Atmos. Chem. Phys., 20, 14139–14162, 2020 https://doi.org/10.5194/acp-20-14139-2020 © Author(s) 2020. This work is distributed under the Creative Commons Attribution 4.0 License. The determination of highly time-resolved and source-separated black carbon emission rates using radon as a tracer of atmospheric dynamics Asta Gregoricˇ1,2, Luka Drinovec1,2,3, Irena Ježek1, Janja Vaupoticˇ4, Matevž Lenarciˇ cˇ5, Domen Grauf5, Longlong Wang2,6, Maruška Mole2,7, Samo Stanicˇ2, and Griša Mocnikˇ 1,2,3 1Aerosol d.o.o., Ljubljana, 1000, Slovenia 2Centre for Atmospheric Research, University of Nova Gorica, Nova Gorica, 5000, Slovenia 3Department of Condensed Matter Physics, Jožef Stefan Institute, Ljubljana, 1000, Slovenia 4Department of Environmental Sciences, Jožef Stefan Institute, Ljubljana, 1000, Slovenia 5Aerovizija d.o.o., Ljubljana, 1000, Slovenia 6School of Mechanical and Precision Instrument Engineering, Xi’an University of Technology, Xi’an, 710048, China 7Quasar Science Resources S.L., Madrid, 28232, Spain Correspondence: Asta Gregoricˇ ([email protected]) Received: 8 October 2019 – Discussion started: 3 December 2019 Revised: 28 September 2020 – Accepted: 8 October 2020 – Published: 21 November 2020 Abstract. We present a new method for the determination of compared to the fraction of its emission rate. Coupling the the source-specific black carbon emission rates. The method- high-time-resolution measurements of black carbon concen- ology was applied in two different environments: an urban tration with atmospheric radon concentration measurements location in Ljubljana and a rural one in the Vipava valley can provide a useful tool for direct, highly time-resolved (Slovenia, Europe), which differ in pollution sources and to- measurements of the intensity of emission sources. Source- pography. The atmospheric dynamics was quantified using specific emission rates can be used to assess the efficiency the atmospheric radon (222Rn) concentration to determine of pollution mitigation measures over longer time periods, the mixing layer height for periods of thermally driven plan- thereby avoiding the influence of variable meteorology. etary boundary layer evolution. The black carbon emission rate was determined using an improved box model taking into account boundary layer depth and a horizontal advec- tion term, describing the temporal and spatial exponential 1 Introduction decay of black carbon concentration. The rural Vipava valley is impacted by a significantly higher contribution to black Black carbon (BC), an important component of fine partic- carbon concentration from biomass burning during winter ulate matter in the atmosphere, significantly contributes to (60 %) in comparison to Ljubljana (27 %). Daily averaged the climate forcing by aerosols (Pöschl, 2005; Bond et al., black carbon emission rates in Ljubljana were 210 ± 110 2013; IPCC, 2013) and is an important air pollutant, associ- and 260 ± 110 µgm−2 h−1 in spring and winter, respectively. ated with undesirable health outcomes (Janssen et al., 2011; Overall black carbon emission rates in Vipava valley were WHO, 2012). Since BC is a chemically inert primary pollu- only slightly lower compared to Ljubljana: 150 ± 60 and tant, it can be used as a good measured indicator of emissions 250 ± 160 µgm−2 h−1 in spring and winter, respectively. Dif- and can provide valuable information to authorities in the im- ferent daily dynamics of biomass burning and traffic emis- plementation and evaluation of air quality action plans, by sions was responsible for slightly higher contribution of indicating the strength of different emissions sources (e.g., biomass burning to measured black carbon concentration, Reche et al., 2011; Titos et al., 2015). On the other hand, emission inventories provide important information for cli- Published by Copernicus Publications on behalf of the European Geosciences Union. 14140 A. Gregoricˇ et al.: Source-separated black carbon emission rates mate models by providing data about the changing pattern A notable contribution of wood smoke was also observed in of BC emissions, its major sources and its historical evo- Slovenian urban (Ogrin et al., 2016) and rural areas (Wang lution. From the perspective of short-term local air quality et al., 2019), responsible for air quality deterioration espe- prediction, improving local diurnal and seasonal patterns of cially in geographically constrained areas such as valleys and BC emissions would greatly benefit the model prediction per- basins. formance. Although atmospheric chemical transport models To assess the efficiency of abatement measures aiming to based on the fundamental description of atmospheric phys- improve air quality, concentration of pollutants is usually ical processes can improve the knowledge about temporal measured before and after the measures are implemented, evolution of BC emissions at the modelled area, they require in order to quantitatively determine the reduction of pollu- comprehensive input data of atmospheric processes (Seinfeld tant concentration. However, this approach can be biased due and Pandis, 2016). to changes of micrometeorology of the planetary boundary Bottom-up emission inventories rely on information on the layer (PBL) which plays an important role in controlling time amount of used fuel combined with fuel-specific emission evolution of pollutant concentration. Therefore, assessment factors (e.g., Bond et al., 2007; Bond et al., 2013; Klimont et of BC emission rate requires decoupling of meteorologically al., 2017). Although current emission inventories agree quite driven variation from the dynamics of the sources. On di- well on the main emission sources and regions, there exist urnal timescales, atmospheric stability and dynamics play a significant uncertainties in the emission factors and activity key role in the variability of primary inert pollutants (e.g., data, used for emission calculation, with recent observation- Quan et al., 2013; McGrath-Spangler et al., 2015; Tang et ally constrained estimations much higher than the ones tra- al., 2016), such as BC, and are affected by them (e.g., Fer- ditionally used (Sun et al., 2019). In contrast to bottom-up rero et al., 2014). The evolution of the planetary boundary emission inventories, top-down constrained methods (such layer (PBL), the lowest part of the troposphere, is driven as inverse modelling) focus on minimizing the difference by convective heat transfer from the ground surface and by between simulated pollutant concentration, based on esti- mechanical mixing (due to wind shear and surface rough- mated emission flux, and measured pollutant concentration ness), which are responsible for the formation of the turbu- (Brioude et al., 2013; Wang et al., 2016b; Guerrette and lent mixing layer (ML). ML grows by entraining the air from Henze, 2017). These methods can provide spatially and tem- above and reaches its maximum depth in the late afternoon. porally better resolved assessment of pollutant emissions, in- The residual layer is formed after the decay of turbulence cluding BC, but they are influenced by different sources of shortly before sunset, with its bottom portion transformed uncertainty, mainly from the insufficient evaluation of long- into a stable nocturnal boundary layer (SNBL) during the range transport of polluted air masses. night. SNBL is characterized by stable stratification with low According to the European Union emission inventory re- mixing. Different approaches exist for mixing layer height port (LRTAP, 2018), 0.2 Tg of BC was emitted in 2016 in the (MLH) determination (Seibert et al., 2000). EU-28 region, with the dominant energy-related emissions An alternative way to overcome the difficulty associated from on-road and non-road diesel engines accounting for with the proper physical interpretation of micrometeorolog- about 70 % of all anthropogenic BC emissions (Bond et al., ical properties of the ML and dispersion characteristics is 2013). A recently updated United States black carbon emis- the use of a tracer method. The naturally occurring noble sion inventory (Sun et al., 2019) pointed out a decreasing radioactive gas radon (222Rn) has been applied in the past trend of BC emissions from 1960 to 2000, dominated by the for different studies. Radon characteristics, its emanation vehicle, industrial and residential sectors. Traffic-related BC from rocks and its transport in rocks, soil (e.g., Etiope and emission primarily dominates particulate matter (PM) emis- Martinelli, 2002) and the atmosphere (e.g., Williams et al., sion, especially in major cities (e.g., Pakkanen et al., 2000; 2011; Williams et al., 2013), were comprehensively studied Klimont et al., 2017). Recently, biomass combustion for res- in the past. Radon was used to study long-range transport idential heating has been promoted under the label of renew- of air masses (Hansen et al., 1990; Crawford et al., 2007), able fuel and additionally increased due to economic crises PBL characteristics (e.g., Griffiths et al., 2013; Williams and increase in other fuel prices (Crilley et al., 2015; De- et al., 2013; Pal et al., 2015; Salzano et al., 2016; Vecchi nier van der Gon et al., 2015; Hovorka et al., 2015; Athana- et al., 2018), microclimate spatial variability (Chambers et sopoulou et al., 2017). Although several studies report a sig- al., 2016; Podstawczynska,´ 2016) and impact assessment of nificant role of wood burning emissions in BC concentra- atmospheric stability on local air pollution (Perrino et al., tions in Alpine valleys (Sandradewi