Distribution of Geochemical Species of P, Fe and Mn in Surface Sediments in the Eutrophic Estuary, Northern Taiwan

: The Danshuei River Estuary (DRE) in northern Taiwan is a seriously eutrophic estuary due to the domestic efﬂuent discharge. Surface sediment samples were collected from the DRE to study the concentrations and spatial distributions of different fractions of phosphorus through the ﬁve-step sequential extraction method which chemically divides the sedimentary P into ﬁve fractions: P SORB , P CDB , P CFA , P DET , and P ORG . The Fe and Mn contents in the extracted solution were also determined. The total organic carbon (TOC) and grain size in sediment samples were analyzed as well. The sedimentary total P (TP) concentrations ranged within 537–1310 mg/kg and mostly exceeded 800 mg/kg, suggesting that the DRE sediments were moderately polluted by phosphorus. The P CDB was the dominant fraction of P, averagely contributing 58% of TP, followed by P DET 31%. The contributions of the P SORB and P CFA fractions to the TP were relatively minor. Two fractions, Fe CDB and Fe ORG, of sedimentary Fe equally shared approximately 70% of total Fe, followed by Fe DET with 22%. The contribution of different fractions of sedimentary Mn followed the sequence: Mn CDB (36%) > Mn CFA (29%) > Mn ORG (14.7%) > Mn DET (14.5%) > Mn SORB (5.3%). The sedimentary P, Fe, and Mn within the DRE are easily mobilized because they were mainly present in the reducible fraction. The concentrations of sedimentary TP positively correlated with the TOC contents and inversely negatively correlated with grain size, suggesting that the TOC and grain size play the crucial roles in inﬂuencing the distribution of sedimentary P within the DRE. Finally, the Fe(III) (hydro)oxides seems to play an important carriers to adsorb dissolved P because P CDB positively correlated with Fe CDB .


Introduction
Phosphorus and nitrogen are the most important essential elements among the trace elements for marine organisms, especially for phytoplankton, and they control the primary productivity in the marine environment [1]. Both elements play a crucial role in influencing the marine carbon cycle and consequently affect the global biogeochemical cycles [1][2][3]. The dissolved nitrogen can be utilized by picophytoplankton, such as Prochlorococcus and Synechococcus, in oligotrophic waters through nitrogen fixation processes. Thus, phosphorus is generally considered as the bio-limiting element in the oligo oceanic environment due to the lack of a P source [2]. However, in the last three decades, estuarine and coastal environments have been frequently reported as being subject to eutrophic or hypereutrophic conditions because most rivers worldwide are more or less polluted by nutrients due to anthropogenic inputs, such as increased land drainage, agriculture fertilization, livestock, and urban/industrial wastewater input [4][5][6][7][8]. The eutrophication in the aquatic environment, especially in estuaries, is recognized as a globally pertinent environmental issue [3,9].
Unlike dissolved nitrogen, phosphorus is not a redox sensitive element, and orthophosphate is the dominant species of dissolved inorganic P (DIP) in estuarine and coastal water [3]. DIP is recognized as a high affinity ion with particles, especially iron (III) (hydro)oxides and hydrous aluminum oxides, which have positive surface charges at slightly acid to neutral pH, inducing high affinities for anions, the important adsorbents for DIP in the water column ( [10], and references cited therein]. Despite that, the desorption of particulate phosphorus (PP) during estuarine mixing is frequently observed in the estuary where the salinity increases and enhances the major anion (Cl − & SO 4 2− ) competition for adsorption sites, a process that promotes PP desorption [11,12]. The PP is not desorbed during the estuarine mixing and settles to the estuarine sediment, where it becomes an important reservoir of P within the sediments [13][14][15]. However, the settled PP in the sediment could be desorbed from the resuspension sediment [16,17] or re-dissolved, accompanying the dissolution of iron (Fe) (oxyhydr)oxides under the anoxic condition [18][19][20]. Sedimentary P cycling in the marine environment is quite complicated due to the speciation transformation. The classical diagram of P recycling after burial in marine sediment is that the non-refractory fraction of sedimentary phosphorus is released to interstitial waters due to the degradation of organic materials or desorption of the redox sensitive phosphorus associated with iron oxyhydroxides. The released P may be adsorbed by the grain surface or captured by iron oxyhydroxides, but P is ultimately taken up in situ in a mineralized form, most likely the authigenic minerals which may include the calcium (Ca)-P mineral carbonate fluorapatite [21][22][23] and vivianite, an Fe(II)-P mineral [24,25], within the sediments. The authigenic carbonate fluorapatite and detrital forms of P are considered as the refractory fraction of sedimentary phosphorus and are not regenerated during the diagenetic processes [26]. Thus, for better understanding the geochemical behavior of sedimentary P during the burial process, it is necessary to quantify the sedimentary P speciation. Many sequential extraction methods, chemically separating the sedimentary P according to different speciation, have been developed [20,[26][27][28][29][30]. One of the most popular sequential extraction methods is the five-step sequential extraction method (SEDEX) developed by Ruttenberg [26], who indicates that the analytical method can successfully separate two of the main categories of authigenic phosphate phases. The method chemically divides the sedimentary P into five fractions: P exchangeable , P CFAP , P Fe, P detrital , and P Organic . A detailed description of the individual P speciation analyzed by the SEDEX method can be found in the work of Ruttenberg [26]. This method is widely used to differentiate the sedimentary P speciation, and to study the P distribution and geochemical cycle in the marine environment [31][32][33][34][35][36].
The Danshuei River Estuary (DRE) in Northern Taiwan ( Figure 1) is characterized by relatively high DIP and DIN (dominated by ammonium) concentrations, exceeding 10 µM and 500 µM, respectively [8,37,38]. The dissolved oxygen in the water column of the estuary generally ranges within 2-5 mg/L, falling in moderately dysoxic to oxic conditions. The upper estuary, at salinity < 5 psu region, frequently approaches the conditions of a hypoxic environment due to the domestic sewage effluent discharge, and the dissolved oxygen gradually increases with salinity due to mixing with the tidal seawater [37,38]. Comprehensive studies focusing on the partitioning, distribution, and species of P and N in the estuarine water were addressed in previous works [8,[37][38][39]. Some of the partitioning characteristics of P found in our previous studies are that the dissolved P was the dominant P (50-89% of total P pool) and inorganic species was the major fraction in both dissolved and particulate phases within the DRE. The higher dissolved organic P concentrations were generally relevant to the higher Chl.a concentrations [8]. The total and inorganic particulate P concentrations correlate well with those of the particulate Fe and Mn elements, suggesting that particulate Fe and Mn played important roles in regulating the particulate P within the DRE [8,39]. Meanwhile, the current understanding of the phosphorus pollution status in the DRE sediment is very limited, and it is well documented that the geochemical cycle of sedimentary P is rather complicated and can be strongly influenced by the speciation of sedimentary P, as mentioned above. In order to understand the geochemical fraction, spatial distribution, and pollution status of sedimentary P within the DRE, this study employed the SEDEX method to quantify sedimentary P speciation in the DRE and to establish such a knowledge in the seriously eutrophic estuary. In addition, the sedimentary Fe and Mn concentrations in the SEDEX extracted solutions were also analyzed because both elements usually play the crucial roles in altering the sedimentary P geochemistry in the marine environment. wastewater treatment plant). R1-R3 represents the investigation stations of the water quality carried out by EPA, Taiwan. These three stations are located in the tributaries, namely the Keelung (R1), the Hsintien (R2) and the Tanhan (R3), of the Danshuei River system.

Study Area
The DRE, located on the outskirts of northern Taipei, is the largest estuary in northern Taiwan with a drainage area of 2726 km 2 . More than six million people, over a quarter of Taiwan's entire population, reside in the catchment area. The main tributaries of the Danshuei River are the Tanhan, Hsintien, and Keelung Rivers. The Hsintien River merges with the Tanhan River in the Banqiao county of the New Taipei City to form the upper estuary of the DRE. The Keelung River joins the DRE at the lower estuary in the Shezidao region of the Taipei City. The monthly average river fluxes, ranging within 1.54-142.794 m 3 /s, of the three major tributaries of the DRE during 2018-2019 are plotted in Figure 2. The monthly average precipitation, ranging within 13.8-439.3 mm, in the catchment area of the DRE during the periods 2018-2019 is shown in Figure 2 (http://www.cwa.gov.tw, accessed on 12 October 2021). Most precipitation occurred during the late May and June, which is influenced by the southwest monsoon. In addition, a heavy precipitation may occur during July to September depending on the tropical typhoon coming or not. The long-term average annual river flux in the Danshuei River system is 6600 × 10 6 m 3 /y with freshwater contributions from the major tributaries, namely 27% from the Keelung River, 31% from the Dahan River, and 37% from the Hsintien River, respectively [37]. Due to the treated and untreated municipal wastewaters discharged into the river system, there is a long record of nutrient pollution in the DRE [8,[37][38][39]. Three wastewater treatment plants (WWTP) are located in the river estuary system (Figure 1). The largest one, namely the Bali WWTP, with daily sewage treatment of approximately 1.05 × 10 6 m 3 /d, is located on the DRE mouth west bank, and the treated effluent is discharged into the seawater through an ocean outfall pipe. The other two WWTPs are located on the upper estuary and discharge the treated effluent into the river. Each of these two WWTPs has daily sewage treatment of approximately 4.0 × 10 5 m 3 /d [40]. Figure 3 shows the historical data of dissolved inorganic nitrogen (DIN, ammonium + nitrite + nitrate) and dissolved total P (DTP) concentrations at the survey stations during the period of 2002-2016 according to the water quality survey data in the three major tributaries of the  Figure 3 also indicates that values of DIN and DTP found at the Tanhan (R3) River were relatively higher than those of the Hsintien (R2) River. The reason for this is probably attributed to the dilution effect of the river fluxes. It can be seen in Figure 2 that the river fluxes of the Hsintien River were significantly higher than that of the Tanhan River during 2018-2019. The municipal wastewaters discharged into the DRE are inducing the DRE to become a poorly eutrophic estuary according to the indicator threshold values of DIP > 0.1 mg/L and DIN > 1 mg/L [9]. The water depth of the whole estuary is generally about 2 m at low tide and the maximum depth may be over 10 m at high tide. Tidal seawater, which can intrude into the upper estuary approximately 25 km from the river mouth, can mix well with the river water during floods but is partially mixed during ebb [42]. The distinctive difference between the upper and the lower estuary of the DRE is that the dissolved oxygen concentration in the upper estuary is generally <3 mL/L and approaches saturation due to mixing with the tidal coastal water [8,[37][38][39]. One of the characteristics of the Taiwanese rivers is that most rivers are small mountainous rivers which have great sediment yields and their values are generally 2-3 orders of magnitude greater than the global average [43,44]. The mean annual sediment deposition rates, sediment load, and sediment yield of the DRE during 1949-1990 were 0.28-2.2 cm/year, 1.45 Mt/year, and 4200 t/km 2 /year, respectively [43,44]. Due to the presence of mountains and narrowness of the river, there is no significant wind-induced current, except for the occasional storm surges induced by the Western Pacific typhoon. The major forcing mechanisms of the barotropic flows are astronomical tide at the river mouth and river discharges at upriver ends. Semi-diurnal tides are the principal tidal constituents, with a mean tidal range of 2.22 m and a spring tidal range of 3.1 m, respectively [42].

Sampling
The DRE estuarine water and surface sediment samples were collected on 18 April, 26 September, 17 December 2018, and 4 April 2019 by employing a small fishing boat. The surficial sediment samples were collected using a grabber and placed in clean plastic bags. Samples were immediately placed in a cooler box on board. The sampling stations along the DRE are shown in Figure 1. After returning to the laboratory, one portion of the sediment sample at each station was gently rinsed with Milli-Q water to remove the interstitial water and was freeze-dried for 5 days using a freeze drier instrument. Rinsing sediment samples with Milli-Q water and the freeze-dried samples should have no influence on the analytical results. Berner and Rao [45] indicate that the sedimentary P measurements were not affected by oxidation during the drying process. The dried sediment samples were ground with a mortar and pestle and placed in glass vials for analysis. One portion of the untreated sediment sample at each station was directly analyzed for grain size.

Analytical Methods
One portion of the dried sediment collected at each station was analyzed for the different P species based on the five-step sequential extraction method (SEDEX) and the analytical procedure, the reaction mechanisms and operationally defined P forms of the SEDEX method can refer to Ruttenberg [26]. Briefly, the SEDEX method chemically divides the sedimentary P into five fractions: loosely sorbed P, hereafter referred to as P SORB ; ferric iron-bound P (P CDB ); authigenic carbonate fluorapatite + biogenic apatite + CaCO 3 -bound P (P CFA ); detrital apatite P (P DET ); and organic P (P ORG ). A briefly analytical procedure is described here. The bulk sediment sample (ca. 0.5 g) was extracted with 30 mL of the chemical reagent used in each fraction. The mixture was centrifuged at 4000 rpm for 5 min after the analytical procedure. Due to the phosphorus content differing from each extractant, a different proportion of the supernatant in each extractant was transferred into the 30 mL polypropylene vials and diluted to 25 mL with Milli-Q water. The measurement of P in the diluted solution was employed the molybdenum blue spectrophotometric method [46]. The dilution of each extractant can also eliminate the interference from the extracting chemicals and the matrix leached from the samples. Prior to the next step of the analysis, the sample was washed with 30 mL of Milli-Q water twice to remove the chemicals residual in the sample.
In addition, the iron and manganese concentrations in the extracted solution in each fraction were also determined using flame atomic absorption spectrophotometry (GFAAS) by a Perkin Elmer Analyst 900, and their analyzed species were also referred to as the same species as P. In order to examine the analytical accuracy of the five-step sequential extraction method, the MESS-3 and BCSS-1 reference materials (National Research Council of Canada) were used to trace the analytical quality. The analytical results are shown in Table 1 and the total concentrations of P, Fe, and Mn measured in the MESS-3 and BSCC-1 reference materials (n = 6, one standard deviation) are as follows: P, 1203 ± 100 mg/kg and 583.3 ± 69.4 mg/kg; Fe, 4.34 ± 0.11 % and 2.43 ± 0.09 %; Mn, 291.6 ± 17.1 mg/kg and 200.1 ± 5.85 mg/kg, respectively. The measured concentration ratios to the certified value and precision (one standard deviation) are as follow: P, 100.2 ± 8.4% and 86.8 ± 8.4%; Fe, 76.6 ± 6.2% and 73.9 ± 2.7%; Mn, 90.0 ± 5.3% and 87.4 ± 2.6 %. The analytical accuracy of P measurement obtained in MESS-3 material is better than that of BCSS-1 material and the analytical accuracy in the both reference materials ranges within 85-100%, suggesting that the analyzed data of the five-step sequential extraction method are reliable. The analyzed results for Fe and Mn obtained in both materials differ slightly. However, the analytical accuracy of Fe in both materials approaches 75% of their certified values, suggesting that the SEDEX method cannot destroy the element present in the residual fraction of the minerals. These results are in good agreement with the literature published by Kryc et al. [28] who use a modified sequential extraction procedure outlined in three published methods to evaluate the analytical accuracy of some reference materials, such as BCSS-1, MESS-1, MAG-1, and NIST-1. Their analytical results indicate that Fe and Mn ratios of the sum of sequential extraction concentrations to the total certified BCSS and MESS-1 values are approximately 0.8 and 0.75, and under-recovery is related to the lithology of sedimentary reference materials. Thus, the analyzed trace metal concentrations will be underestimated when the SEDEX method is used to analyze the sediment samples, as mentioned above (The data of this study can be found in the supplementary materials). The dried sediments were also analyzed for total organic carbon (TOC) using a Horbia carbon analyzer 8210 after the samples were smoked with concentrated HCl acid in a closed container for 48 h to remove the inorganic C content. The detailed TOC analytical procedure can be found in Fang and Hong [47]. The sediment grain size measurement was based on the analytical method published by Folk [48] and the grain size in each sample was divided into four fractions: medium sand (MS, >177 µm); fine sand (FS, 125-177 µm); very fine sand (VFS, 3-125 µm); mud (<63 µm).

Grain Size and TOC
The grain sizes and TOC concentrations in the DRE surface sediments range within 20-470 µm and 0.11-2.00 % during the four surveys and their distributions at the studied stations are plotted in Figure 4, which shows the values were inconsistent and rather scatter during different surveys. However, Figure 4 shows that the grain sizes at the middle estuary stations (station 4-9) were relatively smaller than those found at the upper and the lower estuary stations which were mostly fine sand (125-177 µm), very fine sand (63-125 µm) and mud (<63 µm). The percentages of the four fractions of sediment grain sizes at each station obtained in the four surveys are plotted in Figure 5, and the percentage ranges of each fraction were as follows: MS, 0.5-94.5%; FS, 1.2-91.9%; VFS, 0.1-45.2%; and mud, <0.1-85.8%, respectively. The salinity, data taken from the previous study [8], in surface water at each station was also shown in Figure 5. The salinity ranged within <0.1-32.2 psu and varied in different surveys. A relatively lower salinity, 10-22 psu, was observed at the lower estuary during 17 December, 2018 survey, which is attributed to the falling tide during the sampling period. Figure 5 indicates that the sediment grain sizes at each station during the different surveys were not consistent and varied significantly. The sediment grain sizes at stations of the middle estuary were primarily dominated by fine sand and very fine sand, with exception of 18 April 2018 survey which was dominated by the MS, especially at stations of the upper and lower estuary. The reason for such phenomenon is not clear. However, it can be seen in Figure 2 that the river water fluxes during the period of March-May 2018 were the least among the times of 2018-2019, implying that the flow mechanism could not flush the medium sand out to the lower estuary and medium sand resided in the lower estuary is probably from the Keelung River transportation. This assumption needs further investigation. In contrast to the sediment grain size result, the TOC concentrations found at the middle estuary stations were relatively higher than those found at the upper and the lower estuary stations. Such a contrasting result will be addressed in the discussion section.

Phosphorus
The concentration and percentage ranges of the five fractions and total concentration of phosphorus in DRE surface sediments obtained in the four surveys are tabulated in Table 2. The total P concentrations in DRE surface sediments range within 537-1310 mg/kg (17.33-42.30 µmol/g) and the concentrations found at each station are plotted in Figure 6, which indicates that the concentration variations are significant during the different surveys. The distributions of total P concentrations within the DRE are also scattered and no specific trend can be described. Figure 7 clearly shows that the total P concentrations within the DRE surface sediments are completely dominated by P CDB , accounting for 38-75% (average 58%) of total P pool, followed by P DET , accounting for 15-52% (average 31%) of total P pool. The contributions of P CFA and P ORG to the total P pool are generally <5% and 10%, respectively. The importance of P SORB is relatively minor and is generally <2% of the total P pool.

Iron and Mn
The analyzed results for iron and Mn in DRE sediments obtained in the four surveys are shown in Tables 3 and 4, respectively. The concentrations found at each station are plotted in Figures 8 and 9. The total Fe concentrations in DRE surface sediments ranged within 1.48-4.10%. The distributions also varied with the different stations and no trend could be specified. Figure 8 shows that the total Fe concentrations within the DRE surface sediments are generally dominated by Fe CDB and Fe ORG , each accounting for 23-48% of total Fe pool. The importance of Fe DET and Fe CFA ranked third and fourth and accounted for 16-31% and 4-17% of the total Fe pool, respectively. The concentrations of Fe SORB in all extractants were generally less than 0.1 mg/L and were not detected by the flame AAS (Perkin Elmer Analyst 900).
The total Mn concentrations in the DRE surface sediments ranged within 109-502 mg/kg and were generally dominated by Mn CDB and Mn CFA , accounting for 19-65% (average 36%) and 14-48% (average 29%) of the total Mn pool, respectively. The Mn DET and Mn ORG contributions to the total concentrations nearly equalized and accounted for 5.7-30.3% (average 14.5%) and 5.0-24.7% (average 15%) of the total Mn pool, respectively. The Mn SORB contribution to the total concentrations was minor and generally <5% of the total Mn pool.

Grain size and TOC
The concentration of chemical constituents in marine sediment is generally positively correlated well with organic carbon (OC) contents and is inversely proportion to the sediment grain size (GS) because the OC has the higher adsorption capacity and the finer the GS, the larger the surface area [49]. The plots of TOC against grain size in DRE sediments obtained in the present study are depicted in Figure 10, which shows that TOC concentrations exhibited an inversely exponential decrease with increasing grain size, with a correlation coefficient approaching −0.83 (p < 0.05). A similar result is also seen for the plot of total P concentrations against the grain size, with a good correlation coefficient (r = −0.71, p < 0.05). In addition, total P and P ORG concentrations significantly positively correlated well with the TOC (r = 0.89, p < 0.05 and r = 0.81, p < 0.05, respectively), as shown in Figure 10. These results suggest that the sediment grain size and TOC contents are the important factors controlling the sedimentary total P concentrations within the DRE, in agreement with many studies focused on estuarine and inshore environments [31,32,36,50,51].

Phosphorus
The total P concentrations in DRE sediments ranged within 537-1310 mg/kg (17.33-42.30 µmol/g), with an average value 769 ± 236 mg/kg. Figure 7 shows that the total P concentrations in the DRE surface sediments observed on 18 April 2018 survey were relatively lower than those of the other surveys, probably because the grain size effect. It has been indicated that the chemical compositions in inshore sediment may differ significantly from the grain size, bulk minerals, and heavy mineral fractionation [52]. However, many studies have tried to establish the P value in sediment to classify the sediment quality. The total P values in sediment exceed 700 mg/kg (22.6 µmol/g) representing that the environment is significantly influenced by the anthropogenic impact [53,54]. Another criterion established by Berbel et al. [35] suggests that sedimentary total P values ranging within 495-1300 mg/kg (16-42 µmol/g) are moderately polluted and the values exceeding 1300 mg/kg (>42 µmol/g) are highly polluted with P. The latter guideline seems to be more reasonable because the P contents in the upper continental crust generally range within 665-709 mg/kg [52,55,56]. The sediment P polluted status of the DRE can be classified as the moderately polluted, based on the classification by Berbel et al. [35]. Table 5 compares the total P concentrations and the P species in DRE sediments with eutrophic estuaries worldwide recently reported in the literature. The total P concentrations in the different estuarine sediments may vary 2-4 folds and the P species also significantly differ in the different estuaries. The estuarine sediments are derived mainly from the riverine SPM, and partly from the other sources, such as the atmospheric input, generalized within the estuary and the tidal water intrusion [60]. The phosphorus contents in the riverine SPM in the world's rivers range within 560-2000 mg/kg (18.1-64.6 µmol/g), with a mean value of 1146 mg/kg (37.0 µmol/g) [61]. As a result, the total P concentrations in the estuarine sediment worldwide differ by as much as 2-4 folds regardless of whether the estuary is contaminated or not because the riverine suspended particulate matter (SPM) source varied. However, the riverine SPM influences controlling the sedimentary total P concentrations decrease with increasing seaward. Fang et al. [62] indicate that the total P concentrations in some coastal sediments worldwide, such as the Yellow Seas, the Iberian Sea, the Gulf of Mexico, the Amazon Shelf, and the North Sea, generally remain within a narrow range from 465 mg/kg to 774 mg/ kg (15-25 µmol/g) due to the dilution effect.
It can be seen in Table 5 that the ranking of P species in DRE sediments followed the sequence: P CDB > P DET > P ORG > P CFA > P SORB . Such a sequence is generally inconsistent with the results reported in the literature. The comparable result with the present study is probably the study reported by Hartzell et al. [58] who indicate that the P CDB species entirely governed the total sedimentary P, 53-85% of total P pool, in the Patuxent RE, Chesapeaka Bay sediments. Meanwhile, the contribution of the second important species, P CFA , was relatively minor and occupied 8-22% of total P pool. In fact, it seems that P species in the estuarine sediment generally differs from the different estuaries, as shown in Table 5. For example, the P DET fraction, dominating the sedimentary total P, was reported in the Changjiang RE [31] and the Laizhou Bay [32]. Moreover, the sedimentary total P mainly governed in the P ORG species was seen in the Pearl RE [63], the Qinzhou Bay [32], and the Mondego RE [34]. In contrast, the P CFA species was the most important species in the Seine RE [33]. One of the most common characteristics of P species in the sediment is probably that the contribution of P SORB is relatively minor and generally <5% of the total P pool. However, such a characteristic was not obeyed in the Seine RE [33], the Penze RE [59], the Yamuna River [57], and the Santos-Sao Vicente RE, Brazil [35]. P SORB is the essential species of sedimentary P which can be easily released through the degradation of sediment organic matter and of decaying cells of bacterial biomass [33,64]. However, it is well documented that P SORB is easily desorbed from SPM or suspended sediment as salinity increased due to sulphate competition for adsorption sites, a process which also promotes desorption at higher salinities [11]. Thus, P SORB is generally the least important fraction among the P fractions in the marine sediment. However, the relatively higher percentage of P SORB in marine sediment has been observed in some studies, attributing to the highly adsorbed by Fe oxides [33] and the sewage effluent discharge [35], as shown in Table 5. The latter effect is not seen in the DRE which is also strongly interfered with by the domestic sewage effluent discharge, enhancing the higher DIP concentration (>10 µM) in its upper estuary [8]. P CDB was the most abundant sedimentary P fraction in the DRE and its contribution to the total P pool was significantly higher than the recent literature reports. The mechanism inducing such a result in the present study is not clear. However, it could be relevant to the WWTP effluent discharge, implying that a higher concentration of dissolved total P (>10 µM) commonly exists in the water column of the upper estuary of the DRE [8,38]. It is known that Fe is generally used in the modern wastewater treatment plant to prevent hydrogen sulfide emissions during anaerobic digestion and acts as a coagulant to improve sludge dewatering [10]. Our previous study indicates that the DRE water column enriched in particulate Fe and the concentrations of particulate total inorganic P, extracted by 1 N HCl acid, positively correlated well with particulate Fe within the DRE. This result suggests that particulate Fe plays an important role in influencing the P content in the DRE water column [8].
The P CDB is generally attributed to Fe-oxides which are the important adsorbents for DIP in the aquatic environment [16,19,26,65]. The P CDB is considered as the redox-sensitive sedimentary P fraction because P is generally accompanied with the Fe CDB reduction and is released from surface sediments into the overlying waters under hypoxic or anoxic conditions [63,66,67]. Figure 11 shows that the P CDB with Fe CDB plot concentrations in DRE surface sediments also have a good correlation (r = 0.62), consistent with the particulate Fe and P results in our previous study [8]. The Fe CDB /P CDB mole ratios observed in the DRE surface sediments range within 5.42-20.83 (average 12.3), which is lower than the literature reports indicating that the Fe CDB /P CDB ratio for newly buried iron oxides ranged within 20-26 in the marine environment [20,27,68]. The mole ratio of Fe CDB /P CDB can be used to estimate the spare sorption capacity of Fe oxyhydroxides, and the ratio value, approximately 6.7, is suggested to correspond to surface binding sites supersaturation of nanoscale ferrihydrite, the phase with the highest P sorption capacity [20,69]. The Fe CDB /P CDB ratio in the marine sediments reported in the literature varies widely, ranging within <2-45 ( [70], and references cited therein), suggesting that many factors, such as the sediment texture and grain size, the anthropogenic source, the redox condition of the environment, and the transformation of sedimentary P during the burial processes, may alter the ratio. The relatively lower mole ratio of Fe CDB /P CDB found within the DRE sediments is probably due to the effect of domestic sewage effluents, which elevates the dissolved P concentrations and is adsorbed by SPM, raising the particulate P concentration and reducing the ratio. In spite of that, future work needs to examine whether the P CDB transformation occurs or not during the sediment burial within the estuary.
The P CFA contributions to the total P pool within the DRE sediment are relatively low, ranging within 1.2-9.6% with an average 3.8%. P CFA in sediments is generally derived from detritus of biogenic apatite, such as skeletal materials from fish, authigenic carbonate fluorapatite (CFA) precipitation, and the chemical combination of P with CaCO 3 [26]. In addition, P CFA formation in sediments can be favored at the expense of P ORG and P CDB during the diagenetic process, enhancing a higher proportion of P CFA observed in the marginal and open sea sediments [21,71,72]. The P CFA concentration dominating the sedimentary total P pool is rarely reported in estuarine studies but can be found in the Bay of Seine and in the Mondego RE, and the source of P CFA is derived mainly from marine origin [31] and the adsorption of DIP by calcite [32]. P DET ranked the second most abundant sedimentary P fraction within the DRE and averagely accounted 31% of total P pool. Its contribution to the total P pool was slightly higher than the recent literature reports, as shown in Table 5. However, the relatively higher proportions of P DET , accounting for 62-86 % of sedimentary total P, were observed in the surface sediments of the Changjiang RE and the inner shelf of the adjacent East China Sea where the surface sediments were completely dominated by P DET species and governed by illilite, chlorite, quartz, and feldspar minerals [31]. P DET is mostly derived from igneous and metamorphic rocks and magma and is abundant in surface sediments that are under the substantial influence of riverine input [26]. Many studies suggest that the coarse sandy and sandy marine sediments contain relatively higher proportions of P DET than that of the silt and mud marine sediments, suggesting that the primary minerals may contain a relatively higher concentration of P DET than that of the secondary minerals [31,70]. The present result also finds that the P DET concentration is slightly positively correlated with medium sand (r = 0.43, p < 0.05), but obviously negatively correlated with mud (r = −0.64, p < 0.05). This result is consistent with the literature report. In addition, Fe DET correlates well with the P DET (r = 0.50, p < 0.05) in the present study and the mole ratio of Fe DET /P DET is about 8.7. The molar ratio of Fe DET /P DET in marine sediment is rarely discussed in the literature. The reason for this is probably due to the fact that the monoculture of Fe DET is not as clear as the P DET , due to 1 M HCl dissolving some fractions of the clay minerals but not the most crystalline Fe oxides [19], and that the P DET species, considered as the inert fraction of sedimentary P, is not involved in the redox reaction of iron or during the diagenetic processes [26]. Although P ORG ranked the third most abundant sedimentary P fraction within the DRE, the percentage only accounted for 1.9-11.7 % (average 6.3%) of sedimentary total P pool. These percentages are much lower than those reported in many estuarine studies, such as the Mondego RE [34], the Santos-Vicente RE [35], and the Peel-Harvey RE [73], indicating that the P ORG was generally > 30% of the sedimentary total P pool. Relatively higher percentages (> 50 %) of P ORG have been reported in some estuaries, such as the Qinzhou RE [32], the Seine Bay [33], and the Pearl RE [63], generally attributed to anthropogenic inputs, such as the wastewater, domestic sewage effluent, and agricultural activities. In contrast, our previous studies also indicated that the DRE is seriously impacted by the domestic sewage effluents, inducing its upper estuary to contain relatively higher concentrations of dissolved inorganic P (>10 µM) and dissolved inorganic N (>500 µM). The total dissolved and particulate concentrations of both elements are completely dominated by inorganic species [8,38]. These findings may explain why the P ORG contribution occupied a small amount of sedimentary total P pool within the DRE. P ORG storage in the estuarine sediment is chiefly facilitated by the autochthonous and allochthonous organic matter. The source of autochthonous is calculated according to the Redfield ratio, which is commonly employed to identify the source, provenance, decomposition, and preservation of sedimentary organic matters [70,74,75] and the trans-formation of P in sediments [76,77]. It can be seen in Figure 10 that the P ORG concentration is also positively correlated very well with the TOC (r = 0.81) in the DRE and the whole OC/OP ORG mole ratio during the four surveys is about 443. This ratio value is much higher than the Redfield ratio (106) of marine plankton and falls in the 300-1300 range for terrestrial higher plants with soft tissues [74,78,79]. The OC/OP ORG ratio value obtained in the present study may suggest that the allochthonous source may play an important role in dominating the sedimentary organic matters within the DRE. However, it is worth noticing that the discharge of the domestic sewage effluent within the DRE may interfere with the TOC/P ORG ratio in the sediment, which needs further investigation.

Iron and Mn
It can be seen in Table 3 that the sedimentary total Fe concentration extracted by the SEDEX method within the DRE surface sediments is mainly present in the Fe CDB and Fe ORG spices, each fraction accounting for an average of approximately 35% of total Fe pool, and Fe DET occupies on average nearly 22% of the total Fe pool. Ruttenberg [26] indicated that ferrihydrite, lepidocrocite, and goethite, the most common sedimentary Fe oxyhydroxide phases, and hematite, which is considerably less reactive than oxyhydroxides, are completely dissolved by neutral pH CDB reagent. Such a result is also confirmed by Slomp et al. [18] who studied the percentage of Fe extracted from the common Fe-containing standard minerals using different kinds of chemical reagents and indicated that the CDB reagent is a stronger chemical that totally dissolves the Ferrihydrite (Fe 5 HO 8 ·4H 2 O), Amorphous (FeS), Goethite (α-FeOOH), Hematite (α-Fe 2 O 3 ), and Magneite (Fe 3 O 4 ). In general, Fe CDB is often used as a measure of total Fe-oxides in the sediment [24]. Understanding the Fe CDB fraction present in sediment is important because such a fraction is easily remobilized under anoxic conditions, whether the hypoxic condition takes place seasonally in bottom sediments or the diagenetic process occurs during the sedimentary burial processes. The release of Fe CDB and P CDB into the water column will elevate the dissolved concentration of both elements and could strongly alter the Fe CDB and P CDB fractions and their mole ratio within the sediment profile [24,66,68,80]. The present results provide useful information concerning how many Fe CDB and P CDB fractions exist in the DRE surface sediment when studying the diagenetic processes of the sedimentary P and Fe in the DRE sediment.
In comparison with the Fe CDB, the discussion of Fe DET and Fe ORG in the literature is very limited, probably because the Fe DET fraction is considered an unreactive iron and is not dissolved during the reduction reaction under an anoxic environment [26]. The Fe DET fraction is analyzed by 1 N HCl acid extraction, whereby metal speciation in sediment is regarded as the weakly bound or non-detrital trace metals which correlate well with the biological availability [81]. Although the experimental results suggest that the bioavailability of mineral-bound Fe(III) in intertidal sediment of the microbial utilization is weaker than that of the chemical extraction [82], it should be taken into consideration for Fe DET fraction utilization by benthic organisms during the sediment burial processes. The strong complexation of iron hydroxides with organic compounds (OC) due to the negative surface charge of OC leads to the flocculation/aggregation of dissolved Fe occurs in the estuarine mixing zone and induces the removal behavior of dissolved Fe in the estuary [83,84]. Thus, the Fe concentration generally correlates well the OC concentration, and the fate of iron always accompanies the OC and vice versa in the marine environment ( [85], and references cited therein). Such a phenomenon is also attributed to the grain size effect. The finer the grain size, the greater the surface area for adsorption [49]. The Fe ORG fraction in the DRE sediment ranged within 22-48%, with an average 36%, of total Fe pool. Such a percentage seems to be relatively higher than the literature reports which employed the Tessier sequential extraction method to study the iron fraction in the marine sediments and indicated that the percentage of iron bound to organic matter was generally less than 5% of the total Fe pool [86,87]. However, the total Fe concentrations in the marine sediments differ according to analytical methods and the total concentrations could vary by 20-30% between with and without the total decomposition methods [28]. Thus, the contributions of Fe ORG fraction to the total Fe pool in the marine sediment may vary substantially with the different analytical methods. The percentage variations of Mn species in sediments of the DRE during the four surveys are significant. However, the Mn CDB and Mn CFA species dominated the total Mn concentrations and each fraction accounted for an average of 36% and 29% of the total Mn pool, respectively. The corresponding values for Mn DET and Mn ORG accounted for an average of 15% of the total Mn pool, respectively. These results agree well with the literature regarding the marine sediments, e.g., in the East China Sea [86], the Daya Bay, China [88], the North Sea [89], and the southwest coast of Spain [90], indicating that sedimentary Mn is mainly present in the non-residual fraction which includes the exchangeable, carbonate, (Fe, Mn)-oxhydroxide and sulphides/organic species. The Mn geochemistry characteristics include the fact that Mn is also the redox sensitive element and easily moves from dissolved phase into particulate phase under an oxic environment, and vice versa under an anoxic environment [91]. In addition, the dissolved and particulate Mn in the water column can be adsorbed onto the CaCO 3 surface [92,93]. As a result, these characteristics imply that the sedimentary Mn is mainly present in the non-residual fraction, especially in the (Fe, Mn)-oxhydroxide and carbonate phases, rather than the residual fraction, as observed in the present study and in the literature. It is noticeable that the SEDEX method can extract nearly 90% of total Mn concentrations of the reference materials (MESS-3 and BCSS-1). Thus, the residual fraction of Mn concentration in the marine sediment will be underestimated when the SEDEX method is employed. Figure 12 shows that the five fractions of sedimentary Mn do not have any correlation with the corresponding fraction of sedimentary P in the DRE sediment, suggesting that Mn may play a minor role in influencing the distribution of sedimentary P in the DRE. This result is contradictory to our previous study which indicated that particulate Mn significantly correlated well with particulate P within the DRE [8]. Basically, particulate P and Mn within the DRE exhibited a similar distribution. The higher concentration generally occurred in the upper estuary, at salinity < 5 psu. Afterwards, the concentrations decreased with increasing salinity and remained fairly constant in the lower estuary, at salinity > 20 psu [8]. In contrast, P and Mn within the DRE exhibited different distributions. The reason for this phenomenon is probably that the P CFA was a minor fraction, <5% of total P pool, and Mn CFA was a major fraction, on average approximately 30% of the total Mn pool, respectively, in the sediment samples. The dominant fraction of sedimentary P differs markedly from the sedimentary Mn, indicating that both elements in the sediment deviated from each other.

Conclusions
The present work is a pioneering study on the geochemical speciation of sedimentary P in the DRE sediment. The analyzed results show that the sedimentary total P concentrations ranged within 537-1310 mg/kg (with mean value 769 mg/kg), suggesting that the DRE surface sediments are mildly contaminated by P. The signal of P pollution status in the surface sediment is not as significant as in the water column, which contains relatively higher concentrations of total dissolved P (>10 µM) and DIN (>500 µM), and is recognized as a seriously eutrophic estuary. The most striking finding of the present study is that the sedimentary total P concentrations are completely dominated by P CDB , accounting for an average of 58% of TP. Such a high percentage of P CDB speciation is rarely reported in the marine sediment, which may be attributed to the influence of the wastewater treatment plant effluent discharge. This assumption needs further investigations. It is worth noting that P CDB species is more easily dissolved under the anoxic condition and elevates the dissolved P concentration in the water column. The diffusion flux of dissolved P under the diagenetic processes needs further study. Finally, the sediment grain size and TOC content seem to play crucial roles in determining the sedimentary total P content in the DRE sediment. The river water fluxes may strongly influence the distribution of the sediment grain size within the DRE.