ARTICLE pubs.acs.org/est
Identifying the Causes of Sediment-Associated Toxicity in Urban Waterways of the Pearl River Delta, China W. Tyler Mehler,†,‡ Huizhen Li,†,§ Michael J. Lydy,‡ and Jing You*,† †
State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China ‡ Fisheries and Illinois Aquaculture Center & Department of Zoology, 171 Life Science II, Southern Illinois University, Carbondale, Illinois 62901, United States § Graduate School of the Chinese Academy of Sciences, Beijing 100049, China
bS Supporting Information ABSTRACT: Twenty-one sediments collected in urban waterways of the Pearl River Delta (PRD) in China were screened for acute toxicity using Chironomus dilutus in addition to being examined for potential contributors of sediment toxicity, including 19 organochlorine, five organophosphate, and nine pyrethroid insecticides, 28 polycyclic aromatic hydrocarbons, 27 polychlorinated biphenyls, 15 polybrominated diphenyl ethers, 12 metals, and ammonia. Fifteen of the 21 sediments exhibited acute toxicity to C. dilutus, with 33% of the samples exhibiting 100% mortality. This is one of the first studies in China which directly correlates a broad range of sediment-associated contaminants to ecological effects measured by bioassays. A toxic unit approach showed that pyrethroids contributed most to the observed toxicity; as cypermethrin alone was predicted to cause significant mortality in about half of the sites. Specific toxicity identification evaluation analysis confirmed pyrethroid toxicity. Other contaminants may also be supplemental contributors at a few sites. The current study suggests that pyrethroids are the principal cause of contamination and that future risk assessment and mitigation efforts in this area should focus primarily on pyrethroids but should not disregard other contaminants as potential risk is evident.
’ INTRODUCTION The Pearl River Delta (PRD) is one of the most populous and important economic regions in China. Rapid urbanization has led to the deterioration of aquatic systems and these trends are not expected to decrease in the near future either, as recent work has predicted that from 2002 to 2020, a 3-fold increase in the volume of degraded water resources may occur in this area.1 Studies have suggested that a number of contaminants have contributed to this degradation, including organics,2-4 metals,5 and ammonia.1,5 The majority of studies investigating sediment-associated contamination in the PRD have assessed risk solely based on chemical data, with little information reported regarding biological effects. While these studies provide insight into the occurrence and distribution of sediment-associated contaminants in the PRD, they provide limited estimates in defining the overall risk to aquatic organisms. Furthermore, previous studies on sediment-associated contaminants in the PRD were mainly focused on persistent organic pollutants including organochlorine pesticides (OCPs), polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), and polybrominated diphenyl ethers (PBDEs), and metals.2-5 No study has examined the contribution of currently used insecticides, such as organophosphate (OP) and pyrethroid insecticides, on sediment toxicity in China although they are extensively used and are extremely toxic to aquatic species. Previous studies in the United States (US) showed that pyrethroids were a major cause of sediment toxicity in California,6-9 Texas,10 and Illinois.11 The primary objective of the current study was to evaluate the risk of sediment-associated contaminants in urban waterways in r 2011 American Chemical Society
the three largest cities in the PRD: Guangzhou, Shenzhen, and Dongguan. Acute toxicity testing with the benthic invertebrate Chironomus dilutus coupled with chemical measures of various classes of organics, metals, and ammonia allowed for risk characterization using a weight of evidence approach. Additionally, a toxic unit (TU) approach was used to identify the major contributors to sediment toxicity with confirmation using specific toxicity identification evaluation (TIE) analysis. The current study is intended to provide a framework for future ecological risk assessments in the PRD by defining contaminants of concern and to construct an assessment technique that links chemical- and effect-based analyses.
’ MATERIALS AND METHODS Sampling Locations. Twenty-one surface sediments were collected in urban waterways in the PRD. As mapped in Figure S1 in the Supporting Information (SI), nine, six, and six samples were collected in Guangzhou, Dongguan, and Shenzhen, respectively. Descriptions of the sampling sites and the sampling procedures were detailed in Li et al.12 Soil collected from Baiyu Mountain in Guangzhou was hydrated and used as control sediment (BM), and BM sediment exhibited no toxicity and no target analytes were detected.13 Additionally, three sediments were sampled in an electronic waste (e-waste) recycling site in Qingyuan and used Received: October 28, 2010 Accepted: January 9, 2011 Revised: December 17, 2010 Published: February 03, 2011 1812
dx.doi.org/10.1021/es103552d | Environ. Sci. Technol. 2011, 45, 1812–1819
Environmental Science & Technology to compare the contributions of contaminants originated from e-waste and urban runoff to sediment toxicity. Toxicity Testing. Chironomus dilutus, a benthic invertebrate, was the test organism in the current study, as Chironomus species are native in the PRD. Midges were cultured following US Environmental Protection Agency (USEPA) protocols14 at the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Ten-day sediment toxicity bioassays were conducted using five replicates.14 After being homogenized with a drill equipped with a stainless steel spatula for 2-h, 80 g of wet sediment was distributed into each 400-mL beaker, with 300-mL of moderately hard water as overlying water. After settling the sediment overnight, ten third-instar midge larvae were randomly added to each breaker. An automated water-delivery system was used to change approximately 150-mL of water twice daily, while bioassays were conducted with a 16:8 light:dark cycle at 23 ( l °C. Water parameters including temperature, conductivity, pH, and dissolved oxygen (DO) were monitored daily. At the conclusion of the test, C. dilutus mortality was assessed by sieving the sediment and organisms through a 500-μm sieve. If 100% mortality was evident; the sediment was diluted with control BM soil on a dry weight (dw) basis for a more accurate assessment of toxicity. Analytical Chemistry. Total organic carbon (TOC) was quantified for sediment using an Elementar Vario ELIII (Hanau, Germany) after removing the inorganic carbonates with 1 mol/L HCl. Moreover, sediments were analyzed for a suite of contaminants including nonpolar organics, ammonia, and metals. A complete list of target contaminants are shown in Table S1 in the SI, and their respective analytical methods are described briefly below. Six classes of nonpolar organics including 19 OCPs, five OPs, nine pyrethroids, 28 PAHs, 27 PCBs, and 15 PBDEs were analyzed in sediments following previously developed methods.2,3,13 Sediments were extracted using a CW-2000 ultrasound-assisted microwave extractor (Xintuo Company, Shanghai, China), cleaned with solid phase extraction cartridges or columns, and analyzed using a Shimadzu QP-2010-plus gas chromatograph/mass spectrometer (GC/MS, Shimadzu, Japan). Procedures for sample preparation, detector settings, and QA/QC parameters are detailed in the SI. Furthermore, 12 metals were also assessed on an Agilent 7700 inductively coupled plasma/MS (Santa Clara, CA, USA) after sediments were digested with HNO3 using a MARS microwave digestion system (CEM, Matthews, NC, USA) following USEPA method 3051.15 Ammonia concentrations in sediment porewater were also determined. Porewater was obtained by centrifuging sediment at 2000 g for 30-min, and total ammonia concentrations were measured on a UV-752 spectrophotometer (Jinhua Company, Shanghai, China) using salicylic acid as a color developing agent following a standard method.16 The respective analytical methods for metals and ammonia are detailed in the SI as well. Toxic Unit Evaluation. Two types of TUs were used in the current study to assess the contribution of each contaminant to the observed sediment toxicity. Observed TUs were derived from mean mortality for midges exposed to the sediment samples following eq 1 and reflected the bioassay results.
Observed TU ¼
Observed percent mortality for the midges 50
dilution factor ð1Þ
The dilution factor indicated how many times the tested sediments were diluted with control sediment, with a value of one used for undiluted sediments.
ARTICLE
It should be noted that a modified observed TU approach was used here. Typically, investigations using a TU approach are performed by using a series of dilutions for a site sediment to determine an observed TU, this was not conducted in this study for two reasons. First and foremost, the amount of samples that warranted further investigation prohibited extensive bioassay work due to time and cost restrictions. Additionally, as TOC in site sediments varied dramatically among sites, finding representative reference sediments (so not to alter bioavailability) for dilution purposes in the area was not feasible, as finding sediments with limited contamination in of itself was difficult. While this does pose additional uncertainty in addressing toxicity of various compounds as predicted TUs are based on a concentration-response relationship, it provides a conservative estimate for evaluating the causes of toxicity. On the other hand, predicted TUs represented estimates of sediment toxicity based on measured chemical concentrations in sediments (eq 2). Concentration addition was used for mixture toxicity estimation, and the individual predicted TU was calculated by dividing the contaminant concentration in sediment by the lethal concentration 50 (LC50) values acquired from published literature for midges. It should be noted, however, that in some cases C. dilutus LC50 values were not available, and thus the best available data (sediment benchmarks) were used Predicted TU ¼
Contaminant concentration in sediment ðLC50 or sediment benchmarkÞ ð2Þ
The LC50 values for bifenthrin, lambda-cyhalothrin, permethrin, cypermethrin, esfenvalerate, p,p0 -DDE, p,p0 -DDD, p,p0 -DDT, R-endosulfan, β-endosulfan, endosulfan sulfate, and endrin ketone were from previous studies,6,17 while the LC50 for chlorpyrifos was from Harwood et al.18 Additionally using the extrapolation methods employed by Weston et al.,6 LC50 values for cyfluthrin, fenpropathrin, and deltamethrin were calculated. This method characterized relative sediment toxicity for a compound using criteria from water-only studies by Solomon et al.19 No LC50 values could be found for chlordane and hexachlorocyclohexane (BHC), thus probable effect concentrations (PEC) were used,20,21 and PEC values were normalized to 1% TOC. Meanwhile, it should be noted that R- and γ-chlordane values were summed together for TU analysis, and γ-BHC sediment PEC values were used to quantify β-BHC predicted TUs, as data for β-BHC were not available. Besides the pesticides, LC50 values for PCBs, PBDEs, and PAHs are limited. To evaluate the toxicity contributed from PCB and PBDE congeners and individual PAHs, narcotic modeling was used.22 These methods predict narcosis-related toxicity based on Log Kow values for a similar midge species, Tanytarsus dissimilis. Additionally, the sum PCB concentration was compared to consensus-based PECs for confirmation purposes,20,21 whereas PAHs were evaluated using two additional techniques, equilibrium sediment benchmarks (ESBs23), and threshold effect concentration (TEC) values on a sum concentration basis.24 The LC50 values for porewater-ammonia were obtained from values which take into consideration pH to evaluate toxicity of the un-ionized portion of ammonia to midges.25 Eight of the 12 metals (As, Cd, Cr, Cu, Pb, Hg, Ni, and Zn) were evaluated using PEC values20,21 as LC50 values were not easily obtainable. Moreover, metals were evaluated using effects-range median (ERM) values for coastal waters which included the previous eight metals 1813
dx.doi.org/10.1021/es103552d |Environ. Sci. Technol. 2011, 45, 1812–1819
Environmental Science & Technology
ARTICLE
Table 1. Observed Toxic Units (OTU) Measured by 10-Day Bioassay Using Chironomus dilutus and the Predicted Toxic Units (PTU) Calculated from Concentration of Detected Organochlorine (OCP), Organophosphate (OP), and Pyrethroid Insecticides in Sediments Collected in Guangzhou (GZ), Shenzhen (SZ), Dongguan (DG), and Qingyuan (QY) in the Pearl River Delta, Chinad,e sediment
OTU
Chl
BHC
Chlor
Bif
Cyh
Cyp
Del
Fen
P PTU-OCP
GZ1
0.23
0.05
BRL
0.01
BRL
0.04
0.19
BRL
BRL
0.01
GZ2
0.20
BRL
BRL
0.01
ND
BRL
0.05
BRL
BRL
0.01
GZ3
1.10
BRL
BRL
BRL
BRL
0.14
0.85
BRL
BRL
GZ4
0.30
BRL
BRL
0.2
BRL
BRL
0.57
BRL
GZ5
0.43
BRL
BRL
BRL
ND
BRL
0.73
GZ6
1.20
0.92
BRL
0.24
BRL
BRL
GZ7 GZ8
7.60 1.87
0.21 BRL
BRL 1.52
BRL BRL
BRL BRL
GZ9
1.80
0.35
BRL
BRL
SZ1
8.00
BRL
BRL
0.03
SZ2
8.00
0.03
BRL
SZ3
8.00
0.2
SZ4
8.00
SZ5 SZ6 DG1
P PTU- pyre
P PTU-pest
0.050
0.23
0.29
0
0.05
0.06
0
0
0.99
0.99
BRL
0.20
0
0.57
0.77
BRL
BRL
0
0
0.73
0.73
1.09
BRL
BRL
0.24
0.92
1.09
2.25
BRL BRL
1.02 3.42
BRL BRL
BRL BRL
0 0
0.21 1.52
1.02 3.42
1.23 4.94
ND
BRL
0.26
BRL
BRL
0
0.35
0.26
0.61
0.04
0.06
1.43
BRL
BRL
0.03
0
1.53
1.56
0.02
0.08
BRL
1.24
0.18
BRL
0.02
0.03
1.50
1.55
BRL
BRL
0.53
0.18
1.48
0.46
BRL
0
0.20
2.65
2.85
0.06
BRL
0.1
0.11
0.06
2.22
BRL
BRL
0.10
0.06
2.39
2.55
0.30
BRL
BRL
BRL
BRL
BRL
0.09
BRL
BRL
0
0
0.09
0.09
8.00 1.40
0.1 BRL
BRL 0.06
0.15 0.01
0.7 0.01
0.03 0.08
1.74 0.49
BRL BRL
BRL 0.02
0.15 0.01
0.10 0.06
2.47 0.60
2.72 0.67
DG2
0.47
BRL
0.09
0.02
BRL
0.04
0.25
BRL
0.02
0.02
0.09
0.31
0.42
DG3
1.32
BRL
BRL
BRL
BRL
BRL
0.13
0.05
0.05
0
0
0.23
0.23
DG4
0.77
BRL
BRL
BRL
0.03
0.02
0.84
0.02
0.09
0
0
1.00
1.00
DG5
1.80
0.01
BRL
0.02
0.03
0.22
0.38
0.03
0.15
0.02
0.01
0.81
0.84
DG6
6.13
0.1
ND
0.01
0.04
0.07
0.48
BRL
0.23
0.01
0.10
0.82
0.93
QY1
0.30
ND
ND
0.03
0.01
0.2
0.17
ND
BRL
0.03
0
0.38
0.41
QY2 QY3
0.47 0.17
ND ND
ND ND
0.03 0.01
BRL 0.02
0.07 0.11
0.14 0.22
ND ND
BRL BRL
0.03 0.01
0 0
0.21 0.35
0.24 0.36
mean TU
2.83
0.037
0.15
0.99
1.18
values used to derive
0.085 0.070
0.037
0.067
0.055
0.81
0.031
0.023
1.76a
6.7b
2c
1.3c
1.3c
1.2c
31.5c
0.5a
P
PTU-OP
predicted TU (μg/g OC) a
Based on Probable Effect Concentration (PEC) values.20,21 b Lethal concentration 50 (LC50) value of chlorpyrifos.18 c LC50 values of pyrethroids.6,22 d Pesticides withP predicted TUs lower than 0.1 in all sediment canPbe found in Table S2. e BRL - below reporting P limit, ND - not detected, OC - organic carbon, PPTU-OCP = sum of predicted TUs of OCPs, PTU-OP = sum of predicted TUs of OPs, PTU- pyre = sum of predicted TUs of pyrethroids, PTU-pest = sum of predicted TUs of all three classes of pesticides. Chl = chlordanes, Chlor = chlorpyrifos, Bif = bifenthrin, Cyh = Lambda-cyhalothrin, Cyp = cypermethrin, Fen = fenpropathrin, and Del = deltamethrin. Pesticide concentrations in sediments were listed in a previous study.12
evaluated using PECs as well as Ag.26 The predicted TU was expressed as the mean PEC quotients of the evaluated metals. Specific TIE Analysis. In addition to the TU evaluation, a specific TIE test was conducted to confirm sediments toxicity caused by pyrethroids and it followed the same methods as toxicity testing discussed above, with the following modifications. First, test water was spiked with piperonyl butoxide (PBO, ChemService, West Chester, PA, USA), at 200 μg/L, one day before use. Daily water changes were conducted manually by gently adding test water to each replicate. Additionally, this test was conducted in a four-day time period, rather than 10-d, as significant differences in toxicity between bioassays with or without PBO amendment were noted within four days. Data Analysis. Survival responses among site sediments were compared to the BM control after an arcsine transformation using an analysis of variance (R = 0.05) coupled with a Dunnett’s multiple comparison test using SAS 9.1 software (SAS Institute Inc., Cary, NC, USA). Significant differences between the control sediment and the site sediment indicated toxicity.
’ RESULTS AND DISCUSSION Observed Toxicity. Water quality parameters for all bioassays were within US EPA guidelines.14 Mean ((standard deviation) water quality parameters were temperature 23.3 ( 0.6 °C, DO 5.57 ( 0.84 mg/L, pH 7.56 ( 0.35, and conductivity 380 ( 39 μS/cm. Sediment SZ1 had higher conductivity values (≈598 μS/cm), as this site was closer in proximity to the South China Sea Harbor. Mortality data and observed TUs for all sites are listed in Table 1 and Figure S2 in the SI. Control (BM) and the sediments collected at the e-waste site (QY) all exhibited lower than 0.5 TUs (100% mortality). Geographically, five of the seven sites exhibiting 100% toxicity were located in Shenzhen, with only one site each in Guangzhou and Dongguan. Five of the six sites (83%) in Shenzhen and Dongguan were toxic (>0.5 TUs), while 56% of sediments (five of nine sites) exhibited toxicity in Guangzhou. 1814
dx.doi.org/10.1021/es103552d |Environ. Sci. Technol. 2011, 45, 1812–1819
Environmental Science & Technology
ARTICLE
Table 2. Observed Toxic Units (OTU) Measured by 10-Day Bioassay Using Chironomus dilutus and the Predicted Toxic Units (PTU) Calculated from Concentration of Polycyclic Aromatic Hydrocarbons (PAHs), Polychlorinated Biphenyls (PCBs), Polybrominated Diphenyl Ethers (PBDEs), Metals, and Ammonia (NH3) in Sediments Collected in Guangzhou (GZ), Shenzhen (SZ), Dongguan (DG), and Qingyuan (QY) in the Pearl River Delta, Chinah P
P
PCBs
PAHs
metals P
P
OTU
PECa
ESBb
TECc
PECa
ERMd
NH3e
GZ1
0.23
0.06
0.20
0.37
0.234
0.140
0.12
0.61
GZ2
0.20
0.02
0.11
0.19
0.099
0.062
0.04
0.27
0.33
GZ3
1.10
0.01
0.07
0.15
0.088
0.052
0.2
0.37
1.36
GZ4
0.30
0.01
0.03
0.08
0.201
0.187
0.11
0.35
1.12
GZ5
0.43
0.01
0.05
0.10
0.031
0.021
0.08
0.17
0.90
GZ6
1.20
0.01
0.19
0.46
0.019
0.012
0.54
0.76
3.01
GZ7 GZ8
7.60 1.87
0.01 0.01
0.06 0.09
0.16 0.21
0.051 0.041
0.033 0.026
0.49 0.25
0.61 0.39
1.84 5.33
PTU-ppmn
PTU-all 0.90
GZ9
1.80
0.01
0.14
0.32
0.041
0.029
0.4
0.59
1.20
SZ1
8.00
0.01
0.07
0.17
0.063
0.041
0.05
0.19
1.75
SZ2
8.00
0.01
0.07
0.17
0.055
0.038
0.07
0.21
1.76
SZ3
8.00
0.01
0.10
0.24
0.039
0.027
0.08
0.23
3.08
SZ4
8.00
0.01
0.05
0.12
0.191
0.124
0.03
0.28
2.83
SZ5
0.30
0.02
0.02
0.05
0.071
0.048
0.01
0.12
0.21
SZ6 DG1
8.00 1.40
0.02 0.01
0.07 0.01
0.14 0.03
0.338 0.060
0.214 0.037
0.01 0.07
0.44 0.15
3.16 0.82
DG2
0.47
0.01
0.26
0.61
0.054
0.034
0.33
0.65
1.07
DG3
1.32
0.01
0.40
1.01
0.094
0.062
0.16
0.66
0.89
DG4
0.77
0.01
0.21
0.56
0.084
0.056
0.14
0.44
1.44
DG5
1.80
0.01
0.04
0.11
0.101
0.069
0.14
0.29
1.13
DG6
6.13
0.04
0.08
0.19
0.326
0.210
0.16
0.61
1.54
QY1
0.30
0.50
0.02
0.05
0.159
0.092
0.06
0.74
1.15
QY2 QY3
0.47 0.17
0.56 0.50
0.04 0.02
0.11 0.05
0.238 0.059
0.138 0.036
0.06 0.09
0.90 0.67
1.14 1.03
mean TU
2.83
0.08
0.10
0.45
1.62
values used to derive predicted TU
676 μg/kg dry weight
f
0.24 290 μg/g
0.11
0.07
0.15
f
f
g
organic carbon
a
Based on sum concentration and probable effect concentration (PEC).20,21 b Based on equilibrium sediment benchmarks (ESB).23 c Based on threshold effect concentration (TEC).24 d Based on effects-range median (ERM).26 e Based on values in literature.25 f Toxic units (TU) are based on individual concentrations (of PAHs and metals). Concentration and the respected benchmark or guideline values can be found in Tables S3 hP and S6. g Toxic units (TUs) are based on total ammonia concentrations and PTUP porewater pH concentrations and can be found in Table S7. ppmn = sum of predicted TUs of PAHs, PCBs, metals, and ammonia; PTU-all = sum of predicted TUs of all target contaminants including PAHs, PCBs, metals, ammonia, and pesticides.
The greatest toxicity was observed in the five sediments collected from Shenzhen which showed 100% mortality after diluting the sediment four times indicating >8.0 TUs. Predicted Toxicity. All three classes of insecticides (OCPs, OPs, and pyrethroids) were detected in sediments collected from urban waterways in the PRD, and a more detailed analysis of the occurrence and distribution of the insecticides at these sites and associated trends was discussed in Li et al.12 The current study only analyzed their associated TUs and contributions to sediment toxicity. Insecticides that exhibited >0.1 predicted TUs in at least one site are shown in Table 1, and those that had less predicted TUs but still detectable levels above reporting limits (RLs) are reported in Table S2 in the SI. Additionally, concentrations of PAHs, PCBs, PBDEs, metals, ammonia, and TOC can be found in Tables S3-S7 in the SI as well as toxic benchmarks for each individual compound. Predicted toxicity of the sum of each of these contaminants or group of contaminants is shown in Table 2.
Chlorpyrifos was the only OP detected above RLs, while concentrations of the remaining four OPs were below RLs.12 At four sites, g0.1 predicted TUs was observed for chlorpyrifos with the highest TU value being 0.24 (12% mortality) in GZ6 sediment. Eight of the 19 target OCPs were present above RLs in one or more sediments.12 Six sediments contained chlordanes and one sediment had β-BHC concentrations >0.1 predicted TUs, whereas the predicted TUs for the remaining OCPs were 0.1 predicted TUs for at least one site (Table 1), and the most commonly detected pyrethroid was cypermethrin, which was detected in all sites in the three cities. Additionally, cypermethrin was also detected in the three QY samples. Cypermethrin was not only the most frequently detected insecticide;12 it also had the highest predicted TU of all compounds at all sites (GZ8: 3.4 TUs) and the highest incidence of having predicted TUs >0.5 (57% of samples; 12 of the 21 sites). A recent study evaluating pesticides in eight countries, including China, in the Asia-Pacific region reported that the top two pyrethroids used were cypermethrin and lambda-cyhalothrin.27 Our finding that cypermethrin was the most frequently detected insecticide was consistent with the application pattern of pesticides in China. The predicted TUs indicated that cypermethrin contributed to the toxicity of all toxic sediments collected in the PRD. At most sites, predicted cypermethrin TUs were high enough alone to cause the observed toxicity, and about 70% of the predicted toxicity caused by pesticides could be explained by cypermethrin (Table 1). Collectively, the results suggest that of the insecticides analyzed, pyrethroids were a primary cause of sediment toxicity in the PRD. It should be noted, however, that while OPs and OCPs had low predicted TUs, their contribution to toxicity could still be of concern, as they are still heavily used throughout China (http:// www.chinapesticide.gov.cn/doc09/09112306.html). Other organic contaminants including PAHs, PCBs, and PBDEs were also evaluated in the current study (Tables 2 and S3-S5). Predicted TUs for these three classes of contaminants were first estimated using narcotic modeling based on concentrations of individual compounds,22 and the sum predicted TUs for the three classes of contaminants were all
