2019 Presentation Abstracts and PDFs are listed in order of occurrence at the conference.
Photochemical reaction and diffusion (PRAD) model: The influence of viscosity on photochemistry in single aerosol particles
Summary: It has been recently recognized that aqueous aerosol particles may attain highly viscous, semi-solid, or even glassy states under a wide range of atmospheric conditions. While the impact of reduced mobility in highly viscous particles on dark chemistry has received substantial attention, systematic studies on the effect of high viscosity on photochemical processes are scarce. To fill the gap, we developed a multi-layered photochemical reaction and diffusion (PRAD) model targeted specifically at the iron(III)-citrate/citric acid photochemistry system, to assess how viscosity and diffusivity influence photochemical processes. The PRAD model includes main equilibria, chemical reactions, and diffusivities of major species in the system. It is discretized in time steps as well as spatially by dividing the particle into concentric shells. It allows to simulate concentration gradients of each species inside an aqueous iron(III)-citrate/citric acid particle, and to derive its size and mass changes. While iron(III)-citrate photochemistry is reasonably well established, there are still a number of ill-constrained parameters in the model, such as the diffusivity of CO2 and O2, the re-oxidation pathways and their reaction constants of iron(II) to iron(III). We conducted several dedicated experiments with single aqueous iron(III)-citrate/citric acid particles levitated in electrodynamic balance to determine some of these parameters and applied the model to atmospheric conditions.
Improved parameterizations for neutral and ion-induced H2SO4-H2O particle formation
Summary: Sulfuric acid is a key compound in atmospheric new particle formation. Sulfuric acid-water particle formation is likely responsible for the generation of a persistent aerosol layer in the stratosphere. Global atmospheric models suggest it to be abundant in the free troposphere, extending to the boundary layer in polar regions and providing a significant source of global CCN. Sulfate aerosols are also important components of present and past atmospheres of other planets, such as Venus and Early Mars. Large-scale models for the Earth’s and other planets’ atmospheres need parameterized representations of particle formation for a wide range of atmospheric conditions. We have developed new neutral and ion-induced sulphuric acid-water particle formation parameterizations for this purpose (Määttänen et al., 2018). The new parametrizations are based on improved Classical Nucleation Theory (Merikanto et al. 2016), validated against CLOUD experiments (Duplissy et al., 2016). The neutral parameterization extends the validity ranges of the Vehkamäki et al. (2002) parameterization, and a new parameterization for ion-induced particle formation has been developed. The parameterizations capture the experimental formation rates and provide reliable predictions for a wide range of conditions outside the experimental range. We thank Magnus Ehrnrooth Foundation, Programme national de planétologie, Institut Pierre Simon Laplace, and the ERC - Starting MOCAPAF grant 257360 for funding.
Towards understanding heterogeneous ice nucleation on realistic silver iodide surfaces from atomistic simulation
Summary: Small particles of silver iodide (AgI) are known to have excellent ice nucleating capabilities and have been used in rain seeding applications. It is widely believed that the silver terminated (0001) surface of β-AgI acts as a template for the basal plane of hexagonal ice. However, the (0001) surface of ionic crystals with the wurtzite structure is polar and will therefore exhibit reconstructions and defects. Here, we use atomistic molecular dynamics simulations and to study how the presence of defects on AgI (0001) affects the rates and mechanism of heterogeneous ice nucleation at moderate supercooling at -10 ºC. We first consider AgI (0001) surfaces exhibiting vacancies, step edges, terraces, and pits, and compare them to simulations of the corresponding ideal surface. We find that the presence of defects reduces both the nucleation and growth rates, by up to an order of magnitude, which can be understood from the atomistic details extracted from the simulations. Finally, we consider more realistic AgI (0001) surfaces with 5x5 surface reconstructions that cancel the surface dipole, and report on their ice nucleating abilities.
Efficient model-based retrieval of aerosol properties from composition-resolved thermal desorption measurements
Summary: Using the filter inlet for gas and aerosols (FIGAERO), coupled to a chemical ionization mass spectrometer, composition-resolved thermograms can be measured for a major fraction of an organic aerosol (OA) sample through temperature-controlled evaporation. FIGAERO has been gaining wide-spread use in both lab and field applications, but the quantitative interpretation of the measured thermograms remains a challenge. We developed a detailed model of OA evaporation as it occurs in the FIGAERO, describing the desorption of aerosol constituents, as well as the formation and dissociation of oligomers. The model succeeds in reproducing the wide variety of observed thermogram shapes and provides quantitative interpretations, including volatilities of OA constituents and decomposition processes. As typical OA desorbs as hundreds of individual compositions, a practical issue is the fitting of free model parameters to an equally large number of observed thermograms. To tackle this challenge, we have adapted a covariance matrix adaptation evolution strategy (CMA-ES) as an efficient optimization algorithm. The algorithm adjusts the number of free model parameters dynamically and fits them to the observations. The need of manual user input is greatly reduced, and the simulation of complete FIGAERO datasets becomes practical. We apply our model and optimization routine to OA formation experiments in two substantially different chambers, as well as to ambient observations of boreal-forest OA.
Impacts of phase state on multiphase reactivity and partitioning of secondary organic aerosols
Summary: Secondary organic aerosols (SOA)can occur in amorphous solid or semi-solid phase states depending on chemical composition, relative humidity (RH), and temperature. The phase transition between amorphous solid and semi-solid states occurs at the glass transition temperature (Tg). We have recently developed a method to estimate Tg of pure compounds based on elemental composition. We estimated the characteristic timescale for secondary organic aerosols (SOA) to achieve gas–particle equilibrium under a wide range of temperatures and relative humidities using a state-of-the-art kinetic flux model. Equilibration timescales were calculated by varying particle phase state, size, mass loadings, and volatility of organic compounds in open and closed systems. Model simulations suggest that the equilibration timescale for semi-volatile compounds is on the order of seconds or minutes for most conditions in the planetary boundary layer, but it can be longer than 1 h if particles adopt glassy or amorphous solid states with high glass transition temperatures at low relative humidity. The dependence of equilibration timescales on both volatility and bulk diffusivity provides critical insights into thermodynamic or kinetic treatments of SOA partitioning for accurate predictions of gas- and particle-phase concentrations of semi-volatile compounds in regional and global chemical transport models. By combining kinetic multilayer modeling and laboratory experiments, we demonstrate that phase separation and slow diffusion significantly affect multiphase reactivity of polycyclic aromatic hydrocarbons against ozone and gas uptake of amines.
Vapor pressures of substituted naphthalenes derived from diffusion-controlled evaporation rates of single particles levitated in an electrodynamic balance.
POSTER PRESENTATION: The partitioning of compounds between the aerosol and gas phase is essential for the formation and fate of secondary organic aerosol. To predict atmospheric partitioning based on explicit models for gas phase chemistry, saturation vapor pressures of the relevant compounds need to be estimated. The further development of such models requires reliable data sets of saturation vapor pressures. Often, vapor pressures of semi-volatile and low volatile compounds reported in the literature differ by several orders of magnitude between measurement techniques. Recently, a debate evolved concerning the saturation vapor pressures of substituted naphthalenes with a difference up to 4 orders of magnitude between measurements using a Knudsen effusion mass spectrometer setup (KEMS) and predictions using quantum chemistry calculations (COSMOtherm). By measuring diffusion-controlled evaporation rates of single particles levitated in an electrodynamic balance, we determined pure compound saturation vapor pressures and enthalpies of vaporization for 2-naphtol, 1,3-dihydroxynaphthalene, 2,3-dihydroxynaphatalene and 1,7-dihydroxynaphthalene. Our data indicate that adding one OH-group to 2-Naphthanol lowers the vapor pressure by about 4 orders of magnitude, consistent with the KEMS measurements and in disagreement to COSMOtherm predictions. Interestingly, the relative change in vapor pressure between the different dihydroxynaphathlenes is well captures by COSMOtherm calculations.
Probing the Volatility of α-Pinene SOA: From Molecular Composition to Bulk Volatility
Summary: Secondary organic aerosol (SOA) makes up a significant fraction of the fine particulate matter in the atmosphere, and has a significant impact on human health and the earth’s climate. As a result, the time frame of the aerosol lifecycle must be understood to evaluates its impact. A major constraint of this lifecycle is the volatility of SOA. Previously, there have been many efforts to parametrize volatility based on molecular formula, atomic ratios, or functional groups, however, the tools to determine the molecular composition in real-time for SOA have been lacking. Extractive electrospray ionization time-of-flight mass spectrometry (EESI-TOF) is a soft ionization technique and makes it possible to determine the molecular composition of aerosols without fragmentation with high time-resolution. Detailed chemical information provided by the EESI-TOF makes it possible to test current parameterizations based on molecular information against measurements of volatility. For the base case of SOA formed from α-pinene + O3, parameterizations of volatility based on molecular formula can provide a reasonably good agreement with measurements. However, when measuring the composition during isothermal dilution models predict that molecules with the least number of oxygen should evaporate first, while those with more oxygen should remain in the particle phase. The predicted behavior is observed to some extent during isothermal dilution in an atmospheric simulation chamber, but in a dry evaporation chamber differences are observed. Finally, the composition of α-pinene SOA changes when exposed to UV lights due to photolysis. There is a significant difference in the measured volatility before and after photolysis. After photolysis of α-pinene SOA, there is a dramatic reduction in its volatility. The change in volatility isn’t captured by the changing composition without either changing physical properties (i.e. viscosity) or the molecular parameterizations used. A change in parameterization can be rationalized by the fact that only some functional groups are sensitive to photolysis. A similar change in volatility after photolysis is also found for SOA formed form α-pinene + NO3.
SOA formation in a low NOx biogenic environment perturbed by an urban plume: an explicit modeling study of the GoAmazon 2014/5 field campaign
Summary: The GoAmazon field campaign took place downwind of Manaus (Brazil) in 2014 and 2015. This important urban area is in a unique location isolated in the middle of a vast expanse of the Amazon rainforest. This was the ideal place to focus on the interplay between biogenic air masses and anthropogenic emissions. The goal of this work is to evaluate the ability of the Generator for Explicit Chemistry and Kinetics of Organics in the Atmosphere (GECKO-A) to reproduce observed SOA mass concentrations. GECKO-A is a computing tool designed to systematically generate hyper-explicit chemical schemes describing the atmospheric oxidation of hydrocarbons based on fundamental principles and laboratory data. We estimated biogenic emissions from the rainforest and anthropogenic emissions from the city to run a boxmodel including the full oxidation of 65 primary hydrocarbons (~14 million reactions involving 4 million chemical species) simulating the advection of an air parcel over the rainforest and Manaus. The results show that the explicit model predicts measured SOA mass within a factor of two. It is also able to reproduce the observed anthropogenic enhancement of SOA mass. Comparison with a Volatility Basis Set parameterization developed for this same situation lead us to hypothesize that the discrepancies between the model and the measured mass may come from a misrepresentation of SOA formation from monoterpenes oxidation.
A predictive group-contribution model for the viscosity of aqueous organic aerosol
Summary: Highly viscous organic aerosol particles behave differently than their well-mixed, liquid-like analogues. Viscosity affects molecular diffusion in particles, the characteristic time scale for equilibration of water and organics between the gas and particle phases, heterogeneous reactions, and the ice nucleation potential. Furthermore, viscous organic aerosol phases may impact the interpretation of field and laboratory aerosol measurements given the relatively short residence times in instruments. We have developed AIOMFAC–VISC, a novel predictive mixture viscosity model based on the thermodynamic activity coefficient model AIOMFAC. AIOMFAC–VISC accounts for chemical structure effects on the functional group level and enables predicting the viscosity of organic aerosols covering over 15 orders of magnitude from liquid-like to semi-solid and glassy phase states. In this presentation, we will discuss the theoretical basis of AIOMFAC-VISC, key assumptions, implementation considerations, limitations of our model, as well as its validation by experimental data. The combination of AIOMFAC-based equilibrium calculations with the viscosity model allows for simultaneous predictions of phase compositions and viscosity changes as a function of temperature and relative humidity. We will show that this model successfully predicts the large variations in organic phase viscosity as monoterpene- or toluene-derived aerosols are dried from high to low humidity at different temperatures.
Kinetic modelling of secondary organic aerosol (SOA) formation: connecting the data points
Summary: Despite extensive research over the past decades, the formation of secondary organic aerosol (SOA) remains a major point of uncertainty in quantifying the global aerosol burden. Parameters typically used in these calculations are product branching ratios and volatility distributions derived from environmental chamber experiments. However, chamber experiments often use concentrations that are higher than those typically encountered in the atmosphere, investigate a single precursor at a time and operate at room temperature. To extrapolate the laboratory conditions to the atmosphere, models rely on chemical mechanisms and models that were parameterized using a rather small number of experiments and range of reaction conditions. We postulate that, to encompass the diversity of reaction conditions in the atmosphere and enhance our understanding of SOA, new modelling tools are needed that reconcile and cross-compare large sets of experimental data. Thus, in this presentation, we show the results of a detailed cross-comparison between environmental chamber experiments over a wide range of reaction conditions with the goal of developing a chemical mechanism and kinetic process model that describes the observed SOA yields and composition. We test whether SOA yields are linearly additive in the presence of multiple precursors and assess whether temperature-dependence of SOA yields can be accurately described by temperature-dependent reaction rate coefficients and volatility distributions. We explore avenues how Machine Learning can be used in optimizing the chemical mechanism and model or to reduce computational complexity associated with the model calculations.
An atmospheric aqueous-phase kinetic and CARPAM modelling study of methoxyphenolic species from biomass burning
POSTER DISPLAY Summary: Biomass burning (BB) is a significant air pollution source, with global, regional and local impacts on air quality, public health and climate. Methoxyphenols, which are emitted through biomass burning, are important species in atmospheric chemistry. In the present study, the temperature-dependent aqueous-phase OH radical reactions of six methoxyphenols [guaiacol (2-MP), creosol (Cre), syringol (Syr), 3-methoxyphenol (3-MP), 4-(2-methoxyethyl)phenol (MEP) and 3-methoxycatechol (3-MC)] and two related phenols [(2-ethylphenol (2-EP) and resorcinol (Res)] are investigated through laser flash photolysis and the density functional theory calculations. The rate constants obtained are in a range of 1.1 - 1.9 1010 L mol−1 s−1 with k(3-MC) > k(Cre) ≈ k(Syr) ≈ k(MEP) > k(Res) > k(3-MP) > k(2-EP) ≈ k(2-MP). A structure-function relationship is established through the measurement result, which could be used for predicting unknown rate constants of other phenolic compounds. The obtained kinetic data are preliminary implemented in the chemical aqueous-phase radical mechanism (CAPRAM) to model the aqueous-phase processes of the major methoxyphenolic compounds in the atmosphere. With the development of the oxidation mechanism, a new biomass burring module (addressed as CAPRAM-BB1.0) is being established, which enhances our understanding of the atmospheric process of biomass burning emitted aerosols.
The State of Acidity in the Atmosphere: Particles and Clouds
Summary: Acidity, defined as pH or the negative logarithm of the H+ ion activity, is a central concept of natural water chemistry. In the atmosphere, acidity governs gas-particle partitioning of semivolatile gases as well as chemical reaction rates. It has implications for the lifetime of pollutants, deposition, and human health. Our work is a review and synthesis of the current state of knowledge on the acidity of condensed phases, specifically particles and cloud droplets, in the atmosphere. While pH is rigorously defined by thermodynamics and operationally by scientific organizations based on the activity of H+ ions, a number of approximations to pH are required for the calculation and communication of atmospheric aerosol acidity. This is because aerosol pH estimates are generally based on observationally-constrained model calculations and are thus limited by the thermodynamic treatment of the aqueous phase within the models used. Nevertheless, these approximations of pH are preferred over proxies of particle pH, such as molar ratios and charge balance estimates, as proxies lack modulation by liquid water abundance and can be overwhelmed by limitations in measurement precision. Results from two regional and two global models, as well as over 80 estimates from literature, indicate that acidic fine particles are ubiquitous. Clouds and fogs experience higher liquid water content than particles (and lower ionic strengths), and they have greater but still generally acidic pH that is quite sensitive to anthropogenic emissions of sulfur and nitrogen oxides as well as ammonia. Historical measurements indicate that cloud droplet pH has changed in recent decades in response to controls on anthropogenic emissions, while the limited trend data for particles indicates current emission reductions have not been aggressive enough to substantially change acidity in regions such as the southeast US or Canada. An examination of six current chemical transport models indicates acidity over large areas of the globe is driven by nonvolatile cations and/or the assumptions regarding the mixing state of fine particles, but limited observational data over large portions of the globe limits our ability to properly evaluate current large-scale models.
Trends in Organic Matter and Functional Groups from 2009 to 2016 in the Southeastern Aerosol Research and Characterization (SEARCH) Network
Summary: Organic matter (OM) is composed of thousands of compounds, only a fraction of which can be identified and quantified. Functional groups including aliphatic CH, alcohol OH, carboxylic acids, non-acid carbonyl and oxalates, comprise a large portion of OM and can be summed to estimate OM. Measurements of functional groups may provide a useful measurement for tracking trends in OM mass as well as the level of oxidation and sources of OM. In this work, Fourier transform infrared (FT-IR) spectrometry of Teflon filter samples is used to quantify functional groups in the SEARCH network. FT-IR analysis is quick and non-destructive, enabling the application of this method to large number of samples in air monitoring networks. The methods used here are an extension of FT-IR methods developed for the Interagency Monitoring of PROtected Visual Environments (IMPROVE) network. Based on FTIR measurements from ~5400 SEARCH samples sampled over eight years, a decrease in OM is observed in Southeastern U.S. aerosol between 2010 and 2016, which parallels the decrease in PM2.5 and sulfate concentrations during this time period. Changes in the degree of oxidation of OM, as well as seasonal composition differences will be discussed. This work highlights the value of using FT-IR spectrometry: providing a more direct and sample-specific method of estimating OM concentration and the additional composition information offered by functional group measurements.
How do we represent secondary organic aerosols in Earth System Models?
Summary: Biogenic secondary organic aerosol (SOA) is formed as a result of the atmospheric oxidation of gas-phase biogenic volatile organic compounds (BVOCs). Here, we evaluate the ability of five European Earth System Models (CNRM-ESM2-1, EC-Earth3, IPSL-CM6, NorESM1.2, UKESM1) to capture the amount, and behaviour, of biogenic SOA in the atmosphere. The ESMs cover a range of complexity in terms of their representation of the sources and processing of biogenic SOA (i.e., from a fixed climatology of SOA amount to an interactive BVOC emission scheme followed by atmospheric processing). We combine station measurements of BVOC emission and atmospheric BVOC concentrations with remotely sensed isoprene emission estimates to evaluate the models’ representation of the sources of biogenic SOA. We use organic aerosol mass measurements from a number of forested sites to evaluate the ability of the models to capture the seasonal cycle in the amount of biogenic SOA present. Finally, we combine 15-year present-day ESM simulations with observed relationships between organic aerosol mass and local temperature to explore the models’ capability to capture potential temperature driven changes in atmospheric composition.
Heterogeneous production of nitrate in extreme haze
Summary: Frequent outbreaks of prolonged haze events have become a prevalent challenge to public health and environmental management in many developing countries in Asia. China, in particular, has suffered from wintertime PM2.5 pollution over the last two decades. Some recent observational studies have suggested that nitrate production is becoming more important in haze events after strong regulations on SO2 emissions were introduced in China. Nevertheless, the formation mechanisms of nitrate and their relative importance in polluted air are still poorly understood. This is exemplified by persistent overestimate of nitrate concentration in many air quality models. In this study, we implement several updates to the heterogeneous chemistry scheme in the GEOS-Chem global chemical transport model (GC) and examine the roles of various heterogeneous mechanisms in producing inorganic nitrate in urban air. The model simulations are evaluated with the wintertime measurements of NOy and stable oxygen isotopes in nitrate in northern China. Our preliminary results suggest that the current version of GC overestimates the contribution from NO2 hydrolysis, which is the dominant nitrate formation pathway on polluted days in model, and underestimates N2O5 hydrolysis. After suppressing the uptake rate of NO2 hydrolysis in the model to better match the isotope observations, modeled HONO concentration reduces dramatically. Alternative HONO sources, such as nitrate photolysis or ground surface reactions, are explored. Overall, the updated representation of heterogeneous mechanisms in GC leads to better agreement with measurements of oxygen isotopes and reduces the model’s high bias in simulating nitrate concentration in urban air. We also investigate how these model updates may impact air quality and nitrate formation around the world.
Benchmarking aerosol models on the regional scale using a stochastic particle-resolved approach
Summary: Observational data shows that aerosol populations are complex distributions of particle size and composition. These microscale details are crucial in determining the macroscale aerosol impact on climate but are challenging to incorporate in regional and global models. These models rely on highly simplified aerosol representations that introduce structural uncertainty into the model. To quantify errors due to these simplifying assumptions, we developed a detailed particle-resolved aerosol model on the regional scale. This model is fully coupled to the WRF meteorological driver and tracks the size and composition of individual particles as they are emitted and then transformed by coagulation and condensation. Embedding the particle-resolved representation in WRF resolves the 3D spatial distribution of aerosol mixing state, with particles being transported by high-order accurate, stochastic advection algorithms. This model can be used to benchmark aerosol representations in the WRF-Chem framework, holding other simulation parameters consistent. This way, errors in bulk quantities, such as cloud condensation nuclei concentrations, can be attributed to inherent assumptions about aerosol representation. We present these differences between the particle-resolved simulation results and results using common assumptions (e.g. fully internally mixed or internally mixed within size bins) to illustrate the framework of benchmarking aerosol models on the regional scale.
Heterogeneous cloud and aerosol chemistry in the tropospheric NOy cycle
Summary: Heterogeneous surface and multiphase reactions play an important role in the atmospheric NOy cycle, thereby influencing global oxidant and aerosol concentrations. This presentation examines the roles of clouds and aerosols as sites of heterogeneous reactions in light of recent improvements modeling both. For cloud heterogeneous chemistry, we introduce entrainment-limited uptake as a new, efficient method for simulating these reactions in large-scale atmospheric models that do not spatially resolve clouds. For aerosol heterogeneous chemistry, we adopt uptake coefficients from recent laboratory and field assessments. Both improvements are included in the GEOS-Chem atmospheric chemistry model. We find that NOy reactions in clouds are more important than has been reported in past literature. Moreover, accounting for cloud uptake reduces the sensitivity of atmospheric chemistry to aerosol surface area and uptake coefficient since clouds and aerosols compete for the same NO3 and N2O5. We discuss the implications for global oxidant levels and how the entrainment-limited approach can improve modeling of other aerosol processes.
Urban pollution greatly enhances formation of natural aerosols over the pristine Amazon
Summary: Understanding how anthropogenic emissions have modified natural biogenic secondary organic aerosol (SOA) formation constitutes one of the largest uncertainties in our understanding of the radiative forcing of climate. Due to ubiquitous influence of anthropogenic emissions over most terrestrial locations in the Northern Hemisphere, it is difficult to establish baseline biogenic SOA formation i.e. biogenic SOA that would be formed in the absence of anthropogenic perturbations. The vast Amazon rainforest during its wet season is one of the few remaining places on Earth where atmospheric chemistry transitions between preindustrial-like and present-day polluted conditions and serves as a unique natural laboratory to study anthropogenic-biogenic interactions. We develop insights from several laboratory measurements to simulate SOA formation in the wet-season Amazon using a high-resolution regional model (at 2 km grid spacing) and develop mechanistic insights about the role of anthropogenic emissions in biogenic SOA formation. We perform model simulations using the community regional Weather Research and Forecasting Model coupled to chemistry (WRF-Chem) at cloud-, chemistry-, and emissions-resolving scales i.e. at 2 km grid spacing. Sensitivity simulations that turn the urban emissions on/off are performed to quantify the impacts of anthropogenic-biogenic interactions on SOA formation. We evaluate WRF-Chem simulations using aircraft-based field measurements of SOA during the Green Ocean Amazon (GoAmazon 2014/5) field campaign. Our results show that urban emissions increase concentrations of nitrogen-oxides (NOx), which cause increase in oxidant (ozone and OH radical) concentrations within the otherwise pristine Amazon. Increased oxidant concentrations catalyzed by NOx substantially increase reactions of forest organic carbon, emitted as volatile organic compounds (VOCs include isoprene, monoterpene and sesquiterpene compound classes), and thereby enhance biogenic SOA formation by 60-200% on average in plume-influence regions. Model simulated enhancements agree with those observed by the aircraft which rapidly and concomitantly measures organic aerosols in background and plume-influenced locations using the Aerosol Mass Spectrometer (AMS). Our results provide a clear picture of how anthropogenic emissions might have substantially enhanced natural biogenic SOA formation since preindustrial times on Earth.
Impacts of Spatial Distribution and Spatial Resolution of Emissions on Air Quality Model
POSTER PRESENTATION Summary: Air quality models are widely used to estimate human exposure to air pollution and to predict the health benefits of proposed emissions control programs. Air quality models predict pollutant spatial gradients that influence their overlap with human populations and therefore determine their public health impact. The accuracy of pollutant spatial fields is linked to: 1) the accuracy of the underlying emissions inventory, which often relies on spatial surrogates to determine where emissions occur within each county; and 2) the spatial resolution of the emissions inventory and air quality model, which affect the ability to capture concentration gradients around populated regions.
Here we study the impacts of emissions spatial distribution and spatial resolution on predicted pollutant concentration fields and the resulting air quality model performance. Several major surrogates are updated to improve the spatial distribution of emissions originally at 4km resolution. Surrogates including population, total housing, single-family housing, total employment, industrial employment, agricultural employment and service & commercial employment are created using high spatial resolution socio-economic data. The surrogate for construction equipment is updated using information from the California Water Resource Board NOI records and Caltrans’ on-road construction database. The Longitudinal Employer-Household Dynamics database is used to update the surrogate for industrial equipment. These major surrogates are used to allocate area emissions at 1km resolution.
Simulations were conducted to evaluate if the updated spatial surrogates and downscaled emissions improved the accuracy of predicted particulate matter concentrations during past episodes. Model predictions for future episodes were also evaluated to determine if the two changes significantly affected population exposure to pollutants under various future energy scenarios. The goal of this research is to improve the accuracy of exposure calculations allowing for more detailed analysis of historical air pollution health impacts and more realistic analysis of changes to public health associated with future emissions reductions.
ISORROPIA-MCX: Implementation of the Multicomplex Variable Method into the Aerosol Thermodynamic Model, ISORROPIA
Summary: Sensitivity analysis using atmospheric chemical transport models provides a deeper understanding of how specific emissions affect pollutant concentrations. Given a model with emissions as inputs and pollutant concentrations as outputs, this analysis is achieved by computing the partial derivatives of the underlying functions with respect to their input variables. Implementing higher-order sensitivity calculations can be quite difficult even though they are important to understanding nonlinear processes. A novel approach to sensitivity analysis leverages multicomplex variables to improve accuracy and ease of use over the finite difference method, as it avoids subtractive cancellation and numerical round-off errors. Here, the multicomplex variable method (MCX) is implemented in the inorganic aerosol thermodynamic equilibrium model, ISORROPIA, which treats the Na+ - SO42- - HSO4- - NH4+ - NO3- - Cl- - Ca2+ - K+- Mg2+ - H2O aerosol system (ISORROPIA-MCX). Specifically, the first- and second-order sensitivities of an inorganic species in the aerosol or gaseous phase with respect to the total concentrations are calculated. The ability to compute higher order derivatives is useful when the functions are nonlinear because including the second-order and cross-sensitivity terms in the Taylor Series expansion increases the accuracy of the estimated effect. Implementation of ISORROPIA-MCX is beneficial because there are enough inputs and outputs to demonstrate a main advantage of the method, which is simultaneously calculating multiple sensitivities. Furthermore, since thermodynamics is a nonlinear process and ISORROPIA uses nonlinear functions, we can demonstrate the advantages of calculating higher order sensitivities using MCX. This contribution represents the first application of the multicomplex step approach in an atmospheric chemistry model, and the technique is likely to be valuable in regional or global chemical transport models where ISORROPIA is used.
Trimming the Iterative Fat of Equilibrium Thermodynamic Models Using Neural Networks
Summary: Unidirectional calculations are costly yet common for equilibrium processes, which we tried to circumvent when developing the Binary Activity Thermodynamics model (BAT). The BAT model is a water-sensitive, reduced-complexity activity coefficient model, specifically treating interactions among water and organic compounds. A Volatility Basis Set framework is combined with the BAT model, allowing predictions of RH-dependent organic co-condensation, liquid-liquid phase separation, and cloud droplet activation properties. We will discuss our application of neural networks (NNs) in the VBS+BAT model to shorten and sometimes bypass the iterative numerical solution. We use deep-learning NNs with the BAT model to find the correct organic mole fraction input, since in most applications water activity (a_w) is known but not the organic mole fraction. For example, in atmospheric thermodynamic modeling, RH is a known quantity and, for bulk equilibrium simulations, the RH in the gas phase is equal to the a_w in the liquid phase (when the Kelvin effect is negligible). The resulting BAT-NN inverts the BAT model quite well over the full a_w space up to water activities of ~0.95, above which an iterative refinement is required for good agreement with the targeted a_w. We will also discuss a similar methodology attempted to reduce the computational cost of the VBS+BAT equilibrium solver. The initial guess from the neural network is successful in approximating the non-trivial equilibrium solution, which facilitates using an efficient, though less robust, gradient descent method. This two-step process (first NN, then numerical equilibrium solver) could be applied to additional equilibrium processes that are present in air quality models.
Bayesian state-estimation of nucleation, growth and deposition rates from measured aerosol size distribution dynamics
A Mass-Conserving Machine Learning Algorithm for Atmospheric Chemistry
Summary: Gas phase chemistry and aerosol particle dynamics consume the majority of computer time in Chemical Transport Models (CTMs) of urban and regional air pollution. Likewise, Global Circulation Models (GCMs) heavily parameterize this subgrid process to make them computationally tractable while compromising detailed description of the physics and chemistry. If machine learning techniques could be used to relate the concentrations of the chemical constituents at one operator-splitting time step to those at the next time step, the computational speed of these models would be dramatically increased. CTMs and GCMs use operator splitting to solve the operators that describe the governing physics and chemistry. In CTMs the operator splitting time step ranges typically from 0.1 to 1 hour. If a machine learning algorithm were used to describe the chemistry with 99% accurate, then for a 10-hour run with an 0.1-hour time step, the results could be 100% off. Even if the machine learning algorithm were 99.9% accurate, the errors can come to dominate. This is especially a problem if mass balance is violated. That is, if the 1% error or 0.1% error resulted in systematic creation or destruction of one or more chemical constituents, the answers produced could be dubious at best. In this talk, we will describe a mathematical framework for assuring 100% mass balance regardless of the machine learning algorithm employed or the accuracy of this algorithm.
Coarse-Graining of Aerosol Mixing State Metrics Empowered by Machine Learning
Summary: Atmospheric aerosols vary in their chemical composition, resulting in different “aerosol mixing states”, which we define as the distribution of aerosol chemical species among particles in a population, and the way these species are arranged within the particles. Oversimplified assumptions of aerosol mixing state in atmospheric models can introduce errors in estimations of weather and climate-relevant aerosol properties. A realistic representation of the aerosol mixing state can be achieved in principle with a Particle-resolved Monte Carlo (PartMC) numerical model but at a computational cost that is prohibitive for direct implementation in Earth System Models (ESMs). In this presentation, we will introduce a method for estimating the spatial and temporal distribution of aerosol mixing state in ESMs using the “aerosol mixing state index” as a mixing state metric. We will present a data-driven workflow, leveraging machine learning algorithms to emulate detailed PartMC simulations in terms of the aerosol mixing state quantification. The coarse-grained modeling will be applied to large-scale EMSs such as Community Earth System Model (CESM) to demonstrate the global distribution of aerosol mixing state indexes. This workflow will enable key insights into the importance of machine learning to aerosol research.
JLBox: A Julia implementation of a fast mixed-phase atmospheric chemistry model with adjoint sensitivity analysis
POSTER PRESENTATION: Mechanistic models of atmospheric aerosol particles still require appropriate strategies to mitigate both chemical and numerical complexity. Whilst complexity reduction schemes have been used in the past, retaining a full complexity benchmark framework that can exploit emerging software and hardware paradigms will enable the community to embrace new experimental findings and assess the relevance of parameterisations used in large scale models. In this study we present a new box-model written in the Julia language: JLBox. Julia provides an elegant and expressive new way to write software, without tradeoffs between performance and expressiveness. It enables a differentiable library ecosystem. In this study we demonstrate the simulation of the entire master chemical mechanism, with explicit partitioning to an aerosol phase, including an adjoint sensitivity analysis for quantifying the importance of model parameters and for data assimilations. We discuss how automatic differentiation provides free Jacobian matrices for performing (local) sensitivity analysis and accelerating implicit ODE solvers where sparse matrix operations for gas kinetics enables JlBox to perform simulations on large scale mechanisms. Results show how JlBox is 100% faster than PyBox, a mixed Python-Fortran model, in mixed-phase simulations while producing results with relative differences below 1e-3. JlBox has much less memory consumption and much shorter compiling time for gas kinetics simulations compared to PyBox.
Development of a NOAA Emissions and eXchange Unified System
POSTER PRESENTATION: The past decade has experienced rapid advances in global aerosols and atmospheric composition (AAC) model prediction capabilities. AAC models are key components of unified forecast systems that often employ the Earth System Model Framework (ESMF; i.e., a high-performance, flexible software infrastructure for building and coupling weather, climate, and related Earth science models) for weather and climate predictions. Emissions of trace gases and primary aerosols are a critical component of AAC models and are often the most important component to ensure accurate predictions of trace species distributions. However, developing these emissions inputs to AAC models is often a laborious, time-consuming process, especially to ensure that the datasets are suitable for a range of spatial scales and applications. Furthermore, inventory-based emission inputs are subject to a bottom-up approach that is prepared separately (offline) and suffers distinct time lags from the AAC models, which affects both the timing and accuracy of trace gas predictions. In this work, the Harvard-NASA Emission Component (HEMCO) is serving as the foundation of a new unified emissions modeling framework, which is capable of utilizing numerous emissions datasets (both global and regional), can be run offline (inventory-based) or online (processed-based), is ESMF-compliant, and can be easily linked to satellite data sources. Here we present the initial development of the NOAA Emissions and eXchange Unified System (NEXUS), which will interface both offline and online with different NOAA AAC models, including both global and regional models for both operational and research-oriented applications. Preliminary developments of a comprehensive, adaptable emissions and air-surface exchange (i.e., both emissions and deposition) processing system for use in conjunction with NOAA AAC models will be shown. This includes examples of model-ready anthropogenic emissions using a combination of global and regional anthropogenic emission inventories with the NEXUS platform, and an initial assessment of NOAA AAC model simulations using these emissions. We will also present preliminary results from initial implementation of advanced inline dust and fire emissions using new NOAA products and emission models in NEXUS.
Raoult was Right After All
POSTER PRESENTATION: In 1886, Raoult published his theory of vapor pressure and therefore activity of solutions stating that the vapor pressure is proportional to the mole fraction of the solute. Measurements confirm the validity of this paradigm in dilute solutions but also show that it fails at higher concentrations. In 1908, Callendar extended this theory by positing that solutes are hydrated and that the water of hydration is effectively removed from the free water thereby changing the denominator of the mole fraction calculation. Again, measurements confirmed the validity of this hypothesis but again it fell short at yet higher concentrations. Callendar assumed that the amount of water hydrated to the solute is fixed so presumably this hydration theory failed because at higher concentration, and therefore lower water activity, the solute dehydrates, but Callendar’s theory does not encompass this dehydration. In this work we reformulate Raoult’s and Callendar’s theories by placing the hydrated water in equilibrium with the free water in solution to derive new equations for molality and solute activity as a function of water activity. These resulting equations are valid for the full range of concentration from 100% solvent to 100% solute for organics and electrolytes dissolved in water. These equations provide a robust underpinning for modeling the thermodynamics of atmospheric aerosols.
Appraisal of real-time bioaerosol detection techniques and classification algorithms
Summary: Primary biological aerosol particles (PBAP) are ubiquitous in the atmosphere and have significant impacts on plant, animal and human health. They also may play an important role in cloud-aerosol interactions through ice nucleation processes, thus quantifying PBAP emissions is critical for understanding their impact in a changing climate. Despite this need, accurately classifying and quantifying atmospheric PBAP remains a significant technical challenge, limiting their inclusion in climate and health models. Emerging commercial real-time ultraviolet light induced fluorescence instrumentation offers high spatiotemporal resolution compared to traditional offline bioaerosol detection techniques, however, PBAP classification with these methods requires a careful and considered approach to data processing to produce high quality outputs. Here we present and evaluate a variety of different data processing approaches, including supervised and unsupervised machine learning methodologies. We appraise the current state of technological outputs and provide potential end-users with a road map of future development, focusing on data product quality.
Dilution impacts on aerosol aging in biomass burning plumes: using a novel coupled aerosol, chemistry and large-eddy simulation model to learn about the impacts of dilution rates and cross-plume concentration gradients on smoke aging
Summary: Biomass burning is a major source of atmospheric particulate matter (PM) with implications for health, climate, and air quality. As biomass burning plumes are transported downwind, the particles and vapors undergo chemical and physical aging. Aerosol microphysical processes such as condensation and evaporation of vapors, particle coagulation, and new particle formation control particle number, mass, and size, all of which impact the climate and health impacts of smoke. These processes rely in part upon the plume’s dilution rate and concentrations of ambient aerosol entrained as the plume dilutes. Dilution is both a bulk and fine-scale phenomena: while the entire plume may dilute more or less as a bulk entity, gradients within the plume occur as edges dilute faster than the center of the plume. The SAM-ASP model (the large-eddy simulation model System for Atmospheric Modelling, SAM, coupled to a detailed gas and aerosol chemistry model, the Aerosol Simulation Program, ASP) provides the ability to examine both scales of dilution and their impact on smoke aging. We use SAM-ASP and measurements from recent field campaigns to understand impacts of bulk dilution and plume gradients on aerosol microphysical processes. We also consider the implications of our work within the context of global-scale models, whose grid boxes greatly exceed that of the size of smoke plumes.
Modeling Impacts on Secondary Organic Aerosol Formation from Volatile Chemical Products
Summary: Recently, it has been suggested that volatile chemical product (VCP) emissions can contribute secondary organic aerosol (SOA) formation in cities. Here we construct a VCP emissions inventory for the Continental United States, which has been evaluated with field measurements of volatile organic compounds (VOCs) using a proton transfer reaction time-of-flight mass spectrometer (PTR-ToF-MS) and whole air samples analyzed by gas chromatography-mass spectrometry (GC-MS). Mobile laboratory measurements were made in four US cities, including New York City, Chicago, Pittsburgh, and Denver. We have identified chemical markers for distinguishing VCP emissions from mobile sources, which indicate that VCP emissions comprise a larger fraction of anthropogenic VOC emissions in denser cities. We then model the VCP and mobile source emissions in the Weather Research and Forecasting with Chemistry (WRF-Chem) model, and assess sensitivities on the model for tropospheric ozone and SOA formation across US cities during the summer of 2018.
Representing sub-grid biomass burning processes in regional and global models
Summary: Biomass burning is a significant global source of aerosol number and mass, impacting climate and particulate matter concentrations. In fresh biomass burning plumes, aerosol number, mass, and size can rapidly transform through coagulation, organic aerosol evaporation, and secondary organic aerosol formation. The impact of these processes depends non-linearly on plume aerosol concentrations and dilution rates. Global and regional aerosol chemical transport models cannot explicitly resolve most biomass burning plumes. Hence, these coarse models cannot properly simulate the non-linear plume processes described above. In this presentation, we show (1) the development of a biomass-burning plume-in-grid scheme to represent sub-grid coagulation, (2) the impact of this sub-grid coagulation on aerosol direct and indirect forcing predictions, and (3) ongoing work on the in-plume evolution of aerosol mass and size through coagulation, organic aerosol evaporation, and secondary organic aerosol formation.
Beyond traffic and biogenics: Importance of cooking and volatile chemical products in the urban atmosphere
Summary: The urban atmosphere is spatially and temporally dynamic. PM concentration and composition exhibit both strong gradients at sub-km length scales and large diurnal variations. These spatial and temporal trends are largely driven by fresh emissions. Traffic has historically received significant attention as the major emission source in urban environments, however emissions from cooking may now be the dominant source of urban primary organic aerosol. There also exist modest spatial and temporal variations associated with secondary organic aerosol (SOA) production; some of this SOA is attributable to non-combustion volatile chemical products (VCPs). This presentation will discuss measurements of cooking emissions in three US cities (Pittsburgh, Oakland, Baltimore) as well as direct observations of SOA production from VCPs in Pittsburgh and New York City.
Inferring aerosol sources using multi-pollutant, low-cost air quality sensors
By: David Hagan, MIT
Summary: Low-cost sensors offer the opportunity to measure urban air quality at a spatiotemporal scale finer than ever before. While our understanding of their performance has increased in recent years, most of these devices are still used as stand-in replacements for existing infrastructure – to monitor and report concentrations of gas- and particle-phase pollutants as time series. Considering most low-cost sensor integrations couple multiple gas-phase sensors alongside an optical particle measurement, there exists an opportunity to leverage these measurements against one another to extract potentially useful information. This work highlights the potential for using low-cost sensors to identify types and sources of aerosols via an unsupervised learning approach using data collected in Delhi, India.
Modeling indoor ultrafine particle dynamics
Summary: Ultrafine particles (UFP, < 100 nm) are generated indoors due to human activities such as cooking, cleaning, and using consumer products. Indoor UFP concentrations can vary with indoor activities and transformation mechanisms such as coagulation and deposition. The objective of this study is to develop a model that investigate the effects of coagulation, deposition, and ventilation on UFP size and concentration dynamics for six indoor emission sources: candle, gas stove, clothes dryer, broiled fish, tortilla, and incense. The time- and size-resolved data were collected in previous studies in literature. Although each indoor source had a different observed particle size and particle number concentrations, the data were fitted well with log-normal distributions. Size-resolved coagulation rates were estimated using Brownian motion with the Fuchs correction along with van der Waals and viscous forces. The results show that coagulation and deposition losses are dominant with gas stoves and candles that generate a number of particles smaller than 10 nm. Relative contributions of coagulation, deposition, and air exchange rates to the total particle losses were 58%, 41%, and 0.5% at a high concentration (i.e., 106 cm-3), while they are 18%, 80%, and 2% at a lower concentration (i.e., 3 × 104 cm-3), respectively. However, ventilation and deposition are responsible for 60-95% losses for indoor sources that emit relatively larger particles (e.g., broiled fish and incense). These results highlight the importance of coagulation in indoor ultrafine particle dynamics, as well as aerosol emission and transformation characteristics depending on the indoor source type.
Condensable particulate matter emitted from stationary combustion sources
POSTER PRESENTATION: Emission factors of particulate matters (PM) from stationary combustion sources have been measured without dilution or cooling in Japan and other Asian countries, thus condensable PM were not included in the PM emission inventory. Recently, contributions of condensable PM were analyzed by comparing measured PM concentrations from stationary combustion sources before and after dilution. From our preliminary analysis of these data, we found that condensable PM from stationary combustion sources in industrial or energy sector have critical contributions to emissions and atmospheric concentrations of organic aerosol (OA) over Japan (Morino et al., Environ. Sci. Technol., 52, 8456−8466). OA concentrations drastically increased around urban and industrial areas, including the Tokyo Metropolitan Area, in all the seasons, and model performance of OA was improved by considering condensable PM from stationary combustion sources. For better estimation of condensable PM emissions, we need to improve estimation of emission factors from relevant sectors, volatility distributions, and physicochemical properties of condensable PM. These points are further discussed in this presentation.
Air Pollution Toxicology: My Struggle for Realistic Exposures
The link between air pollution exposure and its health effects is uncovered by epidemiologists and toxicologists working in concert. Epidemiologists investigate statistical associations between air pollution exposure and morbidity and mortality. Toxicologists investigate the pathology underlying these diseases, typically in animal models exposed to models of actual air pollution. Over the last couple of decades, we have taken a number of complementary approaches striving for realistic exposures in collaboration with toxicologists at UC Davis. In this talk, I will review some of these approaches, their strengths and weaknesses, and what we have learned about the toxicology of air pollution and its sources.
Development and Applications of “Air Benefit and Cost and Attainment Assessment System” (ABaCAS)
Summary: A series of collaborative efforts in the development of a next-generation air quality decision support system, or “Air Benefit and Cost and Attainment Assessment System” (ABaCAS), by a team of U.S. and Chinese scientists have been undertaken since 2013. The objective of this ABaCAS system is to provide scientists and policy makers with a user-friendly framework for conducting integrated assessments of air pollution emissions control cost and their associated air quality, health and economic benefits and attainment goals. The “ABaCAS” system includes eight key components: (1) Streamlined edition for integrated policy assessment and analysis (ABaCAS-SE), (2) an optimized edition of ABaCAS (ABaCAS-OE), (3) Real-time air quality response to emissions control tool (RSM/CMAQ), (4) Air quality attainment assessment tool (SMAT-CE), (5) Health and economic benefit tool (BenMAP-CE), (6) International control cost estimate tool (ICET), (7) a multi-scale air quality modeling visualization and analysis tool (Model-VAT), and (8) a monitor and model data fusion tool (Data Fusion Tool). A series of ABaCAS pilot applications over the USA, China, Taiwan, Korea, and other regions have been undertaken to conduct assessment of emissions control strategies and their associated air quality, health, economic and air quality attainment benefits. These ABaCAS case studies and ongoing efforts in extended applications to China and USA will be presented. Further details of the ABaCAS system can be found and the ABaCAS software package can also be freely downloaded at “abacas-dss.com”.
Comparison of multiple PM2.5 exposure products for estimating health benefits of emission controls
Summary: Ambient exposure to fine particulate matter (PM2.5) is one of the top global health concerns. We estimate the PM2.5-related health benefits of emission reduction over New York State (NYS) from 2002 to 2012 using seven different publicly available PM2.5 products that include information from ground-based observations, remote sensing and chemical transport models. While these PM2.5 products differ in spatial patterns, they show consistent decreases of 28% to 37% from 2002 to 2012. We evaluate these products using two sets of independent ground-based observations, the urban New York City Community Air Quality Survey (NYCCAS) Program, and the remote (upstate New York) Saint Regis Mohawk Tribe Air Quality Program. Inclusion of satellite remote sensing improves the representativeness of surface PM2.5 in the remote area. Of the satellite-based products, only the statistical land use regression approach captures some of the spatial variability across New York City measured by NYCCAS. We estimate the PM2.5-related mortality burden by applying an integrated exposure-response function to the different PM2.5 products. The multi-product mean PM2.5-related mortality burden over NYS decreased by 5660 deaths (67%) from 8410 (95% confidence interval (CI): 4570 – 12400) deaths in 2002 to 2750 (CI: 700 – 5790) deaths in 2012. We estimate a 28% uncertainty in the state-level PM2.5 mortality burden due to the choice of PM2.5 products, which is smaller than the uncertainty (130%) associated with the exposure-response function. Overall, we conclude that exposure estimates for PM2.5 combining ground-based measurements, remotely sensed and modeled data hold substantial promise, and are rapidly becoming the state of the art for exposure assessment in epidemiological and health impact studies.
Projecting PM2.5 concentration fields to correspond to just meeting National Ambient Air Quality Standards
Summary: PM2.5 concentration fields that correspond to just meeting National Ambient Air Quality Standards (NAAQS) in areas throughout the U.S. are useful for characterizing exposure in a variety of regulatory and policy assessments. Computationally efficient methods that incorporate predictions from photochemical grid models are needed to realistically project baseline concentration fields for these assessments. Cross validation of hybrid spatial prediction models is also needed to better assess their predictive capability in sparsely monitored areas. In this presentation, a system for generating, evaluating, and projecting PM2.5 spatial fields to correspond with just meeting the PM2.5 NAAQS will be described. The capabilities of the system will be illustrated with results from a recent national-scale application.
The effectiveness of the Air Pollution Prevention and Control Action Plan on the air quality and human health during 2013-2017 in China
Summary: In 2013, China released the “Air Pollution Prevention and Control Action Plan”, which set the roadmap for national air pollution control actions for the period of 2013-2017. A decrease in the PM2.5 concentration may lead to a substantial benefit for human health. This study aims to quantify the relative contributions four factors: emission reductions, changed meteorology, population growth, and a change in baseline mortality rates to the reduced PM2.5-related mortality (PM2.5-mortality) during the 2013-2017 period and evaluate the importance of emission controls for human health protection in China. We used the chemical transport model (i.e., WRF-CMAQ) as well as an integrated scientific assessment system (i.e., ABaCAS) in this study. The estimated total PM2.5-mortality in China was 1,389,000 (95% CI, 1,005,000-1,631,000) in 2013 but was substantially reduced to 1,102,000 (95% CI, 755,000-1,337,000) in 2017. Emission controls contributed 88.7% to this reduction in PM2.5-mortality, while changed meteorology, the change in baseline mortality rates, and population growth during 2013-2017 contributed 9.6%, 3.8% and -2.2%, respectively. The implementation of the Action Plan has significantly reduced the PM2.5 concentration in regions of China where population density is high, dominating the decline in PM2.5-mortality during 2013-2017. However, the health burden of PM2.5 pollution in China is still extremely high compared to that in other developed countries. An aggressive air pollution control strategy should be implemented in densely populated areas to further reduce the health burden.
Predicted Ultrafine Particulate Matter Source Contribution across the Continental United States during Summer Time Air Pollution Events
Summary: The regional concentrations of airborne ultrafine particulate matter mass (Dp < 0.1 µm; PM0.1) were predicted in 39 cities across the United States (U.S.) during summer time air pollution episodes. Calculations were performed using a regional source-oriented chemical transport model with 4 kilometer (km) spatial resolution operating on the National Emissions Inventory created by the U.S. EPA. Measured source profiles for particle size and composition between 0.01 – 10 µm were used to translate PM total mass to PM0.1. Predicted PM0.1 concentrations exceeded 2 µg/m3 during summer pollution episodes in major urban regions across the U.S. including Los Angeles, the San Francisco Bay Area, Houston, Miami, and New York. PM0.1 spatial gradients were sharper than PM2.5 spatial gradients due to the dominance of primary aerosol in PM0.1. Artificial source tags were used to track contributions to primary PM0.1 and PM2.5 from fifteen source categories. On-road gasoline and diesel vehicles made significant contributions to regional PM0.1 in all 39 cities even though peak contributions within 0.3 km of the roadway were not resolved by the 4 km grid cells. Food cooking also made significant contributions to PM0.1 in all cities but biomass combustion was only important in locations impacted by summer wildfires. Aviation was a significant source of PM0.1 in cities that had airports within their urban footprints. Industrial sources including cement manufacturing, process heating, steel foundries, and paper & pulp processing impacted their immediate vicinity but did not significantly contribute to PM0.1 concentrations in any of the target 39 cities. Natural gas combustion made significant contributions to PM0.1 concentrations due to the widespread use of this fuel for electricity generation, industrial applications, residential and commercial use. The major sources of primary PM0.1 and PM2.5 were notably different in many cities. Future epidemiological studies may be able to differentiate PM0.1 and PM2.5 health effects by contrasting cities with different ratios of PM0.1 / PM2.5. In the current study, cities with higher PM0.1 / PM2.5 ratios (ratio greater than 0.10) include Houston TX, Los Angeles CA, Bakersfield CA, Salt Lake City UT, and Cleveland OH. Cities with lower PM0.1 to PM2.5 ratios (ratio lower than 0.05) include Lake Charles LA, Baton Rouge LA, St. Louis MO, Baltimore MD, and Washington DC.
Wood smoke emissions and particulate matter formation in California
POSTER DISPLAY Summary: Despite years of progress, PM2.5 concentrations continue to violate the National Ambient Air Quality Standards (NAAQS) designed to protect human health in California’s San Joaquin Valley (SJV). The highest PM2.5 concentrations in the SJV occur during winter months with significant contributions from primary and secondary organic aerosols (OA). Chemical transport models (CTMs) are used to help design efficient emissions control programs to address PM2.5 pollution, but these models have generally failed to predict base case PM2.5OA concentrations in recent years. This increases uncertainty about the effectiveness of future emissions control strategies for PM2.5 in the SJV.
Recent work suggests that wood smoke emissionsupdated with the help of meteorological variables can play a role to improve CTM performance for OA. Here, we analyze the effect of wood smoke emissions on winter particulate matter in California. Monthly emissions of wood smokeare updated using a statistical model linking wood burning emissions to meteorological variables and the consumption of other fuels. The updated emissions are comparedwith standard emissions inventories provided by the California Air Resources Board (CARB). Episodes spanning the winter months of years 2000 through 2016 are simulated for base and updated wood smoke emissions using theUniversity of California, Davis/California Institute of Technology (UCD/CIT) regional chemical transport model. Predicted concentrations are compared to the measurements to quantify improvements to model performance.
Explicit modelling of gas-phase chemistry and viscosity on secondary organic aerosol formation
POSTER DISPLAY Summary: Secondary organic aerosols (SOA) are major components of atmospheric fine particulate matter, affecting climate and air quality. Mounting evidence exists that SOA can adopt a viscous state, which may impact formation and partitioning of SOA. In this study, we conducted explicit modelling of gas-phase oxidation of linear alkane and subsequent SOA formation using the GECKO-A (Generator of Explicit Chemistry and Kinetics of Organics in the Atmosphere) model (Aumont et al., 2012). Our recently-developed method to predict glass transition temperature and viscosity (DeRieux et al., 2018) was implemented in GECKO-A to predict viscosity of alkane SOA. In addition, the impacts of viscosity on mass accommodation coefficient were explored. The simulated SOA formation yields were compared to chamber experiments (Lim & Ziemann, 2009) to evaluate the role and impacts of gas-phase chemistry and phase state on SOA formation.