Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Stoichiometry, polarity, and organometallics in solid-phase extracted dissolved organic matter of the Elbe-Weser estuary

  • Kerstin B. Ksionzek ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing (KBK); (BPK)

    Affiliations Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany, MARUM–Center for Marine Environmental Sciences, Leobener Straße, Bremen, Germany

  • Jing Zhang,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Writing – review & editing

    Affiliation State Key Laboratory of Estuarine and Coastal Research (SKLEC), East China Normal University (ECNU), Shanghai, China

  • Kai-Uwe Ludwichowski,

    Roles Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – review & editing

    Affiliation Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany

  • Dorothee Wilhelms-Dick,

    Roles Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – review & editing

    Affiliation Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany

  • Scarlett Trimborn,

    Roles Funding acquisition, Resources, Supervision, Writing – review & editing

    Affiliation Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany

  • Thomas Jendrossek,

    Roles Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – review & editing

    Affiliation Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany

  • Gerhard Kattner,

    Roles Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing

    Affiliation Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany

  • Boris P. Koch

    Roles Conceptualization, Formal analysis, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing (KBK); (BPK)

    Affiliations Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven, Germany, MARUM–Center for Marine Environmental Sciences, Leobener Straße, Bremen, Germany, University of Applied Sciences, Bremerhaven, Germany

Stoichiometry, polarity, and organometallics in solid-phase extracted dissolved organic matter of the Elbe-Weser estuary

  • Kerstin B. Ksionzek, 
  • Jing Zhang, 
  • Kai-Uwe Ludwichowski, 
  • Dorothee Wilhelms-Dick, 
  • Scarlett Trimborn, 
  • Thomas Jendrossek, 
  • Gerhard Kattner, 
  • Boris P. Koch


Dissolved organic matter (DOM) is ubiquitous in natural waters and plays a central role in the biogeochemistry in riverine, estuarine and marine environments. This study quantifies and characterizes solid-phase extractable DOM and trace element complexation at different salinities in the Weser and Elbe River, northern Germany, and the North Sea. Dissolved organic carbon (DOC), total dissolved nitrogen (TDN), Co and Cu concentrations were analyzed in original water samples. Solid-phase extracted (SPE) water samples were analyzed for DOC (DOCSPE), dissolved organic nitrogen (DONSPE), sulfur (DOSSPE) and trace metal (51V, 52Cr, 59Co, 60Ni, 63Cu, 75As) concentrations. Additionally, different pre-treatment conditions (acidification vs. non-acidification prior to SPE) were tested. In agreement with previous studies, acidification led to generally higher recoveries for DOM and trace metals. Overall, higher DOM and trace metal concentrations and subsequently higher complexation of trace metals with carbon and sulfur-containing organic complexes were found in riverine compared to marine samples. With increasing salinity, the concentrations of DOM decreased due to estuarine mixing. However, the slightly lower relative decrease of both, DOCSPE and DONSPE (~77%) compared to DOSSPE (~86%) suggests slightly faster removal processes for DOSSPE. A similar distribution of trace metal and carbon and sulfur containing DOM concentrations with salinity indicates complexation of trace metals with organic ligands. This is further supported by an increase in Co and Cu concentration after oxidation of organic complexes by UV treatment. Additionally, the complexation of metals with organic ligands (analyzed by comparing metal/DOCSPE and metal/DOSSPE ratios) decreased in the order Cu > As > Ni > Cr > Co and thus followed the Irving-Williams order. Differences in riverine and marine trace metal containing DOMSPE are summarized by their average molar ratios of (C107N4P0.013S1)1000V0.05Cr0.33Co0.19Ni0.39Cu3.41As0.47 in the riverine endmember and (C163N7P0.055S1)1000V0.05Cr0.47Co0.16Ni0.07Cu4.05As0.58 in the marine endmember.


Dissolved organic matter (DOM) is actively cycling in natural waters and participates in most biogeochemical processes. Assessment of the DOM stoichiometry supports to unravel its origin and fate and to understand its role in different aquatic environments. The biogeochemistry of marine dissolved organic carbon, nitrogen and phosphorus (DOC/N/P) was extensively studied in the past, e.g. [1, 2]. In contrast, the knowledge on quantity, distribution, and the biogeochemical role of dissolved organic sulfur (DOS) in aquatic environments is limited, though not less important. We have previously estimated the global marine DOS inventory to range between 6.7 and 18.6 Pg S [3]. In particular, rivers are known to be important sources of reduced carbon, nitrogen and phosphorus to coastal environments [4]. Riverine DOM can be influenced by different transformation and removal processes along its way into estuarine and marine water: salt-induced flocculation [5, 6], adsorption to particulate matter [7, 8], photo-oxidative remineralization [9], and uptake by heterotrophs [10, 11]. Thus, typical concentrations of DOC and DON decrease over a salinity gradient from riverine to marine water [12, 13]. DOC/DON ratios usually also decrease from land to sea along the salinity gradient in estuaries [14] indicating differences in the stoichiometry of the organic matter precursors. In our previous study, we used existing literature and roughly estimated that the riverine transport of organic sulfur in particulate (POS) and dissolved form combined is about 0.25 Tmol S a-1 (8 Tg S a-1) [3]. In estuarine and marine environments, the concentration of sulfate (up to 29 mmol S L-1) exceeds the concentration of DOS by up to five orders of magnitude. As the analysis of DOS has been analytically hampered, the composition and biogeochemistry of DOS remains widely unknown. Several studies focused primarily on volatile organic sulfur compounds, such as dimethylsulfide (DMS) and carbonyl sulfide (COS), because they are actively involved in climate processes [1517]. However, those climate relevant organic sulfur compounds contribute less than 3% to the total marine DOS pool [3].

Other organic sulfur compounds, such as sulfides and thiols, play an important role as ligands for organic metal complexes [18]. Thiols build strong complexes with copper and account for a major part of the copper complexing ligand pool in surface seawater [19, 20]. Silver and mercury are also known to bind strongly with organic sulfur species [21, 22]. Organic metal-complexing ligands can thus affect the mobility, toxicity, and bioavailability of several trace metals. Some metals or metalloids in aquatic ecosystems, such as As, Co, Cu, Cr, Ni, and V are essential micronutrients to support biological processes [2326], e.g Cu, Co, and Ni are essential for growth and control of marine phytoplankton populations [27]. This has also an indirect effect on bioproductivity, species composition and, in the long term, food web dynamics. However, in high concentrations, these metals can also cause toxic effects [26, 28]. Reduced toxicity was found for some trace metals (e.g. Cu, Pb, Cd) in case of higher DOM complexation rates [29, 30]. Moreover, trace metals can not only trigger the active production of organic ligands but also contribute to their persistence in surface waters: trace-metal complexation has a protective effect against oxidation of DOM-thiol groups [31], whereas the production of copper-binding thiols is enhanced with increasing copper-levels [3235]. Besides quantity, the quality of DOM plays also an equally important role for trace metal complexation [36]. Baken et al. found that increasing aromaticity lead to a higher trace metal affinity of DOM, indicating that aromatic humic substances can act as major metal chelators [36]. Matar et al. [37] analyzed the influence of organic matter polarity on trace metal speciation and bioavailability and revealed that the hydrophobic DOM fraction has a lower binding capacity for Cu than the hydrophilic fraction, suggesting lowered Cu bioavailability in presence of hydrophilic DOM. Although DOM interactions with metals and the distribution and cycling of organic metal complexes are a growing field of interest, the influence of DOM and specifically of DOS compounds on transport, kinetics, bioavailability and toxicity of trace metals remains largely unknown.

Here we present results on the composition and distribution of DOM at different salinities sampled from the rivers Weser and Elbe in northern Germany to the marine waters of the North Sea. Our aim is to improve our knowledge on distribution and composition of organically bound trace metals. The major research questions/hypotheses are:

  1. What is the concentration of solid-phase extractable DOS (DOSSPE) in the rivers Elbe and Weser and how does it change with salinity? How does the stoichiometry (molar elemental ratios) of solid-phase extractable DOM (DOMSPE) change with salinity within the Elbe-Weser-Estuary?
  2. Since some sulfur species, such as thiols, are known as trace-metal ligands, we hypothesize that the amount of organically-bound trace metals correlates with the relative contribution of DOS and DOM. Consequently, changes in DOM concentration with salinity should also be reflected in trace metal concentrations.
  3. How does DOM polarity change along the estuary and is this change connected to dissolved/complexed trace element concentrations?
  4. What is the influence of salinity and different sample pre-treatment conditions (pH 2 versuss pH 8 extraction) on DOM and associated trace element composition?

Materials and methods

Sample collection and processing

Six surface water samples were collected in June and July 2014 from Rivers Weser and Elbe, northern Germany (salinity ~ 0), and in the Southern North Sea (salinity ~33, Fig 1, Table 1). The marine water sample (M1) and samples from the Elbe Estuary (E2, E3) were collected with a rosette sampler connected to a conductivity, temperature, and depth sensor (CTD) (expedition HE426 of R/V Heincke). In total, 36 CTD stations were performed to analyze background parameters, such as temperature and salinity. Other riverine samples from River Elbe (E1) and River Weser (W1, W2) were collected manually in glass bottles. Temperature, conductivity, and pH were measured in situ with a sensor (Cond 340i, WTW). No specific permissions were required for sampling and the field studies did not involve endangered or protected species. The sample processing workflow is presented in Fig 2. All samples were filtered through pre-combusted GF/F filters (Whatman, 450°C, 5 h, 0.7 μm nominal pore size) with a maximum pressure < 200 mbar. Aliquots for DOC and nutrient analyses were stored at -20°C in pre-cleaned high-density polyethylene (HDPE) bottles. Filtered water was either acidified to pH 2 (hydrochloric acid, suprapur, Merck) or processed without acidification (pH ~8). SPE was applied for DOM enrichment and desalting [38] and for each sample 500 mL of filtered water (pH 8 and pH 2) was extracted (PPL, 200 mg, Mega Bond Elut, Varian) in quadruplicates and each eluted with ~1 mL methanol (LiChrosolv, Merck; exact volume was determined by weighing) into pre-combusted glass vials. After extraction, DOMSPE was stored at -20°C until further analysis.

Fig 1. Map of sampling area.

Sampling locations are marked as red dots, colors represent the surface salinity. Black dots represent stations, at which temperature and salinity were measured [39].

Fig 2. Sample processing workflow.

The sample processing steps are represented as white boxes. Measured parameters are specified in grey boxes. DOC and total dissolved nitrogen (TDN) in original water samples were analyzed by high temperature catalytic oxidation (HTCO). The seaFAST analysis was used to determine Co and Cu concentrations in the original filtered water samples. Aliquots of original filtered water samples were solid-phase extracted in quadruplicates for each acidified and non-acidified sample. DOC and DON in solid-phase extracts (DOCSPE, DONSPE) were analyzed by HTCO. All other elements were analyzed by inductively coupled plasma mass spectrometry (ICP-MS).

DOC, TDN and DON analysis

Concentrations of DOC and total dissolved nitrogen (TDN) in filtered water were determined by high temperature catalytic oxidation (HTCO) and subsequent nondispersive infrared spectroscopy and chemiluminescence detection (TOC-VCPN analyzer, Shimadzu). For the determination of solid-phase extractable DOC (DOCSPE, pH 2 and pH 8) and DON (DONSPE, pH 2), 50 μL (250 μL for DONSPE, pH 8) of each methanol extract was evaporated under N2 and subsequently redissolved in 6.5 mL ultrapure water. All samples were acidified in the auto sampler (0.1 M HCl suprapur, Merck) and purged with O2 for > 5 min to remove inorganic carbon. Performance of the instrument was recorded by the analysis of potassium hydrogen phthalate standard solutions and the deep-sea reference samples (DSR, Hansell research lab). Final DOC and TDN concentrations are average values of triplicate measurements. If the standard variation or the coefficient of variation of DOC values exceeded 0.1 μM or 1%, respectively, up to two additional analyses were performed and outliers were eliminated. For DON, outliers of triplicate measurements were eliminated manually. The accuracy was ± 5% for DOC and ± 7% for DON.

ICP-MS analysis

For quantification of DOSSPE, DOPSPE and trace elements (51V, 52Cr, 59Co, 60Ni, 63Cu, 75As), an inductively coupled plasma mass spectrometer (ICP-MS, Element 2, Thermo Fisher Scientific) was equipped with a desolvation nebulizer (Apex Q, Elemental Scientific), a platinum guard electrode, and nickel sampler and skimmer cones. Prior to ICP-MS analysis, 50 μL of the extract was evaporated with N2 gas and redissolved in 2 mL nitric acid (1 M, bidestilled, Merck). 50 μL of 103Rh (50 ppb in the spike solution) were added as internal standard. The samples were sonicated for 10–15 min to ensure that all DOM was redissolved. The instrument was tuned daily for optimized plasma conditions and accurate mass calibration with a multi-element tuning solution (~0.1 ppb in MilliQ). Signals of 32S and 75As were recorded in a resolution of 4000 m/Δm, whereas all other elements were recorded in a resolution of 2000m/Δm, for which the instrument was modified to achieve a flat top peak shape (higher precision). Nitric acid (1 M, double destilled, Merck) was used for analysis blank. If the blank values for SPE were higher than the limit of detection (LOD), the extract concentrations were corrected for the respective blanks. Calibration standards for trace elements were prepared in concentrations of 0.001, 0.01, 0.05, 0.1, 0.5, 1, 5, 10, 50, 100, and 250 μg L-1 from a stock solution (100 mg L-1, multi-element-standard, nonmetals, Spetec). Limits of detection (according to the German industry standard; DIN 32645) are given in S1 Table.

Trace element analysis of filtered seawater samples

We analyzed 59Co and 63Cu in original (filtered) water samples. All labware used for analysis was pre-cleaned according to Dick et al. [40]. Samples for dissolved trace metal analysis were acidified to pH 1.75 using bidistilled HNO3. As organic ligands form complexes with Co and Cu which are relatively blind to the chelating resin and therefore pass through it without being extracted, one half of each sample was UV digested to analyze the total amount of Co and Cu [41]. For UV digestion, samples were filled into pre-cleaned PFA bottles and UV-oxidized for 1.5 h using a 450 W photochemical UV power supply from ACE GLASS (photochemical lamp number 7825; Power Supply number 7830). Two procedural blanks were processed the same way. Prior to analysis, each sample was spiked with Indium as internal standard (final concentration 1 ppb). The multi-element analyses of water samples were performed using a seaFAST system (Elemental Scientific Inc.) as described in Hathorne et al. [42] coupled to ICP-MS (Element 2, Thermo Fisher Scientific). The ICP-MS was optimized every day to achieve oxide forming rates below 0.3%. The ICP-MS was modified to achieve a flat top peak shape (higher precision) with a resolution of R = 2000. Quantification limits are 0.35 ng L-1 for 59Co and 7.35 ng L-1 for 63Cu. To assess the accuracy and precision of the method, the NASS-6 reference standard was analyzed in a 1:2 dilution at the beginning, in between and at the end of a batch run (n = 5). For Cu, we found 208 ng L-1 with a relative standard deviation of 1.5% (certified 248 ± 25 ng L-1). The Student t-Test was used to compare our values to the certified ones (n = 5, 99% significance level) and showed no significant difference. Within the GeoRem database Takano et al. reported Cu concentrations of 224 ng L-1, which is in agreement with our findings [43, 44]. Co is not certified for the NASS-6 standard. For this, only an indicative value is given in the certificate (15 ng L-1). We found 11.2 ng L-1 with a precision of 0.6%.

RP-HPLC analysis

Reversed phase high performance liquid chromatography (RP-HPLC) was performed on a LaChrom EliteTM HPLC-system (Hitachi) equipped with a pump (L-2130), autosampler (L-2200), column oven (L-2300), diode array detector (DAD, L-2450, 210 nm) and fluorescence detector (L-2485; excitation: 260 nm, emission: 430 nm) according to Koch et al. [45]. Of each methanol extract, 100 μL extract were diluted with 400 μL ultrapure water. Original water samples were analyzed without any pre-treatment. For each analysis, 30 μL of the methanol extract and 95 μL of original samples were injected respectively. The separation based on polarity (and molecular size) was performed using a reversed-phase column (4μm Hydro-RP 80 Å, 250 x 4.60 mm; Phenomenex, Synergi) and a solvent gradient (0 to 70 min) from 100% ultrapure water (adjusted with low-concentrated NaOH (Merck, suprapur) to pH 7) to 100% methanol (Merck, LiChrosolv, Table 2). Analysis blanks were performed with 100 μL methanol and 400 μL ultrapure water for the analysis of the extracts and ultrapure water only for the analysis of the original samples respectively. Peak areas of the samples were corrected for the respective blanks. Based on RP-HPLC analyses, we differentiated two major DOMSPE fractions: the polar water-soluble fraction with a retention time < 24 min and the non-polar methanol soluble fraction with a retention time > 24 min. We calculated polar/non-polar ratios (DOCpol/DOCnon-pol) to elucidate changes in DOM polarity with changing salinity. Since nitrate and nitrite can absorb in the DAD210nm range, only fluorescence data were used for the evaluation of the DOM polarity characteristics in the original samples.

Table 2. Gradient for the chromatographic run.

Water (adjusted to pH 7) and methanol were used as eluents.

Data evaluation and statistical analysis

Outliers in quadruplicate measurements of DOMSPE concentrations were defined by Grubbs Test with a significance level of α = 0.5. Statistical analysis was performed with the software R: Analysis of variances between the riverine (E1, W1) and marine (M1) endmember samples) was performed with Mann-Whitney-U-Test (function “wilcoxon.test” in the software R). Correlation of changing DOM, DOMSPE and metal-DOMSPE concentrations, respectively, with changing salinity was analyzed using correlation analysis (function “cor.test” in R). The variables were not manipulated prior to statistical analysis.


Changes in DOM concentration and stoichiometry

The original DOC and TDN concentrations in the water samples decreased with increasing salinity from a maximum of 407 μmol L-1 and 190 μmol L-1, respectively, in riverine water (W1) to 293 μmol L-1 DOC and 101 μmol L-1 TDN in estuarine water (E3), and 97 μmol L-1 DOC and 14 μmol L-1 TDN in the marine sample (M1; Fig 3, Table 3).

Fig 3. Changes of DOC (filled symbols) and TDN (unfilled symbols) concentrations of original water samples with salinity.

Symbols represent sampling locations: River Weser (W1, W2; squares), River Elbe (E1 - E3; diamonds) and the marine station (M1; circles).

Table 3. Average DOMSPE concentrations and molar DOCSPE/DONSPE, DOCSPE/DOPSPE, DOCSPE/DOSSPE ratios at the stations with different salinities.

The values are averages of quadruplicate measurements (except of DOC and TDN). Stations are ordered by increasing salinity.

For DOMSPE, significantly higher concentrations were found for all elements in samples extracted at pH 2 compared to those extracted at pH 8 (p < 0.001). Therefore, this section focusses on the results of pH 2 extracts. Data on pH 8 extracts can be found in Tables 3 and 4.

Table 4. Average metal-DOMSPE concentrations and molar DOCSPE/metal-DOMSPE ratios (/105) at the stations.

The values are average values of quadruplicate measurements. All concentrations are given in nmol L-1.

DOCSPE concentrations of pH 2 extracted samples decreased significantly from 144 ± 4 μmol L-1 in the riverine endmember (W1) to 31 ± 1 μmol L-1 in the marine endmember sample (M1) (p < 0.01, Table 3). Using the DOC concentrations in the original samples and the DOCSPE concentration, we calculated the DOC extraction efficiencies. The average DOC extraction efficiencies at pH 2 were 36 ± 2%. No significant correlation of the extraction efficiency with salinity was found (p > 0.05).

DONSPE concentrations of pH 2 extracted samples decreased significantly with increasing salinity from 6.3 ± 0.1 μmol L-1 in riverine water (E1) to 1.4 ± 0.0 μmol L-1 in seawater (M1) (p = 0.01). Differences of the average molar DOCSPE/DONSPE ratios of 24 ± 1 and 23 ± 1 were insignificant between pH 2 extracted riverine (E1, W1) and marine (M1) endmember (for pH 8 extracted samples however, they decreased significantly from 25 ± 1 to 22 ± 1 in riverine (E1, W1) and marine (M1) samples (p = 0.01)).

Similar to DOCSPE and DONSPE, concentrations of DOSSPE and DOPSPE were higher in pH 2 than in pH 8 extracted samples (p < 0.001, Table 3). DOSSPE concentrations of pH 2 samples decreased significantly from 1.44 ± 0.02 μmol L-1 in riverine (E1) to 0.19 ± 0.02 μmol L-1 in marine water (p = 0.01). To address the influence of mixing of low and high salinity waters, we normalized DOSSPE and DOPSPE concentrations to DOCSPE concentrations and thus calculated molar DOCSPE/DOSSPE ratios and DOCSPE/DOPSPE ratios, respectively. Average DOCSPE/DOSSPE ratios of pH 2 extracted riverine waters (E1, W1) were 103 ± 5 and increased to 162 ± 17 in the marine sample (p = 0.01). No differences in molar DOCSPE/DOSSPE ratios were found between pH 2 and pH 8 extracted samples.

Compared to DOSSPE, DOPSPE concentrations of pH 2 samples were two orders of magnitude lower and decreased significantly from 20.7 ± 0.7 nmol L-1 to 10.5 ± 2.1 nmol L-1 in riverine (E1) and marine water, respectively (p < 0.001). Molar DOCSPE/DOPSPE ratios of pH 2 extracted samples decreased significantly from riverine to marine water (p < 0.01).

Thus, average molar C:N:P:S ratios of pH 2 extracted samples were C106:N4:P0.013S1 for the riverine endmember (W1) and C164:N7:P0.053S1 for the marine endmember.

Changes in DOM polarity

Using reversed-phase chromatography, we found a good relationship of both fluorescence (260/430 nm) and absorption (210 nm) data (total peak areas) with measured DOC and DOCSPE concentrations (R2 = 0.3 and p < 0.01 for DOC concentrations of 0–40 μmol L-1 versus UV peak areas and R2 = 0.6 and p < 0.001 for DOC concentrations > 100 μmol L-1 versus UV peak areas, S1 Fig), confirming that UV absorption in the extracts serves as a suitable predictor of DOC concentration [46]. Lechtenfeld et al. showed that this correlation can also be found in individual chromatographic fractions [46]. In this section, we will use integrated peak areas of fluorescence / adsorption as a proxy for DOCSPE concentration.

Fluorescence data for the original samples confirmed decreasing DOC concentrations (peak areas) with increasing salinity (not shown). For the solid-phase extracted DOM, we found a significant linear correlation of decreasing polar peak areas with increasing salinity (R2 = 0.9; p < 0.01), whereas no significant trend was found for non-polar peak areas. Hence, DOCpol/DOCnon-pol ratios significantly decreased with increasing salinity (R2 = 0.7, p = 0.035).

In addition, absorption (Fig 4) and fluorescence data (not shown) confirmed significantly higher DOC concentrations in pH 2 compared to pH 8 extracted samples (p < 0.001).

Fig 4. UV absorption chromatograms of riverine and marine methanol extracts at 210 nm.

(A) Chromatograms of riverine pH 2 (red) and pH 8 (blue) extracted samples. (B) Chromatograms of marine pH 2 (red) and pH 8 (blue) extracted samples.

Absorption data showed significantly higher DOCSPE concentrations (peak areas) at low salinity compared to high salinity samples (p < 0.01, Fig 4). However, the DOCSPE concentration (peak areas) in the estuarine samples E3 and W3 (salinity of 17.4 and 18.1, respectively) were similar to sample E2 at salinity 9.4 and thus deviated from the linear regression line (R2 = 0.96, p < 0.001). For all extracts, the ratio of DOC derived from polar compounds compared to the total DOC concentration (DOCpol/DOCtot) was significantly higher in pH 2 compared to pH 8 samples (p < 0.001). In pH 2 extracted samples, DOCpol/DOCtot of 0.25 ± 0.02 in riverine samples (W1, E1) was significantly higher compared to 0.06 ± 0.02% in marine samples (p < 0.01). Consequently, we observed a significant decrease of DOCpol/DOCnon-pol in pH 2 samples from 0.36 ± 0.03 in riverine water (W1) to 0.07 ± 0.02 in marine water (p = 0.01, Fig 5A), whereas no significant changes in pH 8 samples occurred. However, if we focus on Elbe samples only, we observed a relative increase of DOCpol/DOCnon-pol ratios in estuarine water (Fig 5A).

Fig 5. Changes in polar/non-polar peak area ratios of DOMSPE samples with salinity changes.

(A) Average polar/non-polar peak area ratios analyzed by UV spectroscopy (DAD210nm) versus salinity and (B) average polar/non-polar peak area ratios analyzed by fluorescence spectroscopy (260/340 nm) versus salinity. The pH 2 extracted samples are indicated by red symbols, pH 8 extracted samples by blue symbols. Symbols represent sampling locations: River Weser (W1, W2; squares), River Elbe (E1 - E3; diamonds) and the marine station (M1; circles).

Fluorescence data showed that DOCSPE concentrations (peak areas) of both pH 2 and pH 8 extracted samples decreased with increasing salinity (p < 0.001). Overall, DOCpol/DOCtot of 0.68 ± 0.07 in pH 2 samples was significantly higher compared to 0.03 ± 0.01 in pH 8 samples (p < 0.001). DOCpol/DOCtot in both pH 2 and pH 8 extracted samples decreased significantly with increasing salinity (p < 0.001). This results in a significant increase in DOCpol/DOCnon-pol of 1.7 ± 0.2 to 2.1 ± 0.3 in pH 2 extracted riverine and marine samples, respectively, and from 0.01 ± 0.002 to 0.04 ± 0.01 in pH 8 extracted riverine and marine samples, respectively (Fig 5B). Comparing the concentration of non-polar DOC (DOCnon-pol) between pH 2 and pH 8 samples, it is noteworthy that in pH 8 samples, DOCnon-pol was only about half of the value in pH 2 samples. By contrast, the contribution of the DOCnon-pol pool to the absorbance of pH 2 or pH 8 extracted samples was similar.

Trace metals

Concentrations of solid-phase extractable trace metals (metal-DOMSPE) were generally higher in pH 2 compared to pH 8 extracts, with the exception of Ni-DOMSPE. In pH 2 extracted samples, Co-DOMSPE, Ni-DOMSPE, Cu-DOMSPE, and Cr-DOMSPE concentrations decreased with increasing salinity, similar to DOCSPE and DOSSPE (Table 4). Although riverine Cu-DOMSPE was significantly higher than in the marine sample (p < 0.05), we found a concentration maximum at E2 followed by a decrease with increasing salinity. V-DOMSPE concentrations also increased in estuarine waters, followed by a decrease in the high salinity marine water, but were highly distributed over all samples. Also, Cr-DOMSPE concentrations were highly distributed. Nevertheless, we could also observe a significant decrease with increasing salinity over all Cr-DOMSPE samples (p < 0.01). No significant differences between riverine and marine endmember concentrations were found for As-DOMSPE and V-DOMSPE. Process blanks were below the detection limit for Co-DOMSPE, Ni-DOMSPE, and Cu-DOMSPE. Blanks for As-DOMSPE, Cr-DOMSPE, and V-DOMSPE were measurable and about factor 2.0 – 5.8 lower for riverine and estuarine samples and by factor 1.2–2.4 lower for marine samples compared to the corresponding metal-DOMSPE concentrations in the samples. To facilitate the subsequent comparison of the relative metal and sulfur content in DOM, we used C/metal ratios (similar to C/S ratios).

Similar to DONSPE, DOPSPE and DOSSPE, the concentrations of metal-DOMSPE were normalized to DOCSPE (Table 4). The decrease in Co-DOMSPE and Ni-DOMSPE concentrations with increasing salinity was a more sensitive function of salinity than that of DOCSPE decrease, resulting in a significant linear increase of the molar DOCSPE/Co-DOMSPE (p < 0.01) and DOCSPE/Ni-DOMSPE ratios (p = 0.01 Fig 6). In contrast, DOCSPE/As-DOMSPE ratios decreased significantly with increasing salinity (p < 0.01). No significant differences between riverine and marine samples were found for DOCSPE/Cr-DOMSPE, DOCSPE/Cu-DOMSPE and V-DOMSPE.

Fig 6. Changes in trace-metal stoichiometry with salinity.

Linear correlation of average molar DOCSPE/metal-DOMSPE ratios of pH 2 extracted samples (except of DOCSPE/Cr-DOMSPE, DOCSPE/Cu-DOMSPE, and DOCSPE/V-DOMSPE) versus salinity. Symbols represent sampling locations: River Weser (W1, W2; squares), River Elbe (E1—E3; diamonds) and the marine station (M1; circles).

In pH 8 extracted samples, most concentrations of As-DOMSPE and V-DOMSPE were below the detection limit (LOD; cf. Table 4). Additionally, a high variance in quadruplicate measurements was found for the low concentrations of all trace metals in the marine samples, therefore those results were excluded from further discussion. Data of metal-DOMSPE concentrations and DOCSPE/metal-DOMSPE ratios of pH 8 extracts can be found in Table 4.

To verify if DOSSPE correlates with trace metals, we plotted DOCSPE/metal-DOMSPE ratios versus DOCSPE/DOSSPE ratios of pH 2 extracted samples (Fig 7). We observed a significant negative linear correlation of DOCSPE/As-DOMSPE (R2 = 0.28, p = 0.01) with DOCSPE/DOSSPE and a positive linear correlation for DOCSPE/Co-DOMSPE (R2 = 0.17, p = 0.05). DOCSPE/Cr-DOMSPE, DOCSPE/V-DOMSPE, DOCSPE/Ni-DOMSPE, and DOCSPE/Cu-DOMSPE showed no correlation with DOCSPE/DOSSPE (Fig 7). For pH 8 extracted samples, no significant correlation was found.

Fig 7. Correlation of trace metal and DOS stoichiometry.

Correlation of DOCSPE/metal-DOMSPE versus average DOCSPE/DOSSPE of pH 2 extracted samples. Linear correlation of DOCSPE/Co-DOSPE and DOCSPE/As-DOMSPE is shown. Symbols represent sampling locations: River Weser (W1, W2; squares), River Elbe (E1 - E3; diamonds) and the marine station (M1; circles).

To assess the extraction efficiencies for metal-DOM, we determined the trace metal concentrations of Co and Cu in original seawater and compared them with Cu-DOMSPE and Co-DOMSPE (Table 5). Co and Cu concentrations in original water samples also decreased with increasing salinity (Fig 8, Table 5). A significant linear correlation with salinity was found for UV-treated Co (CoUV) concentrations only (Fig 8). The application of UV digestion led to 300–500% higher trace element concentrations compared to the untreated samples. Co extraction efficiencies were generally lower for pH 8 extracted samples compared to pH 2 extracted samples. Average extraction efficiencies of pH 2 extracted Co-DOMSPE were 8–45% without UV digestion of original samples and 3–12% with UV digestion, respectively. Co extraction efficiencies of riverine samples were significantly higher compared to marine samples (p < 0.001). However, no correlations with salinity were observed. For pH 8 extracted samples, average Co-DOMSPE extraction efficiencies were 5–19% and 2–4% for samples without and with UV digestion, respectively. No differences between riverine and marine samples were found. Average extraction efficiencies of pH 2 extracted Cu-DOMSPE were 10–13% without UV digestion of original samples and 9–11% with UV digestion. For pH 8 extracted samples, Cu-DOMSPE extraction efficiencies were 5–20% and 6–17% for samples without and with UV digestion, respectively. In contrast to Co-DOMSPE, the extraction efficiencies of pH 8 extracted Cu correlated linearly with salinity (R2 = 0.98, p < 0.001). Interestingly, Cu extraction efficiencies of pH 2 extracted samples showed no differences with increasing salinity, whereas Cu extraction efficiencies of pH 8 extracted samples increased with increasing salinity (Table 5).

Fig 8. Cu (circles) and Co (diamonds) concentrations with changing salinity.

Symbols and correlation line represent pre-treatment conditions of original samples: with UV digestion (filled symbols, solid line) and without UV digestion (unfilled symbols).

Table 5. Dissolved Co and Cu concentrations in original water samples and corresponding average solid phase extraction efficiencies (extr. eff.).


Stoichiometry and polarity characteristics of dissolved organic matter

The high DOM concentrations in the riverine endmember samples of Weser and Elbe River emphasized the importance of rivers as a DOM source to the coastal oceans as previously shown for many other regions from riverine to marine waters [4750]. DOC concentrations in the estuarine and marine samples (E2, E3, M1; Table 3) were typical for the German Bight (76–209 μmol L-1 [51]). DOCSPE, DONSPE, and DOSSPE concentrations as well as DOCSPE/DONSPE and DOCSPE/DOSSPE ratios are in accordance with values published for the North Sea [52] and Atlantic surface waters [3, 53]. Unlike the typical decrease of DOC/DON ratios from land to sea in estuaries (e.g. [14]), our pH 2 extracted samples showed almost constant molar DOCSPE/DONSPE ratios, probably due to an inefficient extraction of nitrogen compared to carbon [54, 55]. Ratios of pH 8 extracted samples were similar, but reflected the typical decrease of molar DOCSPE/DONSPE ratios with increasing salinity. Molar DOCSPE/DOSSPE ratios were higher and molar DOSSPE/DONSPE ratios lower in the riverine compared to the marine endmembers. Our average DOCSPE/DOSSPE ratios were 100 ± 3 and 107 ± 5 in the riverine endmember samples. Previous studies on DOS in freshwater and other aquatic systems are very scarce. Houle et al. found DOS concentrations of ~ 5.8 μmol L-1 (~ 185 μg L-1) and DOC/DOS ratios of ~ 122 in several southwestern Québec lakes [56], values that are higher than those in our riverine samples. However, these lakes are influenced by terrestrial DOM with high DOC and DOS concentrations from forest soils. Nevertheless, spatial differences can also cause differences in DOS concentrations. Some information on sulfur exists also for soil and particulate organic matter (POM), e.g. [57, 58]. Based on a molar C/S ratio of 119 for POC [57], the global flux of DOS and POS from rivers to the ocean accounts for 8 Tg S a-1 [3], a value that would be ~15% higher if we use the C/S ratio from this study. Our numbers for the molar DOCSPE/DOPSPE ratios were very high compared to other studies [59], presumably due to inefficient DOP extraction for the cartridges used. However, this is speculative due to the limited available data.

The chromatographic data reflected also lower DOCSPE concentrations at higher salinity. Since fluorescence represents only a small fraction of DOMSPE, we focus on the absorbance data. In contrast to samples from the Weser River, DOCpol/DOCnon-pol ratios of riverine and estuarine samples of the Elbe River were significantly higher at higher salinity. Possible reasons might be differences in DOM sources or different residence times. Over all samples, however, no significant changes in DOCpol/DOCnon-pol ratios occurred within the group of pH 2 extracted riverine and estuarine samples (W1, W2, E1—E3). Only for the marine endmember, a significant lower DOCpol/DOCnon-pol ratio was found. Riverine DOM is dominated by terrestrial sources, characterized by high polarity due to a high number of carboxyl groups, elevated C/N ratios and higher contribution of aromatic components (e.g. lignin and its degradation products) [60, 61] (resulting in higher DOCpol/DOCnon-pol ratios compared to marine samples, as found also in this study). DOM in marine samples, in contrast, originates mainly from phytoplankton and its degradation products and only 0.7–2.4% appears to be from terrestrial sources [62]. In estuaries, mixing of both, riverine and marine DOM occurs and results in changes of the DOM pool composition: with increasing salinity, molecular weight, carbohydrate and heteroelement content of DOM increases [6, 63], while DOM aromaticity decreases [64]. The highly significant correlation of decreasing DOMSPE concentration with increasing salinity reflects mixing of riverine and marine DOM. However, due to the limited number of samples, precise statements about deviations from conservative mixing, as previously reported (e.g. [50]), would be speculative. To assess if all DOMSPE compounds decreased in a similar range with increasing salinity, we compared the relative changes of DOCSPE, DOSSPE and DONSPE endmember concentrations in riverine and marine samples. We found a similar reduction in DOCSPE and DONSPE concentrations of ~77 ± 1% and a reduction in DOSSPE concentration of 86 ± 1%. These changes indicate that DOSSPE decreased faster than DOCSPE and DONSPE. This is in accordance with previous studies in the oceanic water column, showing preferential depletion of sulfur (and phosphorous) relative to carbon [3, 59]. Similar to sulfur, nitrogen is also removed preferentially to carbon [3]. It might be possible, that in our study, the fast decrease in DONSPE is masked by other DONSPE sources along the estuary (such as agricultural and industrial sources, benthic flux, or microbial activity).

Although the mixing of DOM-rich riverine freshwater with marine water in estuaries is the major factor controlling DOM distribution and composition [48, 50, 65, 66], relative changes in DOM stoichiometry indicate processes beyond estuarine mixing. Different sources and sinks control the amount, composition and reactivity of DOM in aquatic environments: biological release, phyto and zooplankton mediated processes [67], decomposition of riverine DOM by marine bacteria [68], photo-bleaching and photo-degradation [6972] as well as flocculation processes and sorption to sediments [47, 48, 73]. All of these processes might occur simultaneously, and it remains a major challenge to quantify the influence of each process on DOM composition.

Trace metal complexation and DOM composition

Co-DOMSPE, Ni-DOMSPE, and Cu-DOMSPE concentrations decreased with salinity, a result of mixing of trace metal-rich riverine water with trace metal-poor marine water. Higher trace-metal concentrations in the riverine extracts suggest a terrestrial/benthic source for dissolved trace metals [74, 75] and/or differences in trace-metal/organic matter composition. Although statements about conservative or non-conservative mixing would be speculative due to the limited number of samples, we assume that other factors additionally to mixing must occur. Normalization of trace metals to carbon allows us to analyze differences between the decreases in DOCSPE and metal-DOMSPE concentrations. Unlike As-DOMSPE and Cu-DOMSPE, Co-DOMSPE and Ni-DOMSPE decreased disproportionately compared to DOCSPE, similar to DOSSPE. The rapid decrease of trace metal concentrations in the estuarine mixing zone is consistent with previous studies [76, 77] and reasons might be (i) changes in trace metal and/or DOM sources, (ii) changes in DOM quality (polarity) and (iii) consumption. Some trace metals (e.g. V, Cr, and Cu) however, increased in estuarine waters and decreased at high salinity in the marine water.

The distributions of Co-DOMSPE, Ni-DOMSPE, Cu-DOMSPE, and Cr-DOMSPE concentrations with increasing salinity followed that of DOCSPE and DOSSPE concentrations, which implies that they are complexed with organic matter (e.g. via carboxylic, hydroxamate, or thiol groups), whereas V-DOMSPE, and As-DOMSPE distributions lead to assume a lower affinity for organic matter. From a study in the Northeast Pacific Ocean, it was estimated that > 99% of total dissolved Cu in surface water is associated with strong organic complexes [78]. Complexation of Co und Cu is further indicated by relatively high extraction efficiencies for both trace metals. We assume that trace metals, which are not organically complexed would most likely not be captured by our extraction method. The PPL sorbent has previously been shown to achieve high recovery rates for organic Cu [79].

The importance of organometallic complexes in DOM is supported by UV digestion prior to seaFAST analysis yielding 100–120% higher Cu concentrations and up to 300–500% higher Co concentrations in the original sample. In previous studies, Co concentrations increased also but only by 50–160% [41, 80]. These differences could be explained by spatial differences in the availability and composition of organic ligands. Since UV treatment/oxidation is used to destroy even very strong metal-organic complexes, the increase in Co and Cu concentrations after UV digestion indicates that a major part of Cu and Co in aquatic samples is organically complexed.

To explore the role of organic sulfur in organometallic complexes, we compared the values of metal-DOMSPE/DOCSPE and metal-DOMSPE/DOSSPE ratios in the pH 2 extracted riverine (W1) and marine (M1) endmember samples and found an increase in the following order: Cu > As > Ni > Cr > Co and As ≥ Cu > Ni > Cr > Co, respectively. This order is consistent with the Irving-Williams order, which has been used to compare the affinity of (colloidal) trace-metals to organic ligands [76, 81]. It is true for both, the affinity of trace metals to DOCSPE and to DOSSPE. According to our results, Cu has a higher affinity to (S-containing) organic ligands than Co and Cr irrespective of the salinity. Comparing the metal-DOMSPE concentrations in riverine and marine endmember samples (Table 4), we can calculate the relative changes of metal-DOMSPE concentrations as it has been done similarly for DOCSPE and DOSSPE. We found a relative decrease of metal-DOMSPE with increasing salinity in the order Co (90%) > Cr (87%) > Ni (84%) > Cu (82%). Those differences in relative changes of the trace metal concentrations with increasing salinity are similar to DOSSPE and DOCSPE concentrations and cannot be explained by mixing alone. In fact, different transformation and removal processes (as mentioned in the introduction) can influence DOM concentration. The order in relative changes of metal-DOMSPE concentrations with increasing salinity reflects again the Irving-Williams order, indicating that a higher relative decrease in metal-DOMSPE concentration consequently reflects lower affinity to organic ligands. The stronger the affinity of trace metals to organic ligands, the more resistant are the metal-organic complexes against degradation processes.

Trace metal complexation to organic sulfur groups is further supported by the positive correlation of the ratio of DOCSPE/Co-DOMSPE with DOCSPE/DOSSPE (Fig 7). Thus, we found indication for a correlation of Co and sulfur. Comparatively little is known about organic complexation of cobalt in aquatic environments. Studies in the Mediterranean Sea and the Scheldt Estuary suggest partial, but strong complexation of Co to organic ligands [8284]. In organisms, Co and sulfur are coupled via the biosynthesis pathway of methionine: the enzyme methionine synthase is responsible for the regeneration and remethylation of methionine from homocysteinie. In some microorganisms (e.g in E. coli), this enzyme requires the Co-containing cobalamin (vitamin B12) as a cofactor [85].

Following the Irving-Williams order, we assume the affinity of Ni to organic carbon and sulfur groups to be between that of Cu and Co. It is known that about 10 – 60% of Ni in coastal and marine waters is bound by organic ligands [20, 86, 87]. However, it is unclear whether S-containing organic ligands play a role in nickel complexation and we did not find a significant correlation of DOCSPE/Ni-DOMSPE with DOCSPE/DOSSPE ratios in our samples.

Although our results suggest a higher affinity of Cu to sulfur than of Co (Irving-Williams order) we could not find a linear correlation of Cu-DOMSPE with DOSSPE. It is known that dissolved Cu in different aquatic environments is organically complexed by thiols (e.g. [19]). Laglera and van den Berg analyzed copper-thiol complexes in estuarine waters of the Scheldt River, the Netherlands, and found a decrease in copper-thiol complex stability with increasing salinity [19]. However, thiol concentrations in marine waters are usually very low (< 10 nmol L-1) [8890]. Comparing these concentrations with the calculated minimum DOS concentration of 0.34 μmol L-1 in original seawater of the upper East Atlantic Ocean [3], it turns out that thiols contribute to only < 3% of the DOS pool.

We can summarize the differences in riverine and marine trace metal containing DOMSPE by their average molar ratios to be (C107N4P0.013S1)1000V0.05Cr0.33Co0.19Ni0.39Cu3.41As0.47 in the riverine endmember (W1) and (C163N7P0.055S1)1000V0.05Cr0.47Co0.16Ni0.07Cu4.05As0.58 in the marine endmember. Compared to the extended Redfield ratio by Ho et al. of (C95N12P0.8S1)1000Cu0.29Co0.15 for marine phytoplankton [91], we found a considerably higher DOCSPE/DOSSPE ratio in the marine endmember sample, presumably as a result of a more advanced state of degradation. Additionally, we found lower DOCSPE/Cu-DOMSPE and DOCSPE/Co-DOMSPE ratios in our samples compared to marine phytoplankton.

Influence of salinity and sample pre-treatment on extraction and trace element complexes

The DOC extraction efficiencies of 36 ± 2% for pH 2 extracted samples were lower than expected. However, compared to the extraction efficiency for marine DOC of 42 ± 7% (n = 187) found in another study [55], our values are still in the range of uncertainty. Possible reasons for the lower DOC extraction efficiencies might be (i) unknown influence of the source material (ii) too high original DOC measurements or (iii) a problem with the adsorber material. Nevertheless, the results of the quadruplicates of each treatment in our extractions were very consistent, reproducible and invariant, ensuring comparability of our samples.

It has been previously shown that changes in salinity and DOM quality (e.g. at different sample locations in estuaries) can affect DOM recovery via SPE [92, 93]. However, we found no significant effect of salinity on the amount of recovered DOC. Similar DOC extraction efficiencies throughout different salinities suggest little or no fractionation effects as a result of changes in DOM quality. These findings are further supported by an additional experiment, in which low salinity samples from the Weser River were spiked with different concentrations of NaCl and extracted using PPL cartridges (S1 File). The results also showed that the DOC extraction efficiency was not affected by salinity. However, structural changes (indicated by changes in polarity) were observed with changes in salinity, similar to our samples. Thus, we conclude that the polarity of some organic compounds can be reduced by the presence of salt. Structural changes with changes in ionic strength of the medium were also observed for humic acids [94].

Acidification of samples prior to SPE yielded significantly higher DOMSPE and metal-DOMSPE concentrations in the methanol extracts compared with samples extracted at neutral pH and thus gave a more comprehensive picture to discuss changes in DOM stoichiometry and polarity characteristics with changing salinity. It has been shown that acidification leads to higher extraction efficiencies for natural organic matter due to the protonation of functional groups such as organic acids and phenols [38]. Overall, our method is only suitable to extract specific fractions of the natural metal-organic complex pool: the strong acidic fraction in pH 2 extracted samples and the neutral/weak acidic fraction in pH 8 extracted samples, respectively (as defined by Waska et al.) [79]. Cu-DOMSPE extraction efficiencies were similar to those previously reported for acidified and non-acidified PPL extracts [79]. Mills et al. reported decreasing extraction efficiencies of Cu-organic complexes with decreasing pH and mentioned that acidification to pH ≤ 4 did not allow the existence of stabile Cu-DOMSPE complexes [95]. However, another study indicates that also acid-stable Cu-containing compounds can occur in natural aquatic environments [96]. This contradiction reflects that the stability of Cu-organic complexes also depends on the acid-base characteristics of the Cu-binding functional groups and their competitive binding with H+ and possibly other major ions such as Ca2+ and Mg2+. In summary, acidification leads to two competing effects on the recovery of DOMSPE and metal-DOMSPE: (i) an increase in the carbon extraction efficiency due to the protonation of organic matter and (ii) a decrease in the complex extraction efficiency due to reduced stability of protonated/acidified organic complexes. For the riverine samples E1 and W1, acidification led to a relative increase by factor 3–4 in both DOCSPE and DOSSPE concentrations, respectively, compared to the non-acidified samples. For trace metals, the net effect of both processes is reflected in differences in average molar DOCSPE/metal-DOMSPE ratios of pH 2 and pH 8 extracted samples, respectively, which were about a factor of 1.5–2 for Cu and Ni while no net effect was observed for Co. For estuarine and marine samples however, the decrease in trace metals changed by factor 1.3 – 1.7 for Co and 3 – 4 for Ni and Cu, whereas the increase in DOCSPE and DOSSPE concentrations remained similar to those found in riverine samples. Thus, we can conclude that acidification prior to SPE plays an important role, since it improves the recovery of both DOMSPE and metal-DOMSPE. As a result, we found higher extraction efficiencies for pH 2 extracted metal-DOMSPE compared to pH 8 as similarly found for DOMSPE. However, acidification likely leads to changes in the quality of organic ligands. In contrast to DOC and Co, we found a significant correlation of the Cu extraction efficiency with salinity. This can have several reasons: (i) differences in the quality of riverine and marine organic ligands (e.g. a higher binding strength of marine Co-complexing ligands) as indicated by differences in polarity or (ii) ionic strength of the medium. The first assumption is supported by a higher polarity of terrestrial DOM which goes along with lower DOCSPE/metal-DOMSPE ratios. The second assumption might be explained by an increasing amount of inorganic ions with increasing ionic strength of the medium that could compete with the trace metal ions.


In this study we presented the concentration and distribution of DOCSPE, DONSPE, DOSSPE, and dissolved metal-DOMSPE with changing salinity in two rivers draining to the North Sea. With regard to the research question/hypothesis stated in the introduction, we can conclude:

  1. DOMSPE concentrations decreased from riverine to marine waters. The differences in the relative changes in DOCSPE and DOSSPE concentration suggest a preferential removal of DOSSPE over DOCSPE (and DONSPE).
  2. The concentration of some solid-phase extractable trace metals (52Cr, 59Co, 60Ni, 63Cu) was correlated with the DOCSPE and DOSSPE concentrations as a result of the presence of organic complexes. The positive correlations of the DOCSPE/Co-DOMSPE and DOCSPE/Ni-DOMSPE ratios with the DOCSPE/DOSSPE ratio and relatively high extraction efficiencies for Co and Cu suggest complexation of trace metals with organic carbon- and sulfur-containing ligands. Increasing Co and Cu concentrations after UV digestion further supported the presence of strong organic sulfur-trace metal complexes. The affinity of trace metals to (sulfur-containing) organic ligands followed the Irving-Williams order.
  3. DOM polarity reflected typical changes along the estuary from highly polar terrestrial DOM in riverine waters to non-polar DOM compounds in the marine water. This is reflected in a decreasing DOCpol/DOCnon-pol ratio with increasing salinity.
  4. Acidification prior SPE plays an important role and leads to a higher recovery of both DOMSPE and metal-DOMSPE compared to neutral SPE. Higher DOC yield by acidification is more important for the metal yield than the negative effect of acidification on complexation. On a qualitative scale, however, acidification can of course make a big difference for the recovery of different organic ligands.

Supporting information

S1 Fig. UV peak area at 210 nm versus DOCSPE concentrations of all samples.

A significant linear correlation was found for both fractions: the low concentrated fraction (all pH 8 extracted samples and the pH 2 extracted marine sample) with DOCSPE concentrations from 0–40 μmol L-1 (unfilled symbols) and the high concentrated fractions (pH 2 extracted riverine and estuarine samples) with DOCSPE concentrations > 100 μmol L-1 (filled symbols).


S1 Table. Limits of detection for all elements analyzed by ICP-MS, given that solid-phase extraction was performed with an enrichment factor of 430.

These values were calculated according to DIN 32645.


S1 File. This file includes methodical information about the salt-spiking experiment of riverine samples.



We are grateful to the captain and crew of RV Heincke (HE426); I. Stimac is acknowledged for technical support with ICP-MS analysis and Stephan Frickenhaus for his advice regarding statistical questions; we thank W. Geibert for helpful discussions.


  1. 1. Hansell DA, Craig AC. Biogeochemistry of marine dissolved organic matter. Amsterdam: Academic Press; 2002.
  2. 2. Hansell DA, Craig AC. Biogeochemistry of marine dissolved organic matter. 2nd ed. Boston: Academic Press; 2015.
  3. 3. Ksionzek KB, Lechtenfeld OJ, McCallister SL, Schmitt-Kopplin P, Geuer JK, Geibert W, et al. Dissolved organic sulfur in the ocean: Biogeochemistry of a petagram inventory. Science. 2016;354(6311): 456–459. pmid:27789839
  4. 4. Raymond PA, Spencer RGM. Chapter 11—Riverine DOM. In: Hansell DA, Craig AC, editors. Biogeochemistry of Marine Dissolved Organic Matter (Second Edition). Boston: Academic Press; 2015. p. 509–533.
  5. 5. Asmala E, Stedmon CA, Thomas DN. Linking CDOM spectral absorption to dissolved organic carbon concentrations and loadings in boreal estuaries. Estuar Coast Shelf Sci. 2012;111:107–17.
  6. 6. Abdulla HAN, Minor EC, Dias RF, Hatcher PG. Changes in the compound classes of dissolved organic matter along an estuarine transect: A study using FTIR and 13C NMR. Geochim Cosmochim Acta. 2010;74(13): 3815–38.
  7. 7. Gogou A, Repeta DJ. Particulate-dissolved transformations as a sink for semi-labile dissolved organic matter: Chemical characterization of high molecular weight dissolved and surface-active organic matter in seawater and in diatom cultures. Mar Chem. 2010;121(1): 215–23.
  8. 8. Kaiser K, Guggenberger G. The role of DOM sorption to mineral surfaces in the preservation of organic matter in soils. Org Geochem. 2000;31(7): 711–25.
  9. 9. Miller WL, Moran MA. Interaction of photochemical and microbial processes in the degradation of refractory dissolved organic matter from a coastal marine environment. Limnology and Oceanography. 1997;42(6):1317–24.
  10. 10. Sepers ABJ. The utilization of dissolved organic compounds in aquatic environments. Hydrobiologia. 1977;52(1): 39–54.
  11. 11. Elifantz H, Dittel AI, Cottrell MT, Kirchman DL. Dissolved organic matter assimilation by heterotrophic bacterial groups in the western Arctic Ocean. Aquat Microb Ecol. 2007;50(1): 39–49.
  12. 12. Cai Y, Guo L, Wang X, Mojzis AK, Redalje DG. The source and distribution of dissolved and particulate organic matter in the Bay of St. Louis, northern Gulf of Mexico. Estuar Coast Shelf Sci. 2012;96: 96–104.
  13. 13. Abril G, Nogueira M, Etcheber H, Cabeçadas G, Lemaire E, Brogueira MJ. Behaviour of Organic Carbon in Nine Contrasting European Estuaries. Estuar Coast Shelf Sci. 2002;54(2): 241–62.
  14. 14. Ylöstalo P, Seppälä J, Kaitala S, Maunula P, Simis S. Loadings of dissolved organic matter and nutrients from the Neva River into the Gulf of Finland–Biogeochemical composition and spatial distribution within the salinity gradient. Mar Chem. 2016;186: 58–71.
  15. 15. Charlson RJ, Lovelock JE, Andreae MO, Warren SG. Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate. Nature. 1987;326(6114): 655–661.
  16. 16. Lomans BP, van der Drift C, Pol A, Op den Camp HJM. Microbial cycling of volatile organic sulfur compounds. Cell Mol Life Sci. 2002;59(4): 575–588. pmid:12022467
  17. 17. Bentley R, Chasteen TG. Environmental VOSCs—formation and degradation of dimethyl sulfide, methanethiol and related materials. Chemosphere. 2004;55(2): 291–317.
  18. 18. Smith DS, Bell RA, Kramer JR. Metal speciation in natural waters with emphasis on reduced sulfur groups as strong metal binding sites. Comp Biochem Physiol C Toxicol Pharmacol. 2002;133(1–2): 65–74. pmid:12356517
  19. 19. Laglera LM, van den Berg CMG. Copper complexation by thiol compounds in estuarine waters. Mar Chem. 2003;82(1–2): 71–89.
  20. 20. Vraspir JM, Butler A. Chemistry of marine ligands and siderophores. Ann Rev Mar Sci. 2009;1: 43–63. pmid:21141029
  21. 21. Bell RA, Kramer JR. Structural chemistry and geochemistry of silver-sulfur compounds: Critical review. Environ Toxicol Chem. 1999;18(1): 9–22.
  22. 22. Ravichandran M. Interactions between mercury and dissolved organic matter—A review. Chemosphere. 2004;55(3): 319–331. pmid:14987930
  23. 23. Kolber ZS, Barber RT, Coale KH, Fitzwateri SE, Greene RM, Johnson KS, et al. Iron limitation of phytoplankton photosynthesis in the equatorial Pacific Ocean. Nature. 1994;371(6493): 145–149.
  24. 24. Twining BS, Baines SB. The trace metal composition of marine phytoplankton. Ann Rev Mar Sci. 2013;5(1): 191–215.
  25. 25. Twining BS, Baines SB, Fisher NS, Landry MR. Cellular iron contents of plankton during the Southern Ocean Iron Experiment (SOFeX). Deep Sea Res Part I: Oceanogr Res Pap. 2004;51(12): 1827–1850.
  26. 26. Driscoll CT, Otton JK, Ake I. Trace Metals Speciation and Cycling. In: Moldan B, Cerny J, editors. SCOPE 51: Biogeochemistry of Small Cachments A Tool for Environmental Research. 51: John Wiley & Sons; 1994.
  27. 27. Sunda WG. Feedback Interactions between trace metal nutrients and phytoplankton in the ocean. Front Microbiol. 2012;3: 204. pmid:22701115
  28. 28. Paytan A, Mackey KRM, Chen Y, Lima ID, Doney SC, Mahowald N, et al. Toxicity of atmospheric aerosols on marine phytoplankton. Proc Natl Acad Sci. 2009;106(12): 4601–1405. pmid:19273845
  29. 29. De Schamphelaere KAC, Vasconcelos FM, Tack FMG, Allen HE, Janssen CR. Effect of dissolved organic matter source on acute copper toxicity to Daphnia magna. Environ Toxicol Chem. 2004;23(5): 1248–1255. pmid:15180376
  30. 30. Schwartz ML, Curtis PJ, Playle RC. Influence of natural organic matter source on acute copper, lead, and cadmium toxicity to rainbow trout (Oncorhynchus mykiss). Environ Toxicol Chem. 2004;23(12): 2889–2899. pmid:15648764
  31. 31. Hsu-Kim H. Stability of metal−glutathione complexes during oxidation by hydrogen peroxide and Cu(II)-catalysis. Environ Sci Technol. 2007;41(7): 2338–2342. pmid:17438784
  32. 32. Dupont CL, Nelson RK, Bashir S, Moffett JW, Ahner BA. Novel copper-binding and nitrogen-rich thiols produced and exuded by Emiliania huxleyi. Limnol Oceanogr. 2004;49(5): 1754–1762.
  33. 33. Dupont CL, Ahner BA. Effects of copper, cadmium, and zinc on the production and exudation of thiols by Emiliania huxleyi. Limnol Oceanogr. 2005;50(2): 508–515.
  34. 34. Moffett JW, Brand LE. Production of strong, extracellular Cu chelators by marine cyanobacteria in response to Cu stress. Limnol Oceanogr. 1996;41(3): 388–395.
  35. 35. Croot PL, Moffett JW, Brand LE. Production of extracellular Cu complexing ligands by eucaryotic phytoplankton in response to Cu stress. Limnol Oceanogr. 2000;45(3): 619–627.
  36. 36. Baken S, Degryse F, Verheyen L, Merckx R, Smolders E. Metal complexation properties of freshwater dissolved organic matter are explained by its aromaticity and by anthropogenic ligands. Environ Sci Technol. 2011;45(7): 2584–2590. pmid:21405071
  37. 37. Matar Z, Chebbo G, Troupel M, Boudhamane L, Parlanti E, Uher E, et al. Influence of organic matter from urban effluents on trace metal speciation and bioavailability in river under strong urban pressure. In: Xu J, Wu J, He Y, editors. Functions of natural organic matter in changing environment. Dordrecht: Springer Netherlands; 2013. p. 517–521.
  38. 38. Dittmar T, Koch BP, Hertkorn N, Kattner G. A simple and efficient method for the solid-phase extraction of dissolved organic matter (SPE-DOM) from seawater. Limnol Oceanogr: Methods. 2008;6: 230–235.
  39. 39. Koch BP, Rohardt G. Continuous thermosalinograph oceanography along HEINCKE cruise track HE426; 2016. Database: PANGAEA [Internet]. Available from:
  40. 40. Dick D, Wegner A, Gabrielli P, Ruth U, Barbante C, Kriews M. Rare earth elements determined in Antarctic ice by inductively coupled plasma—Time of flight, quadrupole and sector field-mass spectrometry: An inter-comparison study. Anal Chim Acta. 2008;621(2): 140–147. pmid:18573377
  41. 41. Biller DV, Bruland KW. Analysis of Mn, Fe, Co, Ni, Cu, Zn, Cd, and Pb in seawater using the Nobias-chelate PA1 resin and magnetic sector inductively coupled plasma mass spectrometry (ICP-MS). Mar Chem. 2012;130–131: 12–20.
  42. 42. Hathorne EC, Haley B, Stichel T, Grasse P, Zieringer M, Frank M. Online preconcentration ICP-MS analysis of rare earth elements in seawater. Geochem Geophys Geosyst. 2012;13(1): Q01020.
  43. 43. Takano S, Tanimizu M, Hirata T, Sohrin Y. Determination of isotopic composition of dissolved copper in seawater by multi-collector inductively coupled plasma mass spectrometry after pre-concentration using an ethylenediaminetriacetic acid chelating resin. Anal Chim Acta. 2013;784: 33–41. pmid:23746405
  44. 44. Jochum KP, Nohl U, Herwig K, Lammel E, Stoll B, Hofmann AW. GeoReM: A new geochemical database for reference materials and isotopic standards. Geostand Geoanal Res. 2005;29(3): 333–338.
  45. 45. Koch BP, Ludwichowski KU, Kattner G, Dittmar T, Witt M. Advanced characterization of marine dissolved organic matter by combining reversed-phase liquid chromatography and FT-ICR-MS. Mar Chem. 2008;111: 233–241.
  46. 46. Lechtenfeld OJ, Koch BP, Geibert W, Ludwichowski KU, Kattner G. Inorganics in organics: Quantification of organic phosphorus and sulfur and trace element speciation in natural organic matter using HPLC-ICPMS. Anal Chem. 2011;83(23): 8968–8974. pmid:21992549
  47. 47. Benner R, Opsahl S. Molecular indicators of the sources and transformations of dissolved organic matter in the Mississippi river plume. Org Geochem. 2001;32(4): 597–611.
  48. 48. Guo W, Stedmon CA, Han Y, Wu F, Yu X, Hu M. The conservative and non-conservative behavior of chromophoric dissolved organic matter in Chinese estuarine waters. Mar Chem. 2007;107(3): 357–366.
  49. 49. Huguet A, Vacher L, Relexans S, Saubusse S, Froidefond JM, Parlanti E. Properties of fluorescent dissolved organic matter in the Gironde Estuary. Org Geochem. 2009;40(6): 706–719.
  50. 50. Gonçalves-Araujo R, Stedmon CA, Heim B, Dubinenkov I, Kraberg A, Moiseev D, et al. From fresh to marine waters: Characterization and fate of dissolved organic matter in the Lena River Delta Region, Siberia. Front Mar Sci. 2015;2: 108.
  51. 51. Spitzy A. Water chemistry measurements of different Sites; 2005. Database: PANGAEA [Internet]. Available from:
  52. 52. Pohlabeln AM, Dittmar T. Novel insights into the molecular structure of non-volatile marine dissolved organic sulfur. Mar Chem. 2015;168(0): 86–94.
  53. 53. Ksionzek KB, Lechtenfeld OJ, McCallister SL, Schmitt-Kopplin P, Geuer J, Geibert W, et al. Dissolved organic sulfur in solid-phase extracts of water samples obtained during POLARSTERN cruise ANT-XXV (PS73); 2016. Database: PANGAEA [Internet]. Available from:
  54. 54. Flerus R, Lechtenfeld OJ, Koch BP, McCallister SL, Schmitt-Kopplin P, Benner R, et al. A molecular perspective on the ageing of marine dissolved organic matter. Biogeosciences. 2012;9: 1935–1955.
  55. 55. Lechtenfeld OJ, Kattner G, Flerus R, McCallister SL, Schmitt-Kopplin P, Koch BP. Molecular transformation and degradation of refractory dissolved organic matter in the Atlantic and Southern Ocean. Geochim Cosmochim Acta. 2014;126: 321–337.
  56. 56. Houle D, Carignan R, Lachance M, Dupont J. Dissolved organic carbon and sulfur in southwestern Québec lakes: Relationships with catchment and lake properties. Limnol Oceanogr. 1995;40(4): 710–717.
  57. 57. Matrai PA, Eppley RW. Particulate organic sulfur in the waters of the Southern California Bight. Glob Biogeochem Cycles. 1989;3(1): 89–103.
  58. 58. Homann PS, Mitchell MJ, Miegroet HV, Cole DW. Organic sulfur in throughfall, stem flow, and soil solutions from temperate forests. Can J For Res. 1990;20(9): 1535–1539.
  59. 59. Hopkinson CS, Vallino JJ. Efficient export of carbon to the deep ocean through dissolved organic matter. Nature. 2005;433(7022): 142–145. pmid:15650735
  60. 60. Benner R, Pakulski JD, McCarthy M, Hedges JI, Hatcher PG. Bulk chemical characteristics of dissolved organic matter in the ocean. Science. 1992;225: 1561–1564.
  61. 61. Dittmar T, Kattner G. Recalcitrant dissolved organic matter in the ocean: major contribution of small amphiphilics. Mar Chem. 2003;82(1): 115–123.
  62. 62. Opsahl S, Benner R. Distribution and cycling of terrigenous dissolved organic matter in the ocean. Nature. 1997;386(6624): 480–482.
  63. 63. Sleighter RL, Hatcher PG. Molecular characterization of dissolved organic matter (DOM) along a river to ocean transect of the lower Chesapeake Bay by ultrahigh resolution electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Marine Chem. 2008;110: 140–52.
  64. 64. Abdulla HAN, Minor EC, Dias RF, Hatcher PG. Transformations of the chemical compositions of high molecular weight DOM along a salinity transect: Using two dimensional correlation spectroscopy and principal component analysis approaches. Geochimica Et Cosmochimica Acta. 2013;118: 231–46.
  65. 65. Stedmon CA, Markager S. Behaviour of the optical properties of coloured dissolved organic matter under conservative mixing. Estuar Coast Shelf Sci. 2003;57(5–6): 973–979.
  66. 66. Alling V, Sanchez-Garcia L, Porcelli D, Pugach S, Vonk JE, van Dongen B, et al. Nonconservative behavior of dissolved organic carbon across the Laptev and East Siberian seas. Glob Biogeochem Cycles. 2010;24(4): GB4033.
  67. 67. Castillo CR, Sarmento H, Álvarez-Salgado XA, Gasol JM, Marraséa C. Production of chromophoric dissolved organic matter by marine phytoplankton. Limnol Oceanogr. 2010;55(1): 446–454.
  68. 68. Rochelle-Newall EJ, Pizay M-D, Middelburg JJ, Boschker HTS, Gattuso J-P. Degradation of riverine dissolved organic matter by seawater bacteria. Aquat Microb Ecol. 2004;37(1): 9–22.
  69. 69. Helms JR, Stubbins A, Perdue EM, Green NW, Chen H, Mopper K. Photochemical bleaching of oceanic dissolved organic matter and its effect on absorption spectral slope and fluorescence. Mar Chem. 2013;155: 81–91.
  70. 70. Gonsior M, Hertkorn N, Conte MH, Cooper WJ, Bastviken D, Druffel E, et al. Photochemical production of polyols arising from significant photo-transformation of dissolved organic matter in the oligotrophic surface ocean. Mar Chem. 2014;163: 10–18.
  71. 71. Gonsior M, Schmitt-Kopplin P, Bastviken D. Depth-dependent molecular composition and photo-reactivity of dissolved organic matter in a boreal lake under winter and summer conditions. Biogeosciences. 2013;10(11): 6945–6956.
  72. 72. Opsahl S, Benner R. Photochemical reactivity of dissolved lignin in river and ocean waters. Limnol Oceanogr. 1998;43(6): 1297–1304.
  73. 73. Asmala E, Bowers DG, Autio R, Kaartokallio H, Thomas DN. Qualitative changes of riverine dissolved organic matter at low salinities due to flocculation. J Geophys Res: Biogeosci. 2014;119(10): 1919–1933.
  74. 74. Wen L-S, Santschi P, Gill G, Paternostro C. Estuarine trace metal distributions in Galveston Bay: importance of colloidal forms in the speciation of the dissolved phase. Mar Chem. 1999;63(3–4): 185–212.
  75. 75. Santschi PH, Lenhart JJ, Honeyman BD. Heterogeneous processes affecting trace contaminant distribution in estuaries: The role of natural organic matter. Mar Chem. 1997;58(1): 99–125.
  76. 76. Guo L, Santschi PH, Warnken KW. Trace metal composition of colloidal organic material in marine environments. Mar Chem. 2000;70(4): 257–275.
  77. 77. Sholkovitz ER. The flocculation of dissolved Fe, Mn, Al, Cu, Ni, Co and Cd during estuarine mixing. Earth Planet Sci Lett. 1978;41(1):7 7–86.
  78. 78. Coale KH, Bruland KW. Copper complexation in the Northeast Pacific. Limnol Oceanogr. 1988;33(5): 1084–1101.
  79. 79. Waska H, Koschinsky A, Ruiz Chancho MJ, Dittmar T. Investigating the potential of solid-phase extraction and Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) for the isolation and identification of dissolved metal–organic complexes from natural waters. Mar Chem. 2015;173: 78–92.
  80. 80. Milne A, Landing W, Bizimis M, Morton P. Determination of Mn, Fe, Co, Ni, Cu, Zn, Cd and Pb in seawater using high resolution magnetic sector inductively coupled mass spectrometry (HR-ICP-MS). Anal Chim Acta. 2010;665(2): 200–207. pmid:20417331
  81. 81. Mantoura RFC, Dickson A, Riley JP. The complexation of metals with humic materials in natural waters. Estuar Coast Mar Sci. 1978;6(4): 387–408.
  82. 82. Vega M, van den Berg CMG. Determination of Cobalt in Seawater by Catalytic Adsorptive Cathodic Stripping Voltammetry. Anal Chem. 1997;69(5): 874–81. pmid:21639224
  83. 83. Zhang H, Van Den Berg CMG, Wollast R. The determination of interactions of cobalt (II) with organic compounds in seawater using cathodic stripping voltammetry. Mar Chem. 1990;28(4): 285–300.
  84. 84. Ellwood MJ, van den Berg CMG. Determination of organic complexation of cobalt in seawater by cathodic stripping voltammetry. Mar Chem. 2001;75(1): 33–47.
  85. 85. Zydowsky TM, Courtney LF, Frasca V, Kobayashi K, Shimizu H, Yuen LD, et al. Stereochemical analysis of the methyl transfer catalyzed by cobalamin-dependent methionine synthase from Escherichia coli B. J Am Chem Soc. 1986;108(11): 3152–3153.
  86. 86. Saito MA, Moffett JW, DiTullio GR. Cobalt and nickel in the Peru upwelling region: A major flux of labile cobalt utilized as a micronutrient. Glob Biogeochem Cycles. 2004;18(4): GB4030.
  87. 87. Donat JR, Lao KA, Bruland KW. Speciation of dissolved copper and nickel in South San Francisco Bay: a multi-method approach. Anal Chim Acta. 1994;284(3): 547–571.
  88. 88. Al-Farawati R, Van Den Berg CMG. Thiols in coastal waters of the western North Sea and English Channel. Environ Sci Technol. 2001;35(10): 1902–1911. pmid:11393967
  89. 89. Tang D, Hung C-C, Warnken KW, Santschi PH. The distribution of biogenic thiols in surface waters of Galveston Bay. Limnol Oceanogr. 2000;45(6): 1289–1297.
  90. 90. Dupont CL, Moffett JW, Bidigare RR, Ahner BA. Distributions of dissolved and particulate biogenic thiols in the subartic Pacific Ocean. Deep Sea Res Part I Oceanogr Res Pap. 2006;53(12): 1961–1974.
  91. 91. Ho TY, Quigg A, Finkel ZV, Milligan AJ, Wyman K, Falkowski PG, et al. The elemental composition of some marine phytoplankton. J Phycol. 2003;39(6): 1145–1159.
  92. 92. Kruger BR, Dalzell BJ, Minor EC. Effect of organic matter source and salinity on dissolved organic matter isolation via ultrafiltration and solid phase extraction. Aquat Sci. 2011;73(3): 405–417.
  93. 93. Chen M, Kim S, Park J-E, Kim HS, Hur J. Effects of dissolved organic matter (DOM) sources and nature of solid extraction sorbent on recoverable DOM composition: Implication into potential lability of different compound groups. Anal Bioanal Chem. 2016;408(17): 4809–4819. pmid:27117255
  94. 94. Tsutsuki K, Kuwatsuka S. Molecular size distribution of humic acids as affected by the ionic strength and the degree of humification. Soil Sci Plant Nutr. 1984;30(2): 151–162.
  95. 95. Mills GL, Hanson AK, Quinn JG, Lammela WR, Chasteen ND. Chemical studies of copper-organic complexes isolated from estuarine waters using C18 reverse-phase liquid chromatography. Mar Chem. 1982;11(4): 355–377.
  96. 96. Reddy KJ, Wang L, Gloss SP. Solubility and mobility of copper, zinc and lead in acidic environments. Plant Soil. 1995;171(1): 53–58.