- Email: [email protected]
Geochemical baseline level and function and contamination of phosphorus in Liao River Watershed sediments of China
Journal of Environmental Management 128 (2013) 138e143
Contents lists available at SciVerse ScienceDirect
Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman
Geochemical baseline level and function and contamination of phosphorus in Liao River Watershed sediments of China Shaoqing Liu a, Jing Wang b, Chunye Lin a, *, Mengchang He a, Xitao Liu a a b
State Key Joint Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China China Land Surveying and Planning Institute, Key Laboratory of Land Use, Ministry of Land and Resources, Beijing 100035, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 January 2013 Received in revised form 4 April 2013 Accepted 8 May 2013 Available online 1 June 2013
The quantitative assessment of P contamination in sediments is a challenge due to sediment heterogeneity and the lacking of geochemical background or baseline levels. In this study, a procedure was proposed to determine the average P background level and P geochemical baseline level (GBL) and develop P geochemical baseline functions (GBF) for riverbed sediments of the Liao River Watershed (LRW). The LRW has two river systems e the Liao River System (LRS) and the Daliao River System (DRS). Eighty-eight samples were collected and analyzed for P, Al, Fe, Ca, organic matter, pH, and texture. The results show that Fe can be used as a better particle-size proxy to construct the GBF of P (P (mg/ kg) ¼ 39.98 þ 166.19 Fe (%), R2 ¼ 0.835, n ¼ 66). The GBL of P was 675 mg/kg, while the average background level of P was 355 mg/kg. Noting that many large cities are located in the DRS watershed, most of the contaminated sites were located within the DRS and the riverbed sediments were more contaminated by P in the DRS watershed than in the LRS watershed. The geochemical background and baseline information of P are of great importance in managing P levels within the LRW. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Phosphorus Sediment Background Baseline Normalization Contamination
1. Introduction Phosphorus geochemistry is of agronomic and environmental importance. P is required to support aquatic plant growth and a key limiting nutrient in most aquatic and terrestrial ecosystems. However, excess P poses environmental risks because it may lead to the eutrophication of surface water bodies (Conley et al., 2009). P levels in aquatic environments have significantly increased over time due to fertilizer use as well as municipal and industrial wastewater. Many studies have investigated the content, chemical forms, adsorption, and release of P in river sediments (Ruban et al., 1999; Zhang and Huang, 2007; Wang et al., 2008; Zhang and Shan, 2008). In addition, watershed-scale P loadings in water bodies have been widely studied (Sigua and Tweedale, 2003; Rao et al., 2009; Vigiak et al., 2012). On the other hand, the geochemical baseline levels and functions of trace metals in sediments have been studied and developed (Loring, 1991; Daskalakis and O’Connor, 1995; Covelli and Fontolan, 1997; Newman and Watling, 2007). However, studies on the geochemical baseline level (GBL) and geochemical baseline function (GBF) of P in river sediments are limited. It is
* Corresponding author. Tel.: þ86 10 58802078; fax: þ86 10 5880397. E-mail address: [email protected] (C. Lin). 0301-4797/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jenvman.2013.05.012
hypothesized that the P content in sediments is correlated to the particle size distribution or content of particle-size proxy elements in native sediments, and this correlation function can be used as a GBF to distinguish between geogenic and anthropogenic P. Previous studies have often used Al and Fe as a particle-size proxy for trace elements (Schropp et al., 1990; Daskalakis and O’Connor, 1995; Schiff and Weisberg, 1999; Newman and Watling, 2007). The objectives of this study were to elucidate the P geochemical baseline level (GBL) and develop a P geochemical baseline function for sediments in the Liao River Watershed (LRW) to assess and manage P contamination. 2. Materials and methods 2.1. Description of the Liao River Watershed (LRW) The Liao River Watershed (LRW) in Liaoning Province of China consists of the Liao River System (LRS) and the Daliao River System (DRS) (Fig. 1). The RLS mainly drains agricultural areas, while the DRS drains large industrial, urban, and agricultural areas. Major industrial cities are situated near the DRS, including Fushun, Shenyang, Benxi, Anshan, Liaoyang, Panjin, and Yingkou. Over the past decades, large quantities of point and non-point P sources from urban areas, industry, and agriculture have been discharged into the river system. However, the quantitative assessment of P
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
139
Fig. 1. Schematic graph of the Liao River Watershed (LRW) showing Liao River System and Daliao River System, cities and counties, and river sediment sampling sites.
contamination in the LRW sediments is a challenge due to the lack of geochemical background or baseline levels and sediment heterogeneity. 2.2. Sediment sampling and analysis Eighty-eight surface sediment samples (ca. 0e15 cm depth) were collected from the LRW (Fig. 1), using Van Veen grabs (Eijkelkamp; a cable-operated sediment sampler), and then transferred to plastic bags. The sediment samples were freeze-dried in the laboratory, crushed slightly, passed through a 2 mm sieve, and stored in glass bottles. The content of sediment organic matter (OM) was measured by weight loss on ignition at 400 C (Ben-Dor and Banin, 1989). The granularity of the sediment samples was measured by a LS 230 laser diffraction particle analyzer (Beckman Coulter), and the percent contents of clay (<2 mm) and silt (2 mme20 mm) were calculated. The pH values of the sediment samples was analyzed in a 1:10 solid/liquid suspension using a combination pH electrode (Orion, U.S.). Portions of the soil samples were digested with HNO3eHFe HClO4 (Lin et al., 2008). The Al, Fe, Ca, and P contents in the extracts were measured by an inductively coupled plasma atomic emission spectrometer (ICP-AES) (IRIS Intrepid II, Thermo Electron). The accuracy of the method was assessed by replicate analyses of the elements in the certified reference materials GSS17 and GSS25 (Institute of Geophysical and Geochemical Exploration, Chinese
Academy of Geological Sciences), showing that the analytical error ranged from 1.44% to 9.11%. Data statistical analyses were conducted employing statistical soft program, JMPÒPro 10.0.0. First, the distribution histogram of elemental concentration was drawn to identify and remove the outliers. Second, the cumulative distribution function (CDF) of elemental concentration was plotted to identify the bend point cutting the data set into geogenic background and anthropogenically effected concentration sets. Third, the relationship between P and particle-size proxy elements (normalizers) Fe, Al, Ca, and OM for the geogenic background concentration data set was linearly fitted and data points outside 95% confidence band were removed. Finally, the relationship between P and particle-size proxy elements (normalizers) Fe, Al, Ca, and OM for the remaining data set was again linearly fitted to get the geochemical baseline functions (GBF). 3. Results and discussions 3.1. General characterization and elemental concentration distribution of sediments The LRW sediments showed textures with a predominant sandy fraction (80% on average) and low fractions of clay (4.8% on average) (Table 1). However, the sediment texture showed high spatial variation, with a maximal clay content of 19% and maximal sand content of 99.6%. Noting that the specific surface area of clay
140
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
Table 1 Elemental concentrations, organic matter content, pH, and texture of sediment samples in the Liao River Watershed (LRW). P mg/kg Liao River Watershed (LRW) Average 502.6 Median 419.5 Maximum 1975.0 Minimum 52.2 Global crust, shale, sediment, and soil Upper continental crust 665.0 Mean shale 700.0 Mean sediment 670.0 Median soil 800.0
Al %
Fe %
Ca %
OM %
pH
Clay %
Silt %
Sand %
6.02 6.64 12.23 1.85
2.56 1.93 17.91 0.28
1.40 1.32 5.44 0.21
1.84 1.39 8.63 0.01
7.93 8.12 9.10 5.79
4.83 3.55 19.71 0.09
14.58 13.64 42.43 0.24
80.58 82.78 99.65 42.04
7.74 8.80 7.20 7.10
3.09 4.80 4.10 4.00
1.61 1.90 1.67 0.60
particles is one thousand times higher than that of sand particles, the importance of this particle distribution lies in the resulting adsorption properties of the particles for phosphorus. The frequency distributions of elemental concentrations, pH, and OM are summarized in Fig. 2, showing some outliers for the P, Ca, Al, Fe, and OM contents in the sediments. The mean pH of the
sediments was 7.9 (ranging from 5.8 to 9.1), with the lowest values occurring at sampling sites 8 and 78, which are affected by acidic mining wastewater and industrial wastewater, respectively (Table 1, Fig. 1). There were only 9 sampling sites where the pH values of the sediments were below 7.0. The mean OM content was 1.8% (Table 1); OM content was greatest at sampling site 75 (8.6%),
Fig. 2. Distribution histograms of elemental concentrations in the sediments of the Liao River Watershed (n ¼ 88). The vertical lines within the boxes represent the median sample values. The ends of the boxes represent the 25th and 75th quantiles. The lines extending from the ends of boxes are referred to whiskers. Data outside of the range of the whiskers are considered outliers.
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
141
Fig. 3. Distribution histogram and cumulative distribution function (CDF) of P in the sediments plotted by the JMP program. See the explanation in the text.
which is close to the wastewater treatment plant of Benxi City, while values were lowest (<0.1%) in the upstream areas (sites 34 and 37) (Figs. 1 and 2). Low values were attributed to the high sand content of the sediments (>99%). Therefore, the sediments in the
Fig. 5. Scatter plots showing the relationship between P and grain proxy elements Fe and Al after removing the data points falling outside the 95% confidence bands in Fig. 4. The regression lines were computed using the JMP program.
Fig. 4. Scatter plots showing the relationship between P and potential grain-proxy elements. The regression line and 95% confidence band are reported as computed by the JMP program.
142
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
Liao River Watershed were generally characterized by a loamy sand texture, low OM, and alkaline pH. The concentration distribution histograms of P and possible particle-size proxies (Ca, Al, Fe, and OM) in the sediments in Fig. 2 demonstrate the abnormal distribution of these elemental concentrations. One of the major reasons for the abnormal distribution is outliers. The median concentration of P was 419.7 (52.2e1975.0) mg/kg in the LRW sediments, lower than the average content of P in the upper continent crust (665.0 mg/kg) and global sediments (670 mg/kg) (Bowen, 1979; Wedepohl, 1995). The low median concentration of P in the LRW sediments may be mainly due to the loamy sand texture of the sediments. However, the maximal concentration of P in the LRW sediments was approximately three times the P content in the upper continent crust and global sediments. The median concentrations of Ca, Al, and Fe in the LRW sediments were 1.32% (0.21e5.44%), 6.64% (1.85e12.12%), and 1.93% (0.28e17.91%), respectively, slightly lower than those in the upper continent and global sediments due to the loamy sandy texture of the sediments.
used to eliminate the extreme outliers (possibly seriously contaminated site data) and thus to reduce the original data set to a relatively clean data set. This data set shows a relatively normal distribution (Fig. 3(a)). The cumulative distribution function was plotted as a function of the P concentration following the statistical elimination of outliers (Fig. 3(b)). In general, the cumulative distribution function of the P concentration exhibits a linear change as a function of the elemental concentration with two slopes (Fig. 3(b)). The bend at the arrow cuts the data set into two parts: geogenic background and anthropogenically affected concentration sets (Matschullat et al., 2000). Similar to that for trace elements, the P concentration at the bend (the upper limit of the background content) is defined as the GBL. The results presented in Fig. 3(b) show that the GBL of P was approximately 675 mg/kg in the LRW sediments, while the average background concentration of P was approximately 355 mg/ kg in the LRW sediments.
3.2. Geochemical baseline level of P in the sediments
The use of a single geochemical baseline level has the disadvantage of not taking into account the natural geochemical variability of the sediments. The fitted geochemical baseline functions (GBF) of the elemental concentration against the particle-size proxy element concentration were used to normalize the geochemical data and distinguish between the natural and anthropogenic origins of the elements. To develop the GBFs, the
A cumulative distribution curve plotted as a function of elemental concentration is used to identify the geochemical baseline level (GBL) of trace elements in sediments (Matschullat et al., 2000). The same method was used to identify the GBL of P in the LRW sediments. The statistical tests presented in Fig. 2 were first
3.3. Geochemical baseline functions of P in the sediments
Fig. 6. Spatial distribution of sediment sample sites considered outliers and those considered to fall above or below the GBL of the P concentration in the Liao River Watershed (LRW).
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
element concentration should not be influenced by human activities and the covariation of the elemental concentration with the particle-size proxy element concentration should be statistically significant. The close relationship between the environmental fates of phosphorus, iron, calcium, and OM are well known (Brady and House, 1996), suggesting that there may be some correlations between their sediment concentrations if these processes are significant in a particular riverine environment. In addition, House and Denison (2002) used multiple regression analysis to develop the relationships that enabled the total phosphorus content of the sediment to be predicted from a knowledge of the total organic matter, total iron and calcium contents of the sediment for the rivers Wey, Great Ouse and Blackwater. Therefore, the data set with a P concentration below its GBL in the LRW sediments was used to develop the GBF of P, employing Al, Fe, Ca, and OM as potential particle-size proxies (Fig. 4). The correlation coefficients between the proxy elements and P were 0.330 for Al, 0.776 for Fe, 0.193 for Ca, and 0.231 for OM. Though all relationships were statistically significant at p < 0.0001, the correlation coefficient between P and Fe was much higher than that between P and Al, Ca, and OM. The points that fall inside the 95% confidence band are characterized as natural sediments, while a few sampling points that fall outside of the 95% confidence limits are due to the ratio of outliers of P to proxy elements. Thus, the data points falling outside of the 95% confidence bands were removed, and the remaining data points were used to develop the GBF to improve the correlation coefficients (Fig. 5). The correlation coefficients for the formula increased to 0.582 for Al and 0.835 for Fe. Therefore, Fe is the best particle-size proxy element for P. The detailed GBFs of P in the sediments are:
P ðmg=kgÞ ¼ 211:10 þ 95:01 Alð%Þ; R2 ¼ 0:582; n ¼ 69 (1) P ðmg=kgÞ ¼ 39:98 þ 166:19 Feð%Þ; R2 ¼ 0:835; n ¼ 66 (2)
3.4. Spatial distribution of P contamination in the sediments The number of outliers was 5 for the P concentration in the LRW sediments (Fig. 2). The sampling sites with outliers are considered extremely contaminated. In addition, the number of samples with P concentrations above the GBL was 11. These sampling sites are considered less contaminated than the outlier sampling sites. Most of the contaminated sites were located within the Daliao River System (DRS) (Fig. 6). In addition, they are adjacent to large cities (e.g., sites 63, 74, 75, and 82) (Fig. 1), indicate that effluent discharge from large cities might lead to an increase in P content in their adjacent river sediments. However, though sites 8 and 24 are upstream of the LRS, the P content in the sediments in those areas was high (outliers), most likely due to local agricultural activities. As indicated in the introduction, Liaoning Province is a major industrial base in China and most of the surrounding large cities are adjacent to the DRS. Therefore, the sediments near the large city areas were generally more contaminated by P. 4. Conclusions The GBLs and GBFs of P in the watershed sediments can be calculated and developed using statistical and geochemical methods. First, concentration distribution histograms were used to identify outliers. Second, the cumulative distribution function was
143
used to determine the GBL after the removal of outliers. Third, the GBF was developed by fitting the P concentration to the potential particle-size proxy element concentration. This work confirms that Fe can be used as a better particle-size proxy element to construct the GBF of P. The GBLs of P were 675 mg/kg, while the average background level of P was 355 mg/kg. Noting that most surrounding large cities are located in the DLR system, the sediments in the DLR were more contaminated by P. Acknowledgments This study was supported by the National Natural Science Foundation of China (40971058, 40930740) and National Science and Technology Infrastructure Program (2012BAD15B05). References Ben-Dor, E., Banin, A., 1989. Determination of organic matter content in arid-zone soils using a simple ‘‘loss-on-ignition’’ method. Communication in Soil Science and Plant Analysis 201, 1675e1695. Bowen, H.J.M., 1979. Environmental Chemistry of the Elements. Academic Press, London. Brady, P.V., House, W.A., 1996. Surface-controlled dissolution and growth of minerals. In: Brady, P.V. (Ed.), Physics and Chemistry of Mineral Surfaces. CRC Press, London, pp. 225e307. Conley, D.J., Paerl, H.W., Howarth, R.W., Boesch, D.F., Seitzinger, S.P., Havens, K.E., Lancelot, C., Likens, G.E., 2009. Controlling eutrophication: nitrogen and phosphorus. Science 323, 1014e1015. Covelli, S., Fontolan, G., 1997. Application of a normalization procedure in determining regional geochemical baselines. Environmental Geology 30, 34e44. Daskalakis, K.D., O’Connor, T.P., 1995. Normalization and elemental sediment contamination in the coastal United States. Environmental Science and Technology 29, 470e477. House, W.A., Denison, F.H., 2002. Total phosphorus content of river sediments in relationship to calcium, iron and organic matter concentrations. The Science of the Total Environment 282, 341e351. Lin, C., He, M.C., Zhou, Y.X., Guo, W., Yang, Z.F., 2008. Distribution and contamination assessment of heavy metals in sediment of the Second Songhua River, China. Environmental Monitoring and Assessment 137, 329e342. Loring, D.H., 1991. Normalization of heavy-metal data from estuarine and coastal sediments. ICES Journal of Marine Science 48, 101e115. Matschullat, J., Ottenstein, R., Reimann, C., 2000. Geochemical background e can we calculate it? Environmental Geology 39, 990e1000. Newman, B.K., Watling, R.J., 2007. Definition of baseline metal concentrations for assessing metal enrichment of sediment from the south-eastern Cape coastline of South Africa. Water SA 33, 675e691. Rao, N.S., Easton, Z.M., Schneiderman, E.M., Zion, M.S., Lee, D.R., Steenhuis, T.S., 2009. Modeling watershed-scale effectiveness of agricultural best management practices to reduce phosphorus loading. Journal of Environmental Management 90, 1385e1395. Ruban, V., Lopez-Sanchez, J.F., Pardo, P., Rauret, G., Muntauc, H., Quevauviller, P., 1999. Selection and evaluation of sequential extraction procedures for the determination of phosphorus forms in lake sediment. Journal of Environmental Monitoring 1, 51e56. Schiff, K.C., Weisberg, S.B., 1999. Iron as a reference element for determining trace metal enrichment in Southern California coastal shelf sediments. Marine Environment Research 48, 161e176. Schropp, S.J., Lewis, F.G., Windom, H.L., Ryan, J.D., Calder, F.D., Burney, L.C., 1990. Interpretation of metal concentrations in estuarine sediments of Florida using aluminum as a reference element. Estuaries 13, 227e235. Sigua, G.C., Tweedale, W.A., 2003. Watershed scale assessment of nitrogen and phosphorus loadings in the Indian River Lagoon basin, Florida. Journal of Environmental Management 67, 363e372. Vigiak, O., Rattray, D., McInnes, J., Newham, L.T.H., Roberts, A.M., 2012. Modelling catchment management impact on in-stream phosphorus loads in northern Victoria. Journal of Environmental Management 110, 215e225. Wang, S.R., Jin, X.C., Zhao, H.C., Zhou, X.N., Wu, F.C., 2008. Effects of organic matter on phosphorus release kinetics in different trophic lake sediments and application of transition state theory. Journal of Environmental Management 88, 845e852. Wedepohl, K.H., 1995. The composition of the continental crust. Geochimica et Cosmochimica Acta 59, 1217e1239. Zhang, H., Shan, B.Q., 2008. Historical distribution and partitioning of phosphorus in sediments in an agricultural watershed in the Yangtze-Huaihe region, China. Environmental Science and Technology 42, 2328e2333. Zhang, J.Z., Huang, X.L., 2007. Relative importance of solid-phase phosphorus and iron on the sorption behavior of sediments. Environmental Science and Technology 41, 2789e2795.
Contents lists available at SciVerse ScienceDirect
Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman
Geochemical baseline level and function and contamination of phosphorus in Liao River Watershed sediments of China Shaoqing Liu a, Jing Wang b, Chunye Lin a, *, Mengchang He a, Xitao Liu a a b
State Key Joint Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China China Land Surveying and Planning Institute, Key Laboratory of Land Use, Ministry of Land and Resources, Beijing 100035, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 January 2013 Received in revised form 4 April 2013 Accepted 8 May 2013 Available online 1 June 2013
The quantitative assessment of P contamination in sediments is a challenge due to sediment heterogeneity and the lacking of geochemical background or baseline levels. In this study, a procedure was proposed to determine the average P background level and P geochemical baseline level (GBL) and develop P geochemical baseline functions (GBF) for riverbed sediments of the Liao River Watershed (LRW). The LRW has two river systems e the Liao River System (LRS) and the Daliao River System (DRS). Eighty-eight samples were collected and analyzed for P, Al, Fe, Ca, organic matter, pH, and texture. The results show that Fe can be used as a better particle-size proxy to construct the GBF of P (P (mg/ kg) ¼ 39.98 þ 166.19 Fe (%), R2 ¼ 0.835, n ¼ 66). The GBL of P was 675 mg/kg, while the average background level of P was 355 mg/kg. Noting that many large cities are located in the DRS watershed, most of the contaminated sites were located within the DRS and the riverbed sediments were more contaminated by P in the DRS watershed than in the LRS watershed. The geochemical background and baseline information of P are of great importance in managing P levels within the LRW. Ó 2013 Elsevier Ltd. All rights reserved.
Keywords: Phosphorus Sediment Background Baseline Normalization Contamination
1. Introduction Phosphorus geochemistry is of agronomic and environmental importance. P is required to support aquatic plant growth and a key limiting nutrient in most aquatic and terrestrial ecosystems. However, excess P poses environmental risks because it may lead to the eutrophication of surface water bodies (Conley et al., 2009). P levels in aquatic environments have significantly increased over time due to fertilizer use as well as municipal and industrial wastewater. Many studies have investigated the content, chemical forms, adsorption, and release of P in river sediments (Ruban et al., 1999; Zhang and Huang, 2007; Wang et al., 2008; Zhang and Shan, 2008). In addition, watershed-scale P loadings in water bodies have been widely studied (Sigua and Tweedale, 2003; Rao et al., 2009; Vigiak et al., 2012). On the other hand, the geochemical baseline levels and functions of trace metals in sediments have been studied and developed (Loring, 1991; Daskalakis and O’Connor, 1995; Covelli and Fontolan, 1997; Newman and Watling, 2007). However, studies on the geochemical baseline level (GBL) and geochemical baseline function (GBF) of P in river sediments are limited. It is
* Corresponding author. Tel.: þ86 10 58802078; fax: þ86 10 5880397. E-mail address: [email protected] (C. Lin). 0301-4797/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jenvman.2013.05.012
hypothesized that the P content in sediments is correlated to the particle size distribution or content of particle-size proxy elements in native sediments, and this correlation function can be used as a GBF to distinguish between geogenic and anthropogenic P. Previous studies have often used Al and Fe as a particle-size proxy for trace elements (Schropp et al., 1990; Daskalakis and O’Connor, 1995; Schiff and Weisberg, 1999; Newman and Watling, 2007). The objectives of this study were to elucidate the P geochemical baseline level (GBL) and develop a P geochemical baseline function for sediments in the Liao River Watershed (LRW) to assess and manage P contamination. 2. Materials and methods 2.1. Description of the Liao River Watershed (LRW) The Liao River Watershed (LRW) in Liaoning Province of China consists of the Liao River System (LRS) and the Daliao River System (DRS) (Fig. 1). The RLS mainly drains agricultural areas, while the DRS drains large industrial, urban, and agricultural areas. Major industrial cities are situated near the DRS, including Fushun, Shenyang, Benxi, Anshan, Liaoyang, Panjin, and Yingkou. Over the past decades, large quantities of point and non-point P sources from urban areas, industry, and agriculture have been discharged into the river system. However, the quantitative assessment of P
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
139
Fig. 1. Schematic graph of the Liao River Watershed (LRW) showing Liao River System and Daliao River System, cities and counties, and river sediment sampling sites.
contamination in the LRW sediments is a challenge due to the lack of geochemical background or baseline levels and sediment heterogeneity. 2.2. Sediment sampling and analysis Eighty-eight surface sediment samples (ca. 0e15 cm depth) were collected from the LRW (Fig. 1), using Van Veen grabs (Eijkelkamp; a cable-operated sediment sampler), and then transferred to plastic bags. The sediment samples were freeze-dried in the laboratory, crushed slightly, passed through a 2 mm sieve, and stored in glass bottles. The content of sediment organic matter (OM) was measured by weight loss on ignition at 400 C (Ben-Dor and Banin, 1989). The granularity of the sediment samples was measured by a LS 230 laser diffraction particle analyzer (Beckman Coulter), and the percent contents of clay (<2 mm) and silt (2 mme20 mm) were calculated. The pH values of the sediment samples was analyzed in a 1:10 solid/liquid suspension using a combination pH electrode (Orion, U.S.). Portions of the soil samples were digested with HNO3eHFe HClO4 (Lin et al., 2008). The Al, Fe, Ca, and P contents in the extracts were measured by an inductively coupled plasma atomic emission spectrometer (ICP-AES) (IRIS Intrepid II, Thermo Electron). The accuracy of the method was assessed by replicate analyses of the elements in the certified reference materials GSS17 and GSS25 (Institute of Geophysical and Geochemical Exploration, Chinese
Academy of Geological Sciences), showing that the analytical error ranged from 1.44% to 9.11%. Data statistical analyses were conducted employing statistical soft program, JMPÒPro 10.0.0. First, the distribution histogram of elemental concentration was drawn to identify and remove the outliers. Second, the cumulative distribution function (CDF) of elemental concentration was plotted to identify the bend point cutting the data set into geogenic background and anthropogenically effected concentration sets. Third, the relationship between P and particle-size proxy elements (normalizers) Fe, Al, Ca, and OM for the geogenic background concentration data set was linearly fitted and data points outside 95% confidence band were removed. Finally, the relationship between P and particle-size proxy elements (normalizers) Fe, Al, Ca, and OM for the remaining data set was again linearly fitted to get the geochemical baseline functions (GBF). 3. Results and discussions 3.1. General characterization and elemental concentration distribution of sediments The LRW sediments showed textures with a predominant sandy fraction (80% on average) and low fractions of clay (4.8% on average) (Table 1). However, the sediment texture showed high spatial variation, with a maximal clay content of 19% and maximal sand content of 99.6%. Noting that the specific surface area of clay
140
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
Table 1 Elemental concentrations, organic matter content, pH, and texture of sediment samples in the Liao River Watershed (LRW). P mg/kg Liao River Watershed (LRW) Average 502.6 Median 419.5 Maximum 1975.0 Minimum 52.2 Global crust, shale, sediment, and soil Upper continental crust 665.0 Mean shale 700.0 Mean sediment 670.0 Median soil 800.0
Al %
Fe %
Ca %
OM %
pH
Clay %
Silt %
Sand %
6.02 6.64 12.23 1.85
2.56 1.93 17.91 0.28
1.40 1.32 5.44 0.21
1.84 1.39 8.63 0.01
7.93 8.12 9.10 5.79
4.83 3.55 19.71 0.09
14.58 13.64 42.43 0.24
80.58 82.78 99.65 42.04
7.74 8.80 7.20 7.10
3.09 4.80 4.10 4.00
1.61 1.90 1.67 0.60
particles is one thousand times higher than that of sand particles, the importance of this particle distribution lies in the resulting adsorption properties of the particles for phosphorus. The frequency distributions of elemental concentrations, pH, and OM are summarized in Fig. 2, showing some outliers for the P, Ca, Al, Fe, and OM contents in the sediments. The mean pH of the
sediments was 7.9 (ranging from 5.8 to 9.1), with the lowest values occurring at sampling sites 8 and 78, which are affected by acidic mining wastewater and industrial wastewater, respectively (Table 1, Fig. 1). There were only 9 sampling sites where the pH values of the sediments were below 7.0. The mean OM content was 1.8% (Table 1); OM content was greatest at sampling site 75 (8.6%),
Fig. 2. Distribution histograms of elemental concentrations in the sediments of the Liao River Watershed (n ¼ 88). The vertical lines within the boxes represent the median sample values. The ends of the boxes represent the 25th and 75th quantiles. The lines extending from the ends of boxes are referred to whiskers. Data outside of the range of the whiskers are considered outliers.
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
141
Fig. 3. Distribution histogram and cumulative distribution function (CDF) of P in the sediments plotted by the JMP program. See the explanation in the text.
which is close to the wastewater treatment plant of Benxi City, while values were lowest (<0.1%) in the upstream areas (sites 34 and 37) (Figs. 1 and 2). Low values were attributed to the high sand content of the sediments (>99%). Therefore, the sediments in the
Fig. 5. Scatter plots showing the relationship between P and grain proxy elements Fe and Al after removing the data points falling outside the 95% confidence bands in Fig. 4. The regression lines were computed using the JMP program.
Fig. 4. Scatter plots showing the relationship between P and potential grain-proxy elements. The regression line and 95% confidence band are reported as computed by the JMP program.
142
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
Liao River Watershed were generally characterized by a loamy sand texture, low OM, and alkaline pH. The concentration distribution histograms of P and possible particle-size proxies (Ca, Al, Fe, and OM) in the sediments in Fig. 2 demonstrate the abnormal distribution of these elemental concentrations. One of the major reasons for the abnormal distribution is outliers. The median concentration of P was 419.7 (52.2e1975.0) mg/kg in the LRW sediments, lower than the average content of P in the upper continent crust (665.0 mg/kg) and global sediments (670 mg/kg) (Bowen, 1979; Wedepohl, 1995). The low median concentration of P in the LRW sediments may be mainly due to the loamy sand texture of the sediments. However, the maximal concentration of P in the LRW sediments was approximately three times the P content in the upper continent crust and global sediments. The median concentrations of Ca, Al, and Fe in the LRW sediments were 1.32% (0.21e5.44%), 6.64% (1.85e12.12%), and 1.93% (0.28e17.91%), respectively, slightly lower than those in the upper continent and global sediments due to the loamy sandy texture of the sediments.
used to eliminate the extreme outliers (possibly seriously contaminated site data) and thus to reduce the original data set to a relatively clean data set. This data set shows a relatively normal distribution (Fig. 3(a)). The cumulative distribution function was plotted as a function of the P concentration following the statistical elimination of outliers (Fig. 3(b)). In general, the cumulative distribution function of the P concentration exhibits a linear change as a function of the elemental concentration with two slopes (Fig. 3(b)). The bend at the arrow cuts the data set into two parts: geogenic background and anthropogenically affected concentration sets (Matschullat et al., 2000). Similar to that for trace elements, the P concentration at the bend (the upper limit of the background content) is defined as the GBL. The results presented in Fig. 3(b) show that the GBL of P was approximately 675 mg/kg in the LRW sediments, while the average background concentration of P was approximately 355 mg/ kg in the LRW sediments.
3.2. Geochemical baseline level of P in the sediments
The use of a single geochemical baseline level has the disadvantage of not taking into account the natural geochemical variability of the sediments. The fitted geochemical baseline functions (GBF) of the elemental concentration against the particle-size proxy element concentration were used to normalize the geochemical data and distinguish between the natural and anthropogenic origins of the elements. To develop the GBFs, the
A cumulative distribution curve plotted as a function of elemental concentration is used to identify the geochemical baseline level (GBL) of trace elements in sediments (Matschullat et al., 2000). The same method was used to identify the GBL of P in the LRW sediments. The statistical tests presented in Fig. 2 were first
3.3. Geochemical baseline functions of P in the sediments
Fig. 6. Spatial distribution of sediment sample sites considered outliers and those considered to fall above or below the GBL of the P concentration in the Liao River Watershed (LRW).
S. Liu et al. / Journal of Environmental Management 128 (2013) 138e143
element concentration should not be influenced by human activities and the covariation of the elemental concentration with the particle-size proxy element concentration should be statistically significant. The close relationship between the environmental fates of phosphorus, iron, calcium, and OM are well known (Brady and House, 1996), suggesting that there may be some correlations between their sediment concentrations if these processes are significant in a particular riverine environment. In addition, House and Denison (2002) used multiple regression analysis to develop the relationships that enabled the total phosphorus content of the sediment to be predicted from a knowledge of the total organic matter, total iron and calcium contents of the sediment for the rivers Wey, Great Ouse and Blackwater. Therefore, the data set with a P concentration below its GBL in the LRW sediments was used to develop the GBF of P, employing Al, Fe, Ca, and OM as potential particle-size proxies (Fig. 4). The correlation coefficients between the proxy elements and P were 0.330 for Al, 0.776 for Fe, 0.193 for Ca, and 0.231 for OM. Though all relationships were statistically significant at p < 0.0001, the correlation coefficient between P and Fe was much higher than that between P and Al, Ca, and OM. The points that fall inside the 95% confidence band are characterized as natural sediments, while a few sampling points that fall outside of the 95% confidence limits are due to the ratio of outliers of P to proxy elements. Thus, the data points falling outside of the 95% confidence bands were removed, and the remaining data points were used to develop the GBF to improve the correlation coefficients (Fig. 5). The correlation coefficients for the formula increased to 0.582 for Al and 0.835 for Fe. Therefore, Fe is the best particle-size proxy element for P. The detailed GBFs of P in the sediments are:
P ðmg=kgÞ ¼ 211:10 þ 95:01 Alð%Þ; R2 ¼ 0:582; n ¼ 69 (1) P ðmg=kgÞ ¼ 39:98 þ 166:19 Feð%Þ; R2 ¼ 0:835; n ¼ 66 (2)
3.4. Spatial distribution of P contamination in the sediments The number of outliers was 5 for the P concentration in the LRW sediments (Fig. 2). The sampling sites with outliers are considered extremely contaminated. In addition, the number of samples with P concentrations above the GBL was 11. These sampling sites are considered less contaminated than the outlier sampling sites. Most of the contaminated sites were located within the Daliao River System (DRS) (Fig. 6). In addition, they are adjacent to large cities (e.g., sites 63, 74, 75, and 82) (Fig. 1), indicate that effluent discharge from large cities might lead to an increase in P content in their adjacent river sediments. However, though sites 8 and 24 are upstream of the LRS, the P content in the sediments in those areas was high (outliers), most likely due to local agricultural activities. As indicated in the introduction, Liaoning Province is a major industrial base in China and most of the surrounding large cities are adjacent to the DRS. Therefore, the sediments near the large city areas were generally more contaminated by P. 4. Conclusions The GBLs and GBFs of P in the watershed sediments can be calculated and developed using statistical and geochemical methods. First, concentration distribution histograms were used to identify outliers. Second, the cumulative distribution function was
143
used to determine the GBL after the removal of outliers. Third, the GBF was developed by fitting the P concentration to the potential particle-size proxy element concentration. This work confirms that Fe can be used as a better particle-size proxy element to construct the GBF of P. The GBLs of P were 675 mg/kg, while the average background level of P was 355 mg/kg. Noting that most surrounding large cities are located in the DLR system, the sediments in the DLR were more contaminated by P. Acknowledgments This study was supported by the National Natural Science Foundation of China (40971058, 40930740) and National Science and Technology Infrastructure Program (2012BAD15B05). References Ben-Dor, E., Banin, A., 1989. Determination of organic matter content in arid-zone soils using a simple ‘‘loss-on-ignition’’ method. Communication in Soil Science and Plant Analysis 201, 1675e1695. Bowen, H.J.M., 1979. Environmental Chemistry of the Elements. Academic Press, London. Brady, P.V., House, W.A., 1996. Surface-controlled dissolution and growth of minerals. In: Brady, P.V. (Ed.), Physics and Chemistry of Mineral Surfaces. CRC Press, London, pp. 225e307. Conley, D.J., Paerl, H.W., Howarth, R.W., Boesch, D.F., Seitzinger, S.P., Havens, K.E., Lancelot, C., Likens, G.E., 2009. Controlling eutrophication: nitrogen and phosphorus. Science 323, 1014e1015. Covelli, S., Fontolan, G., 1997. Application of a normalization procedure in determining regional geochemical baselines. Environmental Geology 30, 34e44. Daskalakis, K.D., O’Connor, T.P., 1995. Normalization and elemental sediment contamination in the coastal United States. Environmental Science and Technology 29, 470e477. House, W.A., Denison, F.H., 2002. Total phosphorus content of river sediments in relationship to calcium, iron and organic matter concentrations. The Science of the Total Environment 282, 341e351. Lin, C., He, M.C., Zhou, Y.X., Guo, W., Yang, Z.F., 2008. Distribution and contamination assessment of heavy metals in sediment of the Second Songhua River, China. Environmental Monitoring and Assessment 137, 329e342. Loring, D.H., 1991. Normalization of heavy-metal data from estuarine and coastal sediments. ICES Journal of Marine Science 48, 101e115. Matschullat, J., Ottenstein, R., Reimann, C., 2000. Geochemical background e can we calculate it? Environmental Geology 39, 990e1000. Newman, B.K., Watling, R.J., 2007. Definition of baseline metal concentrations for assessing metal enrichment of sediment from the south-eastern Cape coastline of South Africa. Water SA 33, 675e691. Rao, N.S., Easton, Z.M., Schneiderman, E.M., Zion, M.S., Lee, D.R., Steenhuis, T.S., 2009. Modeling watershed-scale effectiveness of agricultural best management practices to reduce phosphorus loading. Journal of Environmental Management 90, 1385e1395. Ruban, V., Lopez-Sanchez, J.F., Pardo, P., Rauret, G., Muntauc, H., Quevauviller, P., 1999. Selection and evaluation of sequential extraction procedures for the determination of phosphorus forms in lake sediment. Journal of Environmental Monitoring 1, 51e56. Schiff, K.C., Weisberg, S.B., 1999. Iron as a reference element for determining trace metal enrichment in Southern California coastal shelf sediments. Marine Environment Research 48, 161e176. Schropp, S.J., Lewis, F.G., Windom, H.L., Ryan, J.D., Calder, F.D., Burney, L.C., 1990. Interpretation of metal concentrations in estuarine sediments of Florida using aluminum as a reference element. Estuaries 13, 227e235. Sigua, G.C., Tweedale, W.A., 2003. Watershed scale assessment of nitrogen and phosphorus loadings in the Indian River Lagoon basin, Florida. Journal of Environmental Management 67, 363e372. Vigiak, O., Rattray, D., McInnes, J., Newham, L.T.H., Roberts, A.M., 2012. Modelling catchment management impact on in-stream phosphorus loads in northern Victoria. Journal of Environmental Management 110, 215e225. Wang, S.R., Jin, X.C., Zhao, H.C., Zhou, X.N., Wu, F.C., 2008. Effects of organic matter on phosphorus release kinetics in different trophic lake sediments and application of transition state theory. Journal of Environmental Management 88, 845e852. Wedepohl, K.H., 1995. The composition of the continental crust. Geochimica et Cosmochimica Acta 59, 1217e1239. Zhang, H., Shan, B.Q., 2008. Historical distribution and partitioning of phosphorus in sediments in an agricultural watershed in the Yangtze-Huaihe region, China. Environmental Science and Technology 42, 2328e2333. Zhang, J.Z., Huang, X.L., 2007. Relative importance of solid-phase phosphorus and iron on the sorption behavior of sediments. Environmental Science and Technology 41, 2789e2795.