Sources and Characteristics of Dissolved Organic Matter (DOM) during the Winter Season in Hangzhou Bay: Insights from Chromophoric DOM and Fluorescent DOM

Abstract

Elucidating the compositions, sources and mixing processes of dissolved organic matter (DOM) is crucial for a gaining deeper understanding of the coastal carbon cycle and global carbon budget. Hangzhou Bay (HZB), a vital estuary in China, receives freshwater inputs in the upper bay, borders the Changjiang River Estuary (CRE) to the north and is adjacent to Zhoushan Islands Region (ZIR) to the east. In HZB, the DOM sources and their compositions in estuaries remain unclear due to the complexity of this dynamic environment. In this study, we aimed to explore the chemical composition and sources of the DOM in the HZB and its adjacent coastal waters based on chromophoric DOM, fluorescent DOM indices and other hydrochemical parameters in the winter. The results showed that the DOM compositions in HZB have significant differences in the upper bay, middle bay and lower bay. The highest concentration of DOC was found in the CRE, close to the northern lower HZB, with high humification index (HIX), low biological index (BIX) and high proportion of humic-like fluorescent component (C1), indicating terrestrial inputs. In contrast, the DOM in the upper bay had high BIX and low HIX, being dominated by protein-like fluorescent components (C2 and C3), indicating an autochthonous source. The DOM in the middle bay showed mixed composition characteristics indicated by the chromophoric DOM (CDOM) and fluorescent DOM (FDOM) indices. Moreover, the terrestrial DOM transported via CDW intrusion accounted for a large proportion of the DOM in Northern HZB. Our study shows that, even in coastal estuaries with very strong hydrodynamics, the DOM composition can still retain its unique source signal, which, in turn, affects its migration and transformation processes. The results of this study provide supplement insights into the global carbon cycle and carbon budget estimation.

1. Introduction

Oceans are the most important carbon reservoir on Earth, absorbing more than 25% of the carbon dioxide (CO₂) emitted by human activities [1]. Although the offshore area only accounts for 7% to 8% of the global ocean area, the primary production accounts for 10% to 30%, and organic carbon burial is as high as 80%, making it one of the most important carbon sinks in the global ocean [2,3]. However, carbon transport fluxes along the land–coastal ocean–open ocean continuum under low carbon emission scenarios are essentially unknown. It is challenging to accurately assess the carbon flux, source-sink pattern and evolution trend in the coastal region, including the estuaries and bays due to the complicated hydrology dynamics and carbon cycle processes [4].

Dissolved organic carbon (DOC) is the largest carbon reservoir in the ocean, impacting the global carbon cycle and climate change. Dissolved organic matter (DOM) is a heterogeneous mixture formed through the secretion and excretion of living organisms [5]. The structure and composition of dissolved organic matter significantly affect the balance of aquatic ecosystems. The DOM in the ocean is generally divided into two categories: terrigenous (humic-like, high aromaticity and high molecular weight) DOM and autochthonous (protein-like, low aromaticity and low molecular weight) DOM [6]. In addition, as a carrier of water pollutants, the source, composition and characteristics of DOM regulate the migration and transformation of pollutants [7].

The DOM chemical composition depends on its parent organic matter and subsequent biogeochemical processes, including photo- and biodegradation [8]. Though many studies interpret DOM at the molecular level, such as sugars amino acids and lignin phenols [9,10], the sources and compositions of most DOM are still largely unexplored [11]. Spectroscopic techniques, including absorbance spectra and fluorescence excitation emission matrices (EEMs), are useful for tracing DOM sources and assessing DOM reactivity by testing chromophoric DOM (CDOM), fluorescent DOM (FDOM) and multiple established optical indices [12]. Combined with parallel factor analysis (PARAFAC) models, the relative contributions of allochthonous and autochthonous FDOM could be semi-quantitatively confirmed [13,14]. Recent research has correlated the EEM data with the maximum fluorescence (F_max) modeled by PARAFAC to examine the accuracy of the fluorescence intensities at the peaks of PARAFAC [15], and the machine learning method combined with PARAFAC model has also been used to identify sources of fluvial DOM in rivers [16]. Moreover, this method was used to study coastal bays under eutrophication conditions and large estuaries.

Through PARAFAC modeling, four fluorescent components, including two protein-like and two humic-like components, were validated in surface water and porewater in Andong Shoal, the southern bank of Hangzhou Bay, in 2016–2018 [17]. Zhao et al. found two humic-like components negatively correlated with salinity and one protein-like component in Xiangshan Bay, a semi-enclosed bay closed to the East China Sea (ECS), in May 2019, indicating the dilution of terrestrial signals by seawater [18]. Five fluorescent DOM components were validated in the Qiantang River Estuary (QRE), the upstream of Hangzhou Bay, in 2016–2017, including three humic-like and two protein-like components [19]. Four fluorescent components, including two protein-like and two humic-like components, were validated along the Changjiang River Estuary-to-ocean continuum in July 2017 [20].

Estuaries and bays link the land and the sea. Hangzhou Bay (HZB), adjacent to the Changjiang River Estuary (CRE) at the northern lower bay, is a vital estuary in China. The Qiantang River, the largest river in Zhejiang Province, China, carries terrestrial materials directly imported into HZB at the upper bay. Changjiang diluted water can also enter the bay directly from the Northern HZB or be carried into the bay through ebb and flow. The Changjiang, the largest river in China, brings large amounts of terrestrial materials, such as nutrients, particles and pollutants, into the CRE, thereby affecting material transport processes in adjacent coastal regions [21]. From the river to the open sea, DOC showed non-conservative mixing and was rapidly removed in low-salinity areas. Incubation experiments of DOC and amino acids conducted by Gou et al. in 2017 in the Changjiang River Estuary and adjacent seas indicated that photodegradation and biodegradation were important DOC removal mechanisms and exhibited specific responses depending on DOC compositions and sources [22]. Changjiang diluted water (CDW), characterized by low salinity and high nutrients, spreads southwestward in the winter, being affected by the northeast monsoon (December–February) [23]. It is generally recognized that the northwest monsoon leads to more significant CDW intrusion into Hangzhou Bay in the winter than in the summer and then strongly influences the hydrography of Hangzhou Bay [24]. It has been shown that, besides Qiantang freshwater (QTW), HZB receives more than half of its materials from CDW, which flows into Hangzhou Bay in the north and leaves via the southern part of the bay [25]. Therefore, CDW should have influenced the sources of DOM in HZB in the winter season, along with characteristic DOM signal invasion.

Recent research on HZB waters has focused on nutrients, particles and hydrodynamics [26,27]. The DOM in HZB waters is also an important carrier of organic carbon and pollutants. However, the studies mainly contained the results for CDOM, lacking the indices for and analysis of FDOM required to classify the exact fluorescent components of DOM. Recently, the combination of bulk carbon isotope, optical techniques and ultra-high-resolution mass spectrometry was used to study the compositions of DOM in CRE, QRE and Xiangshan Bay, respectively [18,19,20]. Studies that regard CRE, QRE and HZB as a whole system remain relatively scarce.

To sum up, research on the DOM compositions and sources in adjacent areas has mainly focused on the CRE–shelf–ECS continuum or the QRE–HZB continuum. Research on how the CDW invades Northern HZB in the winter has focused on nutrients, particles and hydrodynamics studies. How the CDW invasion, along with QTW and SSW, jointly affects the sources and characteristics of DOM in winter HZB is not clear. The objective of this study is to elucidate the chemical composition and sources of the DOM in HZB in the winter and the contribution of CDW inputs based on CDOM absorption, FDOM indices and other hydrochemical parameters collected through a comprehensive investigation survey in HZB and its adjacent sea in the winter.

2. Materials and Methods

2.1. Study Area and Sampling Sites

A comprehensive investigation survey was conducted in HZB and its adjacent seas (Figure 1a) in the winter (27–31 January 2024). In total, 50 stations shown in Figure 1b were investigated during the cruise. Water samples were collected using a Sea-Bird Electronics CTD (CTD—conductivity, temperature and depth; SBE 9, Sea-Bird Scientific, Bellevue, WA, USA) with 6 Niskin sampling bottles. At each station, the hydrologic parameters (temperature, salinity, turbidity, PAR, Chl-a, DO and pH) were constantly monitored. Further processing of raw data was performed using Golden Software Surfer 15 and SBE Data Processing.

2.2. Sample Collection

Seawater samples were collected from different depths (surface 2 m, bottom and every 10 m in the water column) at each sampling station. A total of 152 DOC samples were filtered onboard through 0.2 μm polycarbonate membranes (Millipore, Isopore^TM; Merck KGaA, Darmstadt, Germany) into borosilicate EPA bottles (amber glass, hydrochloride acid precleaned, calcined at 460 °C for 4 h). Samples for the DOC concentration were stored in the dark at −20 °C, and samples for CDOM and FDOM were stored in the dark at 4 °C until laboratory analysis.

2.3. DOC Concentration Analysis

The DOC concentrations were analyzed using a high-temperature catalytic oxidation method with a total organic carbon analyzer (TOC-L, Shimadzu, Kyoto, Japan). Samples were acidified with 6 N HCl to pH ~2 to remove the inorganic carbon before injection. Sample concentrations were quantified using the external standard curve. The Milli-Q water samples (blank) were subtracted as a baseline, and 83.33 μmol/L standards were injected every 10 samples to check the accuracy of the measurements [28]. The analytical error based on the replicated measurements (three measurements per sample) was 0.5 within 5% of the DOC values.

2.4. CDOM and FDOM Optical Analyses

The CDOM absorbance was measured using the absorbance spectrophotometer (Hardware of Aqualog^®, HORIBA Instruments Incorporated (HII), Edison, NJ, USA) with a 1 × 1 cm quartz cuvette. Meanwhile, FDOM excitation emission matrices (EEMs) were recorded using the fluorescence spectrofluorometer (Hardware of Aqualog^®, HORIBA Instruments Incorporated (HII), Edison, NJ, USA) with a 1 × 1 cm quartz cuvette.

2.4.1. CDOM Analysis: Ultraviolet–Visible (UV–Vis) Absorbance

The CDOM absorbance was measured throughout the UV and visible spectral domains (240–500 nm) with a resolution of 3.0 nm. The absorbance (A) was converted into the Napierian absorption coefficient (a) and measured in m⁻¹ using Equation (1) [29]:

$$a_{\lambda }=2.303\times {\displaystyle \frac{A_{\lambda }}{l}},$$

(1)

where A_λ is the absorbance, l is the length of the quartz cuvette, and λ is the wavelength. The a_λ at certain wavelengths can represent the abundance of CDOM in waters [30].

The spectral slope coefficient (S) was calculated using Equation (2) [31]:

$$a_{\lambda 1}=a_{\lambda 2}exp\left[-S\left({\lambda }_1-{\lambda }_2\right)\right],$$

(2)

where a_λ1 and a_λ2 are absorption coefficients at wavelengths λ₁ and λ₂. For example, the spectral range (275–295 nm) was reported to be S_275−295, measured in μm⁻¹. The spectral slope ratio at a specific wavelength (S_R) was calculated using Equation (3) to indicate the molecular weight of the DOM [32]:

$$S_R={\displaystyle \frac{S_{275-295}}{S_{350-400}}}.$$

(3)

The aromaticity index SUVA₂₅₄ was used as the average UV absorption of all DOM molecules, indicating the aromaticity of the DOM [12]. It was calculated using Equation (4) [33,34]:

$${SUVA}_{254}={\displaystyle \frac{a_{254}}{DOC}}.$$

(4)

2.4.2. FDOM Analysis: Fluorescence Excitation Emission Matrices (EEMs)

The FDOM excitation range was set from 240 to 500 nm, while the emission range was set from 245 to 825 nm. The excitation and emission scans were set at 3 nm and 4.66 nm (8 pixel) steps, respectively. Instrument correction, blank subtraction, correction of the inner filter effects, Rayleigh masking of the first and second orders and normalization (Raman scattering units) were performed as detailed below.

Instrument correction: The sensitivity of the water Raman SNR was >20,000, and the peak of water absorption ranged from 397 to 1 nm. Blank subtraction: The EEMs were subtracted by the EEM of Milli-Q water measured under the same conditions. The inner filter effects in the EEMs were corrected using Equation (5) [35]:

$$F_{\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{l}}=F_{\mathrm{o}\mathrm{b}\mathrm{s}}\times {10}^{{\displaystyle \frac{{Abs}_{\mathrm{E}\mathrm{x}}+{Abs}_{\mathrm{E}\mathrm{m}}}{2}}},$$

(5)

where F_ideal is the ideal fluorescence–signal spectrum, and F_obs is the observed fluorescence signal. Abs_Ex and Abs_Em are the measured absorbance values at the respective excitation and emission wavelength coordinates. The Rayleigh and Raman scatter peaks were removed by using the three-dimensional interpolation (R1 ± 10 nm and R2 ± 20 nm) [36]. EEMs were normalized to the water Raman signal [37], and the fluorescence intensities were reported as equivalent water Raman units (R.U).

The fluorescence index (FI), humification index (HIX) and biological index (BIX) were calculated as indices for microbial modification, the humification degree and the contribution of biological and autochthonous DOMs, respectively. The FI represented the relative contribution of microbial and terrestrial sources of the DOM. Higher values indicate an increasing degree of microbial modification. FI was calculated via the ratio of Em intensity at 470 nm divided by 520 nm at Ex 370 nm [38,39]. Higher HIX values indicate an increasing degree of humification. HIX was calculated by the area under the Em spectra 435–480 nm divided by the peak area 300–345 nm at Ex 254 nm [40]. BIX was an indicator of autotrophic productivity. Higher values (>1) corresponded to recently produced DOM of autochthonous origin. BIX was calculated via the ratio of Em intensity at 380 nm divided by 430 nm at Ex 310 nm [41].

2.4.3. The 3D Fluorescence—PARAFAC Analysis

The EEM data were further analyzed using the parallel factor (PARAFAC) analysis with Solo^® Software (9.5R, with PLS_Toolbox, Eigenvector Research; Manson, WA, USA). The n-component model (n = 2, 3, 4 ......) was validated via split-half, random initialization analysis and analysis of residuals.

Four-component PARAFAC analysis had 97.226% of the data explained, but the Core Consistency was only 17, with an only 2.2% similarity measure of splits and the overall model, which means that the four-comp—PARAFAC model is not available. Meanwhile, three-component PARAFAC analysis had 96.746% of the data explained, and the Core Consistency was 87, with an 87.2% similarity measure of splits and the overall model (Figure S1), which meant that the three-component PARAFAC model was reliable. Taking all factors into account, the three-component PARAFAC model was used for DOM composition analysis and further discussion in this study.

2.5. Statistical Analyses

The location of sampling sites, surface distributions and scatter diagrams were mapped via Ocean Data View 5.3.0 software (Schlitzer, Reiner, Ocean Data View, https://odv.awi.de, accessed on 1 March 2025). Correlation analysis (Pearson’s r) and principal component analysis (PCA) in this study were performed using the inbuilt statistics toolbox in OriginPro 2024b software. The results of linear and nonlinear fittings with p < 0.001 were recorded as significant.

3. Results

3.1. Hydrological and Hydrochemical Parameters in HZB and Adjacent Seas

The physical and chemical properties of seawater are primarily affected by the properties of water masses. Three different water sources were identified based on temperature, salinity and turbidity, showing distinct hydrological and hydrochemical properties (

Table 1. Hydrological and hydrochemical parameters in different end member sources in the HZB and adjacent seas in late January 2024 (range values in parentheses).

Water Masses	Temperature /°C	Salinity	Turbidity /FTU	TSM /mg·L−1	Chl-a /μg·L−1	DO /mg·L−1	pH
QTW	7.0 ± 0.1	16.66 ± 0.59	492.42 ± 0.01	(746–4616)	4.40 ± 0.15	10.83 ± 0.08	6.87 ± 0.00
CDW	5.9 ± 0.1	24.39 ± 0.79	451.33 ± 50.25	(256–1475)	2.12 ± 0.50	10.76 ± 0.05	6.86 ± 0.01
SSW	10.2 ± 2.2	31.31 ± 1.80	5.44–293	(6–1184)	0.58 ± 0.18 (0.36–1.37)	9.20 ± 0.61 (8.29–10.38)	6.89 ± 0.03 (6.83–6.97)

). QTW stands for Qiantang Estuary Water; CDW stands for Changjiang diluted water; SSW stands for shelf sea water (salinity > 27.5) in the Zhoushan Islands Region.

Table 1 shows the different water masses with different hydrochemical properties in winter in the study area. Water masses of varying densities and sources are highlighted and circled in different colors in Figure 2a. QTW (blue circle in Figure 2a) had medium temperature, lowest salinity and high turbidity (>400), with the highest DO and Chl-a values. Changjiang diluted water (CDW, orange circle in Figure 2a) had the lowest temperature, medium salinity and high turbidity (>400), with medium DO and Chl-a values. Shelf sea water (SSW, red quadrangle in Figure 2a) had the highest temperature and salinity and low turbidity (<300), with the lowest DO and Chl-a values.

There were 79 water samples with turbidity greater than 300 FTU in the study area, representing mixed waters highly influenced by both CDW and QTW, and 69 water samples with turbidity less than 300 FTU representing shelf sea waters (Figure 2b).

The HZB showed strong vertical mixing characteristics in late January 2024. The temperature and salinity ranged from 5.8 to 13.6 °C and 15.65 to 33.29, with average values of 8.8 ± 2.1 °C and 27.57 ± 4.56, respectively. The spatial distributions of salinity and temperature were consistent between the surface and bottom layers, with low salinity and temperature in the upper HZB and North HZB, and then, both the salinity and the temperature increased gradually southeastward to the Zhoushan Islands Region (ZIR) (Figure 3a–d).

The HZB is known for its strong tidal currents, with a large tidal range accompanied by high turbidity. An obvious turbidity front in the study area was found along the longitude of 122.5 °E. The turbidity was pretty high, ranging from 400 to 500 FTU, at the west side of the front, including the whole HZB. Then, it decreased eastward to the ZIR. On the east side of the front, the turbidity ranged from 5 to 100 FTU. The turbidity of the Northern Changjiang Estuary stations (B3, B4 and B4-0) was slightly lower than that of Hangzhou Bay (Figure 3e,f).

The total suspended matter (TSM) in the HZB and adjacent seas ranged from 113.6 mg·L⁻¹ to 4616 mg·L⁻¹ and 5.6 mg·L⁻¹ to 616 mg·L⁻¹ in the HZB (west of 122.5 °E) and the ZIR (east of 122.5 °E), respectively. The high values of TSM were found in the Qiantang River Estuary (QRE, stations E1–E4) and the northern lower bay near the CRE (Figure 3g,h).

In the HZB stations (from QRE to HZB mouth) of our study area, the turbidity values were very high and varied slightly from 482.75 to 492.47 FTU in all water layers (Figure 4a). In the ZIR stations (from HZB mouth to open ECS), the turbidity values decreased with the longitude and depth in general (Figure 4a). The TSM values showed a similar distribution with turbidity (Figure 4b), decreasing with the longitude and depth.

3.2. Distributions of DOC in the HZB and Adjacent Seas

The concentrations of DOC in the study area ranged from 69.2 to 179.2 μmol·L⁻¹, with an average value of 105.8 ± 25.0 μmol·L⁻¹. High DOC concentrations were observed in both surface and bottom waters near the Changjiang River Estuary (149.2 ± 8.3 μmol·L⁻¹ at stations B3, B4, B4-0 and B4-1) and at station C2-1 (168.3 μmol·L⁻¹) (Figure 5a,b), and then, they gradually decreased southwestward to the HZB and southeastward to the Zhoushan Islands Region (ZIR).

3.3. CDOM and FDOM Properties in the HZB and Adjacent Seas

The highest mean value of DOC concentration was in the CRE (Figure 6a), but the highest mean value of CDOM abundance was in the upper HZB (near QRE), and the CDOM absorption coefficients decreased from the upper bay through the lower bay to the bay mouth (Figure 6b,e). The FDOM relative abundance varied in the same way as a₃₅₅ (Figure 6b,c).

The SR was higher in the ZIR with a mean value 2.01 than in the CRE and HZB (mean value 1.28–1.55) (Figure 6d). On the contrary, the SUVA₂₅₄ mean values decreased from the QRE to the ZIR (from 6.60 to 3.20) from the upper bay to the lower bay to the bay mouth (Figure 6f).

In addition, the highest FI mean value was also found in the QRE (Figure 6g), and the FI decreased from the upper HZB (QRE) eastward to the CRE and the ZIR. The HIX mean values sharply increased from the QRE to the CRE, increasing from 0.47 to 2.19 (Figure 6h). The BIX mean values decreased from the upper HZB (QRE) eastward to the CRE, decreasing from 1.10 to 0.99 (Figure 6i).

3.3.1. CDOM Coefficients

The absorption coefficients of CDOM at different wavelengths were always positively correlated in the coastal estuaries where the DOM was dominated by terrestrial input [43]. The absorption coefficient of CDOM in the HZB and adjacent seas was higher at 254 nm (a₂₅₄) than other, longer wavelengths (Figures S2 and S3). This result was mainly due to the low concentrations of CDOM in the HZB and the shelf region, resulting in an inaccurate absorption coefficient measured at 350–500 nm. Previous studies have also reported that the shorter wavelength absorption fits better than longer wavelength absorption, especially in the open sea [43,44].

The UV absorption coefficients at 254 nm and 355 nm, i.e., a₂₅₄ and a₃₅₅ in the HZB and adjacent seas, ranged from 1.31 to 8.74 m⁻¹ and 0.05 to 3.86 m⁻¹, with an average of 4.77 ± 1.62 m⁻¹ and 1.28 ± 0.69 m⁻¹, respectively. The highest value of a254 was concentrated in the upper HZB and the CRE (surface water stations E1, E2, CE3 and B3) (Figure 7a–d). The S_R ranged from 0.68 to 3.65, with an average of 1.71 ± 0.52. The aromaticity index SUVA₂₅₄ ranged from 1.15 to 9.62, with an average of 3.90 ± 1.62 measured in L·mg⁻¹·m⁻¹ (Figure 7e,f).

The absorption of CDOM at certain wavelengths reveals the relative abundance of CDOM in the study area. There are two high-absorption areas at both 254 nm and 355 nm (UV absorptions) in the surface layer. One is near the CRE (station B3), and the other is near HZB (stations E1, E2 and CE3). At the bottom layer, the distribution of CDOM at 254 nm is similar to that at the surface layer. The highest SUVA₂₅₄ values in the surface and bottom layers were found in the upper HZB, indicating that the DOM was highly aromatic under the influence of the Qiantang River Estuary (QRE) input.

3.3.2. FDOM Indices

The HIX values ranged from 0.41 to 2.93, with lower values near the upper HZB (Qiantang River Estuary) and Southern HZB (from stations E1 to E6 to ZJD12031), while higher values were recorded near the CRE (stations B3 to B8) (Figure 8a,b). The HIX values decreased gradually from the CRE to Southern HZB and from the northern lower bay to the upper bay.

The BIX values ranged from 0.95 to 1.26, with lowest values in bottom water near the CRE (stations B3, B4-0 and B4-1) and higher values in the Southern HZB (stations E1–E6) (Figure 8c,d). The BIX values increased spatially from the Changjiang Estuary southwestward to the Southern HZB and southeastward to the ZIR. The index ranges for salinity are supported in Figure S4.

3.3.3. PARAFAC Analysis Results

Three fluorophore components were identified in the HZB and adjacent seas using PARAFAC (Figure 9), and they were compared with those found in previous studies in the OpenFluor database [45].

Component 1 (C1) displayed the excitation wavelength maxima (Ex) and emission wavelength maxima (Em) at 248 nm and 460 nm, respectively. C1 represents spectral characteristics similar to those of a UVC humic-like component (peak A) [13,46]. Compared with UVA humic-like matter (signature peak C), UVC humic-like DOM has a lighter molecular weight, lower aromaticity and stronger resistance to photodegradation [47].

Components 2 (C2) and 3 (C3) displayed an Ex/Em at 275/318 nm and 287/355 nm, respectively. A second Ex peak of C3 was observed at a wavelength less than 240 nm. Both C2 and C3 were identified as protein-like components. C2 shows typical features of tyrosine-like fluorescence (peak B), and C3 shows typical features of tryptophan-like fluorescence (peak T) [13,48].

Figure 9. The 3D fluorescence spectroscopy results, signature peaks (warm colors) of Ex/Em, and molecular characteristics [49] of the three FDOM components identified in the HZB and adjacent waters. (a) C1: Ex/Em = 248/460 nm, A peak; (b) C2: Ex/Em = 275/318 nm, B peak; (c) C3: Ex/Em = 287/355 nm, T peak.

Figure 9. The 3D fluorescence spectroscopy results, signature peaks (warm colors) of Ex/Em, and molecular characteristics [49] of the three FDOM components identified in the HZB and adjacent waters. (a) C1: Ex/Em = 248/460 nm, A peak; (b) C2: Ex/Em = 275/318 nm, B peak; (c) C3: Ex/Em = 287/355 nm, T peak.

In total, the three components C1, C2 and C3 accounted for 34.7%, 41.5% and 23.8% in the HZB and 56.6%, 23.5% and 19.9% in the ZIR, respectively (Figure 10). The F_total (summary of C1 + C2 + C3) in the HZB and adjacent areas ranged from 0.0270 to 0.0510 R.U. The highest F_total values were concentrated in the QRE and upper HZB (Figure 11a,b). The F_total values of C1, C2 and C3 were 0.0110–0.0187, 0.0050–0.0269 and 0.0043–0.0131 R.U, with average values of 0.0165 ± 0.0016, 0.0096 ± 0.0050 and 0.0070 ± 0.0020 R.U, respectively. The C1 components were higher in the CDW-influenced areas, such as the CRE, the shelf sea area and lower HZB. The C2 and C3 components had the highest values near the QRE in the upper HZB. The distributions of C1, C2, and C3 in the HZB and adjacent sea are shown in Figure 11c–h. There is a strong positive correlation between the two protein-like components C2 and C3 (r² = 0.93, p < 0.0001), while there is a strong negative correlation between the humic-like component C1 and protein-like component C2 when the salinity is lower than 27.5 (r² = 0.96, p < 0.0001), as shown in Figure S3.

4. Discussion

4.1. Composition and Characteristics of DOM in the HZB and Adjacent Seas

The CDOM is a naturally occurring dissolved organic matter that absorbs UV light in water. The intensity of the UV absorption varies with the composition of CDOM, and it is often used in combination with FDOM to indicate the composition and characteristics of DOM [12]. Moreover, FDOM, i.e., DOM with fluorescence characteristics, also has UV–Vis absorption properties, and the absorption intensity could be represented by F_total (summary of C1 + C2 + C3 in unit R.U).

The absolute abundance of FDOM components C1, C2 and C3 in the Qiantang River Estuary (QRE), Hangzhou Bay (HZB), the Changjiang River Estuary (CRE) and the Zhoushan Islands Region (ZIR) showed distinct differences (Figure 12). In the QRE, the main components of DOM were C2 and C3. Then, these DOM decreased from the upper HZB to lower HZB (from the QRE to the CRE and the ZIR), as shown in Figure 12. Both C2 and C3 were identified as protein-like components [13]. C2 shows the typical features of tyrosine-like fluorescence., while C3 shows typical features of tryptophan-like fluorescence [49]. In this study, we analyzed the correlation between RFE corrected at 254 nm [50] and salinity (Figure S4). The result showed a significant positive correlation between the two parameters, indicating that the protein-like components were corrected by the primary production source. Figure S5b showed the biological index (BIX) varies with C2. C2 varied slightly with BIX and varied significantly in the HZB different areas, indicating human inputs of C2 as a possible source in HZB.

In contrast, the highest value of humic-like C1 was found in CRE as a high-value core, and then, it decreased from the CRE westward to the upper HZB and eastward to the ZIR, as shown in Figure 12. The three fluorescent components (C1–C3) were compared with the OpenFluor database [45]. C1 was associated with signatures similar to a photo-oxidized crude oil product [51]. It was associated with the terrestrial humic-like substance in Green Bay [52], a major component of porewater and bottom water in rivers [53], as well as leaf extracts in stream water [54], indicating that C1 is a terrestrial humic-like dissolved organic matter that has undergone photo-oxidization.

A high correlation (p < 0.0001) was revealed between HIX and component C1 (Figure S5), showing that the source of C1 mainly contributed to the degree of humification of DOM. This suggests that the DOC in CRE is dominated by the terrestrial input DOC with a high degree of humification. As Changjiang River carries large amounts of terrestrial materials into the CRE, the DOM in the CRE was mainly carried by Changjiang inputs with a high degree of humification.

Furthermore, the composition and characteristics of DOC indicated by other CDOM parameters and FDOM indices support the above results. Relatively low DOC concentrations, but with high a₂₅₄ and SUVA₂₅₄ values, were found in the upper HZB (the Qiantang River Estuary or QRE in Figure 6a,e), indicating that DOM in the QRE had higher UV absorption per C. Moreover, the highest values of a₃₅₅ and F_total were found at the QRE (Figure 6b,c), indicating that the greatest relative abundance of CDOM and FDOM was in the QRE. The distributions of SUVA₂₅₄, a₃₅₅ and F_total jointly reflect the DOM molecular characteristics of the QTR source with high aromaticity, resulting in high UV absorption and high fluorescence excitation emission peaks. However, the highest DOC concentration with low SUVA₂₅₄ was found at the CRE (Figure 6a,f). Compared to the DOM of QTW inputs, the DOM of CDW showed the characteristic of higher carbon amounts but lower aromaticity.

The FI is often used as a proxy for microbially modifying DOM [39], and the result (Figure 6g) indicated that the microbial modification in the QRE and Southern HZB was more significant than that in the CRE and the ZIR. The HIX is often used to indicate the humification degree of DOM [40]. Since the HIX is an indicator of humic substance content or extent of humification, the highest HIX in CRE (Figure 6h) indicated that an evident terrigenous humic-like DOM was carried by Changjiang River, and then, it spread from the CRE to the surrounding sea areas, including the lower HZB and the ZIR. The higher peak area of the HIX and lower peak area of the BIX in the CRE and Northern HZB showed a significant intrusion from CDW into HZB. The BIX is an indicator of autotrophic productivity, and higher values (>1) always correspond to recently produced DOM of autochthonous origin [41]. The BIX values in HZB (Figure 6i) indicated that the DOM from autotrophic productivity in the QRE and the southern bay was greater than those from the northern bay and the CRE. Although the average value of the BIX in the ZIR is almost equal to the value in the QRE, the significant difference in FI values indicated that there was a significant difference in DOM composition between these two regions, and the DOM in the QRE experienced strong microbial modification.

PCA of different characteristics and indices in HZB and the adjacent seas found two principal components (Figure 13). PC1, with 50.2%, represented positive correlations with CDOM UV absorption, protein-like FDOM (C2 and C3) and turbidity but negative correlations with salinity, the HIX and humic-like FDOM C1. PC1 was a principal component of short-wave absorption, small molecules, photodegradation and non-terrestrial sources. PC2, with 18.6%, represented positive correlations with turbidity and DOC but negative correlations with the BIX and the FI. PC2 was a principal component of freshness and biological modification. PC1 distinguished CRE and ZIR sources of DOM from HZB sources. Moreover, PC2 distinguished SSW from low-salinity, high-turbidity CDW and QTW masses. PCA clearly showed the obvious differences in DOM composition and characteristics in HZB and its adjacent areas.

4.2. Sources of DOM in HZB

Generally, DOM can be divided into autochthonous and allochthonous DOM according to the initial source. The autochthonous source includes phytoplankton synthesis in surface water and the transport of benthic organisms by upwellings. Moreover, autochthonous DOM can come from POM transportation and microbial lysis. Autochthonous DOM have a lower molecular weight, aromaticity and higher UV–Vis absorption in general. In contrast, the allochthonous source includes soil precipitation and plant metabolite leaching as terrestrial sources, as well as pollution sources produced via human activity. The allochthonous DOM may have higher molecular weight, higher aromaticity and be easily photodegraded [55,56].

The humic-like components were mainly terrestrial inputs in bays, estuaries and coastal waters, taking studies in Xiangshan Bay, Changjiang River Estuary and Shelf East China Sea as examples [18,20,22]. The protein-like component DOM is marine autochthonous DOM, which is derived from phytoplankton production. In general, from the bay to the sea, with an increase in salinity, the DOM would be gradually dominated by the in situ-produced composition [57].

It is known that the characteristics of DOM in CRE are mainly modulated by input of the terrestrial DOM. Moreover, the CDW contributes terrestrial, humic-like DOM to the ECS [20]. However, the DOM contribution of CDW invasion into HZB in the winter season remains unclear. In our results, the compositions of DOM in HZB showed contributions from different DOM sources. The highest values of C1 in the CRE and C2 and C3 in the QRE supported the idea that different FDOM composition features could be used to distinguish the sources of DOM in waters in different regions of HZB.

The distinct differences in FDOM components in HZB and its adjacent areas were discussed in Section 4.1. The result, combined with the water mass characteristics, could be used to indicate DOM sources in HZB. The DOM in HZB composed of different sources is controlled by the three different water masses of the QTW, CDW and SSW. Among them, the QTW has the lowest humic-like C1 component due to the modification of high Chl-a linked DOM, and it displays the highest protein-like C2 component of DOM in the upper HZB. In the upper bay near the QRE (stations E1, E2, CE3, E3 and E4), the main component of DOM is C2, and next is C3, the protein-like DOM. Previous studies showed that the Qiantang River carries a mixture of terrestrial humic-rich and anthropogenic protein-rich CDOM and contributes primarily to the CDOM in the upper bay [19]. Moreover, the values of protein-like components in FDOM are notably higher (from 23% up to 90%) than those of the other components in typical bays in China [18]. The main components are either derived from the synthesis of algal production or human sewage discharge [58]. These studies support our results that the highest C2 abundance was in the upper HZB near the QRE and the possible sources of protein-like DOM.

Water samples with turbidity greater than 300 FTU were found in HZB and the CRE areas in winter 2024 (Figure 2b), representing resuspension and material transportation highly influenced by both the CDW and QTW. Water samples in the ZIR had turbidity less than 300 FTU in winter 2024 (Figure 3e,f). It has been reported that the compositions of fluorescent DOM changes with both the salinity and tidal level in the Qiantang River (from 120.1 °E to 121.3 °E, Zhijiang to Zhapu) [19]. In the HZB (from 121.2 °E to 122.1 °E, QRE to HZB mouth) in this study, the turbidity values were very high and varied slightly in the surface and bottom layers (Figure 3a), which showed the strong sediment resuspension and thorough mixing of water layers [59].

Turbidity has been shown to be significantly resistant to the photodegradation and photobleaching of DOM in the bay and estuary [60]. Moreover, the SR is used to indicate the molecular weight (M.W) of DOM. Higher SR ratios indicate small M.W on average. The molecular weight decreased with the DOM decomposition process. Therefore, the SR can not only indicate the characteristics and sources of different kinds of DOM but also indicate the degradation degree of the same species of DOM [32]. In this study, the SR was higher in the ZIR than the CRE and HZB areas (Figure 6d), which indicated that, in SSW with low turbidity, photodegradation happened obviously in the humic-like C1 DOM component.

It has been shown that, besides Qiantang freshwater (QTW), HZB received more than half of its materials from the CDW, which flowed into Hangzhou Bay from the north and left via the southern part of the bay [25]. Therefore, as water mass is the carrier of DOM, the CDW should have contributed to the DOM sources in HZB in the winter, along with humic-like C1 DOM signal invasion. In our results, CDW showed the highest humic-like C1 component due to the strong effects of Changjiang terrestrial DOM inputs. SSW shows the high salinity and medium value of C1. The DOM composition in the northern bay is a mixture of QTW and CDW, with CDW being present in a higher proportion. The DOM composition in the central and southern bay was controlled by the mixing of these three water masses. In Southern HZB (blue dashed section in Figure 1b, Figure 2a and Figure 13), the C1 abundance and salinity (S) showed a significant positive relationship conforming to the formula C1 = 4.18 × 10⁻⁴ × S + 4.44 × 10⁻³ (R² = 0.9862, blue dashed line in Figure 14), but the line was not on the mixing line of QTW and SSW. This further confirms the contribution of the mixing of the three water masses.

The C1 abundance and salinity in the ZIR (salinity > 27.5, red SSW area in Figure 1b, Figure 2a and Figure 13) conform to the following formula: C1 = −3.49 × 10⁻⁵ × S² + 1.77 × 10⁻³ × S − 4.23 × 10⁻³ (R² = 0.5803, red curve in Figure 14). The DOM composition in the ZIR is controlled by the nonlinear mixing of CDW and SSW. This is a reasonable outcome, since the CDW is a mixed water mass made up of Changjiang fresh water and ECS shelf water.

5. Conclusions

The dissolved organic carbon (DOC) concentrations ranged from 69.2 to 179.2 μmol/L in Hangzhou Bay (HZB) and its adjacent seas, with the highest values recorded in the Changjiang River Estuary (CRE) in winter 2024. The high humification index (HIX), low biological index (BIX) and low specific ultraviolet absorbance at 254 nm (SUVA₂₅₄) in Changjiang Diluted Water (CDW) indicated that the terrestrial C1 is characterized by high humification, high molecular weight and low aromaticity because of photodegradation. The high HIX was found in northern HZB (lower bay), while the high BIX was found in Southern HZB. Dissolved organic matter (DOM) coming directly from the Changjiang River made a relatively high contribution to DOM in HZB in the winter. A three-component model was designed via parallel factor analysis (PARAFAC) in the study area, including one humic-like C1 and two protein-like C2 and C3. The highest C1 component was detected in the CRE, indicating high terrestrial DOM carried by Changjiang River into the CRE, while high values of protein-like C2 and C3 components were found in the QRE at the upper HZB. The compositions, sources and mixing behavior of DOM in HZB were determined by three different water masses (CDW, QTW and SSW). The terrestrial DOM transported via CDW intrusion accounted for a large proportion of the DOM in Northern HZB, while the DOC in Central and Southern HZB shows the mixed characteristics of the three water masses. The composition of DOC in the ZIR indicates the mixed sources of CDW and SSW. Our results reveal that DOC compositions can be effectively discriminated against by their spectral characteristics, providing valid insights for investigating offshore carbon cycling and tracing carbon transformation processes.