resulting in the opposite effect from the observed change. Bulk aerosol chemical composition analyses appear to support the role of elevated soil dust particle concentrations released from the fires as a potential INP. However, predictions from the Naval Aerosol Analysis and Prediction System model suggest generally elevated dust mass concentrations during the wildfire periods, as usual in this region during the late Spring period at the sampling location. Thus, it was not possible in all cases to attribute the INP source specifically to the wildfires versus long-range transport.
The chemical compositions of INP were probed more directly via TEM imaging. Single particle analyses of residual INP showed that they comprised various C-containing particles types, but with a higher abundance of mineral and metal oxide containing INP in emissions from flaming phase combustion. Fractal soot was found as an INP type comprising up to 50% of all INP in young smoke emissions from the Georgia prescribed burning of predominately wiregrass fuel. In a series of laboratory combustion experiments, the use of a new instrumental set up, pairing the CFDC with a single particle soot photometer, revealed up to a 60% decrease in active INP after the removal of refractory black carbon from smoke aerosol emitted from highly flaming burn of wiregrass, supporting the idea that soot particles can serve as INP in some fire emissions. The presence of soil minerals were clearly evident in TEM images of sample taken during the wildfires. This result demonstrates that the INP observed in the wildfires were influenced by other sources not active in the laboratory or prescribed burns.