Characteristics of Ambient Ammonia Levels Measured in Three Different Industrial Parks in Southern Taiwan

To characterize concentrations of atmospheric ammonia (NH3) at industrial parks in southern Taiwan, three representative industrial parks and two background locations were selected for this study. The sampling process for gaseous ammonia was performed in accordance with Taiwan EPA NIEA A426.71B. The results revealed that the mean NH3 levels ranged from 70.5 to 153.9 ppb (mean = 100.2 ppb) in Neipu Industrial Park. Both highest mean levels in Neipu Industrial Park during 12:00–13:00 in the daytime and during 20:00–21:00 at night were about 50–150 times greater than ambient background levels. For the whole day, the mean NH3 levels ranged from 43.0 to 114.6 ppb (mean = 72.8 ppb) in Pingtung Industrial Park. In Pingnan Industrial Park, the mean NH3 levels for the whole day ranged from 45.0 to 122.6 ppb (mean = 84.9 ppb). Moreover, the ratio values (Rs) were categorized according to the cover-range (from minimum ratios to maximum ratios) at different sampling periods and the minimum separated boundary lines. The results indicate that the characteristics of ratio values normalized by the background levels can be divided into three main categories in southern Taiwan: (1) R > 200; (2) R > 60; (3) R > 10. Our research data reveal obvious reasons for this clear distinction of ambient NH3 levels in southern Taiwan. Significantly, the results reveal that agricultural NH3 emissions influence local ambient concentrations of NH3.


INTRODUCTION
Emission of ammonia (NH 3 ) is of concern because of its direct contribution to secondary photo-or smog reactions of air pollutants, and for its adverse influences on various ecosystems, such as bodies of water and soil.NH 3 may be either wet-or dry deposited as a gas, or react with sulfuric (H 2 SO 4 ), nitric (HNO 3 ), and hydrochloric (HCl) acids to form ammonium sulfate (NH 4 ) 2 SO 4 , bisulfate (NH 4 HSO 4 ), nitrate (NH 4 NO 3 ), and chloride (NH 4 Cl) aerosols, thereby contributing to inorganic ambient particulate matter (Walker et al., 2004).Like in many other countries, in Taiwan ammonia gas (NH 3 ) is defined as a toxic pollutant under the EPA's Air Pollution Control Act Enforcement Rules (promulgated by Environmental Protection Administration order on July 23, 2003) (EPA Act, 2010).Therefore, considerable attention is paid to emission of NH 3 Corresponding author.Tel.: 886-8-774-0521; Fax: 886-8-774-0256 E-mail address: Lthsieh@mail.npust.edu.tw in Taiwan's industrial parks in order to maintain air quality and reduce pollution.
Many different ammonia sources or sites have been studied to understand their harmful effects on air quality.Bouwman et al. (1997) reported a global high resolution emission inventory for ammonia.Their study indicates that domestic animals are the largest source (22 Tg N/yr, 1 Tg = 10 12 g) of atmospheric NH 3 , comprising approximately 40% of natural and anthropogenic emissions combined, while synthetic fertilizers and agricultural crops together contribute an additional 12.6 Tg NH 3 -N/yr (23% of total emissions) (Bouwman et al., 1997).Carmichael et al. (2003) reported their measurements of ammonia concentrations in Asia, Africa, and South America using passive samplers.Their monthly measurements were obtained at 50 stations in Asia, Africa, South America, and Europe.Their results indicated that median ammonia concentrations range from 20 ppb at Dhangadi, India, to 0.1 ppb at nine other stations.At 27 of the regional stations, the ambient ammonia levels were found to exceed 1 ppb.Walker et al., (2004) presented one year of ambient ammonia (NH 3 ) and other ions' concentrations at three sites in the Coastal Plain region of North Carolina.The three sites, Clinton, Kinston, and Morehead City, are all located in counties with total NH 3 emission densities of 4800, 2280, and 320 kg NH 3 -N/km 2 -yr, respectively (Walker et al., 2004).Their study found that the average NH 3 concentrations were 5.32, 2.46, and 0.58 g/m 3 at Clinton, Kinston, and Morehead City, respectively.In addition, their results showed that NH 3 concentrations were highest during the summer at all sites, with summerto-winter concentration ratios of 2.40, 5.70, and 1.70 at Clinton, Kinston, and Morehead City, respectively.They also indicated that NH 3 concentrations were higher at night at the Clinton site, during the day at the Kinston site, and day vs. night concentrations were similar at the Morehead City site (Walker et al., 2004).To sum up, this earlier study demonstrated that agricultural NH 3 emissions influence local ambient concentrations of NH 3 .Wilson and Serre (2007a) reported that the examination of atmospheric ammonia levels near hog concentrated animal feeding operations (namely, CAFOs), homes, and schools in eastern North Carolina.The data for each phase of sampling was stratified by distance from the nearest hog CAFO.The mean Phase I levels were 16, 8, 7 and 5 ppb for distances < 0.5, 0.5-1, 1-2, and 2 km or more, respectively.The mean levels for Phase II were 29, 16, and 11 ppb for distances < 0.5, 0.5-1, and 1 km or more, respectively (Wilson and Serre, 2007a).The results of the distance stratification provide the strongest evidence that distance from one or more CAFOs is the key variable in controlling weekly NH 3 atmospheric concentration at the community level in Eastern NC (Wilson and Serre, 2007a).
In fact, agriculture and livestock activity contribute significantly to anthropogenic ammonia (NH 3 ) emission to air, the latter ranging from 2.8 to 5.2 Tg (based on N-NH 3 ) per year in Western Europe (ECETOC, 1994).Balsari et al. (2007) reported that their experimental data highlighted a strong effect of both slurry temperature and air velocity over the slurry surface on NH 3 emissions.However, they also indicated that the funnel system allowed a comparison of emissions from different sources (cattle and pig slurry) under different temperature regimes, but, due to the very low wind speed under the funnels, did not allow data to be obtained in a way that would be comparable to real environmental conditions (Balsari et al., 2007).Therefore, it is important that real sampling be performed and that it be undertaken close to the investigated target.
Over the past decade, NH 3 has been intensively studied in many countries (McCulloch et al., 1998;Aneja et al., 2000;Walker et al., 2000aWalker et al., , 2000b;;Aneja et al., 2001;Pryor et al., 2001;Rattray et al., 2001;Robarge et al., 2002;Sather et al., 2008;Cao et al., 2009).However, research into NH 3 in different environmental matrices is still at an early stage in southern Taiwan.Therefore, a detailed and thorough environmental analysis of ammonia is needed to set a baseline for future research in Taiwan.In particular, two distinctive national features call for such an investigation to be conducted.Firstly, in southern Taiwan, many public disputes derive from malodorous toxic pollutants, such as ammonia gas.Moreover, ammonia gas is defined as a toxic pollutant in Taiwan EPA's Air Pollution Control Act Enforcement Rules (promulgated by Environmental Protection Administration order on July 23, 2003).In addition, according to Taiwan EPA's Air Pollution Control Act Enforcement Rules, Article 45 "If an air pollution injury incident in Article 80 of this Act overlaps two special municipalities, counties, or cities, or any combinations thereof, the victims may apply to the central competent authority for appraisal of the reason for their injuries."Therefore, a comprehensive environmental analysis of ammonia is required to improve the EPA's database that contains the characteristics of ambient ammonia levels measured in three different industrial parks in the Kao-Ping Area of Taiwan.Secondly, many unregistered farmyards raise hogs or fowl in southern Taiwan, and thus the lack of knowledge with regard to atmospheric NH 3 at the community level is an important problem, as without it is impossible to develop accurate exposure profiles for residents who live near such farmyards, making it difficult to adequately address any negative health responses in the exposed population.
One of the major preoccupations of environment researchers in southern Taiwan over the past decade has been investigating the characteristics of air-borne toxins.Building on this, the main objective of this study is to measure and characterize the concentrations of NH 3 during industrial activity in southern Taiwan.In addition, meteorological conditions, such as wind speed, wind direction, and temperature, were also measured during the sampling period.These results are expected to generate useful information with regard to evaluating the health effects of NH 3 .

Sampling Strategy
Southern Taiwan is located on the lee side of the central mountain range of Taiwan.This region has a stable atmospheric, clear skies and a distinct pattern of diurnal land-sea breeze circulations (Hsieh et al., 2005).To examine the spatial distribution of NH 3 in different industrial parks, five locations in southern Taiwan were chosen for NH 3 monitoring, three industrial locations plus two reference ones.Firstly, three industrial park sites (namely Neipu, Pingtung, and Pingnan Industrial Parks) were chosen based on both their different industrial and manufacturing characteristics.Secondly, two reference sampling sites near Lao-Pi village, NPUST dormitory (as a background site #1) and a bamboo grove (as a background site #2), were also chosen because there are no registered industrial plants nearby and they are also far from any other registered industrial or agriculture sources.More detailed descriptions of the five sampling locations are presented below.
Neipu Industrial Park: This is an industrial park with a very low population and traffic density.The total area of the park is approximately 99 ha, and the park contains a total of 61 factories, with the buildings few and widely scattered.The types of industry in the park can be divided into three categories: (1) food processing, (2) food manufacturing, and (3) manufacturing electrical appliances (Hsieh et al., 2005;PCIC, 2010).
Pingtung Industrial Park: This is an important industrial park near Pingtung City's main commercial area.This park has high traffic density and there are numerous high rise commercial buildings located nearby.The total area of the park is approximately 113.2 ha, and 146 factories are located within it.The types of industry in this park can be divided into three different categories: (1) machinery manufacturing, (2) metalworking, and (3) food manufacturing (Hsieh et al., 2005;PCIC, 2010).
Pingnan Industrial Park: This is a newly developed area which has slowly industrialized over recent years, close to both Fangliao and Donggang.The area contains a mixture of old low-rise residential buildings and newly built low-rise factory buildings.The total area of the park is approximately 278 ha, and the park contains some 97 factories.The park has two major types of industry: (1) iron and steel and (2) refrigerator manufacturing (Hsieh et al., 2005;PCIC, 2010).
NPUST Dormitory: The National Pingtung University of Science and Technology (NPUST) campus is situated in near Lao-Pi village in Ping-Tung County, with minimal anthrogenic pollution and adverse environmental impact.In this study, NPUST dormitory is considered as background monitoring station #1.In general, motor vehicles in the parking lots around the dormitory are the only potential sources of ammonia emission.
A Bamboo Grove: The bamboo grove is within an area of primeval forests near Lao-Pi village, and driving is strictly forbidden in this area.In this study, the bamboo grove is considered as background monitoring station #2.The major difference between background monitoring stations #1 and #2 is the amount of traffic they experience.

Sampling and Instruments
Fig. 1 displays the locations of five ammonia sampling sites.Field sampling was conducted on two consecutive days per week from September 2003 to December 2004, with rainy days excluded.On each sampling day, samplers were deployed for 14-15 hours at sites at Neipu Industrial Park (the number of samples n, n = 120), Pingtung Industrial Park (n = 120), Pingnan Industrial Park (n = 112), NPUST dormitory (n = 120) and the bamboo grove (n =120).The sampling system operated on 09:00-24:00 daily-cycle at all the sampling sites, with the exception of the site at Pingnan Industrial Park, which was on a 10:00-24:00 daily-cycle.According to EPA NIEA A426.71B (Taiwan EPA, 1997), ambient air should into a sampling duct by using a pump at flow rates under 5 L/min.For NH 3 sampling, the ambient air flows through a prefilter, a front-impinger bottle, a rear-impinger bottle, and a dehumidifier bottle in sequence.During each sampling hour, each impinger bottle contains ~25 mL of 0.1 N H 2 SO 4 .In the field, after starting the pump the ambient air flows through a pre-filter, then the both particle-free and gaseous NH 3 enters the two impinger bottles (including a front-impinger bottle and a rear-impinger bottle) and NH 3 is absorbed in 0.1 N H 2 SO 4 solution.NH 3 concentration is a summation of the levels in the front and rear impinger bottles.In this study, SKC pump samplers were operated at flow rate of 2 L/min for a duration of 1 hr, and then for another 1 hr duration after the replacement of two impinger bottles.Each hour we gathered two samples (by using two sampling instruments side by side) and averaged them to yield 1-h average concentrations for all runs.
The collected NH 3 samples were stored in airtight boxes at 0°C and later analyzed in our EP207 laboratory by the Indophenol-UV Method (Taiwan EPA, 1997).All the chemicals used in this study were of analytical reagent grade, and water purified with a Milli Q Labo (Millipore) was used for the preparation of all solutions.The reagents were added and the mixtures were allowed to react at room temperature (about 25°C) for 30 minutes, after which the absorbance at 630 nm of each sample was measured repeatedly using an UV spectrophotometer (Type: Hitachi 2100).The calculation of the actual ammonia concentrations in the samples was done using a calibration curve based on an ammonia standard solution.

Ambient Ammonia Concentrations in Different Industrial Parks
Fig. 2 presents the daily variation in both meteorological parameters records (Fig. 2(a)-2(d)) and the average ambient NH 3 concentrations (Fig. 2(e)) in Neipu Industrial Park in southern Taiwan.For the Neipu Industrial Park, the mean relative humidity level for the nighttime was approximately 1.5 times higher than the mean level for the daytime (Fig. 2(a)).Fig. 2(b) shows the mean wind speed level for the daytime was approximately 2.7 times higher than the mean level for the nighttime.Fig. 2(d) shows the mean ambient temperature level for the daytime was approximately 1.3 times higher than the mean level for the nighttime.For increased accuracy, we use the value of measured sunlight strength (Fig. 2(e)) to distinguish between the day and night.If the measured sunlight strength is more than zero, then the sampling period is part of the daytime, and otherwise is part of the nighttime.Therefore, from the point of view of the whole day, the mean NH 3 levels ranged from 70.5 to 153.9 ppb (mean = 100.2ppb) in Neipu Industrial Park (Fig. 2(e)).Fig. 2 shows that the mean NH 3 levels ranged from 72.0 to 115.6 ppb (mean = 90.4ppb) and from 70.5 to 153.9 ppb (mean = 114.9ppb) in the day and at night, respectively.For all the sampling duration, the highest mean level of 153.9 ppb was obtained from 20:00 to 21:00.Regarding the maximum level, the highest mean levels of 115.6 and 153.9 ppb were obtained from 12:00 to 13:00 and 20:00 to 21:00, respectively.Both these highest mean levels were about 50-150 times greater than the ambient background levels (1-3 ppb) measured in several previous reports (Sickles et al., 1990;Aneja et al., 1998;Leaderer et al., 1999;Lefer et al., 1999;McCurdy et al., 1999;Walker et al, 2004;Wilson andSerre, 2007a, 2007b).
Fig. 4 shows that the mean NH 3 levels for the whole day ranged from 45.0 to 122.6 ppb (mean = 84.9ppb) in Pingnan Industrial Park.For the duration of sampling, the highest mean level of 122.6 ppb was obtained from 12:00 to 13:00.Moreover, the mean NH 3 levels ranged from 73.0 to 122.6 ppb (mean = 105.2ppb) and from 45.0 to 79.0 ppb (mean = 58.0ppb) in the daytime and at night, respectively.For the daytime and night, the highest mean levels of 122.6 and 79.0 ppb were obtained from 12:00 to 13:00 and 18:00 to 19:00, respectively.Like the case in Pingtung, these two highest mean levels in day and night were about 25-120 times greater than ambient background levels (1-3 ppb) measured in several previous reports (Sickles et al., 1990;Aneja et al., 1998;Leaderer et al., 1999;Lefer et al., 1999;McCurdy et al., 1999;Walker et al, 2004;Wilson andSerre, 2007a, 2007b).To discern the difference between day and night, the mean daytime NH 3 concentration was divided by the mean nighttime NH 3 concentration.The mean daytime to night ratios were 1.8, 1.8, 4.2 and 2.6 in Pingtung, Pingnan, NPUST dormitory, and the bamboo grove, respectively.However, the ratio of 0.79 in Neipu Industrial Park is significantly different to the others being the only one that is under 1, which indicates that there is either an increased emission under the same sources as daytime, or there is a new emission source only at night.At this stage of the investigation, one explanation is NH 3 sampling occurred in the Neipu area with multiple sources, and, while the nearest source does have an impact, we cannot ignore the overall elevation of ammonia in the sampling area due to the contribution of the other sources.

Ambient Ammonia Concentrations in Background Levels
Fig. 5 presents the daily variation in both meteorological parameter records and the average ambient NH 3 concentrations near Ren-Jhai Dormitory on the campus of National Pingtung University of Science and Technology (NPUST) in Neipu Township, in a location that is generally classified as a rural area.The sampling site near the NPUST dormitory is representative of the ambient level of a residential area without industrial sources nearby.Therefore, the NPUST dormitory was selected as the location #1 for the background site without industrial sources.Indeed, the only possible emission of NH 3 is from the transoms of toilets.
For the whole day, the mean NH 3 levels ranged from 13.0 to 101.7 ppb (mean = 52.2ppb) in the ambient air of NPUST dormitory (background site #1).For the duration of sampling, the highest mean level of 101.7 ppb was obtained from 11:00 to 12:00.Moreover, Fig. 5 also shows that the mean NH 3 levels ranged from 17.5 to 101.7 ppb (mean = 69.4ppb) and from 13.0 to 25.1 ppb (mean = 16.6 ppb) in the daytime and at night, respectively.For the day and night, the highest mean levels of 101.7 and 25.1 ppb were obtained from 11:00 to 12:00 and 19:00 to 20:00, respectively.These two highest mean levels in day and night were 25-100 times greater than the ambient background levels (1-3 ppb) measured in several previous reports (Sickles et al., 1990;Aneja et al., 1998;Leaderer et al., 1999;Lefer et al., 1999;McCurdy et al., 1999;Walker et al, 2004;Wilson andSerre, 2007a, 2007b).Fig. 6 presents the daily variation in both meteorological parameters records and the average ambient NH 3 concentrations in a bamboo grove (the background site #2) in southern Taiwan.The NH 3 levels measured in this site compare well with those in other studies on the levels of atmospheric NH 3 at or near hog concentrated animal feeding operations in eastern North Carolina.The mean values obtained for background site #2 in this study were 4.6 ppb (daylong), 5.8 ppb (daytime) and 2.2 ppb (night).These values are comparable to those obtained by Langford et al. (1992), Buijsman et al. (1998) and Wilson and Serre (2007a).When NH 3 values from our study are compared to the mean levels from other studies, it can be seen that the data corresponds well to these earlier measurements.In addition, the data for background site #2 are consistent with what is expected for low ambient NH 3 emission or agricultural sites (Buijsman et al., 1998).The comparison between the data from our work and previous studies also indicates that background site #2 can serve as a reasonable reference to judge the elevated ambient NH 3 measured in the Pingtung area.
The results show that NH 3 concentrations were higher at night in Neipu industrial park, during the day in Pingtung and Pingnan industrial parks, and the day vs. night concentrations were similar in the bamboo grove background site.Table 1 lists the ammonia concentration correlation coefficients (R) during daytime, night time and daylong sampling periods at the five sampling sites.During the daylong periods, significant correlations were found between Pingtung and Pingnan industrial parks and between Pingtung and NPUST dormitory.During the daytime sampling periods, no significant correlations were found among the five sites.However, during night time periods, a significant correlation was found between the Pingtung industrial park and NPUST dormitory.These results can be seen more clearly in Fig. 7.

Ammonia Levels vs. Meteorological Parameters
Table 2 gives the correlations matrix for the NH 3 levels versus their corresponding meteorological values at the five sampling locations during the daylong sampling periods.In addition, Fig. 8 presents the visual interaction   correlations.The information contained in both Table 2 and Fig. 8 is summarized, as follows.
(i) A significant correlation between the NH 3 level and relative humidity was observed in Neipu Industrial Park (Table 2 and Fig. 8(a)).Moreover, high positive correlations between measured ambient temperature, wind speed, and sunlight strength were also observed, although there was a high negative correlation between the ambient temperature level and relative humidity (Table 2 and Fig. 8(a)).
Significant positive correlations between the NH 3 level and ambient temperature, wind speed, and sunlight strength were also observed in Pingtung Industrial Park (Table 2 and Fig. 8(b)), and it is likely that these have significant theoretical importance (for example, in relation to the dispersion of NH 3 ).Furthermore, high positive correlations between the measured ambient temperature level and wind speed and sunlight strength were also observed.These trends are fairly consistent across a wide variety of meteorological parameters.In contrast, high negative correlations between the ambient temperature level and relative humidity and atmospheric pressure were observed (Table 2 and Fig. 8(b)).(ii) Significant positive correlations between the NH 3 level and ambient temperature, wind speed, and sunlight strength were also observed in Pingnan Industrial Park (Table 2 and Fig. 8(c)).Moreover, a significant negative correlation between the NH 3 level and atmospheric pressure was also observed.In contrast, the results showed a high negative correlation between the ambient temperature level and atmospheric pressure.
(iii) Significant positive correlations between the NH 3 level and ambient temperature, wind speed, and sunlight strength were found at NPUST dormitory (Table 2 and Fig. 8(d)).In contrast, the results showed a high negative correlation between the ambient temperature level and relative humidity.(iv) In the bamboo grove, no significant correlation between the NH 3 level and others was found (  and wind speed, and sunlight strength were observed at this site, while a high negative correlation between measured ambient temperature level and relative humidity was also found (Table 2 and Fig. 8(e)).(v) These different correlations highlight the need for further research to investigate many of the ambient meteorological factors, and especially variations between different sampling sites.

Characteristics of Ratio Values Normalized by the Background Levels
It is difficult for local governments to address illegal agricultural practices and the related nighttime emissions in southern Taiwan.Specifically, the facts that there are many unregistered farmyards in this region make it impossible to accurately estimate the emission of NH 3 or odors (such as H 2 S) from such sources.Therefore, we use the characteristics of ratio value based on the standard mathematical normalization process to obtain better information.In this study, ratio values were categorized according to the cover-range (from minimum ratios to maximum ratios.)at different sampling periods and the minimum separated boundary lines.Fig. 8 shows that the characteristics of the ratio values normalized by the background levels can be divided into three main categories: (1) R > 200; (2) R > 60; (3) R > 10.
Fig. 9 shows that the ratio values were highest (R > 200) from both 20:00 to 21:00 and 21:0 to 22:00 in Neipu Industrial Park.The difference at different sampling locations in Fig. 9 is most likely due to the different ammonia sources, such as industrial processes, agriculture, and livestock.The value of R was determined mostly by the effects of the multiple sources at each sampling location, indicating that different distribution zones might authors would also like to thank the National Science Council of the Republic of China, Taiwan, for partly financially supporting this research under Contract No. NSC 93-2211-E-020-004 and NSC 94-2211-E-020-005.

Fig. 2 .
Fig. 2. Daily variation in both meteorological parameter records (from (a) to (d)) and average ambient ammonia concentrations (e) in Neipu Industrial Park in southern Taiwan.

Fig. 3 .
Fig. 3. Daily variation in both meteorological parameter records (from (a) to (d)) and average ambient ammonia concentrations (e) in Pingtung Industrial Park in southern Taiwan.

Fig. 4 .
Fig. 4. Daily variation in both meteorological parameter records (from (a) to (d)) and average ambient ammonia concentrations (e) in Pingnan Industrial Park in southern Taiwan.

Fig. 5 .
Fig. 5. Daily variation in both meteorological parameter records (from (a) to (d)) and average ambient ammonia concentrations (e) in NPUST dormitory in southern Taiwan.

Fig. 6 .
Fig. 6.Daily variation in both meteorological parameter records (from (a) to (d)) and average ambient ammonia concentrations (e) in the bamboo grove in southern Taiwan.

Fig. 7 .
Fig. 7. Correlations in the five different sampling sites during the daylong (a), daytime (b), and nighttime (c) periods in southern Taiwan.

Fig. 8 .
Fig. 8.The visual interaction correlations of the different measured parameters in the five different sampling sites in southern Taiwan.

Table 1 .
Summary of ammonia concentration correlation coefficients (R) during daytime, nighttime and daylong sampling periods in the five sites

Table 2 .
Correlation matrix for NH 3 levels vs. the meteorological values in the five sampling sites.