Source apportionment and source-to-sink transport of major and trace elements in coastal sediments: Combining positive matrix factorization and sediment trend analysis

https://doi.org/10.1016/j.scitotenv.2018.09.198Get rights and content

Highlights

  • Elemental concentrations and enrichment were assessed in coastal marine sediments.

  • PMF was used for source apportionment and seven potential sources were identified.

  • Net sediment transport patterns were established using the STA model.

  • Patterns of elements' contribution coincide with source-to-sink transport processes.

  • Transport of sediment as well as elements is interpreted in terms of hydrodynamics.

Abstract

Major and trace elemental concentrations in coastal marine sediments were incorporated into positive matrix factorization (PMF) to identify potential sources and source contributions. Transport pathways of fine-grained sediments and sediment-bound elements were inferred from sediment trend analysis (STA). The spatial distribution patterns of 21 elements (Co, Cu, Ni, Sr, Zn, V, Ba, Sc, Ga, Pb, Cr, Zr, SiO2, Al2O3, Fe2O3, MgO, CaO, K2O, MnO, TiO2, and P2O5) coupled with sediment grain sizes were investigated. The natural and anthropogenic sources of the elements were distinguished by their medium enrichment factors (EFs). Seven sources were recognized by PMF: weathering products, anthropologic emissions, sand, older sediment, biogenic carbonates, products of siliceous organisms, and mine exploitation. Some land-derived elements, including weathering products, anthropogenic-related elements, and mining-related elements, had a significant positive correlation with sediment silt, clay, and organic carbon contents. The spatial patterns of the land-derived elements' concentrations and source contributions were consistent with the sediment transport pathways inferred from the STA. This result revealed that the delivery of the land-derived elements was determined by marine current flows and the associated sediment transport processes. Conversely, elements originating from marine sources, such as sand and older sediment, and from the biological activities of calcareous and siliceous organisms showed little response to sediment transport and deposition processes. Our study links the outputs of statistically oriented approaches (e.g., PMF) to a process-based understanding of elemental transport in marine environments.

Introduction

Trace metal contamination in estuaries and coastal areas has recently received increasing attention because it may harm humans, living resources, and marine ecosystems. Metals accumulate in sediments and can be absorbed by benthic organisms or released back into the water, causing water quality degradation (Pan and Wang, 2012). The sources of trace metal contaminants in coastal sediments are mostly based on land, while a small fraction of these contaminants originate from marine-based sources (Weis, 2014). Understanding the sources of trace metals and the related characteristics may provide useful information for policy makers to effectively manage coastal resources and habitats.

Multivariate statistical approaches, such as cluster analysis (CA), principal component analysis (PCA), and PCA-related methods, are usually used to infer potential sources from geochemical properties of sediments (Jiang et al., 2017; Owens et al., 2016). Factor loadings and factor score matrices obtained from PCA are interpreted as source composition and source contributions, respectively. These loadings and scores can be positive or negative. Since negative compositions and negative contributions do not actually exist, it is difficult to quantitatively assess the relative contributions from the sources identified.

An alternative approach is imposing nonnegativity constraints on the resulting factors to avoid the effect of the negative values (Paterson et al., 1999). Positive matrix factorization (PMF) is an advanced factor analytical method that fits the nonnegativity criterion (Paatero and Tapper, 1994; Vaccaro et al., 2007). PMF can efficiently handle measurement uncertainties, outliers, values below measurement detection limits, and missing values (Hemann et al., 2009), thus offering a favorable method for environmental applications. Individual data points are weighted so that both small and large data are modeled correctly, and in particular, outliers do not overly influence the forming factors (Bzdusek and Christensen, 2006; Bzdusek et al., 2006). By contrast, outliers are treated as noisy data or non-representative data and are sometimes omitted from PCA. In addition, PMF factors do not need to be orthogonal, which is important for linking the modeled factors to the real sources (Hemann et al., 2009). Numerous studies have quantitatively inferred the sources of trace metals within a wide range of materials, such as atmospheric sediment (i.e., dust and airborne particulate matter), soil, and river deposition (Comero et al., 2012; Comero et al., 2014; Hsu et al., 2016; Jiang et al., 2015; Mazzei et al., 2008; Praẑnikar et al., 2014; Tecer et al., 2012; Zhang et al., 2018).

Once trace elements are released into an aquatic system, they adhere to suspended fine-grained sediment particles (Salomons and Stigliani, 1995) or incorporate into the organic matter fraction (Hirner et al., 1990). Suspended particles are entrained in currents and transported in the same direction as the currents. To obtain further insight into the transport properties of sediments and sediment-bound trace metals, sediment trend analysis (STA) is performed. This method attempts to infer the net sediment transport pattern using grain size trends (McLaren, 1981; McLaren et al., 2007). In most cases, STA reflects the movement of fine-grained sediment in a hydrological system since finer (less dense) grains are preferentially mobilized and transported. Our previous works (Li et al., 2015a; Li et al., 2015b; Li et al., 2013; Li et al., 2014) have demonstrated that there is a direct connection between high trace metal concentrations (e.g., Ba, Co, Cr, Cu, Ni, Pb, V, and Zn) and high percentages of fine-grained sediments, particularly those in silt and clay fractions. Notable decreases in the concentrations of sediment-associated metals occur along the sediment transport pathways. This finding suggests that sediment transport processes have implications for trace metal delivery.

The middle and southern Fujian coastal area, located on the west side of the Taiwan Strait, is an important farming area in Fujian Province. This area faces increasing metal pollution problems due to the rapid industrialization and urbanization that has occurred over the past two decades. Metal pollution may threaten seafood safety, and it was reported that cultured-shellfish products from this area are contaminated with heavy metals (Xi et al., 2017). Many studies have been conducted on the concentrations, distribution, and ecological risk of trace metals from surface sediments in a number of bays (and estuaries) along the middle and southern Fujian coast (Cai, 2010; Cai et al., 2007; Chen, 2008; Chen et al., 2014; Huo et al., 2015; Li et al., 2008; Li et al., 2010; Lin, 2008; Lin, 2012). In these studies, trace metals were assumed to originate from rock weathering and agricultural, mining, and industrial processes. Nevertheless, the compositions and quantitative contributions of the potential sources are not well documented.

Although much research has been invested in identifying and quantifying sources of trace metal contamination, only a few studies have been conducted on estuarine and coastal sediments. In this study, we performed PMF to (1) distinguish the different sources of 21 trace and major elements in the middle and southern Fujian coastal sediments and (2) quantitatively evaluate the composition and relative contributions of each source. Some of the 21 elements, e.g. toxic metals, may be enriched in seafood and enter the human body through the food chain, posing a great threat to human health. Others compounds, such as land-derived nutrients (P2O5), can potentially cause eutrophication and adverse ecological effects. To specifically understand how sediment properties change during transport processes, changes in source contributions were linked to the net sediment transport pattern derived from STA. To our knowledge, this combination of PMF and STA has not been used for the interpretation of the transport dynamics of metals in aquatic settings, and we believe that our approach has valuable applications.

Section snippets

Study area

The study area is situated on the western part of the Taiwan Strait (Fig. 1). The coastline morphology is oriented NW–SE, with the northern section facing N–W. The topography of the sea bottom coincides with the coastline, and the isobaths are generally parallel to the coast (Fig. 1). The geographic and hydrodynamic characteristics of this area have been documented elsewhere (Li et al., 2018). Here, we highlight the river hydrology in the adjacent hinterlands, and most of our data are collected

Sampling and measurements

Surface sediment samples were collected at 1172 sites during four separate marine surveys from 2012 to 2016, and the water depths ranged from 2 m to 80 m (Fig. 1). The sampling locations were determined by a differential global positioning system (DGPS). Sediments were obtained by a stainless steel Petersen grab sampler in shallow water (<40 m) or by a box corer in deep water (>40 m). Only the top ~5 cm of sediment was sampled for laboratory analyses.

Sediments for grain size measurements were

Sediment grain size and organic carbon

The percentages of sediment sand, silt, clay, and organic carbon were determined in this work and their pattern maps are shown in Fig. 2. The continental shelf is predominated by sand with values in the range of 60–100% at most sites. Elevated percentages were also observed near the mouths of the Jiulongjiang and Minjiang Rivers (mean value of 55%). Sand percentages are mostly lower than 15% in the nearshore and coastal areas. Generally, sand percentages increased with increasing water depth

Conclusions

A total of 21 major and trace elements were measured in 1172 surface sediments collected from the middle and southern Fujian coastal areas. The calculated EF values revealed that Ni, Zn, V, Sc, Ga, Pb, and Cr were mainly due to anthropogenic activities, while SiO2, Fe2O3, MgO, CaO, K2O, MnO, TiO2, and P2O5 resulted from natural sources. PMF analysis was performed to identify seven potential sources: weathering products (MnO, K2O, Al2O3, TiO2, MgO, and Fe2O3), anthropologic emissions (Sc, Ni,

Acknowledgments

This work was supported by the National Natural Science Foundation of China (41776112) and the Chinese Special Survey of Marine Geology (DD20160140, DD20189642, DD20160138, GZH201200501, GZH201400208, GZH201400207, 1212011220118). We would like to thank the crew of the marine survey cruise of the western Taiwan Strait for their assistance with sample collection. Special thanks are due to Weiguo Wang of the Third Institute of Oceanography, State Oceanic Administration, for his help in the

References (88)

  • C.-Y. Hsu et al.

    Elemental characterization and source apportionment of PM10 and PM2.5 in the western coastal area of central Taiwan

    Sci. Total Environ.

    (2016)
  • S.Y. Jiang et al.

    Source apportionment and water solubility of metals in size segregated particles in urban environments

    Sci. Total Environ.

    (2015)
  • Y. Jiang et al.

    Source apportionment and health risk assessment of heavy metals in soil for a township in Jiangsu Province, China

    Chemosphere

    (2017)
  • A.P. Karageorgis et al.

    Geochemistry and mineralogy of the NW Aegean Sea surface sediments: implications for river runoff and anthropogenic impact

    Appl. Geochem.

    (2005)
  • T. Li et al.

    Assessment of the impacts of trace metals on benthic foraminifera in surface sediments from the northwestern Taiwan Strait

    Mar. Pollut. Bull.

    (2015)
  • T. Li et al.

    Distribution of trace metals and the benthic foraminiferal assemblage as a characterization of the environment in the north Minjiang River Estuary (Fujian, China)

    Mar. Pollut. Bull.

    (2015)
  • T. Li et al.

    Using self-organizing map for coastal water quality classification: towards a better understanding of patterns and processes

    Sci. Total Environ.

    (2018)
  • T. Li et al.

    Benthic foraminiferal assemblages and trace metals reveal the environment outside the Pearl River Estuary

    Mar. Pollut. Bull.

    (2013)
  • T. Li et al.

    Influence of trace metals in recent benthic foraminifera distribution in the Pearl River Estuary

    Mar. Micropaleontol.

    (2014)
  • M. Lopes-Rocha et al.

    Spatial patterns and temporal trends of trace metal mass budgets in the western Adriatic sediments (Mediterranean Sea)

    Sci. Total Environ.

    (2017)
  • M. Manousakas et al.

    Assessment of PM2.5 sources and their corresponding level of uncertainty in a coastal urban area using EPA PMF 5.0 enhanced diagnostics

    Sci. Total Environ.

    (2017)
  • F. Mazzei et al.

    Characterization of particulate matter sources in an urban environment

    Sci. Total Environ.

    (2008)
  • P. McLaren et al.

    Deriving transport pathways in a sediment trend analysis (STA)

    Sediment. Geol.

    (2007)
  • J. McManus et al.

    Remobilization of barium in continental margin sediments

    Geochim. Cosmochim. Acta

    (1994)
  • Y.M. N'Guessan et al.

    Trace elements in stream bed sediments from agricultural catchments (Gascogne region, SW France): where do they come from?

    Sci. Total Environ.

    (2009)
  • P.N. Owens et al.

    Fingerprinting and tracing the sources of soils and sediments: earth and ocean science, geoarchaeological, forensic, and human health applications

    Earth-Sci. Rev.

    (2016)
  • P. Paatero

    Least squares formulation of robust non-negative factor analysis

    Chemom. Intell. Lab. Syst.

    (1997)
  • K. Pan et al.

    Trace metal contamination in estuarine and coastal waters of China

    Sci. Total Environ.

    (2012)
  • R. Pedreros et al.

    Application of grain size trend analysis for the determination of sediment transport pathways in intertidal areas

    Mar. Geol.

    (1996)
  • S. Peek et al.

    Sr/Ca and Ba/Ca variations in environmental and biological sources: a survey of marine and terrestrial systems

    Geochim. Cosmochim. Acta

    (2012)
  • E. Poizot et al.

    Sediment trend analysis through the variation of granulometric parameters: a review of theories and applications

    Earth Sci. Rev.

    (2008)
  • E. Poizot et al.

    The application of geostatistics in defining the characteristic distance for grain size trend analysis

    Comput. Geosci.

    (2006)
  • X.-H. Song et al.

    Sources of fine particle composition in the northeastern US

    Atmos. Environ.

    (2001)
  • L.H. Tecer et al.

    Metallic composition and source apportionment of fine and coarse particles using positive matrix factorization in the southern Black Sea atmosphere

    Atmos. Res.

    (2012)
  • I. Tubau et al.

    Using EMMA and MIX analysis to assess mixing ratios and to identify hydrochemical reactions in groundwater

    Sci. Total Environ.

    (2014)
  • S. Vaccaro et al.

    The application of positive matrix factorization in the analysis, characterisation and detection of contaminated soils

    Chemosphere

    (2007)
  • D. Wang et al.

    Application of positive matrix factorization to identify potential sources of PAHs in soil of Dalian, China

    Environ. Pollut.

    (2009)
  • L.J. Warren

    Contamination of sediments by lead, zinc and cadmium: a review

    Environmental Pollution Series B, Chemical and Physical.

    (1981)
  • J. Zhang et al.

    Trace elements in PM2.5 in Shandong Province: source identification and health risk assessment

    Sci. Total Environ.

    (2018)
  • J.R. Banta et al.

    Delineation of carbonate dust, aluminous dust, and sea salt deposition in a Greenland glaciochemical array using positive matrix factorization

    Geochem. Geophys. Geosyst.

    (2008)
  • F.K. Barthold et al.

    How many tracers do we need for end member mixing analysis (EMMA)? A sensitivity analysis

    Water Resour. Res.

    (2011)
  • G. Birth

    A scheme for assessing human impacts on coastal aquatic environments using sediments

  • P.A. Bzdusek et al.

    Comparison of a new variant of PMF with other receptor modeling methods using artificial and real sediment PCB data sets

    Environmetrics

    (2006)
  • P.A. Bzdusek et al.

    PCB congeners and dechlorination in sediments of Sheboygan River, Wisconsin, determined by matrix factorization

    Environ. Sci. Technol.

    (2006)
  • Cited by (54)

    • Novel sediment source fingerprinting quantifying erosion-induced total nitrogen and total phosphorus outputs from an intensive agricultural catchment, North China

      2023, International Soil and Water Conservation Research
      Citation Excerpt :

      Sediment source tracing refers to a field-based measurement technique that apportions or un-mixes sediment samples into multiple sources through the use of tracers (Collins et al., 2017, 2020; Haddadchi et al., 2013; Nosrati et al., 2021; Owens et al., 2016; Walling, 2013). Tracers used in sediment fingerprinting studies include color parameters (García-Comendador et al., 2020), plant pollen (Kobe et al., 2021), major and trace elemental composition (Li et al., 2019), rare earth elements (Astakhov et al., 2019; Brito et al., 2018), mineral magnetic characteristics (Wei et al., 2021), clay mineralogy (Verma et al., 2014) and E-DNA approach (Evrard et al., 2019). In particular, geochemical tracer techniques have been used to distinguish sediments in soils formed from different rock types (Gholami et al., 2019; Huang et al., 2020; Li, Gholami, et al., 2020; Nosrati et al., 2021; Torres Astorga et al., 2020).

    View all citing articles on Scopus
    View full text