Flexible Bayesian Ensemble Machine Learning Framework for Predicting Local Ozone Concentrations.

Abstract. In this study, the Community Multiscale Air Quality (CMAQ) modeling system is used to simulate the ozone (O3) episodes during the Program of Regional Integrated Experiments of Air Quality over the Pearl River Delta, China, in October 2004 (PRIDE-PRD2004). The simulation suggests that O3 pollution is a regional phenomenon in the Pearl River Delta (PRD). Elevated O3 levels often occurred in the southwestern inland PRD, Pearl River estuary (PRE), and southern coastal areas during the 1-month field campaign. Three evolution patterns of simulated surface O3 are summarized based on different near-ground flow conditions. More than 75% of days featured interactions between weak synoptic forcing and local sea-land circulation. Integrated process rate (IPR) analysis shows that photochemical production is a dominant contributor to O3 enhancement from 09:00 to 15:00 local standard time in the atmospheric boundary layer over most areas with elevated O3 occurrence in the mid-afternoon. The simulated ozone production efficiency is 2–8 O3 molecules per NOx molecule oxidized in areas with high O3 chemical production. Precursors of O3 originating from different source regions in the central PRD are mixed during the course of transport to downwind rural areas during nighttime and early morning, where they then contribute to the daytime O3 photochemical production. The sea-land circulation plays an important role on the regional O3 formation and distribution over PRD. Sensitivity studies suggest that O3 formation is volatile-organic-compound-limited in the central inland PRD, PRE, and surrounding coastal areas with less chemical aging (NOx/NOy>0.6), but is NOx-limited in the rural southwestern PRD with aged air (NOx/NOy<0.3).

Abstract. In this paper, we present the custom Hong Kong NO2 retrieval (HKOMI) for the Ozone Monitoring Instrument (OMI) on board the Aura satellite which was used to evaluate a high-resolution chemistry transport model (CTM) (3 km × 3 km spatial resolution). The atmospheric chemistry transport was modelled in the Pearl River Delta (PRD) region in southern China by the Models-3 Community Multiscale Air Quality (CMAQ) modelling system from October 2006 to January 2007. In the HKOMI NO2 retrieval, tropospheric air mass factors (AMFs) were recalculated using high-resolution ancillary parameters of surface reflectance, a priori NO2 and aerosol profiles, of which the latter two were taken from the CMAQ simulation. We tested the influence of the ancillary parameters on the data product using four different aerosol parametrizations. Ground-level measurements by the PRD Regional Air Quality Monitoring (RAQM) network were used as additional independent measurements. The HKOMI retrieval increases estimated tropospheric NO2 vertical column densities (VCD) by (+31 ± 38)%, when compared to NASA's standard product (OMNO2-SP), and improves the normalized mean bias (NMB) between satellite and ground observations by 26 percentage points from −41 to −15%. The individual influences of the parameters are (+11.4 ± 13.4)% for NO2 profiles, (+11.0 ± 20.9)% for surface reflectance and (+6.0 ± 8.4)% for the best aerosol parametrization. The correlation coefficient r is low between ground and satellite observations (r = 0.35). The low r and the remaining NMB can be explained by the low model performance and the expected differences when comparing point measurements with area-averaged satellite observations. The correlation between CMAQ and the RAQM network is low (r ≈ 0.3) and the model underestimates the NO2 concentrations in the northwestern model domain (Foshan and Guangzhou). We compared the CMAQ NO2 time series of the two main plumes with our best OMI NO2 data set (HKOMI-4). The model overestimates the NO2 VCDs by about 15% in Hong Kong and Shenzhen, while the correlation coefficient is satisfactory (r = 0.56). In Foshan and Guangzhou, the correlation is low (r = 0.37) and the model underestimates the VCDs strongly (NMB = −40%). In addition, we estimated that the OMI VCDs are also underestimated by about 10 to 20% in Foshan and Guangzhou because of the influence of the model parameters on the AMFs. In this study, we demonstrate that the HKOMI NO2 retrieval reduces the bias of the satellite observations and how the data set can be used to study the magnitude of NO2 concentrations in a regional model at high spatial resolution of 3 × 3 km2. The low bias was achieved with recalculated AMFs using updated surface reflectance, aerosol profiles and NO2 profiles. Since unbiased concentrations are important, for example, in air pollution studies, the results of this paper can be very helpful in future model evaluation studies.

Dynamic evaluation of a regional air quality model: Assessing the emissions-induced weekly ozone cycle

Background and Aims: Ambient monitors are often used to estimate air pollution exposure for epidemiological studies. This approach is efficient and economical but has limitations, including restricted coverage and resolution. Although three-dimensional air quality modeling has its own limitations, this emerging exposure approach has the potential to address some challenges of monitoring networks. We evaluated differences in exposure estimates obtained using monitoring data and simulation results based on the Community Multi-scale Air Quality (CMAQ) modeling system. Methods: Spatially-aggregated (county-level) exposure estimates for PM2.5 were calculated for the eastern two-thirds of the U.S. for years 2002-2006 from three sources: (1) 24hr ambient monitoring data; (2) 12 x 12 km gridded CMAQ simulation results; and (3) concentration estimates from a Bayesian space-time downscaler model “fusing” PM2.5 monitoring data with 12-km gridded CMAQ output (i.e., downscaler-derived exposures). Results: Exposure estimates generated from CMAQ and the downscaler provide greater spatial and temporal coverage compared to estimates from monitor data. The CMAQ and CMAQ-downscaler exposure estimates included 2818 counties and the monitoring approach produced estimates for 190 counties, which contain ~51% of the population within the CMAQ domain. Exposure estimates derived from CMAQ results underestimated PM2.5 levels (normalized mean bias [NMB]=-4.0%), while downscaler-derived exposures slightly overestimated observed PM2.5 (NMB=1.0%). Across time, correlation between monitor-derived exposure estimates and CMAQ- and downscaler-derived exposure estimates were 0.62 and 0.97, respectively. Conclusions: The downscaler approach had superior performance based on several metrics. Epidemiology may benefit from use of regional air quality models, with improved spatial and temporal coverage and ability to study populations far from monitors that may differ from those near monitors.

Abstract. The Community Multiscale Air Quality (CMAQ) modeling system, a state-of-the-science regional air quality modeling system developed by the US Environmental Protection Agency, is being used for a variety of environmental modeling problems including regulatory applications, air quality forecasting, evaluation of emissions control strategies, process-level research, and interactions of global climate change and regional air quality. The Meteorology-Chemistry Interface Processor (MCIP) is a vital piece of software within the CMAQ modeling system that serves to, as best as possible, maintain dynamic consistency between the meteorological model and the chemical transport model (CTM). MCIP acts as both a post-processor to the meteorological model and a pre-processor to the emissions and the CTM in the CMAQ modeling system. MCIP's functions are to ingest the meteorological model output fields in their native formats, perform horizontal and vertical coordinate transformations, diagnose additional atmospheric fields, define gridding parameters, and prepare the meteorological fields in a form required by the CMAQ modeling system. This paper provides an updated overview of MCIP, documenting the scientific changes that have been made since it was first released as part of the CMAQ modeling system in 1998.

Atmospheric nitrogen oxide and nitrogen dioxide (NO + NO2, together termed as NOX) estimates from annual photochemical simulations for years 2002–2016 are compared to surface network measurements of NOX and total gas-phase-oxidized reactive nitrogen (NOY) to evaluate the Community Multiscale Air Quality (CMAQ) modeling system performance by U.S. region, season, and time of day. In addition, aircraft measurements from 2011 Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality are used to evaluate how emissions, chemical mechanism, and measurement uncertainty each contribute to the overall model performance. We show distinct seasonal and time-of-day patterns in NOX performance. Summertime NOX is overpredicted with bimodal peaks in bias during early morning and evening hours and persisting overnight. The summertime morning NOX bias dropped from between 28% and 57% for earlier years (2002–2012) to between −2% and 7% for later years (2013–2016). Summer daytime NOX tends to be unbiased or underpredicted. In winter, the evening NOX overpredictions remain, but NOX is unbiased or underpredicted overnight, in the morning, and during the day. NOX overpredictions are most pronounced in the Midwestern and Southern United States with Western regions having more of a tendency toward model underpredictions of NOX. Modeled NOX performance has improved substantially over time, reflecting updates to the emission inputs and the CMAQ air quality model. Model performance improvements are largest for years simulated with CMAQv5.1 or later and for emission inventory years 2014 and later, coinciding with reduced onroad NOX emissions from vehicles with newer emission control technologies and improved treatment of chemistry, deposition, and vertical mixing in CMAQ. Our findings suggest that emissions temporalization of specific mobile source sectors have a small impact on model performance, while chemistry updates improve predictions of NOY but do not improve summertime NOX bias in the Baltimore/DC area. Sensitivity runs performed for different locations across the country suggest that the improvement in summer NOX performance can be attributed to updates in vertical mixing incorporated in CMAQv5.1.

Abstract. A chemical transport model (CTM) is an essential tool for air quality prediction and management, widely used in air pollution control and health risk assessment. However, the current models do not perform very well in reproducing the observations of some major chemical components, for example, sulfate, nitrate, ammonium and organic carbon. Studies have suggested that the uncertainties in the model chemical mechanism, source emission inventory and meteorological field can cause inaccurate simulation results. Still, the emission source profile (used to create speciated emission inventories for CTMs) of PM2.5 has not been fully taken into account in current numerical simulation. Based on the characteristics and variation rules of chemical components in typical PM2.5 sources, different simulation scenarios were designed and the sensitivity of simulated PM2.5 components to the source chemical profile was explored. Our findings showed that the influence of source profile changes on simulated PM2.5 components' concentrations cannot be ignored. Simulation results of some components were sensitive to the adopted source profile in CTMs. Moreover, there was a linkage effect: the variation in some components in the source profile would bring changes to the simulated results of other components. These influences are connected to chemical mechanisms of the model since the variation in species allocations in emission sources can affect the potential composition and phase state of aerosols, chemical reaction priority, and multicomponent chemical balance in thermodynamic equilibrium systems. We also found that the perturbation of the PM2.5 source profile caused variation in simulated gaseous pollutants, which indirectly indicated that the perturbation of source profile would affect the simulation of secondary PM2.5 components. Our paper highlights the necessity of paying enough attention to the representativeness and timeliness of the source profile when using CTMs for simulation.

Abstract. Regional air quality simulations were conducted for summers 2001–2005 in the eastern US and subjected to extensive evaluation using various ground and airborne measurements. A brief climate evaluation focused on transport by comparing modeled dominant map types with ones from reanalysis. Reasonable agreement was found for their frequency of occurrence and distinctness of circulation patterns. The two most frequent map types from reanalysis were the Bermuda High (22%) and passage of a Canadian cold frontal over the northeastern US (20%). The model captured their frequency of occurrence at 25% and 18% respectively. The simulated five average distributions of 1-h ozone (O3) daily maxima using the Community Multiscale Air Quality (CMAQ) modeling system reproduced salient features in observations. This suggests that the ability of the regional climate model to depict transport processes accurately is critical for reasonable simulations of surface O3. Comparison of mean bias, root mean square error, and index of agreement for CMAQ summer surface 8-h O3 daily maxima and observations showed –0.6±14 nmol/mol, 14 nmol/mol, and 71% respectively. CMAQ performed best in moderately polluted conditions and less satisfactorily in highly polluted ones. This highlights the common problem of overestimating/underestimating lower/higher modeled O3 levels. Diagnostic analysis suggested that significant overestimation of inland nighttime low O3 mixing ratios may be attributed to underestimates of nitric oxide (NO) emissions at night. The absence of the second daily peak in simulations for the Appledore Island marine site possibly resulted from coarse grid resolution misrepresentation of land surface type. Comparison with shipboard measurements suggested that CMAQ has an inherent problem of underpredicting O3 levels in continental outflow. Modeled O3 vertical profiles exhibited a lack of structure indicating that key processes missing from CMAQ, such as lightning produced NO and stratospheric intrusions, are important for accurate upper tropospheric representations.

Through a NERC/MOES funded project, PROMOTE, analysis based on WRF and CMAQ models has been conducted to understand the impact of road transport emissions on air quality over Delhi. NCEP/FNL 1 data was used to drive WRF with four nested domains over India with resolutions of 45km, 15 km, 5 km, and 1.6 km for 2018. EDGAR v5.0 emission inventory (for 2015) 2 and Cam-Chem initial and boundary condition data 3 were used to drive the CMAQ model. In the baseline runs, all domains were considered without any change in emissions, while in Scenario 1, road transport sector was removed in the third domain (5km) covering Delhi region. Model performance for NOx, NO2, PM10, PM2.5 and O3 was evaluated with available observations, recognising that air quality and meteorological datasets were limited for the period analysed. In the later part of the study, OSCAR model 4 was used to predict the high-resolution air quality over Delhi and estimate the contributions from road transport emissions. Relative contributions to Delhi's air quality from local and regional long range transport sources are discussed. As part of a NERC funded COP26 project on Climate Adaptation for India, an overarching goal of this study is to quantify how air quality changes in South Asia in a changing climate under SSP245 (middle-of-the-road scenario) 5. A dynamical downscaling process was implemented and bias-corrected Coupled Model Intercomparison Project Phase 6 (CMIP6) data 6 was used to drive Weather Research and Forecasting (WRF) model simulations. Simulations have been conducted for the South Asia-Cordex domain for 2015 (representative of 2011-2020 years) and 2050 (representative of 2046-2055 years) with a 27 km grid resolution. The Community Multiscale Air Quality (CMAQ) model was driven with an average of ten years meteorology with future land use land cover, initial and boundary conditions, and future emissions 7 for India. To assess the impact of averaged meteorology on the CMAQ performance, the CMAQ model was run with both ten years' averaged meteorology around 2015 (2011-2020) and only 2015 meteorology.Although averaging ten years around the desired year suppresses the diurnal variations, it provides an indication of monthly changes in climate and air quality variables. Under the selected SSP245 scenario, the CMAQ model predicted monthly means of PM2.5 anomalies (2050-2015) range over India between 8 to 41 μg/m3. Significant change is PM2.5, PM10, NOx, and O3 anomalies, especially in the urban regions of India, such as Delhi, before and after the Monsoon months (June to October) have been observed.Financial Support: We acknowledge funding from NERC/MOES (Reference: NE/P016391/1) for the PROMOTE project and NERC funding (Reference: 2021COPA&R48Sokhi) for the COP26 Improving adaptation strategies for climate extremes and air pollution affecting India project.References 1 NCEP/NOAA/U.S. 2015,  https://doi.org/10.5065/D65Q4T4Z.2 Crippa M. et al. (2019): http://data.europa.eu/89h/377801af-b094-4943-8fdc-f79a7c0c2d193 Buchholz, R. R. et al, (2019). https://doi.org/10.5065/NMP7-EP604  Singh V, et al. (2020) Environmental Pollution, 257, 1136235 van Vuuren, D.P. et al. (2011). Climatic Change 109, 5. https://doi.org/10.1007/s10584-011-0148-z6 Xu, Z. et al.  (2021). Sci Data 8, 293. https://doi.org/10.1038/s41597-021-01079-37 SMoG-India v1 2015 and 2050 emissions dataset, NERC project.

A new post-processing method for surface particulate matter (PM2.5) predictions from the National Oceanic and Atmospheric Administration (NOAA) developmental air quality forecasting system using the Community Multiscale Air Quality (CMAQ) model is described. It includes three main components:• A real-time quality control procedure for surface PM2.5 observations;• Model post-processing at each observational site using historical forecast analogs and Kalman filtering;• Spreading the forecast corrections from the observation locations to the entire gridded domain.The methodology is tested using 12 months of CMAQ forecasts of hourly PM2.5, from December 01, 2009 through November 30, 2010. The model domain covers the contiguous USA, and model data are verified against U.S. Environmental Prediction Agency AIRNow PM2.5 observations measured at 716 stations over the CMAQ domain. The model bias is found to have a strong seasonal dependency, with a large positive bias in winter and a small bias in the summer months, and also to have a strong diurnal cycle.Five different post-processing techniques are compared, including a seven-day running mean subtraction, Kalman-filtering, analogs, and combinations of analogs and Kalman filtering. The most accurate PM2.5 forecasts have been found to be produced when using historical analogs of the hourly Kalman-filtered forecasts, referred to as KFAN. The choice of meteorological variables used in the hourly analog search is also found to have a significant effect. A monthly error analysis is computed, in each case using the remaining 11 months of the data set for the analog searches. The improvement of KFAN errors over the raw CMAQ model errors ranges from 50 to 75% for MAE and from 40 to 60% for the correlation coefficient. Since the post-processing analysis is only done at the locations where observations are available, the spreading of post-processing correction information over nearby model grid points is necessary to make forecast contour maps. This spreading of information is accomplished with an eight-pass Barnes-type iterative objective analysis scheme. The final corrected CMAQ forecast over the entire domain is composed of the sum of the original CMAQ forecasts and the KFAN bias information interpolated over the entire domain, and is applied on an hourly basis.

Numerical simulation with a comprehensive chemical transport model of nitrate, sulfate, and ammonium aerosol distributions over east asia

<strong class="journal-contentHeaderColor">Abstract.</strong> Chemical mechanisms describe how emissions of gases and particles evolve in the atmosphere and are used within chemical transport models to evaluate past, current, and future air quality. Thus, a chemical mechanism must provide robust and accurate predictions of air pollutants if it is to be considered for use by regulatory bodies. In this work, we provide an initial evaluation of the Community Regional Atmospheric Chemical Multiphase Mechanism (CRACMMv1.0) by assessing CRACMMv1.0 predictions of surface ozone (O₃) across the Northeast U.S. during the summer of 2018 within the Community Multiscale Air Quality (CMAQ) modeling system. CRACMMv1.0 O₃ predictions of hourly and maximum daily 8-hour average (MDA8) ozone were lower than those estimated by the Regional Atmospheric Chemical Mechanism (RACM2_ae6), which better matched surface network observations in the Northeast US (RACM2_ae6 mean bias of +4.2 ppb for all hours and +4.3 ppb for MDA8; CRACMMv1.0 mean bias of +2.1 ppb for all hours and +2.7 ppb for MDA8). Box model calculations combined with results from CMAQ emission reduction simulations indicated high sensitivity of O₃ to compounds with biogenic sources. In addition, these calculations indicated the differences between CRACMMv1.0 and RACM2_ae6 O₃ predictions were largely explained by updates to the inorganic rate constants (reflecting the latest assessment values) and by updates to the representation of monoterpene chemistry. Updates to other reactive organic carbon systems between RACM2_ae6 and CRACMMv1.0 also affected ozone predictions and their sensitivity to emissions. Specifically, CRACMMv1.0 benzene, toluene, and xylene chemistry led to efficient NO_x cycling such that CRACMMv1.0 predicted controlling aromatics reduces ozone without rural O₃ disbenefits. In contrast, semivolatile to intermediate volatility alkanes introduced in CRACMMv1.0 acted to suppress O₃ formation across the regional background through the sequestration of nitrogen oxides (NO_x) in organic nitrates. Overall, these analyses showed that the CRACMMv1.0 mechanism within the CMAQ model was able to reasonably simulate ozone concentrations in the Northeast US during the summer of 2018 with similar magnitude and diurnal variation as the current operational Carbon Bond (CB6r3_ae7) and good model performance compared to recent modelling studies in the literature.

<strong class="journal-contentHeaderColor">Abstract.</strong> Chemical mechanisms describe how emissions of gases and particles evolve in the atmosphere and are used within chemical transport models to evaluate past, current, and future air quality. Thus, a chemical mechanism must provide robust and accurate predictions of air pollutants if it is to be considered for use by regulatory bodies. In this work, we provide an initial evaluation of the Community Regional Atmospheric Chemical Multiphase Mechanism (CRACMMv1.0) by assessing CRACMMv1.0 predictions of surface ozone (O₃) across the Northeast U.S. during the summer of 2018 within the Community Multiscale Air Quality (CMAQ) modeling system. CRACMMv1.0 O₃ predictions of hourly and maximum daily 8-hour average (MDA8) ozone were lower than those estimated by the Regional Atmospheric Chemical Mechanism (RACM2_ae6), which better matched surface network observations in the Northeast US (RACM2_ae6 mean bias of +4.2 ppb for all hours and +4.3 ppb for MDA8; CRACMMv1.0 mean bias of +2.1 ppb for all hours and +2.7 ppb for MDA8). Box model calculations combined with results from CMAQ emission reduction simulations indicated high sensitivity of O₃ to compounds with biogenic sources. In addition, these calculations indicated the differences between CRACMMv1.0 and RACM2_ae6 O₃ predictions were largely explained by updates to the inorganic rate constants (reflecting the latest assessment values) and by updates to the representation of monoterpene chemistry. Updates to other reactive organic carbon systems between RACM2_ae6 and CRACMMv1.0 also affected ozone predictions and their sensitivity to emissions. Specifically, CRACMMv1.0 benzene, toluene, and xylene chemistry led to efficient NO_x cycling such that CRACMMv1.0 predicted controlling aromatics reduces ozone without rural O₃ disbenefits. In contrast, semivolatile to intermediate volatility alkanes introduced in CRACMMv1.0 acted to suppress O₃ formation across the regional background through the sequestration of nitrogen oxides (NO_x) in organic nitrates. Overall, these analyses showed that the CRACMMv1.0 mechanism within the CMAQ model was able to reasonably simulate ozone concentrations in the Northeast US during the summer of 2018 with similar magnitude and diurnal variation as the current operational Carbon Bond (CB6r3_ae7) and good model performance compared to recent modelling studies in the literature.

Abstract. Systematic analyses of interannual and seasonal variations of tropospheric NO2 vertical column densities (VCDs) based on GOME satellite data and the regional scale chemical transport model (CTM), Community Multi-scale Air Quality (CMAQ), are presented for the atmosphere over eastern Asia between 1996 and June 2003. A newly developed year-by-year emission inventory (REAS) was used in CMAQ. The horizontal distribution of annual averaged GOME NO2 VCDs generally agrees well with the CMAQ results. However, CMAQ/REAS results underestimate the GOME retrievals with factors of 2–4 over polluted industrial regions such as Central East China (CEC), a major part of Korea, Hong Kong, and central and western Japan. The most probable reasons for the underestimation typically over the CEC are accuracy of the basic energy statistic data, emission factors, and socio-economic data used for construction of emission inventory. For the Japan region, GOME and CMAQ NO2 data show reasonable agreement with respect to interannual variation and show no clear increasing trend. For CEC, GOME and CMAQ NO2 data indicate a very rapid increasing trend from 2000. Analyses of the seasonal cycle of NO2 VCDs show that GOME data have larger dips than CMAQ NO2 during February–April and September–November. Sensitivity experiments with fixed emission intensity reveal that the detection of emission trends from satellite in fall or winter has a larger error caused by the variability of meteorology. Examination during summer time and annual averaged NO2 VCDs are robust with respect to variability of meteorology and are therefore more suitable for analyses of emission trends. Analysis of recent trends of annual emissions in China shows that the increasing trends of 1996–1998 and 2000–2002 for GOME and CMAQ/REAS show good agreement, but the rate of increase by GOME is approximately 10–11% yr−1 after 2000; it is slightly steeper than CMAQ/REAS (8–9% yr−1). The greatest difference was apparent between the years 1998 and 2000: CMAQ/REAS only shows a few percentage points of increase, whereas GOME gives a greater than 8% yr−1 increase. The exact reason remains unclear, but the most likely explanation is that the emission trend based on the Chinese emission related statistics underestimates the rapid growth of emissions.

Abstract. A natural emissions inventory for the continental United States and surrounding territories is needed in order to use the US Environmental Protection Agency Community Multiscale Air Quality (CMAQ) Model for simulating natural air quality. The CMAQ air modeling system (including the Sparse Matrix Operator Kernel Emissions (SMOKE) emissions processing system) currently estimates non-methane volatile organic compound (NMVOC) emissions from biogenic sources, nitrogen oxide (NOx) emissions from soils, ammonia from animals, several types of particulate and reactive gas emissions from fires, as well as sea salt emissions. However, there are several emission categories that are not commonly treated by the standard CMAQ Model system. Most notable among these are nitrogen oxide emissions from lightning, reduced sulfur emissions from oceans, geothermal features and other continental sources, windblown dust particulate, and reactive chlorine gas emissions linked with sea salt chloride. A review of past emissions modeling work and existing global emissions data bases provides information and data necessary for preparing a more complete natural emissions data base for CMAQ applications. A model-ready natural emissions data base is developed to complement the anthropogenic emissions inventory used by the VISTAS Regional Planning Organization in its work analyzing regional haze based on the year 2002. This new data base covers a modeling domain that includes the continental United States plus large portions of Canada, Mexico and surrounding oceans. Comparing July 2002 source data reveals that natural emissions account for 16% of total gaseous sulfur (sulfur dioxide, dimethylsulfide and hydrogen sulfide), 44% of total NOx, 80% of reactive carbonaceous gases (NMVOCs and carbon monoxide), 28% of ammonia, 96% of total chlorine (hydrochloric acid, nitryl chloride and sea salt chloride), and 84% of fine particles (i.e., those smaller than 2.5 μm in size) released into the atmosphere. The seasonality and relative importance of the various natural emissions categories are described.