Industrial Source Complex Short-Term Model Research Articles

Emission load distribution and prediction of NO2 and PM10 using ISCST3 and CALINE4 line source modeling

Mysore district is one of the tourist destinations in Karnataka, India, with a population of about 10.25 lakhs. In all category of vehicles, 2-wheelers (2W) constitute more than 50 %. Vehicular count studies showed a maximum number of vehicles during peak traffic hours (7:00-11:00 hrs. and 18:00-22:00 hrs.). The maximum traffic count was observed for 2W (51-56%) followed by 4W (19-28%), LDV (light-duty vehicle), HDV (heavy-duty vehicle) and 3W. However, the percentage contribution of NO 2 and PM 10 emission factor showed, a maximum of HDV (75-89%) followed by LDV, 2W, 3W, and 4W. Emission loads for CO and NO 2 pollutants were found to be high at link 1, 2, 5, 6, 12 and 13 due to heavy traffic densities. However, emission loads for PM 10 , PM 2.5, and SO 2 were notices to be very less in all the links. During winter season Industrial source complex short-term model 3 (ISCST3) and California line source model 4 (CALINE4) predicted NO 2 concentrations were found to be ranging from 9–25 μg/m 3 and 10–40 μg/m 3 , respectively. However, PM 10 concentrations were found to be 18–43 μg/m 3 . The maximum concentration was found near to Receptors (R1-3), which may be due to maximum vehicular density and downwind to the pollution sources. NO 2 and PM 10 predictions made by ISCST3 and CALINE4 model showed a significant variation. It was also observed that CALINE4 model was not applicable to predict pollutant with lower concentrations.

Analysis of air pollution emission patterns in the vicinity of oil refineries in Kuwait

This study focuses on the assessment and analysis of air pollution in the vicinity of oilrefineries in Kuwait. The emission inventory obtained for oil refineries was used topredict pollutant concentrations and distributions using the AERMOD software (V5.1),Industrial Source Complex Short-Term model (ISCST3). Pollutant concentrations andmeteorological parameters were measured at 5-minutes intervals for 24 hours/day inthree regions. The first region was in the Fahaheel residential area (north of the oilrefineries), the second region was in the Umm Alhyman residential area (south of theoil refineries), and the third region was in the Ahmadi residential area (northwest ofthe oil refineries). The measured concentrations of all pollutants are higher than thepredicted concentrations generated by the refinery emissions, which is attributed tothe effect of other pollution sources, e.g., traffic, power plants, and oil fields. Analysisshowed that the strongest contribution comes from the oil fields, followed by trafficor oil refineries.

Pesticides Exposure Assessment of Kettleman City Using the Industrial Source Complex Short-Term Model Version 3

Kettleman City, California, reported a higher than expected number of birth defect cases between 2007 and 2010, raising the concern of community and government agencies. A pesticide exposure evaluation was conducted as part of a complete assessment of community chemical exposure. Nineteen pesticides that potentially cause birth defects were investigated. The Industrial Source Complex Short-Term Model Version 3 (ISCST3) was used to estimate off-site air concentrations associated with pesticide applications within 8 km of the community from late 2006 to 2009. The health screening levels were designed to indicate potential health effects and used for preliminary health evaluations of estimated air concentrations. A tiered approach was conducted. The first tier modeled simple, hypothetical worst-case situations for each of 19 pesticides. The second tier modeled specific applications of the pesticides with estimated concentrations exceeding health screening levels in the first tier. The pesticide use report database of the California Department of Pesticide Regulation provided application information. Weather input data were summarized from the measurements of a local weather station in the California Irrigation Management Information System. The ISCST3 modeling results showed that during the target period, only two application days of one pesticide (methyl isothiocyanate) produced air concentration estimates above the health screening level for developmental effects at the boundary of Kettleman City. These results suggest that the likelihood of birth defects caused by pesticide exposure was low.

Development of a Livestock Odor Dispersion Model: Part II. Evaluation and Validation

ABSTRACT A livestock odor dispersion model (LODM) was developed to predict odor concentration and odor frequency using routine hourly meteorological data input. The odor concentrations predicted by the LODM were compared with the results obtained from other commercial models (Industrial Source Complex Short-Term model, version 3, CALPUFF) to evaluate its appropriateness. Two sets of field odor plume measurement data were used to validate the model. The model-predicted mean odor concentrations and odor frequencies were compared with those measured. Results show that this model has good performance for predicting odor concentrations and odor frequencies. IMPLICATIONS This study provides an evaluation and validation for a newly developed LODM. Two sets of field odor plume measurement data were used. Agreement fractional bias values were applied to compare modeled predicted odor intensities and measured intensities. Fac2 (the percentage of the predictions within a factor of 2 of the observed values) and defined Ran0.1 and Ran0.2 values (the percentage of the predictions within a range of ±0.2 and ±0.1, respectively, of the observed values) were used to validate the model performance of predicting odor frequency.

Air quality modelling study to analyse the impact of the World Bank emission guidelines for thermal power plants in Delhi

Recent strategies for air pollution control in Delhi have largely neglected the emission reduction measures from thermal power plants (TPPs), which are the second most polluting sources. The present study investigates how the ambient air quality of Delhi would improve if the World Bank emission guidelines (WBEG) for the TPPs were to be implemented. To accomplish this, a comprehensive inventory of point, area, and line sources was conducted in the selected study area, primarily aiming to estimate the sectoral emission contributions to ambient air quality. The Industrial Source Complex Short–Term Model, Version 3 (ISCST3) was used to predict the ambient concentrations of total suspended particulates (TSP), sulphur dioxide (SO2), and nitrogen dioxide (NO2) at seven monitoring sites (receptor locations) operated by the Central Pollution Control Board (CPCB) for the period from July 2004 to June 2005. The ISCST3 model predictions for TSP and NO2 were satisfactory at all receptor locations. However, for SO2, the model predictions were satisfactory at only two receptor locations. The vehicles contributed 58% of the total ambient air pollution, followed by TPPs contributing 30%. The study estimates that adoption of WBEG may reduce the ambient air pollution due to TPPs emissions by 56% to 82%, bringing it within the National Ambient Air Quality Standards (NAAQS) set for industrial areas in India, except at one location where TPP’s contribution to ambient air pollution is negligible compared to vehicular emissions.

Soil Suspension/Dispersion Modeling Methods for Estimating Health-Based Soil Cleanup Levels of Hexavalent Chromium at Chromite Ore Processing Residue Sites

The primary health concern associated with exposures to chromite ore processing residue (COPR)-affected soils is inhalation of hexavalent chromium [Cr(VI)] particulates. Site-specific soil alternative remediation standards (ARSs) are set using soil suspension and dispersion models to be protective of the theoretical excess cancer risk associated with inhalation of soil suspended by vehicle traffic and wind. The purpose of this study was to update a previous model comparison study that identified the 1995 AP-42 particulate emission model for vehicle traffic over un-paved roads and the Fugitive Dust Model (FDM) as the most appropriate model combination for estimating site-specific ARSs. Because the AP-42 model has been revised, we have updated our past evaluation. Specifically, the 2006 AP-42 particulate emissions model; the Industrial Source Complex–Short Term model, version 3 (ISCST3); and the American Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD) air dispersion models were evaluated, and the results were compared with those from the previously used modeling approaches. Two sites with and two sites without vehicle traffic were evaluated to determine if wind erosion is a significant source of emissions. For the two sites with vehicle traffic, both FDM and ISCST3 produced total suspended particulate (TSP) estimates that were, on average, within a factor of 2 of measured; whereas AERMOD produced estimates that were as much as 5-fold higher than measured. In general, the estimated TSP concentrations for FDM were higher than those for ISCST3. For airborne Cr(VI), the ISCST3 model produced estimates that were only 2- to 8-fold of the measured concentrations, and both FDM and AERMOD estimated airborne Cr(VI) concentrations that were approximately 4- to 14-fold higher than measured. Results using the 1995 AP-42 model were closer to measured than those from the 2006 AP-42 model. Wind erosion was an insignificant contributor to particulate emissions at COPR sites.

The role of meteorology on predicting SO 2 concentrations around a refinery: A case study from Oman

In this work, SO 2 correlations were developed for the prediction of maximum SO 2 values and their locations around the vicinity of a refinery. The proposed correlations are capable of estimating the hourly maximum SO 2 concentrations from meteorological conditions. Correlation parameters were calculated by multiple regression analysis, using maximum SO 2 concentration as dependent variable and the meteorological parameters as independent variables. The SO 2 data used for the development of these correlations were generated from the industrial source complex short-term (ISCST) model. It was found that wind speed and atmospheric stability class had the most effect on the predicted SO 2 concentration whereas neither mixing height, nor wind direction, nor temperature had an influence on the maximum SO 2 concentration. Therefore, the suggested correlations require only knowledge of the wind speed and stability class parameters. On the other hand, the developed correlations for estimating the locations of these maximum values of SO 2 concentrations contained only one term that describes the dependence of the locations on wind direction. The derived correlations were shown to be statistically significant. They are much simpler to use than the ISCST model. Further, they are invaluable for determining locations at risk of exceeding the SO 2 standard.

Evaluation of the Industrial Source Complex Short-Term Model: Dispersion over Terrain

Terrain around an air discharge source can have several influences on diffusion, the pattern of plume dispersion, the wind flow, and the turbulence characteristics. The Industrial Source Complex Short-Term (ISCST) model contains simple algorithms to attempt to account for the effects of terrain. The model has the ability to analyze concentrations in any type of terrain by using the terrain options available for running the model. In this study, the ISCST model was adopted to predict the concentration of sulfur dioxide (SO2) in and around the Mina Al-Fahal refinery in Oman. The central purpose of the study was to examine the performance of the ISCST model in predicting SO2 concentrations under two different scenarios: (1) when flat terrain was assumed; and (2) when the terrain descriptions were addressed. The results of these two scenarios were validated against SO2 monitoring data. The comparison showed that the model underestimated the observed concentrations for the two scenarios. However, the predicted concentrations of SO2 in the absence of the terrain scenario were in better agreement with the observations. Furthermore, the predicted SO2 concentrations were found to be lower than the World Health Organization guideline values, with the maximum concentrations found to occur relatively close to the sources of emission.

Near-Field Dispersion Modeling for Regulatory Applications

This paper evaluates the application of dispersion models to estimate near-field pollutant concentrations in two case studies. The Industrial Source Complex Short-Term Model (ISCST3) was evaluated with hexavalent chromium measurements collected within 100 m of two facilities in Barrio Logan, San Diego, CA. ISCST3 provided reasonable estimates for higher pollutant concentrations but underestimated lower concentrations. To understand the observed distribution of concentrations in Barrio Logan, a recently conducted tracer experiment was analyzed. The tracer, sulfur hexafluoride, was released at ambient temperature from an urban facility at the University of California at Riverside, and concentrations were measured within 20 m of the source. Modeling results indicated that Industrial Source Complex–Plume Rise Model Enhancement and American Meteorological Society/U.S. Environmental Protection Agency Regulatory Model–Plume Rise Model Enhancement overestimated high concentrations and underestimated low concentrations. A diagnostic study with a simple Gaussian dispersion model that incorporated site-specific meteorology was used to evaluate model results. This study found that incorporating lateral meandering for nonbuoyant urban plumes in Gaussian dispersion models could improve concentration estimates even when downwash is not considered. Incorporating a meandering component in ISCST3 resulted in improvements in estimating hexavalent chromium concentrations in Barrio Logan. Credible near-source concentration estimates depend on accurate characterization of emissions, onsite micrometeorology, and a method to account for lateral meandering in the near field.

Assimilative capacity and pollutant dispersion studies for the industrial zone of Manali

In this study the micro-meteorological parameters have been analyzed and the assimilative capacity of Manali is estimated on the basis of ventilation coefficient for the four seasons of the year 1998. Among the various seasons, summer has the lowest pollution potential. The industrial source complex short-term model has been used to predict the spatial distribution of three pollutants namely sulfur dioxide, oxides of nitrogen and suspended particulate matter. The model has been validated using the data measured at site. Statistical evaluation of the model indicates satisfactory performance. Using the model predictions, isopleths have been drawn. The northeastern part of Manali has been identified as vulnerable to high levels of pollution necessitating mitigation measures to be initiated in the region.

Transfer of dioxin risk between nine major municipal waste incinerators in Taiwan

The objective of this study was to assess site-specific carcinogenic risks of incinerator-emitted dioxins and risk transfers among the areas covered by nine municipal incinerators in Taiwan. We used actual emission data and the industrial source complex short-term model (ISCST3) to determine the dioxin impact areas within the 8×8-km simulation regions surrounding the incinerators. We then used multimedia model to estimate cancer risks in individual impact areas for two exposure scenarios, which were sufficient (SFP) and insufficient food production (IFP) for residents' consumption in each impact area. We also used information of food supply and consumption between impact areas to calculate risk transfers among these nine incinerators. We found that dioxins' carcinogenic risks ranged from 1.4×10 −8 (Incinerator F) to 7.1×10 −5 (Incinerator A) for the nine incinerators under the exposure scenario of SFP, and ranged from 8.7×10 −8 (Incinerator D) to 1.1×10 −6 (Incinerator E) under the exposure scenario of IFP. The food ingestion was the main exposure pathway, which accounted for 64–99% of total dioxin risks among nine impact areas. For the nine major food items consumed by residents in the impact areas, eggs (14–35%) and chicken (11–26%) were two main routes of dioxin exposure in the SFP scenario, while chicken (8–78%) and vegetables (0.2–81%) were two main routes of dioxin exposure in the IFP scenario. Significant risks of dioxins were transferred among incinerators, which accounted for up to 88% among the incinerators. Incinerator E was the major risk-exporting source to six Incinerators C, D, F, G, H, and I. For these six incinerators, Incinerator E accounted for their 51–88% imported risks. We concluded that risk transfers among incinerators through routes of food consumption should be considered in assessing health risks associated with incinerator-emitted dioxins in Taiwan. We should place high priority on implementing control measures to lower dioxin emissions in important food-exporting areas like Incinerator E. We should also emphasize analyzing dioxin contents in eggs, chicken, and vegetables in order to improve dioxin-related health risk assessments in the future.

Patterns of SO 2 emissions: a refinery case study

Air quality modelling is an essential tool for most air pollution studies and the introduction of SO 2 standards creates a need for modelling the dispersion of SO 2. This work deals specifically with the use of the Industrial Source Complex Short Term (ISCST) model at a refinery. The study is performed over a period of 21 days. The first objective of this study was to measure the atmospheric levels of SO 2 and then to compare their values with the international standard limits. The second objective was to evaluate the ISCST model by comparing the calculated and measured concentrations. The third objective was to demonstrate the effect of wind regimes on the dispersion of SO 2 and to determine the spatial distribution of SO 2 over the modelled area. The results showed that the levels of SO 2 were well below the ambient air quality standard. Based on isopleths for SO 2 distribution in the study area (as output from the ISCST model), it can be stated that no health risk is present in areas adjacent to the refinery.

Monitoring of nitrogen dioxide in and around an Oman liquid natural gas plant using passive diffusion tubes: exposure and ISCST model study

Recognising that nitrogen dioxide (NO2) may be the primary air contaminant released from an Oman liquid natural gas (LNG) plant, this study has been designed with emphasis on this pollutant. The dispersion of NO2 was examined and compared with measured data during the period from 23 February to 1 April 2001. The observed concentrations were obtained by using NO2 diffusion tubes to monitor the spatial distribution of NO2 concentrations around at four sites downwind of the plant site. The Industrial Source Complex Short-Term (ISCST) dispersion model was used to predict concentrations at the same sites and for the same time period. Evaluation of the results showed that the model predicted concentrations lower than those observed. However, the comparisons showed that the relationship between the modelled and measured concentrations was consistent. The model also allows the detailed description of predicted concentrations in contour map form. Generally, the predicted concentrations were found to be lower than the permissible limit values, with the maximum ground-level concentrations being found much closer to the plant. On the basis of this study, it can be stated that the Oman LNG plant would not cause any significant deterioration in air quality.

Predicted 1,3-dichloropropene air concentrations resulting from tree and vine applications in California.

The preplant soil fumigant 1,3-dichloropropene (1,3-D) is effective for nematode control and is expected to further replace methyl bromide (MeBr) as MeBr use is phased out. Acute human exposure to soil fumigants is managed in part by using buffer zones between treated fields and occupied structures. The required buffer zone for 1,3-D in California is 91.4 m (300 ft) for all uses. However, a 30.5-m (100-ft) buffer setback is desired for 1,3-D to be an important replacement for MeBr in the orchard and vineyard markets. The Industrial Source Complex Short-Term model, Version 3 (ISCST3) was used to simulate township-wide long-term average and short-term air concentration distributions of 1,3-D. The Gaussian plume model ISCST3 can be used to assess dispersion of air pollutants and pollutant concentrations on receptors from a variety of sources and in diverse airsheds. Long-term and daily-average air concentrations can be compared with the California permitted chronic or acute toxicity endpoints, respectively, to assess the potential risk for individuals living within the township at the proposed buffer setback. Modifications to ISCST3 were made for specific nonpoint-source agricultural constraints and management practices. Chronic and acute air concentration distributions of 1,3-D with a 30.5-m buffer constraint around treated fields are similar to currently permitted air concentration distributions in California. Refinement of exposure as a function of buffer distance, application rate, and field size is possible due to the resolution of the simulation and external post-processing capabilities. Simulated examples of 1,3-D acute and chronic exposure cumulative distributions are presented.

Mapping the air in real-time to visualize the flow of gases and vapors: occupational and environmental applications.

This article describes a new method for measuring and mapping pollutants in air in real-time which can be used for visualizing the flow of gases and vapors in both indoor industrial and outdoor environmental applications. This method uses open-path Fourier Transform Infrared (OP-FTIR) spectrometry and computed tomography for real-time mapping of concentrations of chemicals in air. These maps may be used to evaluate human exposures, source emissions and air dispersion models; thus, this method can be used for both industrial and environmental sampling. It is being developed using computer simulations, and chamber and field studies. Computer simulations used simulated test concentration data to create maps; the original maps of concentrations were compared with the tomographic reconstructed maps. In the chamber studies, tracer gas was released into the chamber and measurements from a tomographic system were compared with point sample measurements taken at the same time. When sulfur hexafluoride was injected in a stable flow field position in the chamber, the concentrations reconstructed by the concentration maps were within +/- 15.9 percent of the measured point samples; overall, they were within +/- 27 percent of the measured point samples. On a 12-foot by 14-foot grid of cells used to model the chamber, the average peak location error was within one foot. The peak location error refers to the error involved in locating the point of highest concentration in the plume. For the field study, field-generated tomographic maps were compared with concentrations estimated using the Industrial Source Complex-Short Term (ISCST) model. Fairly good correlation (R2 = 0.67) was found between the five-minute overall-average cell concentrations in the tomographic and ISCST model maps. Overall, the tomographic map concentrations over-predicted the ISCST model concentrations by 24 percent. Optical remote sensing and computed tomography shows promise as a method to produce spatially and temporally resolved two-dimensional concentration maps indoors and outdoors. These maps would provide near real-time visualization of contaminant generation, movement, concentrations, and emission rates for multiple chemicals simultaneously at low limits of detection.

Testing the USA EPA's ISCST-Version 3 model on dioxins: a comparison of predicted and observed air and soil concentrations

The central purpose of our study was to examine the performance of the United States Environmental Protection Agency's (EPA) nonreactive Gaussian air quality dispersion model, the Industrial Source Complex Short-Term Model (ISCST3) Version 98226, in predicting polychlorinated dibenzodioxins and polychlorinated dibenzofurans concentrations (subsequently referred to as dioxins and furans, or CDD/Fs) in both air and soil near the Columbus Municipal Solid Waste-to-Energy Facility (CMSWTE) in Columbus, OH. During its 11 yr operation, the CMSWTE was estimated to be emitting nearly 1 kg of CDD/F Toxic Equivalents (TEQs) per year, making it one of the highest single emitters of dioxin in the United States during its operation. An ambient air-monitoring study conducted in 1994, prior to its shutdown in December of 1994, clearly identified high dioxin air concentration in the downwind direction during two sampling events. In one of the events, the CMSWTE stack was concurrently monitored for dioxins. A soil sampling study conducted in 1995/1996 was similarly able to identify an area of impacted soil extending mainly in the predominant downwind direction up to 3 km from the CMSWTE. Site-specific information, including meteorological data, stack parameters and emission rates, and terrain descriptions, were input into ISCST3 to predict ground-level 48-h concentrations which could be compared with the 48-h measured air concentrations. Predicted annual average dry and wet deposition of particle-bound dioxins were input into a simple soil reservoir model to predict soil concentrations that would be present after 11.5 yr of emissions, which were compared to measured concentrations. Both soil- and air-predicted concentrations were generally within a factor of 10 of observations, and judged to be reasonable given the small number of observations and the uncertainties of the exercise. Principal uncertainties identified and discussed include: source characterization (stack emission rates of dioxins), meteorological data, and atmospheric and soil fate and transformation processes of the dioxins

Prediction of Methyl Bromide Flux from Area Sources Using the ISCST Model

AbstractGrowing concern about atmospheric exposure of humans to pesticides has led to increased air monitoring in California. Air monitoring data typically consist of measurements made downwind from point or area sources. The utility of monitoring at fixed stations is limited for establishing buffer zones to protect neighboring populations from pesticide exposure. A modeling approach designed to use these data would provide the flexibility needed to establish buffer zones. The Industrial Source Complex Short Term (ISCST) model is a gaussian plume dispersion model that predicts air concentrations around point or area sources using emission rates (flux) and meteorological conditions as model inputs. The flux data, however, are typically not collected in the field nor available in the literature. In order to use the field data currently available, we developed a procedure using methyl bromide (MeBr) concentrations around two area sources and the ISCST model, to back‐calculate flux. In addition, MeBr flux was measured concurrently by cooperators conducting an independent study. Air concentrations, together with actual meteorological data, were used as inputs to the ISCST model. Flux of MeBr was then back‐calculated using a two step process: (i) an arbitrary flux value of 100 µg m−2 s−1 was used as an initial input value. Resultant air concentrations predicted by the model were then regressed on air concentrations measured offsite and (ii) the resultant regression coefficient was then used to adjust the arbitrary flux to a back‐calculated flux. Using another regression analysis, back‐calculated and measured flux rates were found to be significantly correlated, indicating this approach may be suitable for indirect estimation of flux. Implications for the use of this method in establishing buffer zones designed to protect human health are discussed.

Observed Downwash Concentrations Compared to ISCST Predictions in Urban Core

An atmospheric tracer (SF 6 ) was released from an industrial stack in an urban core during nocturnal neutral to stable atmospheric conditions in an effort to determine the dispersive characteristics and the atmospheric transport of stack effluent under these conditions. Tracer concentration data are compared to predictions made by the Environmental Protection Agency's Industrial Source Complex Short Term (ISCST) model. ISCST is the regulatory air-pollution model typically recommended in situations where structure-induced plume downwash is a possibility. Under the meteorological conditions and stack parameters sampled in the present study, the downwash algorithm causes the ISCST model to overpredict surface concentrations below 3.5 m/s or below a wind speed/exit velocity ( u /V e) ratio of 0.25. Model performance was better in neutral than in stable atmospheric conditions. The stack effluent temperature was >450°K, and the model overprediction may be due in part to the treatment of thermal buoyancy in the model treatment of building wake downwash.

Evaluation of the Industrial Source Complex short-term model in a large-scale multiple source region for different stability classes

The original Industrial Source Complex Model (ISCST, 90346) and its latest version, the ISCST2 (92090), are evaluated for the 1- and 24-hour averaging periods, using six statistical parameters. The confidence limits on two parameters were obtained using the bootstrap, jacknife, seductive, and robust resampling techniques. Evaluation is conducted in a multiple point source environment and in all the different stability categories, using a year of meteorological data, emission inventory of the Lucas County, in Toledo, Ohio, and monitoring data from two nearby stations for the year 1987. This was representative of a real time situation in which the ISCST is generally applied for regulatory work. The sensitivity analysis shows that the ISC is a poor performing model in the 1-h and 24-h averaging period, neutral and stable categories, and modifications to it are necessary in order to improve its performance. Its performance changes depending on the chosen paired values. A better relative performance is observed in the unstable category relative to the neutral and stable categories. Improved model performance may be achieved by applying modifications to the physics on which the model is based.

Comparison of the ISCST Model With Two Alternative U.S. EPA Models in Complex Terrain in Hamilton County, Ohio

In order to resolve the inconsistency between SO2 modeled and monitored values in Hamilton County, Ohio, and in response to the U.S. EPA’s requirement to Ohio EPA that the Hamilton County SO2 State Implementation Plan (SIP) be re-evaluated, a coalition of industries who are members of the Greater Cincinnati Chamber of Commerce (GCCC) conducted a technical study. The objective of this study was to compare selected U.S. EPA guideline dispersion models with SO2 monitoring data to determine which model was most accurate in predicting SO2 air quality in Hamilton County. The GCCC developed a protocol to perform this study. The protocol was approved by U.S. EPA and the Ohio EPA. Based upon the geographical proximity and quantity of SO2 emissions, Hamilton County was subdivided into west and east sub-areas. The model comparison was performed for each sub-area. The reference model for this study was the Industrial Source Complex Short-Term (ISCST) model and Rough Terrain Diffusion Model (RTDM) hybrid. Inthis model, ISCST was applied to all receptors upto stack height and RTDM was applied to all receptors above plume height. For the intermediate terrain receptors, i.e., those having elevations between stack height and plume height, the highest concentration from ISCST and RTDM was used. The first alternate model was chosen to be RTDM at all receptors. The second alternate model was chosen to be an ISCST/ RTDM hybrid in which ISCST without downwash was applied to receptors with elevations below stack height and RTDM was applied to receptors with elevations above stack height. Ambient SO2 concentrations were obtained from a network of 13SO2monitors(six in the west sub-area and seven in the east) which were operated for a one-year period beginning November 1, 1990. The monitors were sited based upon dispersion modeling conducted using the ISCST model with and without downwash and the RTDM model. The meteorological data for the one-year study period were obtained from two multi-level meteorological towers,a 100-meter tower in the west sub-area and a 60-meter tower in the eastsub-area. Hourly values of SO2 emission rate, exhaust gas exit temperature, and exhaust gas exit velocity were obtained from the eight major sources which were modeled in the model comparison study. The impact of other SO2, emission sources was accounted for in the hourly background determination. The reference model and the two alternate models were executed using the source and meteorological data to determine predicted SO2 concentrations at each of the 13 SO2 monitor locations. The predicted SO2 data were statistically compared to the observed SO2 data using the methodology referenced in Cox (1987). The test statistic used was the Robust Estimate of the Highest Concentration (RHC). Various RHCs were calculated from which a composite performance measure (CPM) was computed for each model and sub-area. In both sub-areas the winning model was RTDM. Using the Cox “bootstrap” resampling technique, it was shown that, to a 90 percent confidence interval, RTDM was statistically different from the reference model. As discussed in the protocol, the winning model could be used for purposes of either developing a revised Hamilton County SO2 SIP or proposing SO2 SIP revisions for individual facilities in Hamilton County. The protocol included procedures outlining how the winning model would be adjusted if it is shown to underpredicr monitored concentrations. For the west sub-area, there was no underprediction by RTDM at any of the six monitors. For the east sub-area, however, there was under prediction at two low terrain monitors and very small underprediction at one high terrain monitor. Based upon the procedures specified by U.S. EPA, the adjustment factor for RTDM in the east sub-area is 1.68 for 3-hour concentrations and 1.75 for 24-hour concentrations.