Effects of Instrument Precision and Spatial Variability on the Assessment of the Temporal Variation of Ambient Air Pollution in Atlanta,Georgia

Abstract

Data from the U.S. Environmental Protection Agency Air Quality System, the Southeastern Aerosol Research and Characterization database, and the Assessment of Spatial Aerosol Composition in Atlanta database for 1999 through 2002 have been used to characterize error associated with instrument precision and spatial variability on the assessment of the temporal variation of ambient air pollution in Atlanta, GA. These data are being used in time series epidemiologic studies in which associations of acute respiratory and cardiovascular health outcomes and daily ambient air pollutant levels are assessed. Modified semivariograms are used to quantify the effects of instrument precision and spatial variability on the assessment of daily metrics of ambient gaseous pollutants (SO₂, CO, NO_x, and O₃) and fine particulate matter ([PM_2.5] PM_2.5 mass, sulfate, nitrate, ammonium, elemental carbon [EC], and organic carbon [OC]). Variation because of instrument imprecision represented 7–40% of the temporal variation in the daily pollutant measures and was largest for the PM_2.5 EC and OC. Spatial variability was greatest for primary pollutants (SO₂, CO, NO_x, and EC). Population–weighted variation in daily ambient air pollutant levels because of both instrument imprecision and spatial variability ranged from 20% of the temporal variation for O₃ to 70% of the temporal variation for SO₂ and EC. Wind rose plots, corrected for diurnal and seasonal pattern effects, are used to demonstrate the impacts of local sources on monitoring station data. The results presented are being used to quantify the impacts of instrument precision and spatial variability on the assessment of health effects of ambient air pollution in Atlanta and are relevant to the interpretation of results from time series health studies that use data from fixed monitors.     

Particle Mass Emission Rates from Current-Technology,Light-Duty Gasoline Vehicles

ABSTRACT

Now that the U.S. Environmental Protection Agency has promulgated new National Ambient Air Quality Standards for PM_2.5, work will begin on generating the data required to determine the sources of ambient PM_2.5 and the magnitude of their contributions to air pollution. This paper summarizes the results of an Environmental Research Consortium program, carried out under the auspices of the U.S. Council for Automotive Research. The program focused on particulate matter (PM) emissions from representative, current-technology, light-duty gasoline vehicles produced by DaimlerChrysler Corp., Ford Motor Co., and General Motors Corp. The vehicles, for the most part taken from the manufacturer's certification and durability fleets, were dynamometer-tested using the three-phase Federal Test Procedure in the companies' laboratories. The test fleet was made up of a mixture of both low-mileage (2K-35K miles) and high-mileage (60K-150K miles) cars, vans, sport utility vehicles, and light trucks. For each vehicle tested, PM emissions were accumulated over 4 cold-start tests, which were run on successive days. PM emission rates from the entire fleet (22 vehicles total) averaged less than 2 mg/mile. All 18 vehicles tested using California Phase 2 reformulated gasoline had PM emission rates less than 2 mg/ mile at both low and high mileages.     

Measurements of Rural Sulfur Dioxide and Particle Sulfate: Analysis of CASTNet Data, 1987 through 1996

ABSTRACT

The Clean Air Status and Trends Network (CASTNet) was implemented by the U.S. Environmental Protection Agency (EPA) in 1991 in response to Title IX of the Clean Air Amendments of 1990, which mandated the deployment of a national ambient air monitoring network to track progress of the implementation of emission reduction programs in terms of deposition, air quality, and changes to affected ecosystems. CASTNet evolved from the National Dry Deposition Network (NDDN). CASTNet currently consists of 45 sites in the eastern United States and 28 sites in the West. Each site measures sulfur dioxide (SO₂), nitric acid (HNO₃), particle sulfate (SO₄ ⁼), particle nitrate (NO₃ ^- ), and ozone. Nineteen sites collect precipitation samples. NDDN/CASTNet uses a uniform set of site-selection criteria which provides the data user with consistent measures to compare each site. These criteria also ensure that, to the extent possible, CASTNet sites are located away from local emission sources.

This paper presents an analysis of SO₂ and SO₄ ⁼ concentration data collected from 1987 through 1996 at rural NDDN/CASTNet sites. Annual and seasonal variability is examined. Gradients of SO₂ and SO₄ ⁼ are discussed. The variability of the atmospheric mix of SO₂ and SO₄ ⁼ is explored spatially and seasonally. Data from CASTNet are also compared to SO₂ and SO₄ ⁼ data from concurrent monitoring studies in rural areas.     

Analysis of PM10 Trends in the United States from 1988 through 1995

ABSTRACT

Because the U. S. Environmental Protection Agency (EPA) has changed the National Ambient Air Quality Standards (NAAQS) for ambient particulate matter (PM), there is a great deal of interest in determining recent PM trends. This paper examines trends in PM₁₀ (i.e., particulate matter less than 10 micrometers in diameter) for areas of the United States based on their attainment status—for PM₁₀ and ozone nonattainment and attainment areas. The analysis also focuses on urban, suburban, and rural areas, and eastern and western areas. The time period of evaluation is from 1988 through 1995. To shed further light on the ambient PM₁₀ trends, trends in ambient SO₂, NO₂, and volatile organic compounds (VOCs) are also analyzed. Finally, trends in emission inventories of SO₂, NO_x, VOCs, and PM₁₀ are evaluated. Results of the analysis show that widespread and similar reductions in PM₁₀ levels have occurred over the last seven years. Annual reductions range from 3.0% to 3.8%, with the greatest reductions coming in PM₁₀ nonattainment areas, but with very significant reductions also in PM₁₀ attainment areas, ozone attainment areas, and rural areas. The widespread reductions appear to be due to a set of controls or common factors that are having a fairly uniform effect in all of the areas. The consistency of the reductions in different areas suggests that the reductions may also be primarily in the fine particles (i.e., those less than 2.5 micrometers in diameter, or PM_2.5), which are more readily transported than coarse particles.     

Sulfur Dioxide Emission Limitation (Sdel) Program at Tva Power Plants

TVA is conducting a program of intermittent control of SO₂ emissions at nine of its coal-fired power plants.The program is designed to limit SO₂ emissions by reducing generation during adverse atmospheric conditions to avoid exceeding ambient SO₂ standards. Each is identified as a Class I or Class II program—based primarily on its design and operational complexity. The four Class I programs (Paradise, Cumberland, Allen, Gallatin) operate 7 days/week from early morning through mid-afternoon. The five Class II programs (Kingston, Widows Creek, Colbert, Johnsonville, Shawnee) operate 24 hours/day and 7 days/week. Environmental data stations (EDS) are located at each plant site; the staff has the responsibility for the collection and validation of the onsite meteorological data, the SO₂ mobile and fixed monitoring network data, the plant operational data, and the meteorological forecast data from the Meteorological Forecast Center in Muscle Shoals, AL. At the EDS, the data are entered into a computer operating model for scheduling plant generation reductions.     

Assessment of air quality benefits from national air pollution control policies in China. Part I: Background,emission scenarios and evaluation of meteorological predictions

Under the 11th Five Year Plan (FYP, 2006–2010) for national environmental protection by the Chinese government, the overarching goal for sulfur dioxide (SO₂) controls is to achieve a total national emissions level of SO₂ in 2010 10% lower than the level in 2005. A similar nitrogen oxides (NO_x) emissions control plan is currently under development and could be enforced during the 12th FYP (2011–2015). In this study, the U.S. Environmental Protection Agency (U.S.EPA)’s Community Multi-Scale Air Quality (Models-3/CMAQ) modeling system was applied to assess the air quality improvement that would result from the targeted SO₂ and NO_x emission controls in China. Four emission scenarios — the base year 2005, the 2010 Business-As-Usual (BAU) scenario, the 2010 SO₂ control scenario, and the 2010 NO_x control scenario—were constructed and simulated to assess the air quality change from the national control plan. The Fifth-Generation NCAR/Penn State Mesoscale Model (MM5) was applied to generate the meteorological fields for the CMAQ simulations. In this Part I paper, the model performance for the simulated meteorology was evaluated against observations for the base case in terms of temperature, wind speed, wind direction, and precipitation. It is shown that MM5 model gives an overall good performance for these meteorological variables. The generated meteorological fields are acceptable for using in the CMAQ modeling.     

Toxaphene in standard solutions and cleaned biota extracts--results of the first QUASIMEME interlaboratory studies. Quality Assurance of Information for Marine Environmental Monitoring in Europe

Two interlaboratory studies on individual toxaphene congeners have been organised by the project Quality Assurance of Information for Marine Environmental Monitoring in Europe (QUASIMEME). Fifteen laboratories analysed two standard solutions in the first study and 13 laboratories analysed a standard solution and two cleaned biota extracts in the second study. The coefficients of variation obtained for the standard solutions were 6-21% and for the cleaned extracts 16-39%. Although the results were comparable to those of other studies, further improvement in the level of agreement between the participating laboratories was considered possible.     

International field intercomparison measurements of atmospheric mercury species at Mace Head,Ireland

Eleven laboratories from North America and Europe met at Mace Head, Ireland for the period 11–15 September 1995 for the first international field intercomparison of measurement techniques for atmospheric mercury species in ambient air and precipitation at a marine background location. Different manual methods for the sampling and analysis of total gaseous mercury (TGM) on gold and silver traps were compared with each other and with new automated analyzers. Additionally, particulate-phase mercury (Hg_part) in ambient air, total mercury, reactive mercury and methylmercury in precipitation were analyzed by some of the participating laboratories. Whereas measured concentrations of TGM and of total mercury in precipitation show good agreement between the participating laboratories, results for airborne particulate-phase mercury show much higher differences. Two laboratories measured inorganic oxidized gaseous mercury species (IOGM), and obtained levels in the low picogram m^-3 range.     

Estimating Long-Term Ground Level Concentrations of SO2 from Short-Term Peak Data

Maximum concentrations of sulfur dioxide vary approximately inversely as the square root of the averaging time for periods up to a year, a much longer time period than those reported in previously published studies. This finding is based on analyses of SO₂ and tracer gas concentrations measured in the vicinities of a test stack, a petrochemical complex, and eight power plants. Therefore, for a given set of emission conditions, long-term concentrations of SO₂ and other conservative pollutants emitted by well-defined sources can be rapidly estimated from maximum short-term concentrations, and vice versa. The scatter in the data corresponds to an estimation error of about a factor of three for averaging time conversions of one day to one year. This is approximately the same error associated with conventional calculations. Therefore, the use of averaging time conversions is justified as a rapid screening technique to estimate compliance with ambient air quality standards.     

Reactive plume model: Effect of stack exit conditions on gas phase precursors and sulfate formation

A simple form of a carbon-bond smog mechanism has been introduced into a complete set of plume conservation equations. The conservation expressions for mass, momentum, energy and component species have been solved numerically for typical source and neutral or adiabatic ambient conditions to simulate the homogeneous gas phase chemistry for approximately two hours of travel time downwind from the stack exit. The influence of stack exit conditions including the ratios of momentum-to-buoyancy length l_m/l_b, source-to-ambient velocity R and mixing-to-reaction time τ are demonstrated.It was found that homogeneous processes in the plume near-field do not significantly contribute to the overall conversion of SO₂ to SO²⁻₄. In the far-field maximum sulfate formation rates of 2.3% h⁻¹ were predicted for clear summer noon hours, consistent with plume measurements. Variation of stack exit conditions were found to have little effect on the maximum SO₂ conversion rate or the peak OH radical concentration where local HC/NO_x ratios between 16 and 22 were predicted. Parameter changes resulting in greater ambient entrainment rates, however, were found to shift the development of the radical pool closer to the source and to significantly increase the total molar flux of plume sulfate.     

Measurement of Trace Levels of Atmospheric Sulfur Dioxide with a Gold Film Sensor

The mercury translation principle, wherein S(IV) reacts with Hg(l) to produce Hg(ll) and Hg(O), was utilized to measure trace levels of SO₂ with a conductometric gold film sensor. One approach involved impregnated glass fiber filters containing mercurous salts. The second and more successful approach utilized a porous membrane reactor where the reaction occurs in the interstitial liquid-gas interface as the sample gas flows inside and a dilute solution of Hg₂(NO₃)₂ flows outside the porous PTFE membrane tube. The first approach attained a LOD of 200 pptv SO₂(g), however, batch to batch reproducibility of the filters was poor and the filters could not be stored more than two days. The membrane reactor method provided better translation efficiency and highly reproducible results, with an LOD of 50 pptv SO₂(g). An intercomparlson of this method with an independent method relying on completely different chemistry showed good agreement for the measurement of subppbv levels of ambient SO₂. Both mercury translation methods were essentially unaffected by H₂S or CH₃SH.     

Needed: A National R&D Effort To Develop Individual Air Pollution Monitor Instrumentation

The results from the regional air quality analysis for the Four Corners Study are discussed in this paper. This study was one of five regional studies conducted for the National Commission on Air Quality. Potential regional air quality impacts were evaluated through the year 1995 for alternative energy scenarios under current and alternative regulatory policies. The alternative regulatory policies include emission fees, technology standards, emission ceilings, and prevention of significant deterioration class elimination. The alternatives were compared in terms of their impacts on regional visibility and on the ambient concentrations of SO₂, SO₄, and primary fine particulates. The fate of the pollutants was estimated. The alternative regulatory policies were quite different with respect to their control of SO_X emissions and their impacts on regional visibility. Sources located outside of the study region were estimated to have a major impact on regional air quality within the study region.     

An evaluation of artifact SO42− formation on nylon filters under field conditions

The extent of SO₂ conversion on Membrana (Ghia) Nylasorb nylon filters under field conditions has been evaluated and found to be quite variable. The S-SO₄²⁻ loading on the nylon filters is higher at higher SO₂ concentrations, and on a long term basis approaches a saturatio limit of 2.5 μg S-SO₄²⁻ on a 47mm disc, at a dosage of 230 μg SO₂ approximately. The % conversion decreases as the SO₂ concentration increases. On a long term basis, at an SO₂ concentration range of 1.0–7.7 μg m⁻³, the conversion ranges from 8.2% to 2.1%. The dependence of SO₂ conversion on nylon filters on relative humidity displays a diurnal pattern. An expression has been derived to explain the observed % SO₂ conversion on nylon filters as a combined effect of the ambient SO₂ concentration and relative humidity.     

Use of Plants for Air Pollution Monitoring

Leaf injury data from acute and chronic exposure studies of Dare soybean were regressed against the logarithms of exposure time and O₃ and SO₂ concentrations to develop a new two-pollutant leaf injury model (which explains 88% of the variance) and to calculate the parameters of best fit for this new model and a previously developed one-pollutant model. Using the calculated parameters, the percentage of leaf surface Injured over a growing season by O₃, SO₂, or both simultaneously was estimated for an ambient air sampling site located 2 miles from a coal burning power plant. For this site, the one- and two-pollutant models predicted that SO₂ effects would be negligible If SO₂ concentrations never exceeded the National Ambient Air Quality Standard (NAAQS) of 0.50 ppm, averaged over 3 h. However, calculations suggest that O₃ may injure up to 24% of Dare soybean leaf surface over a growing season even though the O₃ NAAQS of 0.12 ppm, averaged over 1 h, is never exceeded. Because the 3 h SO₂ standard is exceeded at very few places, the O₃ model is usually sufficient to estimate Dare soybean leaf Injury. Leaf injury is estimated by taking the logarithm of the summation of each daytime hour’s exponentiated O₃ concentration (c) measured at an ambient air sampling site over a growing season. This is expressed as: z = -0.0828 + 0.4876 in (Σco₃ ^2.618), where z is the Gaussian transform of percent leaf injury. The methods developed in this paper, using Dare soybean data as an example, may apply to other plants.     

The Danish eulerian hemispheric model — a three-dimensional air pollution model used for the arctic

A three-dimensional Eulerian hemispheric air pollution model, the Danish Eulerian Hemispheric Model (DEHM), is in development at the National Environmental Research Institute (NERI). The model has been used to study long-range transport of air pollution in the Northern Hemisphere. The present version of the model includes long-range transport of sulphur dioxide (SO₂) and particulate sulphate (SC₄²⁻. The chemistry in the model is described by a simple linear oxidation of SO₂ to SO₄²⁻, and the wet deposition of SO₂ and SO₄⁻ is estimated based on the amount of precipitation, which is calculated from the contents of liquid cloud water (see Christensen, Air Pollution Modelling and its Applicatioons, Vol. X, pp. 119–127, Vol. XI, pp. 249–256, Plenum press, New York; 1995, Ph.D. thesis, National Environmental Research Institute, Denmark). The model has been used to study the air pollution in the Arctic. Results from yr simulation with an analysis of the results is presented: the model results are verified by comparisons, to measurements not only from the Arctic region but also from Europe and Canada. Some examples of episodes in the Arctic including analysis of the meteorological conditions during the episodes are presented. Finally, the model has been used to estimate the contribution from the different source regions on the northern hemisphere to the Arctic sulphur pollution.     

Shock Wave Cleaning of Air Filters

The trends in and relationships between ambient air concentrations of sulfur dioxide and sulfate aerosols at 48 urban sites and 27 nonurban sites throughout the U.S. between 1963 and 1972 have been analyzed. The substantial decreases in ambient SO₂ concentrations measured at urban sites in the eastern and midwestern U.S. are consistent with the corresponding reductions in local SO₂ emissions, but these decreases have been accompanied by only modest decreases in ambient sulfate concentrations. Large differences in the amounts of SO₂ emitted within individual air quality control regions are associated with much smaller differences in the corresponding ambient sulfate concentrations. Substantial changes in the patterns of SO₂ emissions between air quality regions result in essentially no differences between ambient sulfate concentrations in those air quality regions. Comparisons of several air quality regions in the eastern and western U.S. with similar SO₂ emission levels and patterns of emissions clearly demonstrates the higher ambient sulfate concentration levels in eastern air quality control regions. Relationships between SO₂, sulfates, and vanadium concentrations at eastern nonurban U.S. sites cannot be explained by local emission sources. These various observed results can be best explained by long distance sulfur oxide transport with chemical conversion of SO₂ to sulfates occurring over ranges of hundreds of kilometers. This conclusion has been suggested earlier and the present analysis strongly supports previous discussions. An impact of long range transport of sulfates is to emphasize the need for Consistent strategies for reduction of sulfur oxides throughout large geographical regions. Additions of large capacities involving elevated sources in mid-continental or western regions could result in significant increases in sulfate concentrations well downwind of such sources. Some of the types of research activities required to quantitate crucial experimental parameters are discussed.     

The development of sulfur dioxide and ozone exposure profiles that mimic ambient conditions in the rural southeastern united states

A design for constructing experimental mixed-pollutant exposure profiles that reflect regional O₃ and SO₂ ambient air quality is described. The profiles were developed using hour-by-hour O₃ and SO₂ concentration data from monitoring sites in the southeastern United States where slash pine is indigenous. Only sites designated rural or remote, with co-monitored O₃ and SO₂, and at least 75% of the hourly values reported for the period April– October, were used. Each site was characterized by concentration, frequency of occurrence and duration of concentration values, length of time between episodes, and frequency of cooccurrence. A base profile, a 30-day hour-by-hour concentration regime, was constructed using averaged air quality characteristics from the selected sites. Using the base profile, additional regimes were constructed by increasing the concentration of all hourly values above a designated minimum, or by increasing the frequency of occurrence of selected hourly concentrations.     

Emissions Variability Processor (EMVAP): Design,evaluation, and application

Emissions of pollutants such as SO₂ and NOx from external combustion sources can vary widely depending on fuel sulfur content, load, and transient conditions such as startup, shutdown, and maintenance/malfunction. While monitoring will automatically reflect variability from both emissions and meteorological influences, dispersion modeling has been typically conducted with a single constant peak emission rate. To respond to the need to account for emissions variability in addressing probabilistic 1-hr ambient air quality standards for SO₂ and NO₂, we have developed a statistical technique, the Emissions Variability Processor (EMVAP), which can account for emissions variability in dispersion modeling through Monte Carlo sampling from a specified frequency distribution of emission rates. Based upon initial AERMOD modeling of from 1 to 5 years of actual meteorological conditions, EMVAP is used as a postprocessor to AERMOD to simulate hundreds or even thousands of years of concentration predictions. This procedure uses emissions varied hourly with a Monte Carlo sampling process that is based upon the user-specified emissions distribution, from which a probabilistic estimate can be obtained of the controlling concentration. EMVAP can also accommodate an advanced Tier 2 NO₂ modeling technique that uses a varying ambient ratio method approach to determine the fraction of total oxides of nitrogen that are in the form of nitrogen dioxide. For the case of the 1-hr National Ambient Air Quality Standards (NAAQS, established for SO₂ and NO₂), a “critical value” can be defined as the highest hourly emission rate that would be simulated to satisfy the standard using air dispersion models assuming constant emissions throughout the simulation. The critical value can be used as the starting point for a procedure like EMVAP that evaluates the impact of emissions variability and uses this information to determine an appropriate value to use for a longer term (e.g., 30-day) average emission rate that would still provide protection for the NAAQS under consideration. This paper reports on the design of EMVAP and its evaluation on several field databases that demonstrate that EMVAP produces a suitably modest overestimation of design concentrations. We also provide an example of an EMVAP application that involves a case in which a new emission limitation needs to be considered for a hypothetical emission unit that has infrequent higher-than-normal SO₂ emissions.

ImplicationsEmissions of pollutants from combustion sources can vary widely depending on fuel sulfur content, load, and transient conditions such as startup and shutdown. While monitoring will automatically reflect this variability on measured concentrations, dispersion modeling is typically conducted with a single peak emission rate assumed to occur continuously. To realistically account for emissions variability in addressing probabilistic 1-hr ambient air quality standards for SO₂ and NO₂, the authors have developed a statistical technique, the Emissions Variability Processor (EMVAP), which can account for emissions variability in dispersion modeling through Monte Carlo sampling from a specified frequency distribution of emission rates.     

Assessment of air quality benefits from national air pollution control policies in China. Part II: Evaluation of air quality predictions and air quality benefits assessment

Following the meteorological evaluation in Part I, this Part II paper presents the statistical evaluation of air quality predictions by the U.S. Environmental Protection Agency (U.S. EPA)’s Community Multi-Scale Air Quality (Models-3/CMAQ) model for the four simulated months in the base year 2005. The surface predictions were evaluated using the Air Pollution Index (API) data published by the China Ministry of Environmental Protection (MEP) for 31 capital cities and daily fine particulate matter (PM_2.5, particles with aerodiameter less than or equal to 2.5 μm) observations of an individual site in Tsinghua University (THU). To overcome the shortage in surface observations, satellite data are used to assess the column predictions including tropospheric nitrogen dioxide (NO₂) column abundance and aerosol optical depth (AOD). The result shows that CMAQ gives reasonably good predictions for the air quality.The air quality improvement that would result from the targeted sulfur dioxide (SO₂) and nitrogen oxides (NO_x) emission controls in China were assessed for the objective year 2010. The results show that the emission controls can lead to significant air quality benefits. SO₂ concentrations in highly polluted areas of East China in 2010 are estimated to be decreased by 30–60% compared to the levels in the 2010 Business-As-Usual (BAU) case. The annual PM_2.5 can also decline by 3–15 μg m^?3 (4–25%) due to the lower SO₂ and sulfate concentrations. If similar controls are implemented for NO_x emissions, NO_x concentrations are estimated to decrease by 30–60% as compared with the 2010 BAU scenario. The annual mean PM_2.5 concentrations will also decline by 2–14 μg m^?3 (3–12%). In addition, the number of ozone (O₃) non-attainment areas in the northern China is projected to be much lower, with the maximum 1-h average O₃ concentrations in the summer reduced by 8–30 ppb.     

High-resolution simulation of volcanic sulfur dioxide dispersion over the Miyake Island

After severe eruptions of the volcano at Miyake Island in August 2000, a large amount of volcanic gas was released into the atmosphere. To simulate flows and dispersion of sulfur dioxide (SO₂) over Miyake Island, a set of numerical models was developed. The multi-nesting method was adopted to reflect a realistic meteorological field and to sufficiently resolve the flow over the island with a diameter of 8 km. The outermost model was the Regional Spectral Model (RSM) of the Japan Meteorological Agency (JMA) with a horizontal grid size of 10 km. Finer atmospheric structure was simulated with the nonhydrostatic model jointly developed by the Meteorological Research Institute and the Numerical Prediction Division of JMA (MRI/NPD-NHM) with grid intervals of 2 km, 400 m and 100 m. Realistic topography of the island was represented in the innermost model. The Lagrangian particle method was applied to the dispersion model, which is driven by the meteorological field of the 100 m grid MRI/NPD-NHM. The random walk procedure was used to represent the turbulent diffusion. The model was verified in four cases. Simulated SO₂ concentrations agreed well with observed concentrations at a monitoring station including temporal variation. Under a large synoptic change, however, accurate prediction became difficult. Further numerical experiments have been done to investigate characteristics of the flow and the distribution of SO₂. Steady inflows, classified according to the surface wind speed and direction, were assumed. Simulated SO₂ distribution on the ground apparently depends on the surface wind. Under relatively weak inflow, there is a large diurnal change in SO₂ distribution, affected by the thermally induced flow. SO₂ gas is widely spread downstream in the nighttime but hardly reaches the coastal area in the daytime. On the other hand, SO₂ gas steadily reached the downstream coast with little diurnal variation under the stronger inflow. Ground temperature, as well as the static stability of the inflow, also influences downstream wind, turbulent diffusivity and SO₂ distribution.