Impact of the New South Wales fires during October 2013 on regional air quality in eastern Australia
Smoke plumes from fires contain atmospheric pollutants that can be transported to populated areas and effect regional air quality. In this paper, the characteristics and impact of the fire plumes from a major fire event that occurred in October 2013 (17-26) in the New South Wales (NSW) in Australia, near the populated areas of Sydney and Wollongong, are studied. Measurements from the Fourier Transform InfraRed (FTIR) spectrometer located at the University of Wollongong allowed a calculation of specific emission factors (EFs) in terms of grams per kilogram of dry fuel burned: 1640 g kg-1 of carbon dioxide; 107 g kg-1 of carbon monoxide; 7.8 g kg-1 of methane; and 0.16 g kg-1 of nitrous oxide. These EFs have then been used to calculate daily fire emissions for the NSW fire event using the APIFLAME emissions' model, leading to an increase of 54% of CO emitted compared to calculations with EFs from Akagi et al. (2011), widely used in the literature.Simulations have been conducted for this event using the regional chemistry-transport model (CTM) CHIMERE, allowing the first evaluation of its regional impact. Fire emissions are assumed well mixed into the boundary layer. The model simulations have been evaluated compared to measurements at the NSW air quality stations. The mean correlation coefficients (R) are 0.44 for PM10, 0.60 for PM2.5 and 0.79 for CO, with a negative bias for CO (-14%) and a positive bias for PM2.5 (64%). The model shows higher performance for lower boundary layer heights and wind speeds. According to the observations, 7 days show concentrations exceeding the air quality Australian national standards for PM10, 8 days for PM2.5. In the simulations, 5 days are correctly simulated for PM10, 8 days for PM2.5. For PM10, the model predicts 1 additional day of exceedance (one false detection). During this fire episode, inner Sydney is affected during 5 days by PM exceedances, that are mainly attributed to organic carbon in the model simulations. To evaluate the influence of the diurnal variability and the injection heights of fire emissions, two additional simulations were performed: one with all fire emissions injected below 1 km (CHIM_1 km), since satellite observations suggest low injection for this fire case, and one with a diurnal profile (CHIM_diu) adjusted to best match surface observations closest to the fires. CHIM_1 km displays less bias and root mean square error, and CHIM_diu presents a good agreement for hourly statistics for stations where peaks of PM are well captured, but enhances the differences when a peak is overestimated by the model. This sensitivity analysis highlights significant uncertainties related to these two key fire parameters (which add up to uncertainties on emissions), resulting in variations on concentrations of PM and CO.