DDT) and Its Metabolites in Rivers in Tianjin, China

中国科技论文在线 http://www.paper.edu.cn

Chemosphere 68 (2007) 10–16 www.elsevier.com/locate/chemosphere

Spatial and temporal variations and possible sources of dichlorodiphenyltrichloroethane (DDT) and its metabolites in rivers in Tianjin, China

S. Tao *, B.G. Li, X.C. He, W.X. Liu, Z. Shi

Laboratory for Earth Surface Processes, College of Environmental Sciences, Peking University, Beijing 100871, China

Received 13 July 2006; received in revised form 20 December 2006; accepted 21 December 2006 Available online 9 February 2007

Abstract

Water, suspended solid (SS) and sediment samples were collected from nine water courses in Tianjin, China and analyzed for dichlorodiphenyltrichloroethane (DDT) and its metabolites (DDXs, including o,p0-,p,p0-DDT, DDD and DDE). The average DDX concentrations in water, SS and sediment were 59 ± 30 ng l1, 2690 ± 1940 ng g1 dry wt. and 340 ± 930 ng g1 dry wt., respectively. Due to the termination of the extensive agricultural application and industrial manufacture, DDXs in river sediment decreased by one order of magnitude since 1970’s and low DDT fractions in these sediments were observed. Still, DDXs in the sediments near the outlets of the major manufacturers remained relatively high attributed to the historic input. DDXs in sediment were also positively correlated with organic matter content. Spatial distributions of DDXs in SS and water was different from that in sediment. For SS, a negative correlation between DDX concentration and SS content indicated a dilution effect in many rivers. Dissolved organic carbon content was the major factor affecting DDX concentrations in water phase. Wastewater discharged from dicofol manufacturers and likely illegal agricultural application were the primary reasons causing high DDT (DDE + DDD) ratios in SS and water. 2006 Elsevier Ltd. All rights reserved.

Keywords: DDT; River water; Sediment; Suspended solids; Tianjin; Dicofol

1. Introduction zenemethanol) (Qiu et al., 2005). DDT is used as an inter- mediate to produce dicofol in China and around 5–10% China was a major producer and consumer of dichloro- of DDT has been detected in the ﬁnal product (Wei, diphenyltrichloroethane (DDT) in the past, and also uti- 2002). Since DDXs, especially freshly released DDT, have lized DDT for vector control, resulting in high residual been widely detected in China (Gong et al., 2004; Chen levels of DDT and its metabolites (DDXs) in various envi- et al., 2005; Wu et al., 2005; Peng et al., 2005; Hu et al., ronmental media and ecological compartments (Wong 2005) and some of these compounds were identiﬁed as et al., 2005; Sun et al., 2005; Chen et al., 2005; Liu et al., endocrine disrupters (Soto et al., 1994), public concerns 2006). Although the extensive application in agriculture over the enduring presence of DDT in the environment has been banned since 1992 (Chen, 1990), release of DDT have recently increased. into the environment is continuing, owing to possible illegal As one of the fastest growing areas in northern China, usage and occurrence of DDT as an impurity in other Tianjin, locating in the low reaches of Hai River, has suf- widely employed pesticides, such as dicofol (active ingre- fered from severe contamination of DDT from various dient: 4-chloro-a-(4-chlorophenyl)-a-(trichloromethyl)ben- sources. DDT had been extensively applied in local agriculture for several decades (Zhao and Ma, 2001). A large * Corresponding author. Tel./fax: +86 10 62751938. quantity of DDT was also produced locally in Dagu Chem- E-mail address: [email protected] (S. Tao). ical Company and Tianjin Chemical Company. Moreover,

S. Tao et al. / Chemosphere 68 (2007) 10–16 11

Renmin Pesticide Company and a number of other manu- South Canal (S. Canal), North Canal (N. Canal), South facturers in this area have produced dicofol since 1990’s. Sewer (S. Sewer), North Sewer (N. Sewer) and Beijing Over the years, the wastewater originated from the manu- Sewer (BJ Sewer) are either artificial water courses or facturers has been discharged into the adjacent aquatic open sewers. These water bodies collect the majority of nat- environment without effective treatment. It was reported ural runoff as well as industrial and domestic effluents in recently, that the mean DDX concentration in sediment Tianjin. samples from South Sewer (Nanpaiwu River) was 29 ng The sediment samples were collected using a grab sam- g1 dry wt. (Hu et al., 2005). The measured DDXs in the pler, and the water and SS were sampled using cylinder sediments from the neighboring Hai River and Dagu Sewer samplers. At least three sub-samples were collected for each ranged from 2 to 19 ng g1 dry wt. and 4 to 84 ng g1 dry sample type at each location. The sub-samples were then wt., respectively (Yang et al., 2005). mixed thoroughly into a composite sample to reduce the So far, the data available on DDT contamination in possible random variation. All the samples were frozen aquatic environment of Tianjin were based on a limited immediately after being transported to the laboratory number of sediment samples. To provide an overall under- and kept at 18 C before analysis. The sediment samples standing of the DDT contamination in Tianjin, an exten- were centrifuged (Centrifuge TDL-5 > 3000 r m1), freeze- sive survey was conducted recently, and water, suspended dried (EYELA-FDU-830) and rubbed to pass through a solid (SS) and sediment were sampled concurrently. In 70-mesh sieve. The SS samples were separated from water addition to the levels and compositions of DDXs in the by glass fiber filters (Sartorius, B = 47 mm, 0.5 lm, area, the present study particularly focused on spatial ashed at 450 C). The separated SS samples were also and temporal variations and source identification, in hopes freeze-dried. of offering necessary information for assessing the potential risk on the regional ecosystem and public health, and for 2.2. Extraction and cleanup of sediment and SS samples formulating practical management strategy. The sediment and SS samples were extracted with an 2. Methodology accelerated solvent extractor (ASE-300, Dionex, USA), equipped with 34 ml extraction cells. Around 5 g freeze- 2.1. Sampling dried sediment or 0.5 g SS sample was analytically weighed, mixed with anhydrous sodium sulfate (1:5) and transferred Twenty nine water samples, twenty nine SS samples and into the extraction cell. For the sediment samples, activated sixty bottom sediment samples were collected from various copper powder (1:1 ratio of copper powder to sediment) locations along the nine major rivers and canals (Fig. 1)in was added to desulphurize the extract. The samples were summer 2002. Hai (Hai R.), Chaobai (Chaobai R.) and extracted using dichloromethane as the extraction solvent Yongding (Yongding R.) are natural rivers, while Ji Canal, under 10335 kP at 140 C. The extraction was carried out in one cycle with 7 min heating followed by 5 min static extraction. The extracted solutions were concentrated to

0 20 km N near dryness on a rotary evaporator at 35 C. The concentrated extracts were transferred with 2 ml hexane onto the top of a chromatography column (30 cm · 10 mm i.d.) Beijing filled with 6 g Florisil. The elution procedure consisted of 50 ml hexane and 50 ml 7:3 hexane/dichloromethane (v/v) Tianjin at a rate of 2 ml min1. The eluate was collected and con- Chem. Co. centrated on the rotary evaporator, transferred to a Kud- N. Canal erna–Danish tube and then rinsed three times with BJ Sewer Chaobai R. hexane. The final volume was brought down to 1 ml under Ji Canal. Renmin a gentle stream of nitrogen. Pesticide Co. Tianjin Yongding R. Pesticide Co. 2.3. Extraction of water samples N. Sewer Bohai Bay Hai R. Dissolved DDXs in water samples were extracted by S. Sewer Dagu solid phase micro extraction (SPME). The fiber selected Chem. Co. was a fused-silica rod coated with 100 lm of poly- S. Canal Tianjin dimethylsiloxane (PDMS) from Supelco (Supelco Corp., China USA). The fibers were conditioned according to manufac- turer’s recommendations for 2 h at 250 C before utiliza- Fig. 1. Water courses (thick lines), sampling locations (triangles) and tion. The fiber was exposed to the water sample in a major chemical and pesticide manufacturers (filled squares) in Tianjin. All 4.3 ml Supelco vial for 1 h and the water samples were con- rivers flow from northwest towards southeast. tinuously agitated with a magnetic stir bar on a stir plate 中国科技论文在线 http://www.paper.edu.cn

12 S. Tao et al. / Chemosphere 68 (2007) 10–16 revolving at 1000 rpm (Eyela RCH-3 D with 10 · 3 mm stir methane and hexane, 1:1 v/v) was concentrated to a total bar, Japan) during the extraction. Water that may have volume of 0.5 ml and then measured with GC-ECD to wicked into the needle was removed by vigorous shaking check the solvent background. The procedure blanks were of the fiber assembly and/or by touching the tip of the nee- run simultaneously with every set of the samples. A stan- dle with a tissue paper. The thermal desorption of the fiber dard solution was included for every 20 samples during occurred inside of the injection port at 220 C for 4 min, GC measurements to monitor the stability of the detection. and once the desorption was complete, the split vent was All the samples were measured in two duplicates. opened for the remainder of the chromatographic sepa- Anhydrous sodium sulfate and glassware were heated at ration. 650 C to remove moisture and organic impurities. Florisil was pre-cleaned for 6 h at 650 C and dried in a 130 C 2.4. Sample analysis oven for at least 16 h before use. The quantitative standards were obtained from Agilent, USA. All the solvents The extracted samples were analyzed for DDXs using a used for sample processing were analytical grade (Beijing Agilent GC 6890 (Agilent, USA) equipped with a 30 m Reagent Co., China) and were purified by distillation. HP-5 column (0.25 mm i.d., 0.25 lm film) and a 63Ni elec- The solvents used for GC-ECD analysis were pesticide res- tron capture detector (l-ECD). The GC system was oper- idue grade (Scharlau Chemies, Spain). Active copper pow- ated in a splitless mode with a venting time of 0.75 min. der was prepared by being washed with HCl (1 M) and 1 ll of each sample was injected. The oven temperature methanol. was maintained at 50 C, raised to 150 C at a rate of 10 C min1, then programmed to 240 Cat3C min1, and held for 5 min. The injector and the detector were 3. Results and discussion maintained at 220 C and 280 C, respectively. Nitrogen (purity > 99.999%) was employed as the carrier gas at a 3.1. DDXs in water, SS and sediment flow rate of 1.0 ml min1. GC peaks were identified based on the retention time of individual authentic standards The arithmetic and geometric means, minimum, maxi- (±0.3%). The residues of p,p0-DDT, p,p0-DDD, p,p0- mum, median, and the 5th, 25th, 75th and 95th percentiles DDE, o,p0-DDT, o,p0-DDD and o,p0-DDE were quantified of the total concentrations of the six individual DDT, DDD by the external standard method using a standard mixture and DDE species (DDXs) in water, SS and sediment sam- (Chem Service Inc., USA). The measured DDX concentra- ples are summarized in Fig. 2. The average DDXs in water, tions of SS and sediment samples are presented in a dry SS and sediment were 59 ± 30 ng l1, 2700 ± 1900 ng g1 weight basis. and 340 ± 930 ng g1, respectively. Since the mean content The content of total organic carbon (TOC) in sediment of SS in the water samples was 0.089 g l1, the mean volume and SS samples and dissolved organic carbon (DOC) in concentration of DDXs in SS was as high as 240 ± water samples were determined with TOC analyzer (TOC 170 ng l1, almost three times higher than that in dissolved 5000 A coupled with SSM-5000 A sampler, Shimadzu phase. It appears that most DDXs in water column Corp., Japan). Content of SS in a raw water sample was occurred in SS with less than 20% dissolved in water. The measured by filtering approximate 1 l of water sample measured DDXs in all three media were log-normally dis- through a pre-dried and pre-weighed glass fiber filter. The tributed. As a result, the geometric means were 27 ng l1, gain in mass on the glass fiber filter per unit volume of 2000 ng g1 and 63 ng g1 for water, SS and sediment, water filtered was defined as SS content (Gustafson and respectively, which were much lower than the arithmetic Dickhut, 1997). means. There was a significant difference in DDXs between SS and sediment though they were considered to come from 2.5. Analytical quality control the same origin due to sedimentation and resuspension. In fact, the difference was primarily attributed to an unusually For SPME, the linearity of the method was tested by extracting aqueous standards, in triplicate, over a range 5

1 1 ) 10 between 0.4 ng l and 1000 ng l . Five SPME extractions -1 4 10 were performed using spiked aqueous standard solution max 1 3 p95 (10 ng l each DDXs, ChemService Inc., USA) to deter- ), log(ng g 10 -1 p75 2 geomean mine recoveries and precision of the method. Spiked sedi- 10 p50 mean ments and SS were also used for same purpose. For the 1 10 p25 p target compounds, the detection limits calculated as mean 5 0 min 1 DDXs, log(ng l 10 blank +3 · SD, were 0.24 ng g for 5 g sediment, WaterSS Sediment 0.1 ng l1 for 4 ml water sample and 2.4 ng g1 for 0.5 g Fig. 2. Measured concentrations of DDXs in water, SS and sediment in SS, respectively. The average recoveries of the individual Tianjin. The results are presented as the minimum, maximum, arithmetic

DDX species were 81%, 86% and 92% for sediment, water mean (mean), geometric mean (geomean), median (p50), the 5th (p5), 25th and SS, respectively. A 100 ml mixed solvent (dichloro- (p25), 75th (p75) and 95th (p95) percentiles. 中国科技论文在线 http://www.paper.edu.cn

S. Tao et al. / Chemosphere 68 (2007) 10–16 13

0 1 ) 5 high level of p,p -DDT in SS (2230 ng g in SS compared -1 10 to 140 ng g1 in sediment). 1979 2001

DDT had been widely applied in agriculture in Tianjin ), log(ng g -1 for more than forty years and thus the agricultural soil in 10 3 the area had been severely contaminated (Zhao and Ma, 2001). Based on an extensive survey on 188 samples from

the area, it was reported that the mean concentration of DDXs , log(ng l 10 1 DDXs in surface soil was 56 ± 134 ng g1 with a maximum Water SS Sediment 1 value of 970 ng g (Gong et al., 2004). By average, the lev- Fig. 3. Comparison of DDX concentrations (arithmetic means and els of DDXs in SS and sediment were much higher than standard deviations) in water, SS and sediment from a site immediately those in the terrestrial environment. It was likely that a por- next to an effluent outlet of Tianjin Chemical Company between 1979 and tion of DDXs in rivers was from the direct discharge of efflu- 2001. ents instead of surface runoff. For many years, wastewater containing DDT from the local major chemical companies concentrations were detected in water, SS and sediment was not treated, except for passing through a sedimentation near the company’s effluent outlet in Ji Canal (Mo and tank before discharge. Because of such a heavy input, the Yu, 1983), and are compared with the data collected 23 levels of DDXs in the aquatic environment in Tianjin were years later at the same location (Fig. 3). For all the three historically much higher than those in other places in China. media, the DDX concentrations decreased more than one For example, although Beijing is immediately contiguous order of magnitude. Tao et al. have conducted a dynamic (upstream) to Tianjin, the total concentrations of organo- fate modeling on the change of lindane in various media chlorine pesticides with DDT as the pre-dominant com- in Tianjin, and revealed about one to one and a half orders pound in sediments from Tonghui River in Beijing were of magnitude decreases in the concentrations of lindane in 1 from 1.8 to 14 ng g (Zhang et al., 2004). In Taihu Lake, air, water, soil and sediment from 1980’s to 2001 (Tao which is regarded to be heavily contaminated by pesticides, et al., 2006). The production and application history of the measured DDXs in sediments varied from 0.65 to DDT and lindane were very similar in Tianjin (Ministry 1 38 ng g (Peng et al., 2005). The contamination in Tianjin of Chemical Industry, 1992). With similarities in chemical was also characterized by large variations among sampling property and environmental behavior as persistent and locations and rivers. For instance, DDXs in the five sedi- semi-volatile compounds, similar fates of the two pesticides ment samples collected from Hai River ranged from 2 to in the environment are expected. As suggested by the 1 1100 ng g . results of the dynamic modeling, the concentration of lin- All water courses in Tianjin ultimately discharge into dane would drop another order of magnitude in the next Bohai Bay (Fig. 1). It was reasonable to expect that the two decades or so. It is, therefore, reasonable to expect severe contamination in the surface water would impose the similar change of DDXs in the area. significant influence on the bay. Hu et al. have investigated the offshore sediments of Bohai Bay and reported that in addition to other persistent organic pollutants, DDXs ran- 3.3. Spatial variation ged from 1.6 to 12.3 ng g1 (Hu et al., 2005). According to the results of, a survey conducted by the State Oceanic The sampling scheme of this study covered all major Administration, the mean DDX concentration of the off- water courses in Tianjin. The differences in DDX concen- shore sediments from Bohai Bay near Tianjin was trations in water, SS and sediment among the rivers (or 1.14 ± 1.54 ng g1 with two highest measurements of 3.4 canals) are compared in Fig. 4. Significant differences and 4.6 ng g1, both of which were from the locations close among the water courses are revealed for all the three to the Hai River mouth (Liu et al., 2006). media. DDX concentrations in water, SS and sediment from South Canal were all very low. DDXs in water from 3.2. Temporal change of DDXs in Ji Canal South Sewer, SS from Hai River, and sediment from Ji Canal ranked first in the three media, respectively. In addi- Historical information on DDXs in the aquatic environ- tion to, the differences among the water courses, concentra- ment in Tianjin is scarce, and the only data available were tions at different sampling locations in the same river also the measurements on several samples collected from Ji varied broadly, leading to large standard deviations in Canal in the later 1970’s. Wastewater from Tianjin Chem- many cases (Fig. 4). There was no correlation in DDXs ical Company, one of the largest DDT producers in China, among the three media based on either river average or was discharged directly into Ji Canal (Fig. 1). DDT manu- individual locations. For example, DDXs in sediments facture in the company began in 1953 with annual produc- from Ji Canal, North Canal and Yongding River were con- tion varied from 4000 to 7000 tons. The production was siderably more abundant than those from the other rivers, reduced to 2000 tons per year since 1992 when the agricul- but it was not the case for water and SS. Among all the tural application was banned in China and was totally ter- rivers, the highest DDX concentration in SS was found minated by the end of 2000. In 1979, extremely high DDX in Hai River, while the concentrations in water and 中国科技论文在线 http://www.paper.edu.cn

14 S. Tao et al. / Chemosphere 68 (2007) 10–16

200 6000 3600 Water SS Sediment Ji Canal -1 -1

-1 150 4500 2700

100 3000 1800 N.Canal Yongding DDXs, ng g DDXs, ng l

50 g DDXs,. ng 1500 900

0 0 0 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 91 2 3 4 5 6 7 8 9

Fig. 4. Means and standard deviations of DDXs in water, SS and sediment samples in various water courses. They are: 1. Ji Canal, 2. Chaobai River, 3. North Canal, 4. Beijing Sewer, 5. Yongding River, 6. North Sewer, 7. Hai River, 8. South Sewer and 9. South Canal. sediment from Hai River were all below the averages. The Fig. 5) and six others (the blank symbols in Fig. 5). The partition of DDXs among the three media was not neces- correlation between DDXs and TOC for the second cate- sarily reaching equilibrium and their concentrations might gory is significant at a level of less than 0.00001 (n = 49). be largely under the influence of other factors individually. Similar correlations between the hydrophobic pollutants Among the three media studied, sediment is the best one and organic matter content have been reported in many in terms of recording the historical input. Among the 60 cases (Castilho et al., 2000; Peng et al., 2005). sediment samples surveyed, extremely high DDX levels The variation of DDXs in SS among the rivers was were found at the sites near the major discharge outlets totally different from that in sediment. For instance, DDXs of the local large chemical and pesticide manufacturers. in the SS from the three rivers with severely contaminated Three sediment samples with the highest measured concen- sediment (Ji Canal, North Canal and Yongding River) trations were from immediately downstream of the effluent were not particularly high (Fig. 4, middle). Unlike sedi- outlets of Tianjin Chemical Company (5350 ng g1,Ji ment, no correlation was found between TOC and DDXs Canal), Renmin Pesticide Company (3620 ng g1, North in SS. Instead, a significantly negative correlation was Canal) and Tianjin Pesticide Company (3640 ng g1, revealed between DDXs in SS and SS content in water Yongding River). The extremely high DDX concentrations (Fig. 5, middle). It appears that the majority of the SS came of these samples and a few immediately next to them were from the surface runoff or various point sources with low the primary reason causing considerably high mean values DDXs contents, which diluted the DDXs in SS. This is par- of the three rivers (Fig. 4, right). Such extreme cases also tially true because the samples were collected in wet season bring the mean DDX concentration of all the sediments with frequent precipitations. in Tianjin to a much higher level than those of the other When log-normally transformed DDXs in water are places in China. If these extremes (six in total) were not plotted against DOC for all the samples (Fig. 5, right), the included statistically, the arithmetic mean concentration calculated correlation coefficient is not significant with a of sediment would decrease from 340 ± 930 ng g1 to p-value of 0.158. However, if a few outliers with extremely 110 ± 160 ng g1. In addition to, the effect of the direct low DDXs are excluded (marked filled in Fig. 5, right), input, sediment TOC was found to be another key factor the correlation would become significant at a level of governing the DDX concentrations in sediment. When 0.0005. The DDX concentrations of the six outliers are the measured DDXs are plotted against TOC (Fig. 5, left), more than an order of magnitude lower than those of the positive correlation is demonstrated with a p-value of other samples and they were either from the less contami- 0.00018 (n = 60). Taking the influence of the direct input nated rivers away from the major point sources (Chaobai from the major pesticide producers into account, the nine River, South Canal and North Sewer) or from the upstream water courses can be divided into two categories of the of the contaminated rivers (Yongding River and South three severely contaminated rivers (the filled symbols in Sewer). Like other hydrophobic compounds, DDT and its

4.0 4.0 2.7 ) ) ) -1 3.0 -1 3.5 -1 1.8 2.0 3.0 0.9 1.0 2.5 DDXs, log(ng l DDXs, log(ng g DDXs, log(ng g Sediment SS Water 0.0 2.0 0.0 0 0.5 1.0 1.5 −1.7 −1.2 −0.7 −0.2 0 40 80 120 TOC, % SS, log(g l-1) DOC, mg l-1

Fig. 5. Relationships between DDXs and TOC in sediment, DDXs in SS and SS in water and DDXs and DOC in water. All data except TOC and DOC were log-transformed. For sediment, the ﬁlled symbols represent the samples from Ji Canal, Yongding River and North Canal, which received direct input from the major DDT producers in history, while the blank symbols are the samples from the other six rivers. For water, the six ﬁlled symbols are the outliers with very low DDXs concentrations. 中国科技论文在线 http://www.paper.edu.cn

S. Tao et al. / Chemosphere 68 (2007) 10–16 15

metabolites have a strong aﬃnity to naturally occurring organic matter. It was reported that the binding constant 0 (logKDOC) between p,p -DDT and surface water DOC was 4.83 (De Bruijn et al., 1998). In the case of Tianjin, DOC content was a critical factor aﬀecting DDX concentrations in the water phase.

3.4. DDX proﬁle

Fig. 6 illustrates the arithmetic means, medians and the 5th, 25th, 75th and 95th percentiles of the concentrations of six individual DDX species in water, SS and sediment. For all the three media, p,p0-DDT and its metabolites were dominant, while o,p0-DDXs were one (sediment) to three (water) orders of magnitudes lower than p,p0-DDXs. o,p0- DDXs accounted for only 8% of the total DDXs in SS by average with two exceptions. Unusually high o,p0- DDX concentrations were observed in two SS samples from Yongding River, in which o,p0-DDXs contributed to 75% and 30% of the total DDXs, respectively. The two sampling sites were immediately outside or 10 km downstream of Renmin Pesticide Company, where dicofol was produced. The commercial dicofol product from the company was also measured for DDXs, and over one third Fig. 7. Levels and proﬁles of DDXs in the river sediments. The pie 0 of DDX residue in the product was o,p -isomers. Obvi- diameters are proportional to the total DDXs. ously, wastewater discharged from the company was the primary reason causing the extremely high o,p0-DDXs in high DDT fractions are indicators of recent input since the SS from the two locations. both DDE and DDD are corresponding metabolites of As addressed above, DDXs in the sediment provided DDT (Hitch and Day, 1992). If the individual water evidence on the historical input of DDT, and can be related courses are examined, high DDT/(DDE + DDD) ratios to the major point sources in Tianjin. This is explicitly could be mostly found in the water and SS samples from depicted in Fig. 7. DDT production in Tianjin Chemical Chaobai River, Beijing Sewer, Hai River and South Sewer. Company was not terminated until the end of 2000, while Sampling locations on Chaobai River and Beijing Sewer dicofol production in Renmin Pesticide Company still con- were situated in the rural area, and the recent DDT input tinues. The wastewater discharged from these activities was likely from agricultural activity, leading to high resulted in not only the high DDX levels but also relatively DDT/(DDE + DDD) ratios but not particularly high large DDT fractions in several sediment samples collected DDX concentrations. On the other hand, both DDX con- nearby (Fig. 7). For the majority of the samples collected centrations and DDT/(DDE + DDD) ratios of the water from Tianjin, however, the DDT/(DDE + DDD) value and SS samples from Hai River and South Sewer were was around 0.2, indicating that the historically accumu- high, indicating possible direct discharge from point lated DDT in most sites has gone through a long-term de- sources. In addition to several large chemical companies, gradation (Hitch and Day, 1992). there are many small-scale pesticide manufactures in the In water and SS, the calculated DDT/(DDE + DDD) district. These facilities are often not well monitored and ratios were much higher than that in sediment with mean regulated. Consequentially, their production and discharge values of 15.4 ± 17.3 and 17.1 ± 17.2, respectively. Such are often oﬀ record.

103 104 103 Water SS Sediment )

3 ) 2 p )

-1 95 10 -1 10 mean -1 1 10 p75 2 1 median 10 10 p 25 1 0 − p 10 10 10 1 5 0 −1 DDXs, Log(ng l DDXs, log(ng g

DDXs, log(ng g 10 10

− − −2 10 3 10 1 10 DDE DDD DDT DDE DDD DDT DDE DDD DDT DDE DDD DDT DDE DDD DDT DDE DDD DDT o,p’-DDX p,p’-DDX o,p’-DDX p,p’-DDX o,p’-DDX p,p’-DDX

Fig. 6. Measured DDT and its metabolites as individual species in water, SS and sediment in Tianjin. The results are presented as arithmetic means, medians, the 5th, the 25th, the 75th and the 95th percentiles. 中国科技论文在线 http://www.paper.edu.cn

16 S. Tao et al. / Chemosphere 68 (2007) 10–16

Acknowledgements Ministry of Chemical Industry. 1992. On implementation of the notifica- tion from the office of the State Department on management of The funding of this study was provided by the National pesticides and animal medicines, MCI 1992, 77, Beijing. Mo, H.H., Yu, W.X., 1983. Transportation and Transformation of Science Foundation of China (Nos. 40372132, 40332015), Organochlorine Pesticides and Phosphorous in Water Body of Ji the Ministry of Science and Technology (2007CB407301) Canal. Institute of Geography, Chinese Academy of Science, Beijing. and the Ministry of Education, China. Peng, X.Z., Zhang, G., Zheng, L.P., Mai, B.X., Zeng, S.W., 2005. The vertical variations of hydrocarbon pollutants and organochlorine pesticide residues in a sediment core in Lake Taihu, East China. References Geochem. Explor. Environ. Anal. 5, 99–104. Qiu, X.H., Zhu, T., Yao, B., Hu, J.X., Hu, S.W., 2005. Contribution of Castilho, J.A.A., Fenzl, N., Guillen, S.M., Nascimento, F.S., 2000. dicofol to the current DDT pollution in China. Environ. Sci. Technol. Organochlorine and organophosphorus pesticide residues in the Atoya 39, 4385–4390. River Basin, Chinandega, Nicaragua. Environ. Pollut. 110, 523– Soto, A.M., Chung, K., Sonnenshein, C., 1994. The pesticides endosulfan, 533. toxaphene and dieldrin have estrogenic effects on human estrogen- Chen, L.G., Ran, Y., Xing, B.S., Mai, B.X., He, J.H., Wei, X.G., Fu, sensitive cells. Environ. Health Perspect. 102, 380–383. J.M., Sheng, G.Y., 2005. Contents and sources of polycyclic aromatic Sun, Y.Z., Wang, X.T., Li, X.H., Xu, X.B., 2005. Distribution of hydrocarbons and organochlorine pesticides in vegetable soils of persistent organochlorine pesticides in tissue/organ of silver carp Guangzhou, China. Chemosphere 60, 879–890. (Hypophthalmichthys molitrix) from Guanting Reservoir, China. J. Chen, X., 1990. The pesticides production of China, towards high efficient Environ. Sci. (China) 17, 722–726. and more secure. Pesticides 29, 1. Tao, S., Yang, Y., Cao, H.Y., Liu, W.X., Coveney, R., Xu, F.L., Cao, J., De Bruijn, J., Busser, F., Seinen, W., Hermans, J., 1998. Determination of Li, B.G., Wang, X.J., Hu, J.Y., Fang, J.Y., 2006. Modeling the dynamic octanol water partition-coefficients for hydrophobic organic-chemicals changes in concentrations of c-hexachlorocyclohexane (c-HCH) in with the slow-stirring method. Environ. Toxicol. Chem. 8, 499–512. Tianjin region from 1953 to 2020. Environ. Pollut. 139, 183–193. Gong, Z.M., Tao, S., Dawson, R., Cui, Y.H., Cao, J., Wang, X.J., Shen, Wei, R.J., 2002. Personal Communication. Tianjin Renmin Pesticide Co., W.R., Qing, B.P., Sun, R., 2004. Level and distribution of DDT in Tianjin. surface soils from Tianjin, China. Chemosphere 54, 1247–1253. Wong, M.H., Leung, A.O.W., Chan, J.K.Y., Choi, M.P.K., 2005. A Gustafson, K.E., Dickhut, R.M., 1997. Distribution of polycyclic review on the usage of POP pesticides in China, with emphasis on aromatic hydrocarbons in Southern Chesapeake Bay surface water: DDT loadings in human milk. Chemosphere 60, 740–752. evaluation of three methods for determining freely dissolved water Wu, S.P., Tao, S., Lan, T., Li, B.G., Cao, J., Liu, W.X., Liu, Y., 2005. concentration. Environ. Toxicol. Chem. 16, 452–461. HCH and DDT Residues in the dustfall of Tianjin, China. J. Environ. Hitch, R.K., Day, H.R., 1992. Unusual persistence of DDT in some Sci. Health A40, 1715–1730. western USA soil. Bull. Environ. Contam. Toxicol. 48, 259–264. Yang, R.Q., Lv, A.H., Shi, J.B., Jiang, G.B., 2005. The levels and Hu, J.Y., Wan, Y., Shao, B., Jin, X.H., An, W., Jin, F., Yang, M., Wang, distribution of organochlorine pesticides (OCPs) in sediments from the X.J., Sugisaki, M., 2005. Occurrence of trace organic contaminants in Haihe River, China. Chemosphere 61, 347–354. Bohai Bay and its adjacent Nanpaiwu River, North China. Mar. Zhang, Z.L., Huang, J., Yu, G., Hong, H.S., 2004. Occurrence of PAHs, Chem. 95, 1–13. PCBs and organochlorine pesticides in the Tonghui River of Beijing, Liu, W.X., Chen, J.L., Lin, X.M., Tao, S., 2006. Distribution and China. Environ. Pollut. 130, 249–261. characteristics of organic micropollutants in surface sediments from Zhao, L., Ma, Y.J., 2001. Residual of organochloride pesticides in Bohai Sea. Environ. Pollut. 140, 4–8. agricultural environment. Agri. Environ. Develop. 67, 37–39.