A comprehensive sediment dynamics study of a major mud belt system on the inner shelf along an energetic coast

Globally mud areas on continental shelves are conduits for the dispersal of fluvial-sourced sediment. We address fundamental issues in sediment dynamics focusing on how mud is retained on the seabed on shallow inner shelves and what are the sources of mud. Through a process-based comprehensive study that integrates dynamics, provenance, and sedimentology, here we show that the key mechanism to keep mud on the seabed is the water-column stratification that forms a dynamic barrier in the vertical that restricts the upward mixing of suspended sediment. We studied the 1000 km-long mud belt that extends from the mouth of the Changjiang (Yangtze) River along the coast of Zhejiang and Fujian Provinces of China and ends on the west coast of Taiwan. This mud belt system is dynamically attached to the fluvial sources, of which the Changjiang River is the primary source. Winter is the constructive phase when active deposition takes place of fine-grained sediment carried mainly by the Changjiang plume driven by Zhe-Min Coastal Currents southwestward along the coast.


Quantification of Tidal Influence
The tidal influence is expressed by tidal energy ratio (ER) as follows:

×100%
(1) in which v(t i ) is a variables in a time series, v p (t i ) is the tidal part of the time series expressed as: in which v is mean value of the observed time series; ω k is radian frequency (2π/T k ); T k is period of the k th tidal constituent; α k and θ k are amplitude and phase of the k th tidal constituent respectively; m is the number of data points in the time series.
Basically, ER is the ratio between tidal variance and total variance. The higher the ER, the greater the tidal influence. Some variables measured on the mooring at difference depths are tabulated below to illustrate different degrees of tidal influence on variables. Note: The number after the @ sign is the elevation in meters above the bed (mab).
The above results indicate that the tide was the primary forcing for the water depth changes. The tide also dominated the flow field in the across-shore flow component (ER over 93%), but the alongshore flow was greatly affected by non-tidal processes (ER from 28.6%-44.1%) that included the Zhe-Min Coastal Current, Changjiang buoyant plume, and other coastal forcing. The tidal influence was week in the SSC fluctuations (ER between 27.4%-35.5%). The increase of ER away from the seabed was probably because of reduced wave influence and increased influence of the tidal regime.
The salinity and temperature were the least affected by the tide (ER smaller than 20%). This is not surprising because both were largely determined by the strength of the Changjiang fresh water discharge and other current systems such as ZMCC and KBC. Solar heating and cooling, and weather events also affected the water temperature.
Both tide and the wind field exerted influence on the wave field. The tidal influence can be seen in the temporal changes in the increase of the significant wave height (Hsig) that coincided with the spring tide (Fig. S3). The influence of the wind can be seen from the shift of the wave direction from ENE (0-90 deg.) to SE (90-180 deg.) during episodes of southerly winds. The weight-composition indicates the fine-grained (finer than medium silt) particles dominated the suspended sediment composition (over 80%) in the water column. The higher presence of coarser particles (greater than 63 µm) at the surface was probably due to the presence of biogenic particles such as phytoplankton. The total SSC (mg/l) was 46.9 mg/l at the surface, 46.2 mg/l at mid-depth, and 298.5 mg/l near the seabed. The total SSC was one order of magnitude greater near the bed than at the surface and mid-depth suggesting that after suspended sediments settled into the lower water column, they are prevented from being mixed upward. Also, the high SSC was contributed from resuspension of seabed sediment.

Numerical Simulation of the Observed Water Column Structures
A numerical model with large domain of the Western North Pacific (

Empirical Orthogonal/Eigen Function Analysis of Sediment Trap Samples
Sinking particles accumulated in sediment trap for about 15-cm in height (Fig. S6a). The trap sediment was sampled at 1 cm intervals for various analyses. The entire trap sediment looked homogenous having grayish brown color with some horizontal coloration, no apparent internal structures were visible to the naked eye. However, the x-ray radiograph shows alternating dark-light laminations, whose appearance could be related to the spring tide, which coincided with larger mean grain-sizes (Fig. S6b). The water content (37-55%) and mean grain-size (9-16 µm) trends show mirror images (Fig. S6a, b), which is common in sediment cores. The cumulative percent (by weight) shows a relative composition of about 20% clay and 80% silt. Only a minute amount of sand appeared in the trap and is not visible in the cumulative plot. The TOC content in the trap ranged from 0.5-1.1, most of which were contributed by marine-sourced particles having C/N ratio from 7.4-10 ( Fig. S6d). However, 13 δC-based terrestrial fraction (F t ) suggests about equal contributions from land and sea (Fig.   S6e). Percentages of 7 minerals including mica (ilite), kaolinite, chlorite, quartz, k-feldspar, plagioclase, and calcite are shown in Figure S6f. The result shows that the first mode explains 43.4% of the correlations (normalized co-variance). The 14 variables were grouped into two groups according to sign of the eigenvector. This mode distinguished clay (marked red, in the minus-sign group) from silt and sand (marked blue, in the plus-sign group), which indicates the effect of hydrodynamic sorting in the particle transport process that separated the fines (clay) from the coarse fraction (silt and clay). Co-varied with clay are organics, F t , the three clay minerals, and calcite ( Fig.   S7a). This mode reveals the fundamental nature of the sinking particles that the clay-sized particles were of terrestrial origin, and they were associated with organics, clay minerals, and terrestrial calcite. The second mode sets mica (ilite), kaolinite, and plagioclase apart from other variables (Fig. S7a). This mode points to the provenance contrast of weathering product of granite (marked in yellow) vs. non-granitic sources. In the third mode clastic indicators of sand, total 210 Pb, kaolinite, k-feldspar, and calcite (marked in green) are in the negative group and the organics, F t and silt were in the positive group. This mode might show provenance contrast between marine-sourced coarse clastic provenance vs. silty organics of terrestrial origin.
These modes suggest that physical processes that transported the suspended sediment had the greatest influence on the properties of the sinking particles and the source-signals they carry.
The different provenance in the region and the marine environment also influence the particle properties in the secondary and tertiary degrees.

TS Mud Belt and Locations of Cores G3 and ZS-3 on the CYR
Contours of the mean grain-sizes are extracted from Huh et al. (2011) and were plotted over the bathymetric contours in the TS to show the location of the TS mud belt. Both Cores G3 and ZS-2 were located on the CYR. G3 was located in water depth of 26.3 m, outside the TS mud belt and ZS-2 was located in water depth of 25 m just inside the boundary of the TS mud belt and a little closer to the mouth of Zhuoshui River (Fig S8a). The distance between the two coring sites was 16.9 km and ZS-2 was closer to the Zhuoshui River (Fig.   S8b).

Boundaries of Current Systems Delineated by Sea Surface Temperature Gradients
Composites of multi-year sea surface temperature were acquired by AVHRR (Advanced Very High Resolution Radiometer) on NASA's satellites (https://podaac.jpl.nasa.gov/AVHRR-Pathfinder). Entropy-based edge detection method was than used to obtain gradient magnitude of frontal pixels (Chang and Cornillon, 2015;Shimada et al., 2005). Areas of high-density gradient delineate the boundaries/fronts between different water masses driven by different ocean current systems. The front between ZMCC and the ambient ECS and TS (the black line marked by circled number 1) is better developed in winter (Fig. S10), which was the best developed in February (Fig. S10b). This black line forms a large bend in northern TS, whose apex (marked by the orange arrow) is the position where Taiwan-derived mud belt joins the mud belt along the Zhe-Min Coast. The other frontal system marked by circled number 2 is the KBC, which is also best developed in February (Fig. S10b).