Lagrangian coherent structure analysis on transport of Acetes chinensis along coast of Lianyungang, China*

Kexin WANG, Xueqing ZHANG,**, Qi LOU, Xusheng XIANG, Ying XIONG

1 College of Environmental Science and Engineering, Ocean University of China, Qingdao 266100, China

2 Key Laboratory of Marine Environment and Ecology, Ministry of Education of China, Ocean University of China, Qingdao 266100, China

3 Jiangsu Marine Fisheries Research Institute, Nantong 226007, China

Abstract Spatial heterogeneity or “patchiness” of plankton distributions in the ocean has always been an attractive and challenging scientific issue to oceanographers.We focused on the accumulation and dynamic mechanism of the Acetes chinensis in the Lianyungang nearshore licensed fishing area.The Lagrangian frame approaches including the Lagrangian coherent structures theory, Lagrangian residual current, and Lagrangian particle-tracking model were applied to find the transport pathways and aggregation characteristics of Acetes chinensis.There exist some material transport pathways for Acetes chinensis passing through the licensed fishing area, and Acetes chinensis is easy to accumulate in the licensed fishing area.The main mechanism forming this distribution pattern is the local circulation induced by the nonlinear interaction of topography and tidal flow.Both the Lagrangian coherent structure analysis and the particle trajectory tracking indicate that Acetes chinensis in the licensed fishing area come from the nearshore estuary.This work contributed to the adjustment of licensed fishing area and the efficient utilization of fishery resources.

Keyword: plankton accumulation; hydrodynamic model; Lagrangian particle-tracking model; Lagrangian analysis

1 INTRODUCTION

Aceteschinensis(A.chinensis), a small planktonic shrimp of the genusAcetesin the Sergestidae family,is the most productive marine shrimp resource in China (Qi et al., 2013; Xu et al., 2014).It widely inhabits coastal areas in muddy and sandy shallow waters in China (Li et al., 2021).As prey for a wide variety of fish,A.chinensisis a critical species in the trophic ecosystem (Vase et al., 2021).To protect and effectively utilize the fish resource, since 2020,the fisheries total allowable catch system pilot project forA.chinensisduring the fishing moratorium in China was implemented in Haizhou Bay, Jiangsu Province.

Much literature aboutA.chinensisfocuses on the population characteristics (Amin et al., 2012), genetic laws (Huang et al., 2019), and biochemical properties(Zhang et al., 2021a), especially in food processing(Li et al., 2017) and medicinal value (Kawato et al.,2018).However, there is little research on the spatial and temporal distribution pattern ofA.chinensisand its dynamics.The main research methods are survey and model simulation.In terms of observation, the traditional survey method of fishery resource distribution is trawl method.In recent years, some new approaches, such as underwater visual census techniques (Edgar et al., 2004; Tessier et al., 2013)and acoustic methods (Zeng et al., 2019) show advantages of high efficiency, fast, and real-time.In addition, the vessel monitoring system (VMS) has been effectively used in fishing effort calculation and fishing ground analysis (Fonseca et al., 2008; Li et al., 2022).Based on Beidou VMS data, Li et al.(2021) extracted the net position and fishing effort and then spatially visualized the resource abundances ofA.chinensisin Haizhou Bay.

As to the model simulation, few studies focused on the distribution and its mechanism ofA.chinensis.In the ocean, the distribution of plankton is spatially heterogeneous or in patches, which has always been an attractive and challenging scientific issue to oceanographers.Because of the small size(about 3-4 cm long) and poor swimming ability,A.chinensisis controlled by tidal current in horizontal direction, andA.chinensismakes its way to deep water daytime and rise to the sea surface after the sun sets.This process is known as diel vertical migration (Shi, 1986).A.chinensisis generally classified into zooplankton, which is often considered passive or buoyant floaters (Metaxas, 2011), similar to bioparticles and plumes in the ocean (Xia et al.,2011), driven by wind, tides, or other agents.Therefore, the convection-diffusion equation for continuous material is unsuitable for studying transport ofA.chinensis.The particle tracking model (Mao and Xia, 2020a), especially the individual-based model (IBM) (El Saadi and Bah, 2006; Falcini et al.,2020) is often used to study the transport of the biological particles.However, it is difficult to find the transport pathway and accumulation area and its mechanism from the complex and intertwined trajectories.As we know, transport in the ocean has important implications for biological processes at many trophic levels and is governed by coherent structures on different scales (Martin, 2003; Wiggins,2005).

In this study, the theory of Lagrangian coherent structures (LCSs) is introduced to reveal the source and the accumulation mechanism ofA.chinensisin the research area.LCSs, as a boundary between regions,can be considered as a hidden skeleton of fluid flow(Haller and Beron-Vera, 2013; Peacock and Haller,2013).Their visualization can provide information about the transport paths and barriers in the material transport process (Haller, 2011; Suara et al., 2016;Wang et al., 2017).Harrison et al.(2013) confirmed that attracting LCS (aLCS) would mark the main material lines for marine propagules.In this study,we hypothesize thatA.chinensiswill preferentially concentrate in material lines mapped by attracting Lagrangian coherent structures (aLCSs).In the past decade, the LCSs theory has been widely applied in the flow fields in the atmosphere (Harvey et al.,2021), ocean (Zhang et al., 2021b), and even blood(Mutlu et al., 2021).It can also be used to explore the dynamics of distribution of jellyfish (Dawoodian et al., 2021), coral reefs (Filippi et al., 2021),temperature fronts (Mathur et al., 2019), the impact of typhoon on material transport (Lou et al., 2022),and so on in the ocean.More importantly, many studies have illustrated that LCSs play an essential role in the transport of plankton, larvae, and other propagules (Harrison et al., 2012; Huhn et al., 2012;Vaz et al., 2016; Jonsson et al., 2020) similar in characteristics toA.chinensis.

From the perspective ofA.chinensisfishing,issues that interest us are: What is the driving mechanism of the spatial distribution structure ofA.chinensis? Where did thoseA.chinensiscome from?In this study, we used the Lagrangian coherent structure method to analyze the spatial distribution mechanism ofA.chinensisnear the licensed fishing area.This study is organized as follows.The materials and methods are described in Section 2.In Section 3, the particle-tracking trajectories and the finitetime Lyapunov exponent fields in the study area are calculated, and then LCSs are recognized to analyze the transport characteristics and mechanisms ofA.chinensisin Lianyungang nearshore.Some influencing factors of LCSs are discussed in Section 4.The study is concluded in Section 5.

2 MATERIAL AND METHOD

2.1 Data source

The licensed fishing area forA.chinensiswas located in the area (119°40′E-120°10′E, 34°40′N-34°50′N) near the Lianyungang nearshore in Jiangsu Province, China.Based on the vessel position information provided by Beidou satellite navigation data during fishing period, we estimated the resource production in 2020.The resources distribution ofA.chinensishad a spatially multi-core distribution pattern.The central core area was in 119.8°E-119.9°E,34.6°N-34.7°N, and the secondary one was in 120.1°E-120.2°E, 34.7°N-34.8°N.In 2021, the licensed fishing area expanded to 119°40′E-120°15′E,34°35′N-34°50′N.The predicted distribution ofA.chinensisresources in 2021 is similar to that in 2020, but the fishing yield was significantly lower than that observed in 2020.

2.2 Hydrodynamic model

An unstructured grid, finite-volume, threedimensional primitive equation coastal ocean model(FVCOM) was used to simulate the hydrodynamic.The governing equations consist of the momentum equation, the continuity equation, the diffusion equations of temperature, the salinity, and the density equations; for details please see Chen et al.(2003).In this model, the Smagorinsky formula and the MY-2.5 turbulence closure model are used to calculate the horizontal and vertical diffusion for the momentum.

The model is forced by the sea surface wind and the tide at the open boundary.The tidal elevation is predicted by harmonic constants extracted from TPXO7 (Egbert and Erofeeva, 2002), where four major tidal components (M2, S2, K1, and O1) are considered.The wind data are reanalysis data collected from the European Centre for Medium-Range Weather Forecasts (Hersbach et al., 2020)with temporal resolution of 1 h, and spatial resolution of 1/4 degree.The model computation period is from 1 May 2021 to 30 July 2021, and the output frequency of velocity field is 1 h.

The surface and bottom boundary conditions foruandvare:

The drag coefficientCdis determined by matching a logarithmic bottom layer to the model at a heightzababove the bottom:

whereκ=0.4 is the von Karman constant andz0is the bottom roughness parameter, which is set as constant 0.001 in this study.

The coastline is obtained through Google Earth and the water depth topography is taken from the latest nautical charts.The water depth ranges from 5 m to 20 m near the licensed fishing area forA.chinensis, and there is a shoal with water depth less than 10 m.In the horizontal direction, the unstructured triangular grid is used to fit the irregular coastal boundary of the study area, which contains 47 321 triangles (elements) and 24 270 grid nodes with the resolution of from 600 m nearshore to 6 000 m near the model boundary.Fiveσ-coordinate layers are used in the vertical direction to accommodate the irregular variable bottom topography.

2.3 Lagrangian coherent structure

Several approaches to detect the LCSs have been proposed over the past decade.Hadjighasem et al.(2017) compared 12 detection methods in three examples, among which finite-time Lyapunov exponents (FTLE) is one of the simple and objective algorithms and is widely used.

2.3.1 Computation of the FTLE

For the Lagrangian approach, the researchers study fluid motion by examining particle trajectories,which are given by x(x0,t0,t).The solutions to the following dynamical system are as follows.

where x(x0,t0,t) is the displacement vector of the particle at timet; ẋ is the derivative of the displacement vector x to the timet, v(x(t),t) is the velocity field generated by the FVCOM model.

where the superscriptTrefers to the matrix transposition.AsCt0+τ t0(x0) is symmetric and positive definite, its eigenvaluesλi＞0 are real.

The Lyapunov exponent represents the numerical characteristics of the average exponential dispersion rate of adjacent trajectories in the phase space.According to Haller and Sapsis (2011), the finitetime Lyapunov exponent (FTLE) can be computed as follows:

whereλmaxis the maximal eigenvalue ofCt0+τ t0(x0).FTLE can be computed in both forward (τ＞0) and backward (τ＜0) direction.Integrating the velocity forward in time withτ＞0, the forward FTLE was obtained, while for the backward FTLE field, the negative velocity field was integrated backward in time withτ＜0.

Only the horizontal two-dimensional case was considered in this study, according to the different solutions of the deformation gradient formula(Haller, 2011):

wherexandyare the spatial coordinates and (i,j) is the grid node number.

2.3.2 The rLCSs and aLCSs detection from the FTLE

The LCSs can be considered as the ridges of the largest FTLE field (Shadden et al., 2005), and can be calculated directly from the trajectory (Haller,2001a, 2002).Many algorithms for ridge extraction are effective for a steady flow, but for complex timedependent flow in the real ocean, it is difficult to extract the ridge structure (Allshouse and Peacock,2015).Meanwhile, the structure of the explicit ridge is straightforward to be identified by intuitive recognition from the FTLE field (Huhn et al., 2012;Hadjighasem et al., 2017).The method used to obtain LCSs in this paper is the intuitive recognition,i.e., the local maxima of FTLE is considered as the LCSs

In forward-time calculations, the stretch material lines identified by the FTLE field are called repelling LCSs (rLCSs).In contrast, the fold material lines in the FTLE field are called attracting LCSs(aLCSs) in backward calculations (Haller, 2001b).The rLCSs can be interpreted as the material transport line, and the aLCSs can be understood as the material aggregation line or the boundary of the aggregation region.

3 RESULTS

3.1 Validation of the tide and tidal currents

Tide is the most prominent movement in coastal areas, governing the instantaneous current field (Sun et al., 2000).Therefore, we verified the model results of the tide and tidal currents.The validation data were collected from the in-situ data observed during June 16-23, 2014.The observation stations T1-T3 are shown in Fig.1a, in which the instrument was placed in the middle layer.The wind speed is low during the observation period, and the influence of the wind on the currents is negligible.The comparison between observation data and the model results of tide and tidal currents at the three stations are shown in Fig.2.

The observation data at stations T1-T3 suggest that the semidiurnal tides play a dominant role.It can be seen from Fig.2 that the model results of the tidal level fit well with the observation data.The tides of the model and stations behave almost identically in terms of both high and low waters times and water elevation magnitudes.The average value of Pearson correlation coefficient is 0.96 at the stations, which indicate that the difference between the simulated and measured values was not significant.The average error of the simulated tide elevation is-0.009 m, indicating that the simulation results of the model are accurate.As for the tidal current, the current velocities at stations T1 and T2 are weaker than that at station T3, and the model capture this feature objectively.The current direction curves of the model and observations can also be considered to coincide perfectly.According to the validation results, the hydrodynamic model can provide reliable hydrodynamic fields for the particle-tracking model.

3.2 The distribution of tidal currents

In this study, we focused on the distribution characteristics ofA.chinensisand its mechanism during the 2021 fishing season.Therefore, we analyzed the typical and representative flow field on 15 June 2021.The distribution of the sea surface currents at flood and ebb tides is shown in Fig.3a-b.

At flood tide, in the north of the model area,seawater flows westward and changed southwestward along the coast.In the south of the model area, the seawater flows from the southeast to the northwest alongshore.The current velocities are generally 0.3-0.5 m/s in Lianyungang offshore.The current velocities in the southeastern area of Lianyungang nearshore are 0.4-0.7 m/s; especially, the velocities are larger than 0.7 m/s near the Abandoned Huanghe(Yellow) River estuary.At ebb tide, in the north of the model area, seawater flows alongshore from southwest to northeast and then to the east.In the south of the model area, seawater flows alongshore from northwest to southeast.At this moment, in the northern model area, the current speeds are 0.2-0.4 m/s, the difference in velocity between nearshore and offshore in Qingdao is not significant, and they are smaller than those at high tide.However, in the southern model area, the speeds range from 0.7 to 1.0 m/s, which are greater than those at high tide.

Fig.1 The study area and other specifications

Fig.2 Tide and tidal current validation at stations T1-T3

Fig.3 Distribution of flow fields and Lagrangian residual current

3.3 Lagrangian residual current

The Lagrangian residual current (LRC) is defined as the net displacement of the marker particles divided by the tidal cycle.The calculation formula is as follows:

wherex→(t+T), the displacement vector of the particle at the timet+T, is the solutions to Eq.3,v→ is Lagrangian residual current, andTis the integration time.

LRC can reflect the trend and direction of longterm material transport and is an essential tool for studying material transport.The LRC is a function of the particle release moment and the integration time.Therefore, two typical moments, at high and low water, were taken to calculate the LRC.Here, the time duration for calculating the LRC was 10 days.

Given that particles were released at high water,the LRCs of high water were obtained (Fig.4a).In Haizhou Bay and the east of the Abandoned Huanghe River estuary, the LRC speeds are below 0.01 m/s.In Lianyungang nearshore, the speeds are above 0.025 m/s.Due to the impact of the summer sea surface wind field, the overall distribution of the LRCs is northeastward.Still, near the Lianyungang sea area, the directions of LRCs are southeastward along the coast.A counterclockwise circulation is marked by a red ellipse in Fig.4a.Compared with the LRCs in high water, in the east of the Abandoned Huanghe River estuary, the LRCs speeds in low water are larger than 0.03 m/s.Far from the coast,the directions of the LRCs are northward.Similar to the LRCs of high water, there are three counterclockwise circulations marked by the red ellipses in Fig.4b.These circulation structures might be caused by a combination of tides, summer wind fields, and the prominent topography of the Abandoned Huanghe River delta.

3.4 The FTLE and LCSs

The particles were uniformly released on the sea surface at a resolution of 100 m in the Lianyungang sea area (119.18°E-121.15°E, 33.87°N-35.38°N).The integration time was set from June 15 to July 15, 2021.Based on the particle-tracking trajectories,LCSs were detected by the ridges of the FTLE fields.The forward FTLE and backward FTLE are shown in Fig.5.

Fig.4 Lagrangian residual current field of particle released at high (a) and low (b) water

Fig.5 Distribution of FTLE fields

According to the previous definition of LCSs,they can be considered as the pathways and barriers of material transport.As shown in Fig.5a, there is a“Γ” shape of rLCSs near the Lianyungang sea area.In the Guanhe River estuary, several rLCSs pass through the licensed fishing area forA.chinensis.Far from the coast, there are several southeastnorthwest banded rLCSs, and near the Abandoned Huanghe River estuary, the rLCSs extend to the licensed fishing area forA.chinensis.Considering the rLCSs as the pathways, combined with the LRC distribution, we deduce thatA.chinensisin the licensed fishing area might come from the Guanghe River estuary and the southeast of the Abandoned Huanghe River estuary.

Additional to the transport pathways ofA.chinensis, its aggregation is another interesting issue.In LCSs theory, the aLCSs detected from the backward FTLE indicate the material aggregation.The elliptical-shaped structures of the aLCSs in the licensed fishing area and the banded aLCSs in the east of Haizhou Bay are shown in Fig.5b.These aLCSs can preventA.chinensisfrom leaving the vicinity of the licensed fishing area.Considering the rLCSs and aLCSs,A.chinensiswere transported to the licensed fishing area when they came from the Guanhe River estuary or the Abandoned Huanghe River estuary, they would encounter flow barriers and accumulate.

LCSs could be made to visualize material transport and aggregation.It is also possible to explain the sources, pathways and aggregation areas ofA.chinensisfishing area.These structures agree with the distribution of the fishing production ofA.chinensisin 2020 and 2021 (Fig.1), demonstrating the feasibility of the LCSs theory for analyzing the distribution and transport ofA.chinensis.

3.5 Sources of A.chinensis in licensed fishing area

Aceteschinensisfeeds on phytoplankton and organic detritus.Estuaries and nearshore waters are the central locations for the distribution of these materials.A.chinensisprefer to shallow water with large amounts of freshwater input and shallow muddy bottom waters (Chen et al., 2022).Therefore,judging by the habitat environment, we think thatA.chinensisin the fishing area came from the estuaries and nearshore waters near Lianyungang.

The origins ofA.chinensiswere explored by particle trajectory tracking to determine which estuaryA.chinensiscame from.There are several rivers along the Lianyungang coast, including Xiuzhen River,Xingzhuang River, Linhong River, Shaoxiang River,Guposhanhou River, Guanhe River, Zhongshan River,and the abandoned Huanghe River.Among these rivers, Xiuzhen River, Xingzhuang River, and Linhong River are within Haizhou Bay.The other rivers are distributed outside Haizhou Bay.From the rLCSs in Fig.5, there is no obvious material transport pathway pointing to the licensed fishing area within Haizhou Bay.Thus, we considered only the supplemental effects of the rivers numbered 4-8 on the licensed fishing area without considering the ones numbered 1-3 (Fig.6).Therefore, we divided the nearshore area from Lianyungang to the abandoned Huanghe River estuary into six zones and tracked the destination of particles (Fig.6).It was found that Zone A had no complementary effect on the fishing area.A small part of the particles in Zone B entered the licensed fishing area.In contrast, almost all the particles in Zones C-F were involved in the licensed fishing area.Among previously mentioned rivers, Shaoxiang River is in Zone A, the Guposhanhou River is in Zone B, the Guanhe River is in Zone C, the Zhongshan River is in the middle of Zones D and E, and the Abandoned Huanghe River is in Zone F.This means thatA.chinensisproduction is high in such places(Guanhe River estuary, Zhongshan River estuary,and abandoned Huanghe River estuary), as shown in Fig.1, coming from Guanhe River estuary and its southeastern nearshore areas.This result is consistent with the statements of experienced fishers.

In Fig.5, we also find an interesting phenomenon.We tracked the particles near the estuaries and shore in the forward time and the particle distribution results (Fig.6b) match well with the backward FTLE field (Fig.5b).This similarity suggests that particles do tend to accumulate around the aLCSs, which implies that aLCSs act as hidden barriers in the ocean, resulting in the transport of matter along its edges, and eventually behave as the similarity between particle distribution and backward FTLE field.Tracing the particle sources showed that the results are consistent with those of the LCSs,indicating the reliability of the transport characteristics derived from the LCSs.

Fig.6 Particles destination tracking in Lianyungang nearshore

4 DISCUSSION

Many Lagrangian descriptions are sensitive to the initial release moment of the particles and the integration time.In this section, the robustness of LCSs is discussed.

4.1 Sensitivity of the FTLE to the initial time

Due to the sensitivity of Lagrangian fluids to initial conditions, their motion is inherently unstable, but LCS is a robust skeleton of material surfaces (Haller et al., 2015).Here, we compared the FTLE fields obtained at different initial particle release moments to explore the robustness of the LCSs.Four typical moments (high water, low water,ebb tide, and flood tide) were selected as the initial time to calculate the two tidal periods (Fig.7).The FTLE fields obtained at high and low water and flood and ebb tide moments are very similar near the fishing area in the water environment affected by tidal currents.Therefore, for the FTLE integrating multiple cycles, the initial moment has little effect on the results.The FTLE and LCSs are less dependent on the initial time than the other Lagrangian methods.This conclusion is particularly critical for using LCSs to predict the material distribution in the real ocean because, most of the time, marine sampling is not synchronous, and we show that LCSs can approximately ignore the differences in sampling moments.

4.2 Sensitivity of the FTLE to integration time

Fig.7 The FTLE fields for particles released in typical moments

The forward FTLE fields were obtained with six integration times: 1, 5, 10, 15, 20, and 25 days to investigate the sensitivity of the LCSs to the integration time.From Fig.8a, it can be seen that there is a prototype of the ridges, and as the integration time increased, the whole structure become more and more apparent.By day 15, the entire structure is distinct.The FTLE value decreases with the integration time,but the LCSs remain robust from 15 to 25 days.

Therefore, it is suggested that 15 days of integration time for the FTLE was optimal in the sea area with semi-diurnal tides.In Bohai Bay and Laizhou Bay with semi-diurnal tides, the authors also indicated that the optimal integration time for the FTLE is 15 days (Feng et al., 2020; Zhang et al., 2021b).

4.3 Effect of the wind fields on FTLE

Fig.8 Forward FTLE fields with different integration duration

As we know, currents are crucial in driving particle transport and material exchanges between the adjacent coastal ocean (Fitzenreiter et al., 2022),the sea surface wind field has an important effect on the current and residual current fields, does the wind field also have a substantial effect on FTLE? In this section, four wind fields were used to test the effect(Fig.9), i.e., no wind (0 m/s) and southeast wind(1.12, 2.06, and 4.03 m/s).The structures of the forward FTLE fields are similar to each other at different wind velocities.There is almost no significant difference near the estuary of the Guanhe River, where there is a clear vortex-shaped structure.The results indicate that tide is the main dynamic forcing factor in the sea area.It was noted that the vortices were slightly different in shape when the wind velocity is 4.03 m/s.The comparison shows that the material line would almost appear in the same position at a low wind velocity or a statistical wind field.At high wind speeds, the FTLE in this sea exhibits a different shape, which affects the material transport pathways here a little, but the overall position of LCSs is hardly changed.The results indicate that the LCS is a persistent structure hidden in the fluid (Maquet et al., 2007; Mangiarotti et al., 2019), exhibiting the advantage of LCSs in studying the transport of materials.

4.4 Effect of topography on the FTLE

The effect of topography on hydrodynamics has been well studied from both observation and simulations.Form drag will occur in the ocean when currents flow over rough topography (McCabe et al., 2006), and the horizontal flow separation will produce an eddy in the lee of the headland or prominent topography.For example, the topography plays a more important role in the gyre circulation in the southwestern shore of Lake Michigan (Mao and Xia, 2020b).

Fig.9 The distribution of the FTLE fields

In the nearshore area, due to the effect of zigzag shoreline, there are some residual vortices,especially the counterclockwise circulation induced by the abandoned Huanghe River delta.The FTLE field also shows more significant ridges in the area where residual changed significantly, which can be explained by the influence of the southeast wind in summer from a dynamic perspective.When a flow from the abandoned Huanghe River estuary was transported to the northeast, flow separation would occur due to the blocking effect of the prominent topography.The coastal current flows southeastward along coast, resulting in a counterclockwise circulation nearshore in Lianyungang.This special circulation structure should be attributed to the Γ-shaped rLCSs and the elliptical aLCSs.

In the licensed fishing area forA.chinensis, there is an underwater shoal and a narrow channel between shoal and coast.The LRC field (Fig.4), the FTLE field(Fig.5), the tidal currents, residual currents, and FTLE fields are all structurally associated with this micro-topography (Fig.1d).Accordingly, the two special topographic features cause some material transport pathways and aggregation areas, which are the main driving mechanisms ofA.chinensistransport.

4.5 Effect of diurnal vertical migration on the FTLE

Aceteschinensisis poor in swimming and controlled by tide horizontally, while they have an autonomous diurnal migration ability in vertical direction.Therefore, the diurnal migration behavior is considered in the particle-tracking model.The particles migrate from the surface to the bottom as the sun rises and from the bottom to the surface as the sun sets.The duration of the model was 1 month.The FTLE field drawn from the particletracking results is shown in Fig.10.

It can be seen in Fig.10 that in the forward FTLE field, there are several rLCSs extending from Guanhe River estuary to the licensed fishing area, forming a ring-shaped rLCS in the licensed fishing area, which can indicate thatA.chinensisin the estuary can be transported to the licensed fishing area.In the backward FTLE field, the aLCSs in the licensed fishing area show a semi-enclosed elliptical shape,which means thatA.chinensiscan aggregate in the licensed fishing area.A comparison with the FTLE field in Fig.5 shows that the LCSs differ in shape but are essentially the same in position.Therefore,the diurnal vertical migration activity ofA.chinensisdoes not affect their retention in the licensed fishing area significantly.

5 CONCLUSION

In this paper, the flow field was simulated using the FVCOM model.Then, Lagrangian particle tracking, LRC, and LCSs were combined to study the distribution ofA.chinensisin the Lianyungang nearshore area.

Fig.10 Distribution of FTLE fields with diurnal migration

The results of the LRCs in the study area show that the overall residual current direction is northward due to the effect of the summer wind field.The results indicate that the material in the study area is transported northward generally in summer.The counterclockwise circulation induced by the shoreline and the local micro-topography near the licensed fishing area forA.chinensisexplain whyA.chinensiscame from the Guanhe River estuary and the Abandoned Huanghe River estuary.The results also explain why they easily accumulated in the licensed fishing area.The effect of vertical migration behavior on the aggregation ofA.chinensisis not significant.

The LCSs could maintain relatively robust structures in the sea area, and are influenced by the circulation induced by the shoreline and the local micro-topography.The LCSs vividly and reasonably describe the transportation characteristics ofA.chinensisin Lianyungang sea area.The rLCSs near the estuary of Guanhe and Zhongshan River point to the licensed fishing area imply thatA.chinensisfrom this area could be transported to the licensed fishing area.The aLCSs have an elliptical structure near the licensed fishing area that allowed the retention ofA.chinensishere.With particle destination tracing along coast of Lianyungang, the estuaries of the Guposhanhou River, Guanhe River,Zhongshan River, and Abandoned Huanghe River could provide recruitment forA.chinensisresources in the licensed fishing area.Specifically, the Guanhe River estuary and the Zhongshan River estuary contributed more.

6 DATA AVAILABILITY STATEMENT

The raw data supporting the conclusions of this article are available by contacting the corresponding author, without undue reservation, to any qualified researcher upon request.

7 ACKNOWLEDGMENT

The authors thank the anonymous reviewers.