measuring and modeling of particulate dispersion from the cement ...
1
MEASURING AND MODELING OF PARTICULATE DISPERSION FROM THE CEMENT PLANTS Anchaleeporn W. Lothongkum1*, Kanyanee Seangkiatiyuth1 and Vanisa Surapipith2* 1*
2*
Department of Chemical Engineering, Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand ([email protected])
Air Quality and Noise Management Bureau, Pollution Control Department, Ministry of Natural Resources and Environment, Bangkok 10400, Thailand ([email protected])
Abstract This work focuses on the dispersion of particulate matters in the atmosphere around key cement plants, situated about 100 km northeast of Bangkok, Thailand. It is aimed at identifying the impact of emission from four cement manufacturers by both measuring and modeling analyses. The daily average concentrations of total suspended particulate matters (TSP) and particulate matters with diameter less than 10 µm (PM10) within the distance of 15 km from the cement plant were measured for a week time period every six months at 12 receptors by high-volume volume samplers. The PM10 concentration at further distance, i.e., about 25 km away from the cement plant was also monitored routinely at 2 automatic monitoring station of the Pollution Control Department (PCD), Thailand. Two Gaussian based models, AERMOD and CALPUFF, were then applied to estimate the dispersion of the particulate matters (PM). The estimated results by both AERMOD and CALPUFF showed that the model-predicted PM concentration at 12 receptors and 2 automatic monitoring stations were lower than the observed concentration. The CALPUFF model showed higher values and frequently found the PM peaks than the AERMOD. Keywords: cement plant, TSP, PM10, AERMOD, CALPUFF
2
1. Introduction Particulate matters with diameter of 10 µm or smaller (PM10) have been concerned worldwide because they can pass through human’s respiratory system to the lungs causing serious health effects, for example, aggravated respiratory symptoms, deteriorating lung function, asthma, development of chronic bronchitis, irregular heartbeat and nonfatal heart attacks. In case of the particulate matters with diameter of 2.5 µm (PM2.5), it can impair visibility and reduce solar radiation.[1] In Thailand, two categories of particulate matters (PM): total particulate matters (TSP) of which its diameter is 100 µm or smaller, and PM10 have been regulated. Cement industry is well-known as a significant emission source of the PM. Generally, PM released from the cement manufacturer ranges from 0.12-100 µm.[2] PM of which the range impacts to human’s welfare and the environmental assessment has been investigated by measurement and modeling method. The Gaussian models were reported the PM pattern from stacks of the Kerman cement in Iran, and the cement complex in India.[3-4] The reports showed that the results from the models agreed with those of the measurement. The released PM from the stacks of the cement plant in Iran was a very significant effect on the respiratory problems. During the year of 2006, the Pollution Control Department (PCD), Ministry of Natural Resources and Environment, Thailand carried a pilot project to establish an emission inventory in order to investigate the impact of fine particulates around the industrial areas located in Nah Phra Laan; a part of Chaloem Phra Kiat district, Saraburi province. One of a key results indicated that the major source of PM10 was from crushing activity.[5] Similar to the PCD’s results, in 2006, R. Sivacoumar et al. reported that the dust generated from stone crushing activities contained a significant amount of fine inhalable matters. Not only the sizes but also sources of fine particulate matters are of important to be concerned.[6] Based on the results from the PCD, the Hybrid Particle and Concentration Transport model (HYPACT) with Lagrangian option and the Regional Atmospheric Modeling System (RAMS) was consequently used in evaluation of PM10 dispersion by Sittichai et al.[7] They found that the modeling results were in good agreement with the observations by the PCD. It was revealed that high PM10 concentration appeared in the source area and near source area in downwind direction. Kaengkoi district, situated about 21 km southeast of Chaloem Phra Kiat district, is of interest area as it is
the location of cement complex and the activities in the area are rather similar to those of Nah Phra Laan. This paper focuses on the PM emission from 14 stacks of the four cement plants located near the mountain areas as shown in Figure 1. From the Environmental Impact Assessment (EIA) reports of the four model plants, the total PM emission is 4,286 ton/year, based on the total emission rate of 136 g/s. Considering the total emission of the cement industry in Nah Phra Laan area, the emission from in Kaengkoi is higher (total stack PM10 emission from Nah Pra Laan is about 600 ton/year). Therefore, in some areas, though stacks are not the major PM source, the continuous release may contribute to a significant impact. However, by setting up the electrostatic precipitator with high efficiency of 99.99% at the stack, it can be assumed that the PM emission from the stacks are PM10, and considered the released PM as PM10. As it is impossible to evaluate the overall PM emission by measurement only, two dispersion models which are recommended by the United States Environmental Protection Agency (U.S.EPA): the American Meteorology Society-Environmental Protection Agency Regulatory Model (AERMOD) and the California Puff Dispersion Model (CALPUFF) are applied by using the surface and upper meteorological observation data input from the stations.
3
⊗
Figure 1. Locations of the study areas.
2. Methodology 2.1 Study areas Four target cement plants are located in Kaeng Khoi district, Saraburi province in the middle of Thailand as shown in Figure 1. The eastern part of the province is covered by high plains and plateaus, while the western part is mostly low flat plains. The study domain coordinates cover UTM X from 701000E to 751000E, and UTM Y from 1593000N to 1643000N. The domain area is located in time zone 47P with WGS84 datum type. The weather in this area is sultry hot in summer from February to April, heavy rain from May to September and cool in winter from October to January.
2.2 Emission and ambient air measurement The PM emission from the stacks of the four cement plants were reported as TSP in the each EIA report. The emission was collected in accordance with the Code of Federal Regulations, 40 CFR Part 60, Appendix A.[8] The volumetric flow rate of the effluent gas from the stack was measured by a pitot tube at
4 several points in the duct. The emission rates of each cement plant which are used in the AERMOD and CALPUFF programs are shown in Table 1. Table 1. Emission rates of four cement plants. Cement plant number 1 2 3 4
Emission rate (g/s) 32.9 36.8 63.8 2.4
The PM in ambient air was collected at 12 receptors, which were selected according to the guidelines in the EIA report of the cement plant 1. TSP and PM10 were continuously collected for a week (168 hours) by a high-volume sampler, which was 2-m above ground level with a flow rate of gas stream of 10 L/min. After running continuously for 24 hours, the glass fiber filters of the sampler were changed. At the end of the sampling period, the filters were removed, dried and reweighed. The concentration of the TSP, reporting in micrograms (or milligram) per cubic meter, was calculated from the weight gained by the filters divided by the volume of air sampled. According to the guidelines in the EIA report, the PM must be measured twice a year as summarized in Table 2. The data which used in AERMOD and CALPUFF in this work were taken from the 6-month EIA report of each cement plant, which was approved by the National Committee by the Office of Environment at Policy and Planning. Continuous monitoring of PM10 concentration in ambient air was detected by 2 automatic monitoring stations of the PCD at Kao Noi and Nah Phra Laan. The PM10 at Kao Noi station was measured by equivalent method, and at Nah Phra Laan station by the Tapered Element Oscillating Microbalance (TEOM) method. The observation periods in this study are also shown in Table 2.
2.3 Meteorological data The meteorological data were received from the following 4 meteorological stations of the PCD and the Thai Meteorological Department (TMD). Kao Noi and Nah Phra Laan are both meteorological and monitoring stations of PCD where no observation of cloud cover and ceiling height are available. Since the main purpose of both Kao Noi and Nah Phra Laan stations is for air pollution monitoring applications, therefore, the stations are located near the residential areas. Pakchong, a station of TMD located nearby Kao Yai Natural Park, is weather monitoring for agriculture. It is surrounded by high mountains thus the meteorological data are influenced by mountain-valley wind. Bangna, another station of TMD, provides both surface and upper air data. Although Bangna station situates over 100 km away in southwest of the study areas, it is selected for upper air data because it is the nearest radiosonde balloon launching site. The surface air data in this work is received from other 3 stations mentioned above. Details of the meteorological stations and the relevant data are presented in Table 3. The locations of surface meteorological stations can be seen in Figure 1.
5 Table 2. Periods of ambient air monitoring. Receptors 1 2 3 4 5 6 7 8 9 10 11 12 Kao Noi Nah Phra Laan
Period 1 Start End 2/03/07: h13 9/03/07: h13 2/03/07: h10 9/03/07: h10 2/03/07: h12 9/03/07: h12 2/03/07: h16 9/03/07: h16 9/03/07: h14 16/03/07: h14 9/03/07: h8 16/03/07: h8 10/03/07: h12 17/03/07: h12 10/03/07: h13 17/03/07: h13 17/03/07: h11 24/03/07: h11 17/03/07: h18 24/03/07: h18 17/03/07: h16 24/03/07: h16 17/03/07: h12 24/03/07: h12 2/03/07: h1 24/03/07: h24 2/03/07: h1 24/03/07: h24
Period 2 Start End 12/10/07: h15 19/10/07: h15 12/10/07: h17 19/10/07: h17 12/10/07: h14 19/10/07: h14 20/10/07: h15 27/10/07: h15 20/10/07: h9 27/10/07: h9 20/10/07: h18 27/10/07: h18 20/10/07: h12 27/10/07: h12 28/10/07: h16 4/11/07: h16 28/10/07: h9 4/11/07: h9 12/10/07: h17 19/10/07: h17 28/10/07: h12 4/11/07: h12 28/10/07: h9 4/11/07: h9 12/10/07: h1 4/11/07: h24 12/10/07: h1 4/11/07: h24
Table 3. Details of the meteorological stations. Meteorological station
Collected data by
Meteorological parameter
Distance from the cement plant 1 (km)
Direction to the cement plant 1
Kao Noi
PCD
20.73
SW
Nah Phra Laan
PCD
24.13
WNW
Pakchong
TMD
25.14
E
Bangna
TMD
surface: G, P1, RH, T 10 m: WD, WS1 (hourly basis data) surface: G, P1, RH, T 10 m: WD, WS1 (hourly basis data) surface: P2, RH, T 10 m: WD, WS2 upper air data: CH, CV (3-hour basis data) upper air data: DWPT, H, P, RH, T, WD, WS2
118.93
SSW
Notes: CH CV DWPT G H P1 P2 RH T WD WS1 WS2
= = = = = = = = = = = =
ceiling height (m) cloud cover (tenths) dew point temperature (˚C) global radiation (Wh/m2) height above sea level (m) pressure (mmHg) pressure (hPa) relative humidity (%) dry bulb temperature (˚C) wind direction (degree) wind speed (m/s) wind speed (knot)
2.4 Modeling AERMOD is a steady-state plume model that incorporates air dispersion based on planetary boundary layer turbulence structure and scaling concepts, including treatment of both surface and elevated sources, and both simple and complex terrain. AERMOD is recommended to evaluate the pollutants dispersion in surrounded distance within 50 km from the emission source. It is composed of three parts: AERMOD Meteorological Preprocessor (AERMET), AERMOD Terrain Preprocessor (AERMAP) and AERMOD. The AERMET processes the hourly surface data. The AERMAP is used to process the terrain data in conjunction with a layout of receptor and sources for AERMOD control files. The AERMOD is a
6
Figure 2. Data flow in AERMOD.
Figure 3. Data flow in CALPUFF.
dispersion model. In this work, the commercial interface, ISC-AERMOD View (Version 4.2.6) (Lakes Environmental Software; Waterloo, Ontario, Canada) was run. The surface meteorological data obtained from Kao Noi station while the cloud cover and ceiling height were from Pakchong station. The upper meteorological data obtained from Bangna station. The data flow of the program is shown in Figure 2. The ISC-AERMOD View was run with the modeling domain covered the smaller area than that of CALPUFF. The maximum grid number which ISC-AERMOD View can operate is 7,200 grids. The modeling domain is expected to be large enough to see plume characteristics, therefore, it is designed to covers 39.5 x 39.5 km2 on UTM 47P coordinates between 701000E to 740500E and 1600000N to 1639500N. The grid spacing between the receptors is 500 x 500 m2. CALPUFF is non-steady-state puff dispersion model that simulates the effects of time- and space-varying meteorological conditions on pollution transport, transformation and removal. CALPUFF can be applied for pollution transport from the distance between 100 m – 100s km and for complex terrain. [9] CALPUFF modeling system includes three main components: CALMET, CALPUFF and CALPOST. CALMET is a meteorological model that develops hourly wind and temperature fields on a 3-D gridded modeling domain and the generated meteorological data can be displayed by PRTMET, a postprocessing program. CALPOST is used to process the files from CALPUFF, producing tabulations that summarize the results of the simulation, identifying the highest and second highest 1-hour, 3-hour, 24-hour or period averaged concentration at each receptor. The data flow in CALPUFF is shown in Figure 3. In this work, the CALPUFF PROfessional Beta 5.2.0 was run with free version which could be downloaded from the website of Earth Tech, Inc (www.src.com/calpuff1.htm). The surface meteorological data obtained from three stations: Kao Noi, Nah Phra Laan and Pakchong. The upper meteorological data obtained from Bangna station. The modeling domain covers 50 x 50 km2 on UTM 47P coordinates between 701000E to 751000E and 1593000N to 1643000N. Both discrete and grid modes (grid spacing 500 x 500 m2) were selected to find the PM concentration at ground level of 12 receptors and the impact areas.
3. Results and Discussion 3.1 Meteorological analysis The daily climate of Kaengkoi, the of interest area, is influenced by the seasonal winds and mountain-valley breeze. For the seasonal winds, the southwesterly monsoon brings moisture from the Indian Ocean to Asian Continent during March to September, whereas the northeasterly monsoon blows from the cool Asian land mass towards the Indian Ocean during October to February. The valley breeze is a warm wind occurring in the daytime or late afternoon. The clouds which are formed over the hills in the afternoon are the result of the condensation taking place when the air rises to cooler heights over the mountains. At night, the valley walls cool down resulting to a cool air surface. Therefore, winds flow downwards the slope of the mountains known as night breeze.[10] At receptor areas, during 2 to 24 March 2007 the average wind speed was 0.4 m/s or lower at the night to the morning (7 p.m.-10 a.m.); then, the wind speed increased but not over 4.5 m/s during the day (10 a.m. – 7 p.m.). During 12 October to 4 November 2007, the wind was gentle from the evening to the morning (5 p.m.-10 a.m.) after that the wind speed increased but not higher than 4.5 m/s. Wind statistics at Kao Noi, Nah Phra Laan and Pakchong stations are summarized in terms of wind roses, as shown in Figures 4(a)-(c), respectively. The wind rose shows wind strength, direction and frequency.
7 The percentage of calm wind is represented by the size of the center circle; the bigger the circle, the frequent higher calm wind was found. Each branch of the wind rose represents wind-coming direction, identifying the north at the top of the diagram. Each segment length within a branch is proportional to the frequency of winds blowing within the corresponding range speeds from that direction. Figure 4(a) shows that during 2-24 March 2007 the wind at Kao Noi station is dominated by westerly wind. The strongest wind is about 3.5 m/s. The major wind direction of Nah Phra Laan station is southerly as shown in Figure 4(b). The wind speed is light of 1-2 m/s. Figure 4(c) shows that the wind at Pakchong station is dominated by east-southeasterly wind. The wind is rather strong and the strongest wind speed is 9.3 m/s.
Calm wind 3.84%
Calm wind 5.98%
(a) Wind speed (m/s)
Calm wind 52.72%
(b) ≥ 6.0
4.0-6.0
(c) 3.0-4.0
2.0-3.0
1.0-2.0
0.5-1.0
Figure 4. Wind rose during 2-24 March 2007 (a) Kao Noi, (b) Nah Phra Laan and (c) Pakchong.
Calm wind 5.93%
Calm wind 26.04%
Calm wind 2.86%
Wind speed (m/s)
(c)
(b)
(a) ≥ 6.0
4.0-6.0
3.0-4.0
2.0-3.0
1.0-2.0
0.5-1.0
Figure 5. Wind rose during 12 October-4 November 2007 (a) Kao Noi, (b) Nah Phra Laan and (c) Pakchong.
8 The wind roses of the three stations during 12 October-4 November 2007 are shown in Figures 5(a)-(c). During this period, the major wind at Kao Noi station are still westerly with higher speed. The strongest wind speed is 5.7 m/s. The wind at Nah Phra Laan station in this season has turned to northeasterly. The wind speed is varied from 0.5 to 6 m/s. The dominated wind at Pakchong station is still east-southeasterly. The wind speed is varied from calm wind to 12.3 m/s. In Figures 6(a)-(b), the wind speeds of the three stations were plotted as the wind class distribution. It is clearly seen that wind speed at Pakchong station is either calm or strong (over 3 m/s) during both periods. While the winds at Kao Noi and Nah Prah Laan stations are within 1 – 3 m/s most of the time.
Percent (%)
Percent (%)
The non-uniform wind direction and speed at the three stations as shown through the analysis clarifies that the winds in the study areas are strongly influenced by mountain-valley breezes.
Wind speed (m/s)
Wind speed (m/s)
(a) (b) Figure 6. Wind class distribution at Kao Noi, Nah Phra Laan and Pakchong stations: (a) during 2-24 March 2007 and (b) during 12 October – 4 November 2007.
3.2 Wind field pattern evaluated by CALMET CALMET is a meteorological model which includes a diagnostic wind field generator containing objective analysis and parameterized treatments of slope flows, kinematic terrain effects, terrain blocking effects, a divergence minimization procedure, and a micro-meteorological model for overland and overwater boundary layers.[5] Wind speeds and directions in this work which were simulated by CALMET of the selected days were plotted to evaluate the wind field pattern during two episodes. Figure 7 shows wind field on 9 March 2007 representing 2 - 24 March 2007. It is the output of the CALMET hourly wind speeds and directions which were generated by using input from the surface meteorological data recorded above the ground level of 10 meters from three stations; Kao Noi, Nah Phra Laan and Pakchong, and from the upper meteorological data from Bangna station. Figure 8 shows CALMET wind field on 1 November 2007 representing 12 October - 4 November 2007. The wind fields show that most of the time the wind speed was southwesterly. Seasonally, the southwesterly monsoon should be found during March to September, and the northeasterly monsoon during October to February. In fact, the season wind affects to the wind profile from CALMET directly, especially the wind direction should be in accordance with the season wind direction. However, in this study the wind profile direction has not changed when the season wind direction changed. This might be because of (i) effect of mountain-valley breeze and (ii) the upper meteorological data are not relevant to the wind field calculation. This part will be further study.
9
7h UTM Y (km)
16 h UTM Y (km)
UTM Y (km)
UTM Y (km)
22 h
19 h
UTM X (km)
UTM X (km)
UTM X (km)
UTM X (km)
UTM Y (km)
UTM Y (km)
UTM Y (km)
UTM X (km)
UTM X (km)
13 h
10 h UTM Y (km)
4h
1h
UTM X (km)
UTM X (km)
Figure 7. Wind field on 9 March 2007 generated by CALMET.
7h
UTM X (km)
16 h
UTM Y (km)
22 h
UTM Y (km)
19 h
UTM Y (km)
UTM Y (km)
UTM X (km)
UTM X (km)
UTM X (km)
UTM X (km)
13 h
10 h
UTM Y (km)
UTM Y (km)
UTM Y (km)
4h
UTM X (km)
UTM Y (km)
1h
UTM X (km)
Figure 8. Wind field on 1 November 2007 generated by CALMET.
UTM X (km)
10
3.3 PM measurement From the EIA reports, PM emissions of 4 cement plants were within the permissible emission of 50 mg/m3. The PM emissions from 4 cement plants are shown in Figure 9. It can be seen that cement plant 3 released the largest volume of PM whereas the cement plant 4 released the lowest. PM10 concentration was measured routinely by the two automatic monitoring stations of the PCD away from the cement plant 1 about 25 km. Figures 10(a)-(b) and 11(a)-(b) show the PM10 concentrations monitoring at Kao Noi and Nah Phra Laan stations. The results of Nah Figure 9. PM emissions from the stacks Phra Laan station, which is located around by stone and of 4 cement plants. mining industries, shows that the 24-h avg peak of PM10 concentration in 5 days during 2 to 24 March 2007, and 11 days during 12 October to 4 November 2007 are over the Thai permissible value of 120 µg/m3 In contrast, none at Kao Noi station are over the permissible value. Based on the 1-h avg PM10 concentration, Nah Phra Laan station was frequently found peaks of PM10 concentration more than 120 µg/m3 while Kao Noi was not. Particularly, during 2-24 March 2007, both 1-h avg and 24-h avg peaks of PM10 concentration at Nah Phra Laan station is rather high resulting in significant impact to human’s health. Taking into account the twelve receptors near 4 cement plants, it was found that the 24-avg PM10 and TSP concentrations of all receptors were not over Thailand’s permissible value of 120 µg/m3 for PM10 and of 330 µg/m3 for TSP as presented in Figures 12 and 13. In common, other influences on twelve receptors, apart from the stack emission, are possible PM sources in the areas, e.g., motor vehicles, open mining, power utilities of other industries, combustion products from open burning, and residential cooking. Table 4 presents the characterization of the receptors. It was found that the difference between the measured TSP and PM10 concentrations in two periods (2 to 24 March 2007 and 12 October to 4 November 2007) was not significant, especially at receptors 3 and 8 during 12 October to 4 November 2007. Based on the information by the PCD’s two automatic monitoring stations, some over-limit 24-h avg PM10 are found. This is likely due to the emission activities around the stations, where campaign of emission reduction is going on. The Nah Phra Laan district has been designated as Pollution Control Zone since 2005, and air quality has been improving by restricted rules for all trucks and mining activities in that area to clean up the dust throughout the year. Table 4. Characterization of the receptors. Receptor 1 2 3 4 5 6 7 8 9 10 11 12 Kao Noi Nah Phra Laan
Characterization of receptors on the area of the temple and near the road on the area of the temple and near the road on the area of the temple, near the residential areas, agricultural area and the road on the area of the school, near the residential areas and the road on the area of the school, near the residential areas and the road on the area of the school, near the residential areas, agriculture and the road on the are of electrical substation, near the residential areas and the road on the area of animal conservation station, near agricultural area, not far from the residential areas and the road on the area of the school, near the residential areas, agricultural area and the road on the area of the school, near the residential areas, agricultural area and the road on the road side, near the residential areas and agricultural area on the area of the school, near the residential areas, agricultural area and the road near the residential areas near the road and open mining
PM10 concentration (µg/m3)
11
Kao Noi
no data
Date PM10 concentration (µg/m3)
(a) no data
Nah Phra Laan
Date
(b)
PM10 concentration (µg/m3)
Figure 10. PM10 concentration monitoring during 2-24 March 2007: ( ) 1-h avg and (•) 24-h avg. no data
Kao Noi
Date
PM10 concentration (µg/m3)
(a) no data
Nah Phra Laan
Date
(b) Figure 11. PM10 concentration monitoring during 12 October – 4 Novemberh 2007: ( ) 1-h avg and (•) 24-h avg.
24-h avg PM concentration (µg/m3)
Receptor 7
Date
Receptor 10
Date Receptor 5
Date
Date
24-h avg.PM concentration (µg/m3)
Date 24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Date
Receptor 8
Receptor 11
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 2
24-h avg PM concentration (µg/m3)
Date 24-h avg PM concentration (µg/m3)
Receptor 4
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 1
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
12
Date
Receptor 3
Date
Date
Receptor 9
Date
Receptor 12
Date
Figure 12. TSP and PM10 concentrations at 12 receptors during 2-24 March 2007: (•) TSP and (+) PM10.
24-h avg PM concentration (µg/m3)
Receptor 7
Date
Receptor 10
Date Receptor 5
24-h avg PM concentration (µg/m3)
Date
Receptor 8
Receptor 11
Date 24-h avg PM concentration (µg/m3)
Date
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 2
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 4
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 1
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
13
Date
Receptor 3
Date Date
Receptor 6
Date Date
Receptor 9
Date
Receptor 12
Date
Figure 13. TSP and PM10 concentrations at 12 receptors during 12 October-4 November 2007: (•) TSP and (+) PM10.
14 3.4 Modeling results by AERMOD and CALPUFF In this work AERMOD and CALPUFF models were used to evaluate the impact results of the PM emissions from the stacks of 4 cement plants. The stack emission samplings in the EIA reports were noted in terms of TSP concentration, considering as PM10. The model-predicted PM concentrations at Kao Noi and Nah Phra Laan stations during the two periods are shown in Figures 14 (a)-(b) and 15(a)-(b). It can be concluded from the plots that AERMOD and CALPUFF predict the PM concentration in a similar range. AERMOD generated more peaks at Kao Noi station during March than October to November. On a contrary, CALPUFF generated more peaks during October to November than March. Model predictions at Nah Prah Laan during March by the two models are similar, whereas, AERMOD rarely generated peaks during October to November.
PM10 concentration (µg/m3)
The model-predicted PM concentrations at the twelve receptors are shown in Figures 16 and 17. Clearly, the similar trend by the PDC monitoring stations is seen. AERMOD generated many peaks during March, while CALPUFF generated mostly during October to November. Of the 12 receptors, only the receptor 3 CALPUFF generated more peaks in March.
Kao Noi
Date
PM10 concentration (µg/m3)
(a)
Nah Phra Laan
Date
(b) Figure 14. Model-predicted PM10 concentration at 2 automatic PCD stations during 2-24 March 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
PM10 concentration (µg/m3)
15
Kao Noi
Date
PM10 concentration (µg/m3)
(a) Nah Phra Laan
Date
(b) Figure 15. Model-predicted PM10 concentration at 2 automatic PCD stations during 12 October4 November 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
Receptor 1
3 3 3
Receptor Receptor88
3
PM concentration (µg/m )
3
3
Time (h)
Receptor 9
Time (h)
Time (h) PM concentration (µg/m )
Time (h)
Receptor 6
Time (h) PM concentration (µg/m )
3
PM concentration (µg/m )
3
PM concentration (µg/m )
Receptor 5
Time (h) Receptor 77 Receptor
Receptor 3
Time (h) PM concentration (µg/m )
3
PM concentration (µg/m )
3
PM concentration (µg/m )
Receptor 4
Time (h)
PM concentration (µg/m )
Receptor 2
Time (h)
Time (h)
Receptor 10
PM concentration (µg/m )
3
3
PM concentration (µg/m )
PM concentration (µg/m )
16
Receptor 11
Time (h)
Receptor 12
Time (h)
Figure 16. Model-predicted PM concentration at 12 receptors during 2-24 March 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
Receptor 10
Time (h)
3 3
3
PM concentration (µg/m )
Receptor 9
Time (h) 3
3
3
Receptor 8
Time (h) PM concentration (µg/m )
Time (h)
Receptor 6
Time (h)
Time (h) 3
Receptor 7
PM concentration (µg/m )
3
Receptor 5
PM concentration (µg/m )
Time (h)
Receptor 3
Time (h)
Time (h) PM concentration (µg/m )
3
PM concentration (µg/m )
Receptor 4
3
PM concentration (µg/m )
Receptor 2
PM concentration (µg/m )
Receptor 1
Time (h)
PM concentration (µg/m )
PM concentration (µg/m )
3
PM concentration (µg/m )
3
PM concentration (µg/m )
17
Receptor 11
Time (h)
Receptor 12
Time (h)
Figure 17. Model-predicted PM concentration at 12 receptors during 12 October – 4 November 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
Hour
Modeled PM concentration (µg/m3)
Measured PM concentration (µg/m3)
Modeled PM concentration (µg/m3)
18
1.16 0.67
0.60 0.19
Hour
Figure 19. Simulated PM10 concentration at Figure 18. 1-h avg measured and simulated receptor 4 during 5-7 March 2007: PM10 concentration at Nah PhraLaan on 20 March 2007: ( ) AERMOD, ( ) 1-h avg by AERMOD, (•) 24-h avg ( ) CALPUFF and ( ) measured. by AERMOD, ( ) 1-h avg by CALPUFF and (×) 24-h avg by CALPUFF.
To clarify the performances of AERMOD and CALPUFF, time series of the measured and model-predicted PM10 concentrations at Nah Pha Laan station on 20 March 2007 are plotted in Figure 18. Even though the PM concentrations from the models are very low comparing to the measured results but the peak pattern during the day can be seen. CALPUFF can calculate the peaks which appear about 7, 16 and 21 hrs at the local station time (LST), while AERMOD can calculate only one peak about 14 hrs at LST. This reveals that AERMOD gives the statistical calculation whereas CALPUFF can detect real time values. Figure 19 shows the results at receptor 4 indicating that CALPUFF generated several PM concentration peaks while AERMOD could detect only one peak at 14 hrs. The magnitude of the peak generated by AERMOD was lower than that of CALPUFF.
3.5 The maximum PM concentration Table 5 shows the average model-predicted maximum PM concentrations by AERMOD and CALPUFF at different periods; 1 h, 24 h and 1 year. As seen, 1-h avg values by CALPUFF were much higher than those by AERMOD. The difference between the generated PM concentrations decreased with long range period such as 24 h and 1 year. It should be noted that the 24-h avg PM concentrations predicted by both AERMOD and CALPUFF are in Thailand’s ambient air quality standard of PM10 (120 µg/m3), and slightly low comparing to the TSP standard of 330 µg/m3. The possible maximum PM concentrations are on different spots at the receptors around the cement plants 1, 2 and 3 as far as 2-7 km, as shown in Figure 20.
D2,D5,D6 G4,G6 G1 G3,G5
D3,D4 G2 D1
Figure 20. Possible maximum PM concentration locations by AERMOD and CALPUFF: ( ) stack and (+) receptor.
19
D1 G2 CALPUFF AERMOD
D2 G3 D3
24-h avg
G4 CALPUFF AERMOD
D4 G5 D5
1-year avg
G6 CALPUFF
D6
Distance to the cement plant 1 (km)
Grid Discrete (Receptor 5) Grid Discrete (Receptor 11) Grid Discrete (Receptor 8) Grid Discrete (Receptor 8) Grid Discrete (Receptor 11) Grid Discrete (Receptor 11)
Local station time
G1
PM concentration 3 (µg/m )
1-h avg
Receptor types
AERMOD
Receptor
Model
Max. PM concentration
Table 5. The maximum PM concentrations obtained by AERMOD and CALPUFF.
95.26
4/01/07 at 10 a.m.
5.90
47.57
28/08/07 at 9 a.m.
2.35
262.97
19/10/07 at 9 p.m.
2.55
105.69
17/03/07 at 11 p.m.
2.47
11.43
25/05/07
6.25
5.40
24/07/07
3.90
39.68
19/11/07
4.02
11.12
16/10/07
3.90
2.88
-
6.25
1.03
-
2.47
8.23
-
4.02
2.09
-
2.47
4. Conclusions Both the U.S. EPA recommended AERMOD and CALPUFF models can be applied to the mountain-valley areas where cement complex is located. The CALPUFF model showed higher values and frequently found the PM peaks than the AERMOD. In short, particulate matters are an important concern. TSP and PM10 emissions must be monitored twice a year at receptors nearby residential areas around the cement plants to minimize the impact on human’s health. The obtained result is not valid if this research evaluates the model performance using only the existing PM emission input from the cement stacks since there are various PM sources. This case proves that PM emissions from the stacks contributed very little to the level of PM concentrations found at the receptors. Near future research is to incorporate more PM emission sources including to figure out what is the best strategy to control PM in the long run.
5. Acknowledgements This work was supported by the Royal Golden Jubilee Ph.D. program (IUG50K0021), Thailand Research Fund (TRF). The authors wish to thank the Air Quality and Noise Management Bureau, Pollution Control Department, Ministry of Natural Resources and Environment, the Thai Meteorological Department, Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang.
6. References [1] US.EPA. http://www.epa.gov/air/particlepollution/index.html. [2] Heinsohn, R.J., and Kabel, R.L., 1999. Sources and Control of Air Pollution. United States of America: Prentice-Hall, Inc. [3] Baroutian, S., Mohebbi, A., and Soltani Goharrizi, A., 2006. Measuring and Modeling Particulate Dispersion: A Case Study of Kerman Cement Plant. Journal of Hazardous Materials A136, 468-474. [4] Goyal, P., Singh, M.P., and Gulati, A., 1996. Air Quality Assessment of the Environment over a Cement Industrial Complex. Atmospheric Environment 30 (7), 1159-1166.
20 [5] Pollution Control Department, 2006. Project of studying the problem and impact of PM10 in Saraburi; surveying, measuring and developing database of PM10 emission sources in Nah Phra Laan. (Published in Thai language) [6] Sivacoumar, R., Jayabalou, R., Swarnalatha, S., and Balakrishnan, K., 2006. Particulate Matter from Stone Crushing Industry: Size Distribution and Health Effects. Journal of Environmental Engineering, 132 (3), 405-414. [7] Pimonsree, S., Wongwises, P., Rudklao, P.A., and Meigen, Z., 2008. Dispersion modeling of PM10 during winter episode over a mineral products industrial area in Saraburi, Thailand. Proceedings of the International Conference on Environmental Research and Technology, Penang, Malaysia. 28 – 30 May 2008. pp. 637-641 (available online at http://www.ppti.usm.my/ICERT_website/index.html). [8] Code of Federal Regulations, 40 CFR Part 60, Appendix A, Office of the Federal Register. National Archives and Records Administration, Washington, D.D., July 1, 1993. [9] Earth Tech, Inc., A User’ Guide for the CALPUFF Dispersion Model (Version 5). 2000 (available online at www.src.com) [10] Hidore, J.J. and Oliver, J.E., 1993. Climatology and Atmospheric Science. Macmillan, New York.
MEASURING AND MODELING OF PARTICULATE DISPERSION FROM THE CEMENT PLANTS Anchaleeporn W. Lothongkum1*, Kanyanee Seangkiatiyuth1 and Vanisa Surapipith2* 1*
2*
Department of Chemical Engineering, Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand ([email protected])
Air Quality and Noise Management Bureau, Pollution Control Department, Ministry of Natural Resources and Environment, Bangkok 10400, Thailand ([email protected])
Abstract This work focuses on the dispersion of particulate matters in the atmosphere around key cement plants, situated about 100 km northeast of Bangkok, Thailand. It is aimed at identifying the impact of emission from four cement manufacturers by both measuring and modeling analyses. The daily average concentrations of total suspended particulate matters (TSP) and particulate matters with diameter less than 10 µm (PM10) within the distance of 15 km from the cement plant were measured for a week time period every six months at 12 receptors by high-volume volume samplers. The PM10 concentration at further distance, i.e., about 25 km away from the cement plant was also monitored routinely at 2 automatic monitoring station of the Pollution Control Department (PCD), Thailand. Two Gaussian based models, AERMOD and CALPUFF, were then applied to estimate the dispersion of the particulate matters (PM). The estimated results by both AERMOD and CALPUFF showed that the model-predicted PM concentration at 12 receptors and 2 automatic monitoring stations were lower than the observed concentration. The CALPUFF model showed higher values and frequently found the PM peaks than the AERMOD. Keywords: cement plant, TSP, PM10, AERMOD, CALPUFF
2
1. Introduction Particulate matters with diameter of 10 µm or smaller (PM10) have been concerned worldwide because they can pass through human’s respiratory system to the lungs causing serious health effects, for example, aggravated respiratory symptoms, deteriorating lung function, asthma, development of chronic bronchitis, irregular heartbeat and nonfatal heart attacks. In case of the particulate matters with diameter of 2.5 µm (PM2.5), it can impair visibility and reduce solar radiation.[1] In Thailand, two categories of particulate matters (PM): total particulate matters (TSP) of which its diameter is 100 µm or smaller, and PM10 have been regulated. Cement industry is well-known as a significant emission source of the PM. Generally, PM released from the cement manufacturer ranges from 0.12-100 µm.[2] PM of which the range impacts to human’s welfare and the environmental assessment has been investigated by measurement and modeling method. The Gaussian models were reported the PM pattern from stacks of the Kerman cement in Iran, and the cement complex in India.[3-4] The reports showed that the results from the models agreed with those of the measurement. The released PM from the stacks of the cement plant in Iran was a very significant effect on the respiratory problems. During the year of 2006, the Pollution Control Department (PCD), Ministry of Natural Resources and Environment, Thailand carried a pilot project to establish an emission inventory in order to investigate the impact of fine particulates around the industrial areas located in Nah Phra Laan; a part of Chaloem Phra Kiat district, Saraburi province. One of a key results indicated that the major source of PM10 was from crushing activity.[5] Similar to the PCD’s results, in 2006, R. Sivacoumar et al. reported that the dust generated from stone crushing activities contained a significant amount of fine inhalable matters. Not only the sizes but also sources of fine particulate matters are of important to be concerned.[6] Based on the results from the PCD, the Hybrid Particle and Concentration Transport model (HYPACT) with Lagrangian option and the Regional Atmospheric Modeling System (RAMS) was consequently used in evaluation of PM10 dispersion by Sittichai et al.[7] They found that the modeling results were in good agreement with the observations by the PCD. It was revealed that high PM10 concentration appeared in the source area and near source area in downwind direction. Kaengkoi district, situated about 21 km southeast of Chaloem Phra Kiat district, is of interest area as it is
the location of cement complex and the activities in the area are rather similar to those of Nah Phra Laan. This paper focuses on the PM emission from 14 stacks of the four cement plants located near the mountain areas as shown in Figure 1. From the Environmental Impact Assessment (EIA) reports of the four model plants, the total PM emission is 4,286 ton/year, based on the total emission rate of 136 g/s. Considering the total emission of the cement industry in Nah Phra Laan area, the emission from in Kaengkoi is higher (total stack PM10 emission from Nah Pra Laan is about 600 ton/year). Therefore, in some areas, though stacks are not the major PM source, the continuous release may contribute to a significant impact. However, by setting up the electrostatic precipitator with high efficiency of 99.99% at the stack, it can be assumed that the PM emission from the stacks are PM10, and considered the released PM as PM10. As it is impossible to evaluate the overall PM emission by measurement only, two dispersion models which are recommended by the United States Environmental Protection Agency (U.S.EPA): the American Meteorology Society-Environmental Protection Agency Regulatory Model (AERMOD) and the California Puff Dispersion Model (CALPUFF) are applied by using the surface and upper meteorological observation data input from the stations.
3
⊗
Figure 1. Locations of the study areas.
2. Methodology 2.1 Study areas Four target cement plants are located in Kaeng Khoi district, Saraburi province in the middle of Thailand as shown in Figure 1. The eastern part of the province is covered by high plains and plateaus, while the western part is mostly low flat plains. The study domain coordinates cover UTM X from 701000E to 751000E, and UTM Y from 1593000N to 1643000N. The domain area is located in time zone 47P with WGS84 datum type. The weather in this area is sultry hot in summer from February to April, heavy rain from May to September and cool in winter from October to January.
2.2 Emission and ambient air measurement The PM emission from the stacks of the four cement plants were reported as TSP in the each EIA report. The emission was collected in accordance with the Code of Federal Regulations, 40 CFR Part 60, Appendix A.[8] The volumetric flow rate of the effluent gas from the stack was measured by a pitot tube at
4 several points in the duct. The emission rates of each cement plant which are used in the AERMOD and CALPUFF programs are shown in Table 1. Table 1. Emission rates of four cement plants. Cement plant number 1 2 3 4
Emission rate (g/s) 32.9 36.8 63.8 2.4
The PM in ambient air was collected at 12 receptors, which were selected according to the guidelines in the EIA report of the cement plant 1. TSP and PM10 were continuously collected for a week (168 hours) by a high-volume sampler, which was 2-m above ground level with a flow rate of gas stream of 10 L/min. After running continuously for 24 hours, the glass fiber filters of the sampler were changed. At the end of the sampling period, the filters were removed, dried and reweighed. The concentration of the TSP, reporting in micrograms (or milligram) per cubic meter, was calculated from the weight gained by the filters divided by the volume of air sampled. According to the guidelines in the EIA report, the PM must be measured twice a year as summarized in Table 2. The data which used in AERMOD and CALPUFF in this work were taken from the 6-month EIA report of each cement plant, which was approved by the National Committee by the Office of Environment at Policy and Planning. Continuous monitoring of PM10 concentration in ambient air was detected by 2 automatic monitoring stations of the PCD at Kao Noi and Nah Phra Laan. The PM10 at Kao Noi station was measured by equivalent method, and at Nah Phra Laan station by the Tapered Element Oscillating Microbalance (TEOM) method. The observation periods in this study are also shown in Table 2.
2.3 Meteorological data The meteorological data were received from the following 4 meteorological stations of the PCD and the Thai Meteorological Department (TMD). Kao Noi and Nah Phra Laan are both meteorological and monitoring stations of PCD where no observation of cloud cover and ceiling height are available. Since the main purpose of both Kao Noi and Nah Phra Laan stations is for air pollution monitoring applications, therefore, the stations are located near the residential areas. Pakchong, a station of TMD located nearby Kao Yai Natural Park, is weather monitoring for agriculture. It is surrounded by high mountains thus the meteorological data are influenced by mountain-valley wind. Bangna, another station of TMD, provides both surface and upper air data. Although Bangna station situates over 100 km away in southwest of the study areas, it is selected for upper air data because it is the nearest radiosonde balloon launching site. The surface air data in this work is received from other 3 stations mentioned above. Details of the meteorological stations and the relevant data are presented in Table 3. The locations of surface meteorological stations can be seen in Figure 1.
5 Table 2. Periods of ambient air monitoring. Receptors 1 2 3 4 5 6 7 8 9 10 11 12 Kao Noi Nah Phra Laan
Period 1 Start End 2/03/07: h13 9/03/07: h13 2/03/07: h10 9/03/07: h10 2/03/07: h12 9/03/07: h12 2/03/07: h16 9/03/07: h16 9/03/07: h14 16/03/07: h14 9/03/07: h8 16/03/07: h8 10/03/07: h12 17/03/07: h12 10/03/07: h13 17/03/07: h13 17/03/07: h11 24/03/07: h11 17/03/07: h18 24/03/07: h18 17/03/07: h16 24/03/07: h16 17/03/07: h12 24/03/07: h12 2/03/07: h1 24/03/07: h24 2/03/07: h1 24/03/07: h24
Period 2 Start End 12/10/07: h15 19/10/07: h15 12/10/07: h17 19/10/07: h17 12/10/07: h14 19/10/07: h14 20/10/07: h15 27/10/07: h15 20/10/07: h9 27/10/07: h9 20/10/07: h18 27/10/07: h18 20/10/07: h12 27/10/07: h12 28/10/07: h16 4/11/07: h16 28/10/07: h9 4/11/07: h9 12/10/07: h17 19/10/07: h17 28/10/07: h12 4/11/07: h12 28/10/07: h9 4/11/07: h9 12/10/07: h1 4/11/07: h24 12/10/07: h1 4/11/07: h24
Table 3. Details of the meteorological stations. Meteorological station
Collected data by
Meteorological parameter
Distance from the cement plant 1 (km)
Direction to the cement plant 1
Kao Noi
PCD
20.73
SW
Nah Phra Laan
PCD
24.13
WNW
Pakchong
TMD
25.14
E
Bangna
TMD
surface: G, P1, RH, T 10 m: WD, WS1 (hourly basis data) surface: G, P1, RH, T 10 m: WD, WS1 (hourly basis data) surface: P2, RH, T 10 m: WD, WS2 upper air data: CH, CV (3-hour basis data) upper air data: DWPT, H, P, RH, T, WD, WS2
118.93
SSW
Notes: CH CV DWPT G H P1 P2 RH T WD WS1 WS2
= = = = = = = = = = = =
ceiling height (m) cloud cover (tenths) dew point temperature (˚C) global radiation (Wh/m2) height above sea level (m) pressure (mmHg) pressure (hPa) relative humidity (%) dry bulb temperature (˚C) wind direction (degree) wind speed (m/s) wind speed (knot)
2.4 Modeling AERMOD is a steady-state plume model that incorporates air dispersion based on planetary boundary layer turbulence structure and scaling concepts, including treatment of both surface and elevated sources, and both simple and complex terrain. AERMOD is recommended to evaluate the pollutants dispersion in surrounded distance within 50 km from the emission source. It is composed of three parts: AERMOD Meteorological Preprocessor (AERMET), AERMOD Terrain Preprocessor (AERMAP) and AERMOD. The AERMET processes the hourly surface data. The AERMAP is used to process the terrain data in conjunction with a layout of receptor and sources for AERMOD control files. The AERMOD is a
6
Figure 2. Data flow in AERMOD.
Figure 3. Data flow in CALPUFF.
dispersion model. In this work, the commercial interface, ISC-AERMOD View (Version 4.2.6) (Lakes Environmental Software; Waterloo, Ontario, Canada) was run. The surface meteorological data obtained from Kao Noi station while the cloud cover and ceiling height were from Pakchong station. The upper meteorological data obtained from Bangna station. The data flow of the program is shown in Figure 2. The ISC-AERMOD View was run with the modeling domain covered the smaller area than that of CALPUFF. The maximum grid number which ISC-AERMOD View can operate is 7,200 grids. The modeling domain is expected to be large enough to see plume characteristics, therefore, it is designed to covers 39.5 x 39.5 km2 on UTM 47P coordinates between 701000E to 740500E and 1600000N to 1639500N. The grid spacing between the receptors is 500 x 500 m2. CALPUFF is non-steady-state puff dispersion model that simulates the effects of time- and space-varying meteorological conditions on pollution transport, transformation and removal. CALPUFF can be applied for pollution transport from the distance between 100 m – 100s km and for complex terrain. [9] CALPUFF modeling system includes three main components: CALMET, CALPUFF and CALPOST. CALMET is a meteorological model that develops hourly wind and temperature fields on a 3-D gridded modeling domain and the generated meteorological data can be displayed by PRTMET, a postprocessing program. CALPOST is used to process the files from CALPUFF, producing tabulations that summarize the results of the simulation, identifying the highest and second highest 1-hour, 3-hour, 24-hour or period averaged concentration at each receptor. The data flow in CALPUFF is shown in Figure 3. In this work, the CALPUFF PROfessional Beta 5.2.0 was run with free version which could be downloaded from the website of Earth Tech, Inc (www.src.com/calpuff1.htm). The surface meteorological data obtained from three stations: Kao Noi, Nah Phra Laan and Pakchong. The upper meteorological data obtained from Bangna station. The modeling domain covers 50 x 50 km2 on UTM 47P coordinates between 701000E to 751000E and 1593000N to 1643000N. Both discrete and grid modes (grid spacing 500 x 500 m2) were selected to find the PM concentration at ground level of 12 receptors and the impact areas.
3. Results and Discussion 3.1 Meteorological analysis The daily climate of Kaengkoi, the of interest area, is influenced by the seasonal winds and mountain-valley breeze. For the seasonal winds, the southwesterly monsoon brings moisture from the Indian Ocean to Asian Continent during March to September, whereas the northeasterly monsoon blows from the cool Asian land mass towards the Indian Ocean during October to February. The valley breeze is a warm wind occurring in the daytime or late afternoon. The clouds which are formed over the hills in the afternoon are the result of the condensation taking place when the air rises to cooler heights over the mountains. At night, the valley walls cool down resulting to a cool air surface. Therefore, winds flow downwards the slope of the mountains known as night breeze.[10] At receptor areas, during 2 to 24 March 2007 the average wind speed was 0.4 m/s or lower at the night to the morning (7 p.m.-10 a.m.); then, the wind speed increased but not over 4.5 m/s during the day (10 a.m. – 7 p.m.). During 12 October to 4 November 2007, the wind was gentle from the evening to the morning (5 p.m.-10 a.m.) after that the wind speed increased but not higher than 4.5 m/s. Wind statistics at Kao Noi, Nah Phra Laan and Pakchong stations are summarized in terms of wind roses, as shown in Figures 4(a)-(c), respectively. The wind rose shows wind strength, direction and frequency.
7 The percentage of calm wind is represented by the size of the center circle; the bigger the circle, the frequent higher calm wind was found. Each branch of the wind rose represents wind-coming direction, identifying the north at the top of the diagram. Each segment length within a branch is proportional to the frequency of winds blowing within the corresponding range speeds from that direction. Figure 4(a) shows that during 2-24 March 2007 the wind at Kao Noi station is dominated by westerly wind. The strongest wind is about 3.5 m/s. The major wind direction of Nah Phra Laan station is southerly as shown in Figure 4(b). The wind speed is light of 1-2 m/s. Figure 4(c) shows that the wind at Pakchong station is dominated by east-southeasterly wind. The wind is rather strong and the strongest wind speed is 9.3 m/s.
Calm wind 3.84%
Calm wind 5.98%
(a) Wind speed (m/s)
Calm wind 52.72%
(b) ≥ 6.0
4.0-6.0
(c) 3.0-4.0
2.0-3.0
1.0-2.0
0.5-1.0
Figure 4. Wind rose during 2-24 March 2007 (a) Kao Noi, (b) Nah Phra Laan and (c) Pakchong.
Calm wind 5.93%
Calm wind 26.04%
Calm wind 2.86%
Wind speed (m/s)
(c)
(b)
(a) ≥ 6.0
4.0-6.0
3.0-4.0
2.0-3.0
1.0-2.0
0.5-1.0
Figure 5. Wind rose during 12 October-4 November 2007 (a) Kao Noi, (b) Nah Phra Laan and (c) Pakchong.
8 The wind roses of the three stations during 12 October-4 November 2007 are shown in Figures 5(a)-(c). During this period, the major wind at Kao Noi station are still westerly with higher speed. The strongest wind speed is 5.7 m/s. The wind at Nah Phra Laan station in this season has turned to northeasterly. The wind speed is varied from 0.5 to 6 m/s. The dominated wind at Pakchong station is still east-southeasterly. The wind speed is varied from calm wind to 12.3 m/s. In Figures 6(a)-(b), the wind speeds of the three stations were plotted as the wind class distribution. It is clearly seen that wind speed at Pakchong station is either calm or strong (over 3 m/s) during both periods. While the winds at Kao Noi and Nah Prah Laan stations are within 1 – 3 m/s most of the time.
Percent (%)
Percent (%)
The non-uniform wind direction and speed at the three stations as shown through the analysis clarifies that the winds in the study areas are strongly influenced by mountain-valley breezes.
Wind speed (m/s)
Wind speed (m/s)
(a) (b) Figure 6. Wind class distribution at Kao Noi, Nah Phra Laan and Pakchong stations: (a) during 2-24 March 2007 and (b) during 12 October – 4 November 2007.
3.2 Wind field pattern evaluated by CALMET CALMET is a meteorological model which includes a diagnostic wind field generator containing objective analysis and parameterized treatments of slope flows, kinematic terrain effects, terrain blocking effects, a divergence minimization procedure, and a micro-meteorological model for overland and overwater boundary layers.[5] Wind speeds and directions in this work which were simulated by CALMET of the selected days were plotted to evaluate the wind field pattern during two episodes. Figure 7 shows wind field on 9 March 2007 representing 2 - 24 March 2007. It is the output of the CALMET hourly wind speeds and directions which were generated by using input from the surface meteorological data recorded above the ground level of 10 meters from three stations; Kao Noi, Nah Phra Laan and Pakchong, and from the upper meteorological data from Bangna station. Figure 8 shows CALMET wind field on 1 November 2007 representing 12 October - 4 November 2007. The wind fields show that most of the time the wind speed was southwesterly. Seasonally, the southwesterly monsoon should be found during March to September, and the northeasterly monsoon during October to February. In fact, the season wind affects to the wind profile from CALMET directly, especially the wind direction should be in accordance with the season wind direction. However, in this study the wind profile direction has not changed when the season wind direction changed. This might be because of (i) effect of mountain-valley breeze and (ii) the upper meteorological data are not relevant to the wind field calculation. This part will be further study.
9
7h UTM Y (km)
16 h UTM Y (km)
UTM Y (km)
UTM Y (km)
22 h
19 h
UTM X (km)
UTM X (km)
UTM X (km)
UTM X (km)
UTM Y (km)
UTM Y (km)
UTM Y (km)
UTM X (km)
UTM X (km)
13 h
10 h UTM Y (km)
4h
1h
UTM X (km)
UTM X (km)
Figure 7. Wind field on 9 March 2007 generated by CALMET.
7h
UTM X (km)
16 h
UTM Y (km)
22 h
UTM Y (km)
19 h
UTM Y (km)
UTM Y (km)
UTM X (km)
UTM X (km)
UTM X (km)
UTM X (km)
13 h
10 h
UTM Y (km)
UTM Y (km)
UTM Y (km)
4h
UTM X (km)
UTM Y (km)
1h
UTM X (km)
Figure 8. Wind field on 1 November 2007 generated by CALMET.
UTM X (km)
10
3.3 PM measurement From the EIA reports, PM emissions of 4 cement plants were within the permissible emission of 50 mg/m3. The PM emissions from 4 cement plants are shown in Figure 9. It can be seen that cement plant 3 released the largest volume of PM whereas the cement plant 4 released the lowest. PM10 concentration was measured routinely by the two automatic monitoring stations of the PCD away from the cement plant 1 about 25 km. Figures 10(a)-(b) and 11(a)-(b) show the PM10 concentrations monitoring at Kao Noi and Nah Phra Laan stations. The results of Nah Figure 9. PM emissions from the stacks Phra Laan station, which is located around by stone and of 4 cement plants. mining industries, shows that the 24-h avg peak of PM10 concentration in 5 days during 2 to 24 March 2007, and 11 days during 12 October to 4 November 2007 are over the Thai permissible value of 120 µg/m3 In contrast, none at Kao Noi station are over the permissible value. Based on the 1-h avg PM10 concentration, Nah Phra Laan station was frequently found peaks of PM10 concentration more than 120 µg/m3 while Kao Noi was not. Particularly, during 2-24 March 2007, both 1-h avg and 24-h avg peaks of PM10 concentration at Nah Phra Laan station is rather high resulting in significant impact to human’s health. Taking into account the twelve receptors near 4 cement plants, it was found that the 24-avg PM10 and TSP concentrations of all receptors were not over Thailand’s permissible value of 120 µg/m3 for PM10 and of 330 µg/m3 for TSP as presented in Figures 12 and 13. In common, other influences on twelve receptors, apart from the stack emission, are possible PM sources in the areas, e.g., motor vehicles, open mining, power utilities of other industries, combustion products from open burning, and residential cooking. Table 4 presents the characterization of the receptors. It was found that the difference between the measured TSP and PM10 concentrations in two periods (2 to 24 March 2007 and 12 October to 4 November 2007) was not significant, especially at receptors 3 and 8 during 12 October to 4 November 2007. Based on the information by the PCD’s two automatic monitoring stations, some over-limit 24-h avg PM10 are found. This is likely due to the emission activities around the stations, where campaign of emission reduction is going on. The Nah Phra Laan district has been designated as Pollution Control Zone since 2005, and air quality has been improving by restricted rules for all trucks and mining activities in that area to clean up the dust throughout the year. Table 4. Characterization of the receptors. Receptor 1 2 3 4 5 6 7 8 9 10 11 12 Kao Noi Nah Phra Laan
Characterization of receptors on the area of the temple and near the road on the area of the temple and near the road on the area of the temple, near the residential areas, agricultural area and the road on the area of the school, near the residential areas and the road on the area of the school, near the residential areas and the road on the area of the school, near the residential areas, agriculture and the road on the are of electrical substation, near the residential areas and the road on the area of animal conservation station, near agricultural area, not far from the residential areas and the road on the area of the school, near the residential areas, agricultural area and the road on the area of the school, near the residential areas, agricultural area and the road on the road side, near the residential areas and agricultural area on the area of the school, near the residential areas, agricultural area and the road near the residential areas near the road and open mining
PM10 concentration (µg/m3)
11
Kao Noi
no data
Date PM10 concentration (µg/m3)
(a) no data
Nah Phra Laan
Date
(b)
PM10 concentration (µg/m3)
Figure 10. PM10 concentration monitoring during 2-24 March 2007: ( ) 1-h avg and (•) 24-h avg. no data
Kao Noi
Date
PM10 concentration (µg/m3)
(a) no data
Nah Phra Laan
Date
(b) Figure 11. PM10 concentration monitoring during 12 October – 4 Novemberh 2007: ( ) 1-h avg and (•) 24-h avg.
24-h avg PM concentration (µg/m3)
Receptor 7
Date
Receptor 10
Date Receptor 5
Date
Date
24-h avg.PM concentration (µg/m3)
Date 24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Date
Receptor 8
Receptor 11
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 2
24-h avg PM concentration (µg/m3)
Date 24-h avg PM concentration (µg/m3)
Receptor 4
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 1
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
12
Date
Receptor 3
Date
Date
Receptor 9
Date
Receptor 12
Date
Figure 12. TSP and PM10 concentrations at 12 receptors during 2-24 March 2007: (•) TSP and (+) PM10.
24-h avg PM concentration (µg/m3)
Receptor 7
Date
Receptor 10
Date Receptor 5
24-h avg PM concentration (µg/m3)
Date
Receptor 8
Receptor 11
Date 24-h avg PM concentration (µg/m3)
Date
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 2
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 4
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
Receptor 1
24-h avg PM concentration (µg/m3)
24-h avg PM concentration (µg/m3)
13
Date
Receptor 3
Date Date
Receptor 6
Date Date
Receptor 9
Date
Receptor 12
Date
Figure 13. TSP and PM10 concentrations at 12 receptors during 12 October-4 November 2007: (•) TSP and (+) PM10.
14 3.4 Modeling results by AERMOD and CALPUFF In this work AERMOD and CALPUFF models were used to evaluate the impact results of the PM emissions from the stacks of 4 cement plants. The stack emission samplings in the EIA reports were noted in terms of TSP concentration, considering as PM10. The model-predicted PM concentrations at Kao Noi and Nah Phra Laan stations during the two periods are shown in Figures 14 (a)-(b) and 15(a)-(b). It can be concluded from the plots that AERMOD and CALPUFF predict the PM concentration in a similar range. AERMOD generated more peaks at Kao Noi station during March than October to November. On a contrary, CALPUFF generated more peaks during October to November than March. Model predictions at Nah Prah Laan during March by the two models are similar, whereas, AERMOD rarely generated peaks during October to November.
PM10 concentration (µg/m3)
The model-predicted PM concentrations at the twelve receptors are shown in Figures 16 and 17. Clearly, the similar trend by the PDC monitoring stations is seen. AERMOD generated many peaks during March, while CALPUFF generated mostly during October to November. Of the 12 receptors, only the receptor 3 CALPUFF generated more peaks in March.
Kao Noi
Date
PM10 concentration (µg/m3)
(a)
Nah Phra Laan
Date
(b) Figure 14. Model-predicted PM10 concentration at 2 automatic PCD stations during 2-24 March 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
PM10 concentration (µg/m3)
15
Kao Noi
Date
PM10 concentration (µg/m3)
(a) Nah Phra Laan
Date
(b) Figure 15. Model-predicted PM10 concentration at 2 automatic PCD stations during 12 October4 November 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
Receptor 1
3 3 3
Receptor Receptor88
3
PM concentration (µg/m )
3
3
Time (h)
Receptor 9
Time (h)
Time (h) PM concentration (µg/m )
Time (h)
Receptor 6
Time (h) PM concentration (µg/m )
3
PM concentration (µg/m )
3
PM concentration (µg/m )
Receptor 5
Time (h) Receptor 77 Receptor
Receptor 3
Time (h) PM concentration (µg/m )
3
PM concentration (µg/m )
3
PM concentration (µg/m )
Receptor 4
Time (h)
PM concentration (µg/m )
Receptor 2
Time (h)
Time (h)
Receptor 10
PM concentration (µg/m )
3
3
PM concentration (µg/m )
PM concentration (µg/m )
16
Receptor 11
Time (h)
Receptor 12
Time (h)
Figure 16. Model-predicted PM concentration at 12 receptors during 2-24 March 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
Receptor 10
Time (h)
3 3
3
PM concentration (µg/m )
Receptor 9
Time (h) 3
3
3
Receptor 8
Time (h) PM concentration (µg/m )
Time (h)
Receptor 6
Time (h)
Time (h) 3
Receptor 7
PM concentration (µg/m )
3
Receptor 5
PM concentration (µg/m )
Time (h)
Receptor 3
Time (h)
Time (h) PM concentration (µg/m )
3
PM concentration (µg/m )
Receptor 4
3
PM concentration (µg/m )
Receptor 2
PM concentration (µg/m )
Receptor 1
Time (h)
PM concentration (µg/m )
PM concentration (µg/m )
3
PM concentration (µg/m )
3
PM concentration (µg/m )
17
Receptor 11
Time (h)
Receptor 12
Time (h)
Figure 17. Model-predicted PM concentration at 12 receptors during 12 October – 4 November 2007: ( ) 1-h avg concentration by AERMOD, (•) 24-h avg concentration by AERMOD, ( ) 1-h avg concentration by CALPUFF and (×) 24-h avg concentration by CALPUFF.
Hour
Modeled PM concentration (µg/m3)
Measured PM concentration (µg/m3)
Modeled PM concentration (µg/m3)
18
1.16 0.67
0.60 0.19
Hour
Figure 19. Simulated PM10 concentration at Figure 18. 1-h avg measured and simulated receptor 4 during 5-7 March 2007: PM10 concentration at Nah PhraLaan on 20 March 2007: ( ) AERMOD, ( ) 1-h avg by AERMOD, (•) 24-h avg ( ) CALPUFF and ( ) measured. by AERMOD, ( ) 1-h avg by CALPUFF and (×) 24-h avg by CALPUFF.
To clarify the performances of AERMOD and CALPUFF, time series of the measured and model-predicted PM10 concentrations at Nah Pha Laan station on 20 March 2007 are plotted in Figure 18. Even though the PM concentrations from the models are very low comparing to the measured results but the peak pattern during the day can be seen. CALPUFF can calculate the peaks which appear about 7, 16 and 21 hrs at the local station time (LST), while AERMOD can calculate only one peak about 14 hrs at LST. This reveals that AERMOD gives the statistical calculation whereas CALPUFF can detect real time values. Figure 19 shows the results at receptor 4 indicating that CALPUFF generated several PM concentration peaks while AERMOD could detect only one peak at 14 hrs. The magnitude of the peak generated by AERMOD was lower than that of CALPUFF.
3.5 The maximum PM concentration Table 5 shows the average model-predicted maximum PM concentrations by AERMOD and CALPUFF at different periods; 1 h, 24 h and 1 year. As seen, 1-h avg values by CALPUFF were much higher than those by AERMOD. The difference between the generated PM concentrations decreased with long range period such as 24 h and 1 year. It should be noted that the 24-h avg PM concentrations predicted by both AERMOD and CALPUFF are in Thailand’s ambient air quality standard of PM10 (120 µg/m3), and slightly low comparing to the TSP standard of 330 µg/m3. The possible maximum PM concentrations are on different spots at the receptors around the cement plants 1, 2 and 3 as far as 2-7 km, as shown in Figure 20.
D2,D5,D6 G4,G6 G1 G3,G5
D3,D4 G2 D1
Figure 20. Possible maximum PM concentration locations by AERMOD and CALPUFF: ( ) stack and (+) receptor.
19
D1 G2 CALPUFF AERMOD
D2 G3 D3
24-h avg
G4 CALPUFF AERMOD
D4 G5 D5
1-year avg
G6 CALPUFF
D6
Distance to the cement plant 1 (km)
Grid Discrete (Receptor 5) Grid Discrete (Receptor 11) Grid Discrete (Receptor 8) Grid Discrete (Receptor 8) Grid Discrete (Receptor 11) Grid Discrete (Receptor 11)
Local station time
G1
PM concentration 3 (µg/m )
1-h avg
Receptor types
AERMOD
Receptor
Model
Max. PM concentration
Table 5. The maximum PM concentrations obtained by AERMOD and CALPUFF.
95.26
4/01/07 at 10 a.m.
5.90
47.57
28/08/07 at 9 a.m.
2.35
262.97
19/10/07 at 9 p.m.
2.55
105.69
17/03/07 at 11 p.m.
2.47
11.43
25/05/07
6.25
5.40
24/07/07
3.90
39.68
19/11/07
4.02
11.12
16/10/07
3.90
2.88
-
6.25
1.03
-
2.47
8.23
-
4.02
2.09
-
2.47
4. Conclusions Both the U.S. EPA recommended AERMOD and CALPUFF models can be applied to the mountain-valley areas where cement complex is located. The CALPUFF model showed higher values and frequently found the PM peaks than the AERMOD. In short, particulate matters are an important concern. TSP and PM10 emissions must be monitored twice a year at receptors nearby residential areas around the cement plants to minimize the impact on human’s health. The obtained result is not valid if this research evaluates the model performance using only the existing PM emission input from the cement stacks since there are various PM sources. This case proves that PM emissions from the stacks contributed very little to the level of PM concentrations found at the receptors. Near future research is to incorporate more PM emission sources including to figure out what is the best strategy to control PM in the long run.
5. Acknowledgements This work was supported by the Royal Golden Jubilee Ph.D. program (IUG50K0021), Thailand Research Fund (TRF). The authors wish to thank the Air Quality and Noise Management Bureau, Pollution Control Department, Ministry of Natural Resources and Environment, the Thai Meteorological Department, Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang.
6. References [1] US.EPA. http://www.epa.gov/air/particlepollution/index.html. [2] Heinsohn, R.J., and Kabel, R.L., 1999. Sources and Control of Air Pollution. United States of America: Prentice-Hall, Inc. [3] Baroutian, S., Mohebbi, A., and Soltani Goharrizi, A., 2006. Measuring and Modeling Particulate Dispersion: A Case Study of Kerman Cement Plant. Journal of Hazardous Materials A136, 468-474. [4] Goyal, P., Singh, M.P., and Gulati, A., 1996. Air Quality Assessment of the Environment over a Cement Industrial Complex. Atmospheric Environment 30 (7), 1159-1166.
20 [5] Pollution Control Department, 2006. Project of studying the problem and impact of PM10 in Saraburi; surveying, measuring and developing database of PM10 emission sources in Nah Phra Laan. (Published in Thai language) [6] Sivacoumar, R., Jayabalou, R., Swarnalatha, S., and Balakrishnan, K., 2006. Particulate Matter from Stone Crushing Industry: Size Distribution and Health Effects. Journal of Environmental Engineering, 132 (3), 405-414. [7] Pimonsree, S., Wongwises, P., Rudklao, P.A., and Meigen, Z., 2008. Dispersion modeling of PM10 during winter episode over a mineral products industrial area in Saraburi, Thailand. Proceedings of the International Conference on Environmental Research and Technology, Penang, Malaysia. 28 – 30 May 2008. pp. 637-641 (available online at http://www.ppti.usm.my/ICERT_website/index.html). [8] Code of Federal Regulations, 40 CFR Part 60, Appendix A, Office of the Federal Register. National Archives and Records Administration, Washington, D.D., July 1, 1993. [9] Earth Tech, Inc., A User’ Guide for the CALPUFF Dispersion Model (Version 5). 2000 (available online at www.src.com) [10] Hidore, J.J. and Oliver, J.E., 1993. Climatology and Atmospheric Science. Macmillan, New York.
