Grain size and organic geochemistry of recent sediments in Lingding Bay, Pearl River Delta, China: implications for sediment dispersal and depositional processes perturbed by human activities

: Lingding Bay (LDB), on the Pearl River Delta (PRD) of southern China, is a typical example of a large river mouth that is strongly affected by anthropogenic perturbations that have changed the boundary conditions of hydro- and sediment dynamic processes. An analysis of recent sedimentary patterns can shed light on the role of anthropogenic impacts on delta evolution. In this study, we collected surficial sediments from the LDB in December 2016 (dry season) and August 2017 (flood season) to analyze their grain size and organic geochemical compositions, with the aim of investigating recent depositional patterns in the bay and evaluating human impacts. The results reveal two major mud depocenters in the northeastern and southwestern parts of the bay, which are characterized by high values of grain-size end member 1 (EM1) and increased contributions of terrestrial organic carbon in the flood season. We propose that this sedimentary pattern is a manifes- tation of a system regime shift due to the strengthening of the fluvial function in fluvial-tide interactions and associated changes in the suspended sediment dispersal routine. We suggest that these changes are a result of recently intensified human activities, such as coastal land reclamation and sand mining. Coarsening of the surficial sediments in the LDB in the dry season and a marked increase in the terrestrial organic contribution at the mouth of the LDB indicates the redistribution of fine-grained sediments by waves and currents and increased mud export from the LDB in response to the shallowing of the bay.


Introduction
Intensive human activities at the coast have increasingly impacted fluvial sediment source-to-sink processes and associated geomorphological characteristics in estuarine, coastal, and continental shelf environments. Therefore, many deltas are facing system regime shifts (Syvitski 2008;Moufaddal 2013;Fan et al. 2019;Gao et al. 2019;Nienhuis et al. 2020). The reduction in sediment supply and rising relative sea levels have increased the risk of flooding and coastal erosion of densely populated deltas (Syvitski et al. 2009;Giosan et al. 2014;Wu et al. 2020), which is exacerbated by land subsidence (Herrera-García et al. 2021). Human activities at river mouths, including land reclamation, channel dredging, and sand mining, directly perturb hydro-morpho-dynamic and sedimentary processes (Xie et al. 2017), thus affecting the evolution of deltas. The Pearl River Delta (PRD, Fig. 1) has been subjected to intensive human activities (Zhang et al. 2015a;Wu et al. 2018) due to population pressure and rapid economic development. Indeed, the PRD has emerged as one of the largest urban agglomerations and most economically active zones in the world (Jiao et al. 2019); thus, it is a key site for the study of human impacts on mega-delta evolution. Lingding Bay (LDB) is the largest sheltered embayment of the Pearl River mouth and is proximal to the megacities of the PRD. Over the past decades, human activities in the fulfillment of demands for land resources and navigation (Zhang et al. 2015a;Wu et al. 2018) have intensively changed the geometry and subaqueous topography of the bay. Therefore, it is important to investigate the response of sedimentary processes to human impacts in the LDB.
Intensive human activities in the PRD, especially in recent decades, have altered the discharge of freshwater through each outlet of the distributary channels (Fig. 1c), accelerated seaward advancement of the shoreline and subaqueous delta, and initiated tremendous changes in the hydro-dynamics and associated geomorphological environment in the LDB (Lu et al. 2007;Luo et al. 2007;Wu et al. 2018;Yang et al. 2019). First, excavation of sand in the river channels on the delta plain changed the ratios of freshwater discharge in different distributaries and increased the amount of freshwater discharged into the LDB (Zhang et al. 2009;Yuan and Zhu 2015;Zhang et al. 2015b). Second, substantial reclamation of the estuary mouth shoals caused rapid seaward extension of the outlets in recent decades at an average rate of 19.3 m/year in 1976-2006, which resulted in a rapid seaward progradation of the subaqueous delta in the western LDB and also changed the direction of freshwater flow into the bay (Zhang et al. 2015a). Third, dredging in the navigation channels (He et al. 2020) and sand mining of shoals in the east and central LDB (Wu et al. 2016Yang et al. 2019) markedly altered the subaqueous topography, which increased the tidal amplitude in the inner LDB and strengthened the intrusion of higher-salinity shelf water ). These various anthropogenic activities have altered the boundary conditions of hydrological and sediment dynamics and intensified both fluvial and tidal functions in the LDB He et al. 2020;Wang et al. 2020). Given that such interventions will inevitably change the sediment dispersal and depositional processes and influence the evolution of geomorphological and ecological environments of the LDB, further investigation of the emerging pattern of depositional processes as recorded in recent sediments is needed to evaluate the impacts of high-intensity human activities.
Accordingly, in this study, we present an analysis of surficial sediments in the LDB collected in December 2016 (dry season) and August 2017 (flood season) to compare the spatial and seasonal variations in grain size and organic elemental and isotopic compositions. Grain-size end-member modeling (Paterson and Heslop 2015) and Gao-Collins grain-size trend analysis (Gao and Collins 1992;Gao 1996) were applied to investigate the sediment dispersal and trapping patterns in the LDB. By comparing these results with the sediment distribution recorded in the 1970s and the early 2000s, when human activities were much less prominent, we aim to shed light on the changes in the depositional processes under anthropogenically perturbed boundary conditions, thereby providing important information for the management of the LDB.

Study area
The PRD is characterized by a complex network of distributaries on the delta plain, eight outlets, and two major, partly filled embayments (the LDB in the east and the Huangmaohai in the west; Fig. 1b). Three major rivers, namely the West, North, and East rivers, flow into the PRD from catchments that experience a subtropical monsoon climate (Huang et al. 1982;Zhao 1990). Limestone bedrock dominates in the catchment (Zhao 1990), resulting in a relatively small mean annual sediment load of 64.9 m/year during 1957(MWR 1957. The West River is the largest of the three main rivers flowing into the PRD Pearl River Delta (PRD) with its eight outlets. There are four outlets into Lingding Bay (LDB): a, Humen; b, Jiaomen; c, Hongqili; d, Hengmen. Other outlets: e, Modaomen; f, Jitimen; g, Hutiaomen; h, Yamen. LDB is divided into an inner and outer bay separated by the inner Lingding Island. The maps were generated with ArcGIS 10.6 software (www.esrichina.com.cn) using the topographic dataset provided by the Geospatial Data Cloud site, Computer Network Information Center, Chinese Academy of Sciences (http://www.gscloud.cn). (c) Subaqueous topography of LDB in 2015 (after Wu et al. 2018). The red line represents 5-m bathymetry in 2017, which significantly advanced southeastward from 2015 to 2017. The black line represents the plume front during the dry season in 2011 (Lai et al. 2015). I, II, and III indicate three major hydro-and sediment dynamic regimes including the freshwater-dominated regime in the northwestern LDB, the tide-dominated regime in the northeastern LDB, and the high-salinity shelf-water regime in the outer LDB, respectively (Miao et al. 1988;Wei and Zhu 2019). Surficial sediments in this study were located in four Zones (I, II, III-1, and III-2) according to the position of the river plume front in the dry season (Lai et al. 2015) and the Lingding Channel. ①-④ represent islands in the LDB and at its mouth, including Qi'ao Island, inner Lingding Island, Luhuan Island, and Dayu Mountain, respectively. and contributes the greatest volume of both freshwater and sediment. The mean annual freshwater discharge from the Pearl River system as a whole was 282.6 × 10 9 m 3 during 1957(MWR 1957. The flood season (April to September) accounts for approximately 80% of the total annual discharge and 95% of the annual sediment load (Xia et al. 2004).
The LDB is a sheltered embayment from the Humen Outlet to Guishan Island in the eastern PRD (Fig. 1b). The bay is bell-shaped and occupies an area of approximately 2000 km 2 . Currently, the water depth is mostly less than 5 m (Fig. 1c). The bay is divided into inner and outer bays, as shown in Fig. 1b, which has four major distributary outlets, namely Humen, Jiaomen, Hongqili, and Hengmen from northeast to southwest (Figs. 1b,1c). Prior to the 1980s, the total annual freshwater discharge and sediment load through these four outlets accounted for 53.4% and 47.7% of the Pearl River, respectively, while the remaining freshwater and suspended sediments were discharged through the other four outlets in the west PRD (Fig. 1b;Pearl River Water Resources Commission 1987). The four northern outlets accounted for 18.5%, 17.3%, 6.4%, and 11.2% of the total annual freshwater discharge into the PRD, and 9.3%, 18.1%, 7.3%, and 13.0% of the annual sediment load, respectively (Pearl River Water Resources Commission 1987). Since the 1980s, sand excavation within the channel network of the delta has altered the slope of the riverbed (Zhang et al. 2009;Yuan and Zhu 2015), thereby leading to increases in both freshwater (accounting for 63.5% of the Pearl River) and suspended sediment discharge (accounting for 56.8%) through the four outlets into the LDB (Huang and Zhang 2005;Liu et al. 2019). In addition, two waterways have formed at the Jiaomen Outlet, one in the north (Fuzhou Waterway) and one in the south (Jiaomen Waterway), resulting from shoal reclamation in the late 1960s to the early 1970s (Fig. 1c). The Fuzhou Waterway discharges ∼70% of the Jiaomen freshwater into the mouth region of the Humen Outlet (Li 2004).
In the 1980s, when human activities were less prevalent, the subaqueous topography of the LDB was characterized by three shoals (West, Middle, and East) where the water depth was shallower than 5 m, and two channels, the West Trough (also named as the Lingding Channel) and the East Trough (also named as the Fanshi Channel) with water depths between 10 and 20 m (Xu et al. 1985;Zhang et al. 2015a;Fig. 1c). Dredging of the Tonggu Channel in 2011 ) divided the Middle Shoal into two parts, resulting in the South Shoal. To accommodate increasing shipping demand, the Denglong Waterway ( Fig. 1c) offshore of the Hongqili Outlet was dredged in 2002 and extended southeastward to connect with the outer Lingding Channel (He et al. 2020). In addition, sand mining of the Middle Shoal since 1995 resulted in the formation of numerous deep pits, substantially reducing the size of the Middle Shoal and causing the East Trough to become deeper and wider ( Fig. 1c; Wu et al. 2018).
The tidal regime in the LDB is semi-diurnal and somewhat irregular, with an average range of 1.11-1.63 m and a maximum range of 2.48-3.39 m at the four outlets (Xu et al. 1993;Mao et al. 2004). Flood tidal currents intrude into the LDB in a north-northwest direction, forming a flood-dominated trough (the East Trough), while ebb currents flow in a south-southeast direction and form the ebb-dominated West Trough (Zhao 1990), leading to a major southwestward dispersal of the river plume ). Based on the tidal ranges and the ratio between freshwater discharge and the tidal prism, Luo et al. (2002) suggested that the Humen mouth is tide-dominated, while the Jiaomen, Hongqili, and Hengmen mouths are fluvial-dominated. In addition, higher-salinity shelf water intrudes into the LDB through the bottom parts of the stratified water columns flowing through the Lantau and Urmston channels during both the flood and dry seasons (Fig. 1b;Miao et al. 1988;Zhao 1990;Mao et al. 2004). Previous studies suggest that there are three major sediment dynamic regimes (Fig. 1c) in the LDB: a fluvial-dominated regime in the northwestern part, a tide-dominated regime in the northeastern part, and a high-salinity shelf water regime in the southern part (Miao et al. 1988;He and Li 2005). Correspondingly, the morphology of the northwestern LDB is characterized by estuary mouth shoals and subaqueous distributary channels, while tidal channels and a linear ridge system have developed with muddy tidal flats in the northeastern LDB. Coarse-grained sediments with a high content of medium and coarse sand and some fine gravel prevail in the tidal channels and ridges (Xu et al. 1985). Deep troughs, such as the Lantau and Urmston channels, occupy the outer LDB where shelf water prevails, and the surficial sediments are dominated by silty mud (Zhao 1981).
Partial mixing of the river water plume and shelf water characterizes the LDB at both the tidal and seasonal scales. However, the water column is highly stratified in neap tides during the flood season when there are large volumes of freshwater discharge but is fully mixed during spring tide conditions during the dry season (Xu et al. 1985;Su et al. 1992;Dong et al. 2004). Shen et al. (1995Shen et al. ( , 2001 suggested that the intrusion of shelf water is important for the formation of turbidity maxima in the LDB.

Material and methods
We used a grab sampler (DS-QNC6) and collected 130 and 120 surficial sediments (approximately 500 g for each sample, 5-10 cm thick) from the LDB in December 2016 (dry season) and August 2017 (flood season), respectively (Fig. 1c). Surficial sediments are likely to integrate material deposited over a period of 1-10 years in addition to the seasonal deposition, based on a sedimentation rate of ∼1-5 cm/year during 1995-2015 in most parts of the LDB ). All samples were immediately stored in cool conditions (0-4°C) after being transported to the laboratory. The grain size and organic geochemistry of surficial sediments sampled during the 2016 dry season were reported by Yuan et al. (2019). In this study, we used identical pretreatment methods and instrumentation as in Yuan et al. (2019) to analyze the grain size and organic elemental and isotopic compositions of the surficial sediments collected in August 2017. Grain size was measured using a Mastersizer 2000 laser diffraction particle size analyzer (Malvern, UK) with a measurement range of 0.02-2000 μm. Samples containing fine gravel were measured using a Camsizer X2 particle size and shape analyzer (Retsch Technology, Germany). Total organic carbon (TOC) and total nitrogen (TN) were measured with a vario EL III element analyzer (Germany), using the national geochemical reference standard GSD-9, with an accuracy of 0.5%. Stable organic carbon isotope (δ 13 C) measurements were carried out using a Delta Plus XP isotope ratio mass spectrometer (Thermo Finnigan, USA) with reference standards caffeine (IAEA-600), cellulose (IAEA-CH-3), and black carbon (GBW04407 and GBW04408) with an accuracy of <0.1‰. Both grain size and organic geochemical analyses were performed at the State Key Laboratory of Estuarine and Coastal Research, East China Normal University.
End-member modeling of grain size is widely used to identify dynamic processes, such as near-bottom hydrological conditions, ocean currents, and paleo-storm events, which control sediment dispersal and trapping (Prins et al. 2002;Hamann et al. 2008;Parris et al. 2010;Dietze et al. 2014). This is based on the principle that different sedimentation processes are associated with particular granularity characteristics, viz. dynamic populations, and that these are distinguishable as end-members (Weltje 1997;Weltje and Prins 2003). In this study, MATLAB R2019a software was used to run the end-member modeling program AnalySize (Paterson and Heslop 2015) to examine the major end-members of grain size and their spatial distributions for surficial sediments collected during both dry and flood seasons. We determined the number of end-members by examining the linear correlations between the original dataset and the dataset represented by the end-members ( Fig. 2; R 2 and deviation angle, see Paterson and Heslop 2015). The modeling showed that when four end-members were used for both flood-and dry-season samples, the median value of R 2 was close to 1 and the mean value exceeded 0.95 (Figs. 2a, 2c), whereas the increase in R 2 was limited when five or six end-members were used. Moreover, both the median and mean values of the angle of deviation were below 10 in the four end-member solutions (Figs. 2b,2d), and there was minimal decline in the deviation angle when using five or six end-members. It was thus concluded that a four end-member solution is most appropriate for surficial sediments in both the flood and dry seasons in this study.
Gao-Collins grain-size trend analysis extracts the direction of net transport of sediment in the grain-size trend vector images, based on the hypothesis that the direction should be related to spatial variations in grain-size parameters such as mean grain size, sorting coefficient, and skewness coefficient (Mclaren 1981;Gao and Collins 1992;Gao 1996). This approach has been widely applied in estuaries (Mallet et al. 2000;Duck et al. 2001;Pascoe et al. 2002;Friend et al. 2006), harbors (Gao and Collins 1992;Duman et al. 2004), continental shelf (Duc et al. 2007;Liu et al. 2014), and other sedimentary environments. We applied this technique to surficial sediments in the LDB during both sample seasons. First, ArcMap 10.6 was used to determine the characteristic distance (D cr ) (Gao and Collins 2001;Poizot et al. 2008) of the surficial sediments, which was calculated as 6827 and 8862 m in the dry and flood seasons, respectively. We then determined the net transport direction for the grain-size trend analysis following the method of Gao (1996). Previous research revealed that gravel-bearing sandy sediments in the channels were formed during the late Pleistocene and were subsequently exposed at the seabed due to dredging (Yuan et al. 2019). Therefore, these sediments were not included in the grain-size trend analysis.
The surficial sediments were classified according to their distribution across four zones (Fig. 1c) identified based on the subaqueous topography and sediment dynamics suggested by previous studies, that is, the Lingding Channel and the boundary of the river water plume front in the dry season, as reported by Lai et al. (2015). Sediments in Zone I are from the northwestern and western LDB bordered by the Lingding Channel to the east and the plume front to the south, where it receives freshwater from the Jiaomen, Hongqili, and Hengmen outlets and is therefore influenced by low salinity discharge (Miao et al. 1988;Lai et al. 2015;Wei and Zhu 2019). Zone II is located in the northeastern LDB bordered by the Lingding Channel to the west and the plume front to the south, where tidal currents predominate (Miao et al. 1988;Wei and Zhu 2019). Zone III-1 is located in the eastern part of the outer LDB bordered by the plume front to the north and the Lingding Channel to the west. Zone III-2 is located in the western part of the outer LDB and is bordered by the Lingding Channel to the east and the plume front to the west. Higher-salinity shelf water prevails in Zones III-1 and III-2 (Miao et al. 1988;Wei and Zhu 2019).

Spatial and seasonal variations in end-member grain size
First, similar end-member (EM) compositions were evident for the surficial sediments in the dry season of 2016 and the flood season of 2017 (Fig. 3). For surficial sediments during the dry season in 2016, EM1 represents very fine silt (peak size at 7.19 μm) and EM2 represents coarse silt (peak size at 34.20 μm). The peak for EM3 lies at 136.79 μm (fine sand), with a secondary peak at 5.08 μm (very fine silt). EM4 has a main peak at 460.09 μm (medium sand) and a secondary peak at 34.20 μm (coarse silt). In the flood season of 2017, the peak value of EM1 is 7.19 μm (very fine silt) and EM2 is 48.36 μm (coarse silt). A major peak at 162.67 μm (fine sand) and a secondary peak at 6.05 μm (very fine silt) occurred in EM3. EM4 revealed an initial peak at 386.89 μm (medium sand) and a secondary one at 4377.17 μm (medium gravel).
Surficial sediments in the LDB are predominantly characterized by EM1 and EM2, while the contributions of EM3 and EM4 are usually below 10% in both the flood and dry seasons (Fig. 4). EM1 prevailed in the sediments of Zones I and II (Figs. 4a, 4b), with median values of ∼85% recorded in the flood season of 2017 (Fig. 5a). EM2 exhibited increased proportions in Zones III-1 and III-2 (Figs. 4c, 4d), particularly in the dry season of 2016, with median values of 35%-45% (Fig. 5b). The highest values for EM3 were found in individual samples retrieved from the shoal between the Jiaomen and Hongqili outlets in the northwestern LDB and the linear sand ridges in the northeastern LDB (Figs. 4e, 4f), which reflects the jet flow of the freshwater discharge superimposed by the ebb current at the river mouths (Wei et al. 2020) and the influence of tidal currents in the northeastern LDB. High EM4 values are only exhibited by very few samples from the channels and adjacent shoals (Figs. 4g, 4h), possibly reflective of late Pleistocene riverbed deposits exposed on the seabed by channel dredging (Yuan et al. 2019). Accordingly, in this study, we focused on the spatial distribution and variation in EM1 and EM2.
Statistical analysis revealed that, spatially, the contribution of EM1 in Zones I and II was markedly higher than that in Zones III-1 and III-2, while the situation is reversed in the case of EM2. EM1 (EM2) in all zones exhibited generally higher (lower) proportions in the 2017 flood season than in the 2016 dry season (Fig. 5a). In particular, the seasonal variation of EM1 was significant (0.01 < p < 0.05) and the seasonal variation of EM2 was very significant (p < 0.01) in Zone III-2 (Fig. 5).

Grain-size trend analysis
The analysis of the Gao-Collins grain-size trend (Fig. 6) revealed consistent net transport directions of the LDB surficial sediments for December 2016 and August 2017, with two similar major sediment convergence centers, that is, depocenters, in the northeast (northern part of Zone II) and southwest (southwestern part of Zone I). There is clear seasonal variation in the net transport vector, which was significantly greater during the flood season than during the dry season. A strong northeastward net transport trend was observed from the Jiaomen and Hongqili outlets during the flood season, leading to a more significant depocenter in the northeastern LDB (Fig. 6b). The net transport trend of sediments at the mouth of the LDB was mainly northward during the dry season and westward during the flood season, which resulted in a stronger southwestward transport trend in the outer LDB during the flood season. There was also a weak depocenter in the dry season to the west and southwest of the inner Lingding Island, where the Denglong Waterway converges with the Lingding Channel (Fig. 6a), and a weak convergence appeared at the South Shoal in the southeastern LDB during the flood season (Fig. 6b).

Spatial and seasonal variations in organic geochemical composition
The TOC contents of surficial LDB sediments in December 2016 and August 2017 mostly exceeded 0.6%, with the highest values (>2.0% in the flood season) near the northeast coastline (Figs. 7a, 7b). During the flood season, TOC values were slightly lower than in the dry season, with markedly lower values (<0.4%) in the northwestern LDB, Middle Shoal, Lingding Channel, and Urmston Channel (Figs. 7a, 7b, 8a). C/N ratios mostly exceeded 8.0 in both flood and dry seasons (Figs. 7c, 7d), although a greater proportion of sediments with dry season C/N ratios <8 was sampled at the southeastern inner LDB and southwestern outer LDB. This contrasts with the situation in the northwest shoal, Urmston Channel, and shelf, where only occasional samples had lower C/N ratios. Marked seasonal variations in C/N ratios of sediments from Zones I and III-2 were observed, especially in the latter zone, with a slight increase in the flood season for Zone II (Fig. 8b). Spatial variations in δ 13 C values proved to be statistically significant among the four zones, whereby samples from Zone II were the most strongly depleted (−25‰ to −28‰) (Figs. 7e, 7f, 8c). δ 13 C values in Zone I were also depleted, but the values increased southward. The δ 13 C values of sediments from Zones III-1 and III-2 ranged mainly between −23‰ and −24‰, while individual samples on the shelf exhibited more depleted values in the dry season. Zones II and III-1 exhibited seasonal variation in δ 13 C characteristics, whereby in the flood season values were significantly more negative in Zone II and slightly more enriched in Zone III-1 (Fig. 8c).  EM1 (a, b), EM2 (c, d), EM3 (e, f), and EM4 (g, h) for the surficial sediments collected from Lingding Bay (LDB) in December 2016 and August 2017. The black line in each inset represents the river plume front in Fig. 1c. The maps were generated with Surfer 13 software.
In addition, flood season δ 13 C values in Zone I were more depleted in the area southeast of Qi'ao Island (Fig. 7f). Figure 9 shows the C/N versus δ 13 C plots for the LDB surficial sediments, which can help discriminate organic carbon sources (Lamb et al. 2006). Sediments from Zone I in both seasons plot within the ranges of marine algae/particulate organic carbon (POC) and freshwater algae/POC, with some transitional between algae/POC and C 3 plants. Samples for the flood season were more closely aligned with the ranges of freshwater algae/POC and terrestrial C 3 plants. Most samples of Zone II in both seasons, but particularly in the flood season, indicate ranges consistent with freshwater algae/POC and transitional to C 3 plants, reflecting a prominent terrestrial organic source (Fig. 9b). Figures 9c and 9d indicate that sediments from Zones III-1 and III-2 lie in the ranges of marine algae/POC and are influenced mainly by the contributions of marine carbon. Minor seasonal variation was observed in the C/N versus δ 13 C plots for Zone III-1 (Fig. 9c), suggesting that flood season samples were more influenced by marine algae/POC than in the dry season. In contrast, flood season samples from Zone III-2 (Fig. 9d) are seen to be concentrated in the transitional range Fig. 4. (continued) between marine algae and C 3 plants, reflecting an increased terrestrial contribution. Overall, the flood season C/N versus δ 13 C plots indicate an increase in terrestrial contribution of organic carbon in surficial sediments of Zones I, II, and III-2 and a slight increase in the marine organic carbon contribution in those of Zone III-1.  5. Box-and-whisker plots for EM1 (a) and EM2 (b) with dry-season samples in red and flood-season in black. Very significant (p < 0.01) and significant (0.01 < p < .05) differences, as examined by t-tests, are indicated by ** and *, respectively.   Fig. 1c. The maps were generated with Surfer 13 software. Fig. 8. Box-and-whisker plots for the total organic carbon (TOC) content (a) and organic elemental (b) and isotopic (c) compositions, with dry-season samples in red and flood-season in black. I, II, III-1, and III-2 refer to the zones in Fig. 1c. Very significant (p < 0.01) and significant (0.01 < p < .05) differences, as examined by t-tests, are indicated by ** and *, respectively.

The depositional pattern in the LDB
The comparison of end-member grain size between LDB samples collected in December 2016 and August 2017 indicates the marked deposition of fine-grained sediments (EM1) during the flood season, particularly in the northeast and southwest parts (Fig. 4). The increased terrestrial organic contribution in the flood season (Figs 7,9) is further evidence of the trapping of fluvial sediments in the LDB. Previous investigations revealed that the median diameter (d 50 ) of flood season Pearl River suspended sediments lay in the range of 2-9 μm (Dong et al. 2006), which concurs with the peak value of 7.19 μm for EM1 recorded here. Furthermore, the consistent occurrence of two major depocenters in the northeastern and southwestern LDB derived from samples of both dry and flood seasons (Fig. 6) suggests that this is also a long-term depositional pattern in addition to the clearly evident seasonal pattern. Fig. 9. C/N versus δ 13 C plots (after Lamb et al. 2006) to indicate organic carbon sources for the surficial sediments collected in Lingding Bay (LDB) in both flood and dry seasons. I, II, III-1, and III-2 refer to the zones in Fig. 1c. In contrast, the secondary depocenters in December 2016 (i.e., the area between the Denglong Waterway and Lingding Channel, Fig. 6a) and August 2017 (the region of Zone III-1, Fig. 6b) had lower C/N ratios (Fig. 7c) and less depleted δ 13 C values (Figs. 7e, 7f, 8c), indicating an increase in the contribution of marine-sourced organic carbon and inferred that the deposition of fine-grained sediments was induced by the intrusion of shelf water and associated gravitational circulation (Cui et al. 2018).

Formation mechanism of two major depocenters and their connections to human impact
The occurrence of two major depocenters of fine-grained sediments in the northeast and southwest LDB (Fig. 6) has not been previously reported and may therefore be reflective of a recent or novel sediment dispersal and depositional pattern in the bay (Fig. 10). We propose that the new depositional pattern reflects a system regime shift induced by changes in the boundary conditions in the LDB, including the geometry and hydro-dynamic processes that resulted from intensive human activities over the past decades. Evidence for a mud depocenter in the northeastern LDB, particularly prominent in the flood season (Figs 4b, 6), contrasts markedly with previous reports of sand dominance in the 1970s ( Fig. 10a; Tang et al. 2013). It has been noted that tidal reworking and associated ridge-and-trough morphology prevailed in this part of the bay ( Fig. 10d; Xu et al. 1985;Miao et al. 1988). However, based on evidence for the trapping of significant volumes of terrestrially sourced fine-grained sediments in the northeastern LDB during the flood season (Figs. 4b, 9b), we suggest that present-day depositional processes here have shifted to mud convergence at the interface of the river and tidal flow due to the strengthening of the fluvial function (Fig. 10e). Increased fluvial function may occur for two reasons. First, it has been shown that sand excavation in the river channels of the delta plain has led to an increase in the volume of freshwater discharged into the LDB (Zhang et al. 2009;Yuan and Zhu 2015;Zhang et al. 2015b), thereby amplifying the fluvial influence in the bay. Second, the Fuzhou Waterway (i.e. northern branch of the Jiaomen Outlet; Fig. 10e) was constructed due to reclamation of the mouth shoal in the late 1960s to the early 1970s and led to the discharge of approximately 70% of the freshwater in the Jiaomen Channel flowing into the mouth region of the Humen Outlet in the northeastern LDB (Li 2004). Such an increase in freshwater volume would have altered the interactions between fluvial and tidal characteristics in the northeastern part of the LDB during the flood season, resulting in an intensification and seaward shift of the mud convergence zone between river-dominated and tidedominated tracts (Dalrymple and Choi 2007;Gugliotta and Saito 2019), thereby fashioning the fine-grained depocenter of Zone II (Fig. 4b).
Moreover, the grain-size trend analysis reported here indicates that the fine-grained suspended sediments from the northwestern outlets augment the mud depocenter in the northeastern bay. This also contrasts with previous reports that the suspended sediment load of the Hongqili and Hengmen outlets was deposited mainly in the shoal west of the Lingding Channel (Xu et al. 1985;Miao et al. 1988). Indeed, the simulation by Zhang et al. (2019) indicated an eastward dispersal sub-direction of freshwater discharge from these outlets. We further suggest that this eastward dispersal resulted from extensive reclamation at the river mouths of the Jiaomen, Hongqili, and Hengmen channels in the northwestern LDB (Figs. 10d-10f;Yang et al. 2019). Accordingly, the accommodation space in the northwestern LDB has been substantially reduced, which, together with the increased freshwater discharge due to sand excavation in the delta plain channels (Zhang et al. 2009;Yuan and Zhu 2015;Zhang et al. 2015aZhang et al. , 2015b, resulted in the eastward and cross-channel dispersal of the fluvial plume. Compared to 2005-2006, markedly depleted δ 13 C values in the northeastern LDB reported here (Figs. 10g-10i) provide evidence for increased terrestrial input, which can be attributed to both higher freshwater discharge into the northeastern LDB and the strengthened eastward and cross-channel dispersal of the fluvial plume from the northwestern LDB.
The mud depocenter close to the coastline of the southwestern LDB (Fig. 6), which exhibits >90% EM1 content and increased terrestrial organic contribution in the flood season (Figs. 4b, 7) has also not been recorded in previous studies (Tang et al. 2013). We suggest that this is indicative of a change in sediment dispersal and depositional processes in the western LDB. A previous study by Zhang et al. (2019) suggested major trapping of fine-grained sediments southeast of Qi'ao Island in the wet season of 2007. In contrast, this study revealed limited contributions from EM1 in that region (Fig. 4b), which are suggestive of vigorous hydrological energy conditions, although the C/N ratios and δ 13 C values in the flood season (Figs. 7d, 7f) are consistent with elevated terrestrial input. We propose here that the reclamation at the shoreline and river mouth and rapid shallowing of the LDB's West Shoal during recent decades, reflected in the rapid southeastward advance of the shoreline and 5-m bathymetry (Figs. 10d, 10e), markedly altered the hydro-dynamic circulation around Qi'ao Island, thus changing the routine of sediment dispersal from the channel on the eastern side of the island to the channel on the western side. Previous field investigations suggested a clockwise circulation that dominated in the 1980s by ebb flows on the eastern side and flood flows on the western side of the island ( Fig. 10d; Xu et al. 1985). However, a numerical simulation by Lai et al. (2015) showed that a considerable volume of freshwater discharge from the Hongqili and Hengmen outlets was dispersed seaward via the channel on the western side of Qi'ao Island, suggesting that this channel evolved in response to a change from flood-dominated to ebb-dominated flow conditions in the recent past (Fig. 10e). Based on this, and on the evidence of surficial sediments presented here, it is proposed that the seaward advance of the shoreline and river mouth due to reclamation strengthened the ebb flow and associated transport of fluvial sediments through the channel west of Qi'ao Island, while the shallowing of the West Shoal, as indicated by the southeastward migration of 5-m bathymetry (Figs. 10d, 10e), weakened the flood tidal current by increasing bottom friction. Thus, during the flood season, the fluvial plume passing through the channel west of Qi'ao Island, together with the prevailing easterly wind and the Coriolis force Liu et al. 2019), resulted in nearshore trapping of finegrained suspended sediments in the southwestern LDB. Furthermore, shallowing of the West Shoal may also have promoted seaward migration of the plume front, as inferred by the occurrence of the secondary dry season mud depocenter located seaward of the position reported by Lai et al. (2015) (Fig. 6a).

Dry-season export of sediments out of the LDB
The coarsening of surficial sediments in the dry season (Figs. 4,5) may indicate that the present-day LDB serves as a temporary depositional sink for fine-grained suspended fluvial sediments during the flood season. We suggest that human activities play a role in seasonal sediment dispersal in two respects. First, intensive dredging and sand mining in the bay have maintained the overall volume of water in the bay, even though its surface area has been reduced by reclamation ( Fig. 10f; Yang et al. 2019). As a consequence, tidal amplitude increased in the inner LDB, which may intensify the tidal redistribution of fine-grained sediments in the dry season. Second, previous studies have suggested that wind-driven currents in the LDB are negligible because of their sheltered environment (Owen 2005). Seasonal changes in the grain size of the surficial sediments in the LDB in this study, however, suggest augmented dry season redistribution. Therefore, we propose that the shallowing of the West Shoal due to human activities promotes sediment resuspension by waves and subsequent longshore transport of fine-grained sediments to the continental shelf beyond the LDB in the winter season (Wong et al. 2003;2004).
Such redistribution of fine-grained sediments due to seasonal variations in freshwater discharge and hydrodynamics has been widely reported for the open coast of other tide-dominated and monsoon-influenced subaqueous deltas, such as the Yangtze and Mekong (Nguyen et al. 2000;Guo et al. 2003;Nittrouer et al. 2017), but less commonly in the associated embayments. We contend that terrestrial sediments exported from the LDB have increased in recent decades. This is supported by evidence presented here of elevated terrestrial organic carbon content in surficial sediments at the mouth of the LDB compared to 2005. The implications of dry-season transport of fine-grained sediments out of the LDB, for example, on the ecological environment of the continental shelf, warrants further consideration.

Conclusions
This study reports on the analysis of surficial sediments from the LDB collected in the dry season (December 2016) and flood season (August 2017) to investigate the spatial and temporal variations in their grain size and organic geochemical compositions. Grain-size end-member modeling and trend analysis are applied to the reconstruction of sediment dispersal and trapping patterns and processes in the LDB under the new boundary conditions imposed by significant levels of anthropogenic disturbance. The main conclusions are as follows: • Two major depocenters of fine-grained sediments characterized by high EM1 content in the northeastern and southwestern LDB across both flood and dry seasons were revealed as a recently emerged depositional pattern in the LDB, which we speculate is indicative of a hydro-and sediment dynamic system regime shift induced by intensive human impact over recent decades.
• The mud depocenter and depleted δ 13 C values in the northeastern LDB indicate a transition of depositional processes from predominantly tidal reworking to mud convergence in the interface zone between the river and tidal flows, which we infer as a response to strengthened fluvial function induced by human activities, including sand excavation in the river channels of the delta plain and reclamation in the river mouth.
• The major depocenter of fine-grained sediments in the southwestern LDB is characterized by increased terrestrial contributions of organic carbon during the flood season. It is proposed that the formation of this depocenter is associated with the evolution of a channel west of Qi'ao Island. The channel appears to have evolved from being flood-tide dominated to an ebb-dominated freshwater conduit owing to reclamation and associated shoaling of the western LDB.
• The coarsening of surficial sediments in the dry season in the LDB, together with the increase in the terrestrial organic carbon contribution at the mouth of the bay in recent years suggests an increase in mud exported from the bay to the continental shelf, possibly induced by the shallowing of the western LDB.