Chemical Composition of Water Soluble Inorganic Species in Precipitation at Shihwa Basin, Korea

Weekly rain samples were collected in coastal areas of the Shihwa Basin (Korea)


Introduction
Rain is the most effective process transporting soluble gases and particles from the atmosphere to the ground [1,2].Precipitation chemistry plays an important role in understanding the air quality in a study area, because the concentrations and distribution of chemical components in rain depend on a variety of emission sources including sea spray, soil particles, and industrial pollutants [3][4][5][6][7].
In coastal areas, the Na + /Cl − ratio, SO4 2− , Mg 2+ , and Ca 2+ are useful for evaluating both anthropogenic and natural influences [8].Soil particles in northern China, Africa, and the Middle East play an important role in neutralizing the acidic components of rain [9].The concentrations of anthropogenic components are related to energy consumption and the use of fertilizers [10,11].
In many regions of Asia, rapid economic growth has led to an increase in pollutant emissions and seriously compromised air quality [21].In addition, air quality has also been affected by wind-blown soil particles originating from the deserts of Mongolia and northern China [22][23][24].Several studies conducted in this region have found that the rain is acidic, with H + concentration ranges similar to that reported for Europe [25][26][27][28].
In Korea, previous studies have shown that most precipitation events are acidic and even in background areas of Korea precipitation pH is <4.9 [29][30][31].Some studies have suggested that air masses from the Asian continent increase the concentration of anthropogenic components of rainwater in Korea [31,32].Choi et al. (2015) [33] studied precipitation chemistry in the area near Shihwa Lake (located in the mid-western area of the Korean peninsula), using both urban and background stations to compare precipitation chemistry.To date, precipitation chemistry in the industrial area has not been studied yet.
This study aimed to provide a detailed evaluation of the chemical composition of precipitation in the Shihwa Basin and to evaluate the relative contributions of various sources.Because the Shihwa Basin includes areas with industrial, agricultural, and coastal land use, the study are has experienced serious problems with air pollution [34][35][36][37].

Sampling Sites
The Shihwa Basin is located 60 km southwest of Seoul, the capital of South Korea.The basin of Shihwa Lake has an area of about 475 km 2 and a population of more than 0.8 million people.The Shihwa and Banwol industrial complexes are located in Shihwa Basin.Three sampling stations were selected for precipitation monitoring, covering the southern (Hwasung), northern (Banwol), and western (Daeboo) areas of the Shihwa Basin (Table 1).Hwasung station is located inland about 20 km east of Shihwa Lake, in a residential and rural area.Banwol station is located 5 km east of Shihwa Lake and is important because the Banwol industrial complex is located there.This complex is a site for intensive chemical, leather, and metal processing industries, which have led to air pollution episodes, including pollutants such as ammonia, amines, hydrogen sulfide (H2S) and mercaptans.Daeboo station is located on an island, about 2 km west of Shihwa Lake.The locations of the Shihwa Basin and the three sampling stations are shown in Figure 1.The Shihwa Basin is strongly influenced by air masses driven by westerly winds.According to data from the Incheon meteorological administration (about 20 km north of Shihwa Lake), the average annual rainfall from June 2000 to November 2007 was 1187 mm•year −1 .The average temperature and wind speed were 15 °C and 2 m•s −1 , respectively.In the summer (June to August), the mean temperature was 24.5 °C and about 61% of the annual precipitation occurred during this season.In winter, the average temperature was 3.5 °C with only 5% of the annual precipitation occurring during winter.The spring season is characterized by the influence of Asian dust, with an average wind speed of 3.67 m•s −1 , 58% higher than that in winter.According to AWS data from Korea Meteorological Administration near the three sites, the prevailing wind directions were north and south at Hwasung station, east and west at Banwol station and northwest at Daeboo station.At Hwasung, the wind during spring and summer was north, but south during fall and winter.At Banwol and Daeboo stations, the prevailing wind was constant throughout the year.

Sample Collection
Precipitation samples were collected weekly using wet and dry collectors from June 2000 to November 2007.Butler and Likens (1998) [38] found that daily samples may provide a better estimate of actual rain chemistry than weekly samples.Gilliland et al. (2002) [39] also suggested that considerable monthly variation existed between daily and weekly data for H + , NH4 + and SO4 2− .All of the samplers were placed 1 m from the ground, and the wet sample collector was covered with a lid to prevent any contamination from dry deposition.The wet collector uses a conductivity sensor, which automatically opens at the onset of each precipitation event and closes again when a rain event stops.The automated samplers consisted of two Teflon buckets 28 cm in diameter.The average precipitation intensity for onset was 40 ± 59 mm/week during the study period.Before sampling, the buckets were carefully rinsed with deionized water several times until the conductivity of the water was <1.5 µS•cm −1 .Precipitation samples were collected in a 30-mL high-density polyethylene bottle.The samples were unrefrigerated during sampling but stored in a freezer at −20 °C until chemical analysis.Prior to analysis, the precipitation samples were filtered using 0.45-µm membrane filters (Millipore).

Sample Analysis
After the precipitation samples were collected, the weight of the bucket was measured to calculate the volume-weighted mean (VWM) concentrations of the atmospheric components.The pH and conductivity of the unfiltered precipitation were also measured immediately with a pH and conductivity meters (Fisher Scientific) [40].The pH meter was calibrated before each measurement using standard 4.00 and 7.01 buffer solutions.The conductivity meter was also calibrated with a standard solution.For this study, 528 rain samples were analyzed.Major anion and cation concentrations were determined using ion chromatography (Waters, USA).For anions (Cl − , NO3 − , and SO4 2− ), AS14 or AS11 columns were used; CS12A or CS14 columns were used to measure cations (Na + , NH4 + , K + , Mg 2+ , and Ca 2+ ).The detection limits were 0.08 µeq•L −1 for Cl − , 0.12 µeq•L −1 for NO3 − , 0.11 µeq•L −1 for SO4 2− , 0.54 for Na + , 2.32 µeq•L −1 for NH4 + , 0.2 µeq•L −1 for K + , 0.44 µeq•L −1 for Mg 2+ and 2.87 µeq•L −1 for Ca 2+ .Detection limits were calculated by dividing the standard deviation of the response by the slope of the calibration and then multiplying by 3.3.

Quality Assurance
Of the 678 precipitation samples collected, about 7% were discarded because the amount of precipitation was insufficient to perform the chemical analyses.An additional 12% of the rain samples were discarded due to noticeable contamination by dry deposition (soil, leafs and insects) or because of sampler malfunction [41].The remaining samples were subjected to a quality check based on ionic balance and conductivity balance.When the pH was above 5.6, the concentration of HCO3 − (in µeq•L −1 ) was calculated using the formula (HCO3 − ) = 10 (pH − 11.24) [42][43][44].According to Okuda et al. (2005) [15], the acceptable ion range (∑ cation/∑ anion) and conductivity (∑ measured conductivity/∑ anion conductivity) ratio for a rain sample is 0.67-1.5.Data points outside this range were excluded.The percentage of samples excluded at each sampling station was 38% at Hwasung, 21% at Banwol, and 32% at Daeboo, averaging 30% overall.Linear regression of the relationship between cation sum vs. the anion sum showed an r 2 = 0.95 and slope of 0.95 (Figure 2a,b).Edmonds et al. (1991) [45] suggested that anion deficit means of the amount of unmeasured organic acids.Kim et al. (2013) [32] found that the organic acids (CH3COO − , HCOO − ) contributed to 12.4% of the acidity in precipitation at Jeju Island in Korea.The relationship between the calculated and measured conductivity was also highly correlated with a r 2 = 0.96 and slope of 0.92.

Positive Matrix Factorization (PMF)
Positive Matrix Factorization is a multivariable factor analysis tool to identify the contributions of various emission sources.PMF decomposes a speciated data matrix X of n by m dimensions (n number of samples and m chemical species) into factor profiles (g) and factor contributions (f) based on the correlation between the different components (Equation ( 1)).The objective of the PMF solution is to minimize the object function Q based on the uncertainties (u) as follows [46]: where xij are the measured concentrations (in µeq•L −1 ), uij are the estimated uncertainty values (in µeq•L −1 ), n is the number of samples, m is the number of species and p is the number of factors included in the analysis [47].PMF needs two input files: one for the measured concentrations of the species and the other for the estimated uncertainties of the concentrations.The uncertainty greater than the detection limit was calculated using the concentration and detection limit (Equation ( 2)).If the concentration was less than or equal to the detection limit, the uncertainty was calculated with a fixed fraction (Equation ( 3)).Missing data and their uncertainty value were replaced with the value of the species-species median and four times the median value, respectively [48].
= (Error fraction × Concentration) + (0.5 × Detection limit) (2) A variable was defined to be weak if its S/N was between 0.2 and 2.0.Most of the scaled residuals were between −3.0 and 3.0.Fpeak values between −1 and 1 (in steps of 0.1) were examined to find out the most appropriate solution.Error estimate with bootstrap results showed 100% mapping for five factors at three sites.

Acidity
SO2 and NOx are the major precursors of acidity in precipitation.Wind-blown soil particles and sea spray species atmosphere play an important role in neutralizing acidic constituents.In uncontaminated precipitation, the equilibrium concentration of CO2(aq) generates a pH of approximately 5.6, which serves as a reference value.Figure 3 shows the frequency distribution for pH in the Shihwa Basin.Among the samples, 84% of the pH values were <5.6.The pH of uncontaminated precipitation is considered to be 5.0-5.6;this acidity originates from natural levels of atmospheric CO2, NOx, and SOx [49].About 66% of the rain samples had a pH < 5.0, indicating that the rain samples were affected by additional acidic components.At Banwol station in the center of two industrial complexes, 69% of the samples had a pH of <5.0 and 14% had a pH < 4.0.At Hwasung station, 59% of the samples had a pH of <5.0 and 7% had a pH < 4.0.At Daeboo, 78% of the samples had a pH lower than 5.0 and 15% of the samples had a pH < 4.0.Overall, the pH of the precipitation ranged from 3.4 to 7.7 (Table 2).The VWM pH was 4.84, indicating slight acidity in the study area.For Hwasung, located east of Banwol, the VWM pH was 4.96, higher than at the other sites (p < 0.001).At Banwol, the pH was 4.77, lowest among all the stations (p < 0.001), reflecting its proximity in the industrial complex.Daeboo had the highest VWM pH (4.81), similar to the VWM pH at Banwol.

Concentration of Inorganic Species
The minimum and maximum concentrations of chemical components in the individual precipitation samples and their VWMs are shown in Table 2. SO4 2− and NH4 + were the most abundant ions in precipitation at the three stations.At Hwasung, the concentrations of anthropogenic species were lower than those at the other stations.The VWM concentration of sea salts was also lowest in Hwasung, as it is located farthest from the seashore.The highest concentrations of SO4 2− , NH4 + and NO3 − were measured at Banwol, related to anthropogenic sources associated with two industrial complexes and motor vehicles.The VWM concentration of H + was also highest in Banwol.In Daeboo, the precipitation acidity was similar to that in Banwol due to the low concentration of NH4 + .The VWM conductivities in Hwasung, Banwol, and Daeboo were 18, 24, and 23 µS•cm −1 with VWM total ions of 181, 253, and 235 µeq•L −1 , respectively.
For the combined data collected in the Shihwa Basin, the ion concentrations fell in the order SO4 2− > NH4 + > H + > NO3 − > Cl − > Ca 2+ > Na + > Mg 2+ > K + .Among the anions, SO4 2− had the highest concentrations at all stations (54%).The next most abundant anion was NO3 − at nearly 24%.The precipitation chemistry in the Shihwa Basin was mainly dominated by NH4 + with an average VWM of µeq•L −1 , representing 40% of all cations.Considering the sampling period, the concentration of NH4 + can be as high as 49.5 µeq•L −1 (10%).The next most abundant cations were H + , Ca 2+ , and Na + , representing 24%, 15%, and 12% of the total, respectively.NH4 + in the Shihwa Basin is associated with agricultural activities and an industrial wastewater treatment facility.Ca 2+ is related to resuspension of dust from the soil and Asian dust originating from China mainly in the spring.[50].These alkali components neutralize and decrease the acidity of the rainwater [51,52].The ratios of individual ions to Na + are higher than seawater ratios, which indicates that contributions from anthropogenic and soil sources are important, while the marine contribution is negligible [53,54].The Cl − /Na + ratio in the combined precipitation samples over six years was 45% higher than that of seawater Cl − /Na + (1.165).This suggests that the Cl − originated from industrial activities in the area such as coal combustion and incineration of polyvinyl chloride [55,56].The SO4 2− /Na + , K + /Na + , Mg 2 /Na + , and Ca 2+ /Na + ratios for precipitation in the Shihwa Basin were 4.02, 0.32, 0.45, and 1.25, respectively, higher than the seawater ratios (0.12, 0.23, 0.04, and 0.02, respectively).This suggests that anthropogenic influences were present from local (two industrial complexes) or regional (Asian continent) sources.

Temporal Variation
During the study period, there were 26 episodic events in which the total ion concentration precipitation was >2000 µeq•L −1 .There were 14 events in winter, 11 in spring, and one in autumn.In summer, there were no events when the total ion concentration was >2000 µeq•L −1 .Rastogi and Sarin (2007) [67] reported that low-solute events are associated with heavy amounts of rain or successive events.The concentrations of most of the chemical components in the wet-only samples were inversely related to precipitation amount.About 74% of the rain in the Shihwa Basin fell from June to September (Figure 4).Concentrations during the rainy period (June to August) were significantly lower (p < 0.05 for H + and p < 0.01 for other ions) than those in the dry period (September to May).The concentrations of the inorganic species decreased due to dilution during the rainy period, thus ion concentrations were lower during the rainy period than ion concentrations in the dry period.The concentrations of soil-derived species (Ca 2+ and Mg 2+ ) in the dry season were about six times higher than those in the rainy season.In March, influenced by dust from Asian continental deserts such as the Taklamakan, Gobi, and Loess plateau, the concentrations of all chemical components were highest.The monthly VWM concentrations of SO4 2− and NO3 − in the wet-only samples were correlated with concentrations of SO2 (r = 0.466, p < 0.001) and NO2 (r = 0.559, p < 0.001) in atmosphere (Figure 5).The NO2 and SO2 data are available from the National Institute of Environmental Research (NIER) in the same administrative district (Ansan city).Tu et al. (2005) [19] also reported that the SO4 2− concentration had a temporal trend corresponding to SO2 in atmosphere in China.Similar results have been found in a number of studies [15,68].Figure 6 shows the annual VWM concentrations of the major components in precipitation.Anthropogenic SO2 and NO2 emitted from mobile and power plant sources contributed to the increasing of SO4 2− and NO3 − in the wet-only samples [69].In 2002, power plants (450 MW) near the study area expanded to Boryeong thermal power plant located about 100 km south of Daebudo station.Thermal power facilities (1600 MW) were also added to the Yeungheung thermal power plant located 13 km west from Daebudo station in 2004.In 2005, the Incheon thermal power plants located 20 km north of the Banwol industrial complex expanded to a 342 MW power plant facility.Additional power generating capacity (5275 MW) has been added to these three power plants between 2008 and 2014.During the study period, the NH4 + /SO4 2− ratio decreased from 1.01 to 0.68.The NH4 + /NO3 − ratio also decreased from 2.39 to 1.20.Results from studies of wet-only precipitation in Austria, Brazil, Canada, USA, Japan, and China have shown significant decreasing trends in the concentrations of SO4 2− and H + over long-term periods [15,16,[70][71][72].Table 3 shows VWM concentrations in recent precipitation studies in East Asia.The concentrations of water-soluble components in the Shihwa Basin were lower than those found in other Asian continental countries.The average concentrations of NO3 − and SO4 2− ranged from 1.8 to 2.0 and were 2.0 to 2.4 times higher than average concentrations in Seoul [33].

Statistical Analysis
Table 4 shows the correlation coefficients among chemical species for this study.Significant and strong correlations were found among Na + , Cl − , Mg 2+ and Ca 2+ , indicating that this region is heavily affected by sea salts and soil.Ca 2+ and Mg 2+ correlations with Cl − , NO3 − , and SO4 2− indicate that CaCl2, MgCl2, Ca(NO3)2, Mg(NO3)2, CaSO4, and MgSO4 were the primary soil-derived components in the wet-only precipitation [73].The correlation among NH4 + , NO3 − , and SO4 2− indicate that they originated from similar anthropogenic sources.The significant correlation between NH4 + and NO3 − (r = 0.76, p < 0.001) and between NH4 + and SO4 2− (r = 0.83, p < 0.001) indicate that NH4NO3, (NH4)HSO4, and (NH4)2SO4 were the major forms of NH4 + in precipitation.Although positive correlations between acidic anions and H + were observed, their correlations were weaker due to neutralization effect by basic soils components [74].Additional details on the patterns in the chemical constituents were analyzed using PMF.Similar studies have also used this technique to identify sources of air pollutants.Anttila et al. (1995) [75] determined the sources of bulk wet deposition in Finland using PMF analysis.Recently, Kitayama et al. (2010) [76] used PMF to identify the sources of SO2 in Japan.The number of factors was determined by the optimized values.Five factors were chosen (Figure 7).The factors were identified as sea salts (Na + , ss-Cl − ), Asian dust (Mg 2+ , Ca 2+ ), sulfuric acid (H + , SO4 2− ), ammonium salt with nitrate (NH4NO3), and ammonium salt with sulfate (NH4HSO4 and (NH4)2SO4)components.The first factor was characterized by high concentrations of sea salts and its contribution to ions in precipitation was 18%-26%.At the Banwol site, some anthropogenic components (NO3 − , NH4 + , and SO4 2− ) were also included in this factor.The second factor, accounting for 17% of the total ions, was dominated by Asian dust at all three sites, indicated by high concentrations of Mg 2+ and Ca 2+ .This factor explained 70%-79% of the variation in Ca 2+ and 42%-74% of that in Mg 2+ .The average contribution of the Asian dust factor was especially high during the spring.The third factor represents H + and SO4 2− , accounting for 16%-23% of the total.The acidity of the precipitation is largely affected by the dissolution of SO2 and NO2 as precursors of the acidic components.As noted above, the high SO4 2− /NO3 − ratio in the Shihwa basin indicates that emissions of SO2 from the industrial complex were dominant sources for acidity.SO4 2− accounted for 93%, 52%, and 57% of this acidity factor at Hwasung, Banwol, and Daeboo, respectively.The fourth factor was characterized by ammonium nitrate, which explained 17%-21% of the total and originated from either local industrial or farmland sources.The presence of nitrate with ammonium ions indicates dissolutions of ammonium nitrate from aerosols.The fifth factor represents high loadings of NH4 + and SO4 2− .The highest NH4 + concentrations were found in Hwasung where sizeable NH3 sources, such as farmland, exist.The relative contributions of five factors to measured components varied largely by the month of year (Figure 8).The acidity and secondary aerosol factors were mainly associated with the temperature and precipitation amount.Their contribution was highest in summer.However, Asian dust influences were dominant during the spring.

Figure 1 .
Figure 1.Locations of the study area and sampling stations.

Figure 2 .
Figure 2. Linear regression (a) between the sum of cations and anions, and (b) between measured and calculated conductivity.

Figure 4 .
Figure 4. Monthly VWM concentrations of ions in rainfall and rainfall amount.

Figure 5 .
Figure 5. Trends with Lowess smoothing (span = 0.2): (a) monthly VWM SO4 2− concentration in wet-only precipitation and the monthly mean atmospheric SO2 concentration and (b) monthly VWM NO3 − concentration in wet-only precipitation and the monthly mean atmospheric NO2 concentration.

Figure 6 .
Figure 6.Annual variations of major inorganic VWM concentrations in the precipitation.

Figure 7 .
Figure 7. Source contributions of factors 1-5 at the three stations.

Figure 8 .
Figure 8. Temporal variations in the five source contributions.

Table 1 .
Characteristics of sampling stations in the Shihwa Basin.

Table 3 .
Volume-weighted mean (VWM) concentrations of the chemical components in wet-only precipitation in East Asia (µeq•L −1 except pH).

Table 4 .
Correlations between ionic components in wet-only precipitation over seven years in the Shihwa Basin (p < 0.001 for all correlation).