W. L. Freddy Kho, et al.

Computer simulated versus observed NO2 and SO2 emitted from elevated point source complex

1*W. L. Freddy Kho; 2J. Sentian; 2M. Radojevi; 2C. L. Tan; 1P. L. Law; 1S. Halipah

1Department of Civil Engineering, Universiti Malaysia Sarawak (UNIMAS) 94300 Kota Samarahan, Sarawak, Malaysia

2School of Science and Technology, Universiti Malaysia Sabah (UMS) Locked Bag No. 2073, 88999 Kota Kinabalu, Sabah, Malaysia

Int. J. Environ. Sci. Tech., 4 (2): 215-222, 2007ISSN: 1735-1472© Spring 2007, IRSEN, CEERS, IAU

Received 4 August 2006; revised 25 February 2007; accepted 6 March 2007; available online 20 March 2007

*Corresponding author, Email: frexwl@gmail.com Tel.: +6082-456877; Fax: +6082-456577

INTRODUCTIONSignificant emissions of air pollutants, particularly

in industrialised areas have always been a concern toin plant workers and nearby population in terms of airquality management. This is mainly due to thecomplexity of emissions sources, load of emissions,and type of emissions, variability of the localmeteorological and terrain conditions as well as thepresence of sensitive receptors in the surroundings ofthe air pollution sources. Air dispersion models (ADM)have been widely used to investigate the dispersionpatterns and behaviour of air emissions in such areas(Mehdizadeh and Rifai, 2004), and also to assess thepotential hazards to the human health (Zhou, et al.,2003; James, et al., 1995). ADMs are also used in airquality impact assessment on specific industrialfacilities or a group of industry to assess the cumulativeimpact on the downwind pollutant concentrations andto predict future air quality in the surrounding forenvironmental management planning purposes. Theyare also used as a tool in environmental auditingexercises on the assessment audit of air quality impact

as well as on the air pollution control efficiency onspecific facilities. The model generated results couldfacilitate the respective authorities to make appropriateactions in accordance to the requirements of the relevantenvironmental laws and regulations. Point sourceemissions from various industrial sources have beenthe target for investigating the pollutants dispersionpattern by using various types of ADMs. For example,in power plants emission, ADMS 3.1 dispersion modelhad been used to model the dispersion of SO2(Carruthers, et al., 1997; Bennet and Hunter, 1997;Carslaw and Beevers, 2002), ISCST3 dispersion modelfor CO and NOx (Venegas and Mazzeo, 2005), andCALPUFF dispersion model for PM2.5 and other gases(Zhou, et al., 2003). Air dispersion model such asSCREEN was also used to study the dispersion of VOCsfrom various types of industries and their impact tothe surrounding residential areas (James, et al., 1995).

MATERIALS AND METHODSLocality of study area

The objectives of this study were to model or predictthe maximum ground level concentrations of SO2 and

ABSTRACT: ISC-AERMOD dispersion model was used to predict air dispersion plumes from an diesel powerplant complex. Emissions of NO2 and SO2 from stacks (5 numbers) and a waste oil incinerator were studied to evaluatethe pollutant dispersion patterns and the risk of nearby population. Emission source strengths from the individualpoint sources were also evaluated to determine the sources of significant attribution. Results demonstrated thedispersions of pollutants were influenced by the dominant easterly wind direction with the cumulative maximumground level concentrations of 589.86 μg/m3 (1 h TWA NO2) and 479.26 μg/m3 (1 h TWA SO2). Model performanceevaluation by comparing the predicted concentrations with observed values at ten locations for the individual airpollutants using rigorous statistical procedures were found to be in good agreement. Among all the emission sourceswithin the facility complex, SESB-Power (diesel power plant) had been singled out as a significant source of emissionthat contributed >85% of the total pollutants emitted.

Key words: ISC-AERMOD, plumes, diesel-fired, power plant complex, emission, stacks

W.L., Freddy Kho, et al.

216

NO2emitted from Melawa Power Plant Facility Complex

(MPPFC) as shown in Fig. 1 in order to assess thecumulative impact of emission to the surroundingenvironment using ISC-AERMOD View. Potentialimpact of pollutants exposure on selected discretereceptors within the 2 km x 2 km domain coverage wasalso assessed. The study also sought to determine thepollutants sources strength from emission sourceswithin the MPPFC. MPPFC (latitude 6° 8’N; longitude116° 13’E) is located within the Kota Kinabalu IndustrialPark (KKIP). KKIP is an area established by the StateGovernment for development of various types ofindustrial activities ranging from light to heavyindustries, housing, commercial, institutions andconservation areas. As the development of KKIP areais still at its early stage, currently the only significantstationary source of air pollutants in this area is fromthe MPPFC. The MPPFC comprises of two power plants(SESB-Power and ARL-Power) and a waste oilincinerator (ARL-Incinerator). The facility is locatedon a 4.4-hactare land with elevation at approximately15 meters above mean sea level. It is located about 23km northeast of Kota Kinabalu City centre. The MPPFCis surrounded by industrial complexes on the east, andresidential houses and institutions on the west, northand south (Fig. 1). Within the domain coverage of 3 kmby 3 km, the topography is generally flat with ridgesranging from 5 m to 100 m in elevation.

Fig. 1: Locality of the Melawa Power Plant Facility Complex (MPPFC) and discreet receptors

Description of dispersion modelISC-AERMOD View is the latest version of ISC

models (Industrial Source Complex). It is a completeand powerful air dispersion modelling package whichseamlessly incorporates the popular United StatesEnvironmental Protection Agency (USEPA) models intoone interface: ISCST3, ISC-PRIME, AERMOD andAERMOD-PRIME. ISC-AERMOD view hasincorporated the Plume Rise Model Enhancements(PRIME) building downwash algorithms, which providea more realistic handling of downwash effects thanprevious approaches (MEO, 2003). PRIME model wasdesigned to incorporate two fundamentals featuresassociated with building wash which are (a) theenhanced plume coefficients due to the turbulent wake,and (b) the reduced plume rise caused by combinationof the descending streamlines in the lees of buildingsand the enlarged entrainments in the wakes. ISC-AERMOD packaged is supported by Building ProfileInput Program (BPIP) View which is a graphical userinterface designed to speed up the work involved insetting up input data for the BPIP. BPIP was used toperform building downwash analysis for point sourcesin order to run the model. Data input for location anddimension of buildings (structures) and stacks can bedone either by graphic, where the buildings and stackscan be digitised on the screen with the mouse, or bytext mode, where information can be added in aspreadsheet-like input form.

0 600km

W.L., Freddy Kho, et al.

217

Int. J. Environ. Sci. Tech., 4 (2): 215-222, 2007

ISC-AERMOD View has also incorporated essentialuser-friendly options such as complete RAMMET andAERMET meteorological data pre-processing andmultiple pollutant utilities for modelling multiplepollutants in a single ISCST3, ISC-PRIME or AERMODrun (Jesse, et al., 2000). The AERMET program is ameteorological pre-processor, which prepares hourlysurface data and upper air data in the modelling.AERMET processes meteorological data in threestages. Firstly, the meteorological data extracted fromarchive data files and processed the data throughvarious quality assessment checks. Secondly, all theavailable data for 24 h. periods were merged and storedin a single file. Lastly, the merged data was read andboundary layer parameters were estimated for use inthe AERMOD model. Wind rose was also plotted usingWRPLOT View based on the analysed meteorologicaldata.

The ISC-AERMOD View interface uses six pathwaysthat compose the run stream file as the basis for itsfunctional organization (Jesse, et al., 2000). Thesepathways are:• Control Pathway (CO): Where the modelling scenariois specified, and control of the overall modelling run;• Source Pathway (SO): Where the sources ofpollutant emissions are defined;• Receptor Pathway (RE): Where the air qualityimpacts on specific receptors locations are defined;• Meteorological Pathway (ME): Where theatmospheric conditions of the area being modelled aredefined;• Terrain Grid Pathway (TG): Where the option ofspecifying gridded terrain data to be used in calculatingdry deposition in elevated or complex terrain isavailable, and• Output Pathway (OU): Where the output results aredefined.

Data acquisitionMeteorological data

The meteorological parameters data input for thismodelling were obtained from the surface weatherobservatory station at Kota Kinabalu InternationalAirport located 40 km south of the MPPFC. A detailedanalysis of the meteorological data such as ceilingheight, wind speed, wind direction, air temperature, totalcloud opacity and total cloud amount has been madeover five years (1996-2001). The dominant winddirection was of easterly wind. The wind speed was inthe range between 1.3 and 3.6 m/s with an average of

1.97 m/s. A multi-directional wind was taken intoaccount for this study to calculate the overalldispersion of pollutant concentration surrounding thepoint sources area. About 35.5% of the time, thiseasterly wind speed was within the range between 0.5m/s and 3.6 m/s. Referring to the annual 24 h winddirection, the frequencies of other wind directions wereall below 10%. In determining atmospheric stability,Turner’s Stability Classification Method was used. Inthis study, for short time periods, atmospheric stabilityof the atmosphere was assumed to be constant andbased on the mean velocity of the wind speed, theatmospheric stability of Class B was chosen which isdefined to be moderately not stable. Movement of airin the vertical dimension is enhanced by verticaltemperature differences. The steeper the temperaturegradient is, the more vigorous the convective andturbulent mixing of the atmosphere. The larger thevertical column in which turbulent mixing occurs, themore effective the dispersion process, which is definedby the mixing height. Mixing height used in this studywas 1,000 m.

Building and stack dataFor modelling purposes, the physical and emission

data of each stack within the MPPFC site were obtainedfrom the database supply by MPPFC operators.Location and dimension of stack, building and anymajor structure within the complex were also obtainedas an input for BPIP in calculating and analysing thebuilding downwash effects. BPIP data input for locationand dimension of buildings (structures) and stacks canbe done either digitising graphically on the screen oradding structures information in a spreadsheet. Table1 summarises the operational characteristics and stackand emission properties of the MPPFC.

Air quality monitoringThe current status of air quality within the study

domain was determined by on site ground levelmeasurement. Ground level measurements of pollutantsat ten sites within the domain coverage wereestablished. The ground level monitoring sites (R1-R10) are situated at strategic locations mainly inresidential areas as well as institutional and industrialsites as shown in Fig. 1. The monitoring sites werealso identified as discrete receptors locations. Theground level concentrations data of pollutants wereused to validate the predicted value from the emissionmodelling. Table 2 summarises the ground level

W.L., Freddy Kho, et al.

218

Computer simulated versus observed No2 and So2 emitted from elevated...

pollutant measurements, instrumentation andoperational specifications.

Receptors informationIn evaluating the exposure of pollutants emission

from the MPPFC to the surrounding population, tendiscrete receptors (R1-R13) were selected within thestudy area (Fig. 1). Most of the receptors are residentialand institutional areas (Table 2). In assessing thepotential impact on public health, the predicted valueswere compared with/to the Malaysian Ambient AirQuality Guideline Standard (DOE Malaysia, 2000) forthe respective pollutants.

Model performance criteriaThe applications of rigorous statistical procedures

(Hanna, 1989; Kumar, et al., 1999) were used toquantitatively evaluate the performance of the model.Generally, the model would be considered acceptableif it meets the following performance criteria:

Table 1: Operational characteristics, stack and emission properties

SESB-Power ARL-Power ARL-Incinerator No. of Stack 3 1 (4 in 1) 1 Stack Identification 1,2 & 3 4 5 Stack Height 15 m 35 m 17 m Stack Diameter 1.0 m 1.6 m 0.46 m Exhaust Temperature 573K 623K 580K Exhaust Velocity 25.0 m/s 30 m/s 30 m/s Fuel Diesel Fuel

Oil Medium fuel oil Diesel fuel oil

NO2 31.0 g/s 62.80 g/s 0.04 g/s Emission Rate: SO2 28.0 g/s 56.80 g/s 0.08 g/s

Table 2: Receptors identification within the domain coverage

Receptors Description Direction from MPPFC Approximate distance (km) R1-Kayu Madang (Landfill) Landfill Northeast 2.5 R2-Kg. Norowot Village Northeast 3.0

R3-SIRIM Institution East 2.5

R4-Kg. Salut Village East 2.0

R5-Cemetery Cemetery East 1.6

R6-Politeknik KK Institution North 1.0 R7-KKIP Resettlement Scheme Residential North 0.8

R8-Powertron Residential Southeast 0.2

R9-Kg.Malawa/Tmn Sapangar Residential Southwest 0.8

R10-ILP R11-MARDI R12-LPPB Housing R13-Tmn Seri Maju

Institution Institution Residential Residential

South South South South

1.0 1.8 2.3 2.4

(a) Fractional Bias (FB) which is the mean error thatdefines the residual of the observed and the predicted

FB = (<Co>-<Cp>)/0.5(<Co+<Cp>)

Where Co is observed concentrations and Cp ispredicted concentrations. Note that a < > bracketindicates an average over all points in the group. Themodel is considered acceptable when, -0.5d”FBd”0.5. (b) Normalized mean square error (NMSE) whichemphasizes on the scatter of the entire data set. Zerovalue of NMSE indicates an ideal model performance.It is express as:

NMSE = < (Co-Cp) 2>)/ (<Co><Cp>)

Where Co is the observed concentrations and Cp isthe predicted concentrations. Note that a < > bracketindicates an average over all points in the group. Themodel is considered acceptable when, NMSE d”0.5.(c) Factor of 2 (FAC2) which is the percentage ofpredictions, where the predictions and observed valuesare within a factor of 2 from one another (0.5d” (Cp/Co)d”2). The model is considered acceptable when,FAC2³ 0.8.

concentrations. It has the value of zero for an idealmodel and varies between -2 and 2 and express as:

W.L., Freddy Kho, et al.

219

Int. J. Environ. Sci. Tech., 4 (2): 215-222, 2007

RESULTSCumulative and individual impacts assessment

Cumulative impact on ground level concentrationsof pollutants, emitted from combined sources wasmodelled and the results are depicted in Figs. 2 and 3 .Similar dispersion patterns were also observed forindividual source (SESB-Power, ARL-Power and ARL-Incinerator) but with varied ground level concentrationsas shown in Table 3. Based on worse-case scenario,where both power plants and incinerator were inoperation, the predicted maximum 1 h averaging groundlevel concentration of NO

2 was 613.851μg/m3

,which was

observed at approximate distance of 1.0 km to thesouthwest (coordinate x-430312.72, y-5299302.00).Maximum ground level concentration of SO

2 was also

observed at approximately similar distance and directionof that NO

2, which was 554.48 μg/m3 (coordinate x –

430312.72, y – 5299302.00). The ground levelconcentrations of the pollutants at receptors inresidential/villages and institutional areas are shown inTable 3. During the study period, the SESB-Power was

not in operation (note: the power plant was not fully inoperation since a year ago due to technical and economicreasons). Based on the current scenario (without SESB-Power), the ground level concentrations were expectedprimarily from ARL-Power with some attribution fromARL-Incinerator. Potential health impact on receptorswas assessed by comparing the predicted ground levelconcentrations with/to the Malaysian Ambient AirQuality Guideline (DOE Malaysia, 2000). Based on theworse scenario receptors such as Tmn Sapangar (R9),KK Politeknik (R6) and Kg. Salut (R4) were exposed tohigh concentration of NO

2, which exceeded the limit of

320 μg/m3 (1h, TWA).SO

2 was also observed to exceed the limit of 350 μg/

m3 (1h, TWA) at Kg. Malawa/Tmn Sapangar. However,considering that the SESB-Power was not in operationduring the study time, the ground level concentrationsof NO

2 and SO

2 at all receptors observed were well

below the allowable limits. Based on the currentscenario, ground level pollutant concentrations wouldnot impose any significant impact to the nearby area.

Table 3: Ground level concentration of pollutants at selected receptors from individual and combined sources

NO2 (μg/m3) SO2 (μg/m3) SESB-

Power ARL-Power

ARL-Inc.

Combined SESB-Power

ARL-Power

ARL-Inc.

Combined

Guideline (μg/m3) 320.00 350.00 R2 – Kg. Norowot 0.01 0.03 <0.01 0.04 0.01 0.03 <0.01 0.04 R3 – SIRIM 9.12 0.83 <0.01 9.96 8.24 0.75 <0.02 9.17 R4 – Kg. Salut 372.93 60.44 0.53 373.05 336.83 54.66 1.05 337.08 R6 – Politeknik KK 310.51 82.65 0.95 366.67 280.46 74.76 1.90 330.79 R7 – KKIP Resettlement Scheme

135.49 0.19 <0.01 140.03 122.38 0.17 <0.01 126.49

R9 – Kg. Malawa / Tmn Sapangar 530.38 72.03 1.08 589.86 479.06 65.15 2.17 479.26

R10 – ILP 103.88 63.48 0.65 219.93 93.82 57.41 1.31 133.32 R11 – MARDI 86.90 50.10 0.36 189.28 78.49 45.32 0.73 116.15 R12 – LPB Housing 156.08 64.19 0.56 260.89 140.97 58.06 1.12 175.02

Fig. 2: Ground level concentrations of NO2 (μg/m3) from the combined sources.

W.L., Freddy Kho, et al.

Fig. 3: Ground level concentrations of SO2 (μg/m3) from the combined sources

Model performance evaluationTo evaluate the performance of the model, the

concentrations of the observed and predictedpollutants were statistically measured and the resultsof the model performance analysis are presented inTable 4. Based on the Fractional Bias (FB) values, theperformance of the model for NO

2 and SO

2at each

monitoring sites/receptors location were within theacceptable ranges (-0.5d”FBd”0.5) except for NO

2 at

R10. Performance measures in term of Normalised MeanSquare Error (NMSE) were also found to be satisfyingthe criteria (d”0.5). The results showed reasonablygood agreement between the predicted and observedvalues. Meanwhile, the performance measurementswere based on Factor of 2 (FAC2) with more than 90%of the hourly average values within a factor of two(with 0.8). Three locations namely, R10 (for NO

2 and

SO2) and R5 (for NO

2) were observed to have less than

the acceptable limit of 0.8, but still within a factor oftwo. Traffic emissions particularly NO

2 could partly

attribute this result.

Emission source strengthsThis study was also aimed to identify the emission

source strength of each source of emission within theMPPFC. Relative emission-source strength has beendetermined at locations where recorded maximumground level concentrations of pollutants bycalculating the percentage of attribution. Table 5 showsthe attribution of each source to the ground levelconcentrations. Based on worse-case scenario, it hadbeen observed that SESB-Power has attributed morethan 85% of ground level concentration of NO

2 and

SO2. Meanwhile the emission attributions from ARL-

Power and ARL-Incinerator for all pollutants weregenerally less than 15%. Apart from the characteristicsof the emission load of pollutants from the SESB-Power,which are relatively higher, other possible explanationcould be due to the building downwash effect. Theheights of the three stacks of SESB-Power are relativelyshorter (9 m) than the height of the power plant building(20 m). Buildings and other structures near a relativelyshort stack can have a substantial effect on plumetransport and dispersion, and on the resulting groundlevel concentrations that are observed. The stackheight factor has induced the occurrence of buildingdownwash phenomena within the facility complex,which had drawing the plume to the ground near thesource (Venkatram, et al., 2004). When the airflow meetspower plant building, it is forced up and over thebuilding. On the lee side of the building, the flowseparates leaving a closed circulation containing lowerwind speeds. Farther downwind, the turbulent wakezone was created where the airflow forced downwardagain and in addition with the creation of shear, whichinduced more turbulence. The plume emissions fromthe SESB-Power most likely caught in the cavity and ifthe plume escaped the cavity, but remains in theturbulent wake, it may be carried downward anddispersed more rapidly by the turbulence. Thiscondition could result in either higher or lowerconcentrations than would occur without the building.It is depending on whether the reduced stacks heightor increased turbulent diffusion has the greater effect.In minimising the building downwash effect installationof at least twice the height of adjacent building as theoperating rule should be sufficient to minimise the effectof building downwash (Wark, et al., 1998), or

220

W.L., Freddy Kho, et al.Int. J. Environ. Sci. Tech., 4 (2): 215-222, 2007

application of good engineering practice of stackheight (MEO, 2003). The higher stacks enabledpollution to be dispersed away from the source areascompetently compare to the shorter stacks (Downerand Radojevic, 1992).

DISCUSSION AND CONCLUSIONThis study has focused on ground level

observations and modelling analyses of NO2 and SO

2

dispersion from various elevated point sources withinthe MPPFC in elucidating pollutants dispersionpatterns, emission source strengths and emissionpotential risks on selected receptors within the domaincoverage. Modelling analysis using ISC AERMODView showed that the two pollutants were disperseddominantly to the west in corresponding with theeasterly dominant wind direction. Small fractions werealso dispersed to the east and north directions. Analysison the emission source strengths showed that emissionfrom SESB-Power has attributed to the emission of morethan 85% of NO2

and SO2. This suggests that SESB-

Power had significantly attributed to the ground levelconcentration of pollutants in the study area. One of

the main factors that could lead to this result was theheight of the stacks, which were found to be shorterand not conformed to the stack height requirement.This condition has allowed the building downwashphenomena to occur which had drawing the plume tothe ground near the source. A total number of discretereceptors, mainly residential area within the domaincoverage were assessed in terms of exposure to airpollutants by comparing with the Malaysian AmbientAir Quality Standard Guidelines. The computersimulated results based on worse case scenario showedthat receptors such as Kg. Malawa/Tmn Sapangar (R9),KK Politeknik (R6) and Kg. Salut (R4) were exposed tohigh concentration of NO2

, which exceeded the limit of320 μg/m3 (1h, TWA). SO

2 was also observed to exceed

the limit of 350 μg/m3 (1h, TWA) at Kg. Malawa/TmnSapangar. However, by considering that the SESB-Power was not in operation during the study period,the ground level concentrations of NO

2 and SO

2 at all

receptors were observed to be well below the allowablelimits. Based on the current scenario, the pollutantsground level concentrations would not impose anysignificant impact to the surrounding area.

Table 4: Model performance evaluation results

Monitoring Sites/Location of Receptors

Fractional Bias (FB) Normalized Mean Square Error NMSE) Factor of 2 (FAC2)

NO2 SO2 NO2 SO2 NO2 SO2

R1 0.14 0.04 0.019 0.002 0.87 0.96 R2 0.09 0.06 0.008 0.004 0.92 0.94 R3 0.19 0.07 0.037 0.005 0.83 0.93 R4 0.09 0.03 0.008 0.001 0.91 0.97 R5 0.38 0.18 0.147 0.033 0.68 0.83 R6 0.22 0.25 0.051 0.063 0.8 0.78 R7 0.12 0.19 0.015 0.036 0.88 0.83 R8 0.01 0.06 0 0.004 0.99 0.94 R9 0.19 0.17 0.035 0.028 0.83 0.85

R10 0.52 0.47 0.256 0.236 0.65 0.62

Table 5: Emission source strengths of the individual source for pollutant ground level concentrations(based on worse case scenario*)

Emission Sources SESB-Power (Ie-SESB)

ARL-Power (Ie-ARL-Power)

ARL-Incinerator (Ie-ARL-inc)

Total

Maximum Conc. (μgm-3) 613.82

101.92 1.30 717.04 NO2

Attribution (%) 85.60 14.21 0.19 100.00

Maximum Conc. (μgm-3) 554.42

92.19 2.60 649.21 SO2

Attribution (%) 85.40 14.20 0.40 100.00

221

*Worst case scenario- all power plants and the incinerator are in operation

W.L., Freddy Kho, et al.

AUTHOR (S) BIOSKETCHESFreddy Kho, W.L., B.Sc., is a postgraduate student in environmental engineering, Department of CivilEngineering, University of Malaysia Sarawak (UNIMAS), Kota Samarahan, Sarawak, Malaysia.Email: frexwl@gmail.com

Sentian, J., B.Sc., M.Sc., is a lecturer in the Environmental Science Program, School of Science and Technology,University Malaysia Sabah (UMS). Email: jsentian@ums.edu.my

Radojevic, M., B.Sc., M.Sc., Ph.D., is an associate professor in the Environmental Science Program, School ofScience and Technology, University Malaysia Sabah. Email: miroslav@ums.edu.my

Tan, C.L., B.Sc., is a postgraduate student in the Environmental Science Program, School of Science andTechnology, University Malaysia Sabah (UMS). Email: chewleng99@hotmail.com

Law, P.L., B.SCE., M.SCE., M.Sc., Sc.D., is an associate professor in environmental engineering, Departmentof Civil Engineering, University of Malaysia Sarawak (UNIMAS), Kota Samarahan, Sarawak, Malaysia. Email: puonglaw@feng.unimas.my

Computer simulated versus observed No2 and So2 emitted from elevated...

REFERENCESBennet, M., Hunter, G.C., (1997). Some comparisons of Lidar

estimates of peak ground-level concentrations with thepredictions of UK-ADMS, Atmos. Environ., 31 (2), 429-439.

Carruthers, D.J., Edmunds, H.A., Bennet, M., Wood, P.T.,Milton, M.J.T., Robinson, R., Underwood, B.Y., Franklin,C.J., Timmis R., (1997). Validation of the ADMS dispersionmodel and assessment of its performance relative to R-91and ISC using archived LIDAR data. Int. J. Environ. Pollut.,8 (3-6), 264-278.

Carslaw, D.C., Beevers S.D., (2002). Dispersion modellingconcentrations for transient emissions from elevated pointsources, Atmos. Environ., 36 , 3021-3029.

Department of Environment Malaysia. (2000), A Guide ToAir Pollution Index In Malaysia, Ministry of Science,Technology and Environment, Kuala Lumpur.

Downer, R.M., Radojevic, M., (1992). Technologies ofControl, in Atmospheric Acidity, Eds. M. Radojevic, R.M.Harrison, Elsevier Science Publishers Ltd, U.K., 467-494.

Hanna, S.R., (1989). Confidence limits for air quality modelevaluations as estimated by bootstrap and jackkniferesampling methods, Atmos. Environ., 23 , 1385-1398.

James, K.J., Cherry, M., Stack, M.A., (1995), Assessment ofchemical plant emission on an urban environment: A newapproach to air quality measurements, Chemosphere, 31 ,(7), 3741-3751.

Jesse, L., Cristiane L., Michael A., (2000). User’s Guide ISC-AERMOD View, Volume I & Volume II, Lakes EnvironmentalPt. Ltd., Ontario, Canada.

Mehdizadeh, F., Rifai H.S., (2004). Modelling point sourceplumes at high altitudes using a modified Gaussian model,Atmos. Environ., 38 , 821-831.

Ministry of Environment Ontario (MEO), (2003). Proposedguidance for air dispersion modelling [online]. Ministry of

Environment Ontario. Available from http://www.ene.gov.on.ca/envision/techdocs/3614e02.pdf>

Kumar, A., Bellman N.K., Sud A., (1999). Performance of anindustrial source complex model: predicting long-termconcentrations in an urban area. Environ. Prog., 18(2). 93-100.

Mazzeo, N.A., Venegas, L.E., (2000). Practical use of theISCST3 model to select monitoring site locations for airpollution control. Int. J. Environ. Pollut., 14 (2), 1-6.

Rama Krishna, T.V.B.P.S, Reddy, M.K., Reddy, R.C., Singh,R.N., (2004). Assimilative capacity and dispersion ofpollutants due to industrial sources in Visakhapatnam bowlarea. Atmos. Environ., 38 , 6775-6787.

Venegas, L.E., Mazzeo, N.A., (2005). Modelling of urbanbackground pollution in Buenos Aires City (Argentina).Environmental Modelling & Software.

Venkatram, A., Isakov, V., Yuan, J., Pankratz, D., (2004).Modeling dispersion at distances of meters from urbansources. Atmos. Environ., 38 , 4633-4641.

Wark, K., Warner, F.C., Davis, T.W., (1998), Air Pollution.Its Origin and Control, 3rd. Ed. Addision-Wesley,Longman.Inc.

Yassin, M.F., Kato, S., Ooka, R., Takahashi, T., Kouno R.,(2003). Study on effects of fluctuations in the wind directionon pollutant diffusion in urban areas. in Proceedings of the11th. International Conference on Modeling, Monitoring andManagement of Air Pollution, Eds. C.A. Brebbia & F. Patana,WIT Press, U.K., pp 143-151.

Zhou, Y., Levy, J.I., Hammitt, J.K., Evans, J.S., (2003).Estimating population exposure to power plant emissionsusing CALPUFF: A case study in Beijing, China. Atmos.Environ., 37 , 815-826.

This article should be referenced as follows:Freddy Kho, W.L., Sentian, J., Radojevic, M., Tan, C.L. Law, P.L., (2007). Computer simulated versus observedNo2 and So2 emitted from elevated point source complex. Int. J. Environ. Sci. Tech., 4 (2), 215-222.

222