Tianli Li, Tonglin Liu, Ruhan Liu, Lijing Chen,＊

aNational Demonstration Center for Experimental Fisheries Science Education, Shanghai Ocean University, Shanghai, 201306, China

bKey Laboratory of Freshwater Aquatic Genetic Resources, Ministry of Agriculture, Shanghai Ocean University, Shanghai, 201306, China

cShanghai Collaborative Innovation for Aquatic Animal Genetics and Breeding, Shanghai Ocean University, Shanghai, 201306, China

Keywords:

ABSTRACT Zooplanktons are important food organisms of fishes, and their spatial and temporal distribution pattern has an important influence on the growth and reproduction of fishes in local habitats. The Yangtze River Estuary of China is rich in fishery resources, but the knowledge on the distribution pattern of zooplankton in this area is still very limited. The diel vertical migration (DVM) patterns and seasonal variations of the crustacean community,and their relationship with environmental factors were investigated in the south branch of the Yangtze Estuary,from June 2017 to May 2018. The results showed that the dispersion patterns of dominant crustacean taxa in the Yangtze Estuary were mostly aggregated. Migration amplitudes of crustacean were greater in spring and summer than in autumn and winter, and different migratory patterns occurred in different seasons. Sinocalanus dorrii,Schmackeria inopinus, Mesmackeria leuckarti, Bosmina longirostris, and B. fatalis, unlike S. forbesi, showed a significant DVM. The DVM patterns of crustaceans was possibly influenced by light dentisy, temperature, and total dissolved solids in the water column, and the fish predation effect; whereas, the nutrient concentration and pH were not supposed to be key factors.

1.Introduction

Zooplankton is the natural food of fish; its community plays a vital role in regulating the abundance of phytoplankton, maintaining the stability of fishery resources as well as environmental monitoring (Finney, Gregory-eaves, Douglas, & Smol, 2002). The spatial and temporal distribution patterns of the zooplankton community have strongly influenced the species diversity and community structure of fishes(Damotharan et al., 2010). Zooplankton in estuarine areas has important habitat adaptation mechanisms and they have important potential roles in maintaining biodiversity, ecosystem balance, and fish productivity in this freshwater-seawater ecotone (Schmitt, Devreker, Dur, & Soussi,2011; Ueda, Kuwatani, & Suzuki, 2010).

Vertical migration, most is diel vertical migration (DVM), is one of the important adaptation and survival mechanisms of zooplankton in deep waters (eg. oceans, estuaries, and lakes). It profoundly affects nutrient transport and nutrient relationships of marine and estuary ecosystems (Hays, 2003; Ringelberg, 2010, chap. 10). Passon and Carlson (2012) found that the DVM of zooplankton could cause carbon transport from the surface water layer to deeper layers, which effectively regulate the carbon circulation of water and maintain the stability of the mid-ocean ecosystem. Planktonic cladoceran and copepod are undoubtedly one of the most abundant estuarine zooplankton group and a major player of DVM. Many factors influence the DVM of crustaceans,among them, light intensity is proximate factor regulating migrating behavior (Cohen & Forward, 2009). Furthermore, Devreker, Souissi,Molinero, and Nkubito (2008) have observed tidally-induced vertical migration (TVM) of crustacean existing in the estuaries. While long-term sampling investigations, conducted in months or years, are the most common study modes for estuarine zooplankton, fine-scale sampling studies (at hourly intervals or less) are still limited for studying the DVM of zooplankton in estuaries (Goncalves et al., 2012).

The Yangtze River is the third largest river worldwide, which transports large amounts of sediment and particulate organic carbon to the estuary and its adjacent coast, resulting in high concentrations of nutrient and suspended solid (Fan & Song, 2014; Zhou, Yu, & Zhou,2009). As an important water source in Shanghai, the Yangtze Estuary has many key functions, such as navigation, aquaculture, fisheries, and supplying drinking water (Chen, 2018; Zhu & Ruan, 2004). Unfortunately, according to the latest study, the fishery resources in this area significantly declined, that dominant taxa have been shifted from carnivorous to planktonic fish during the 2016 and 2017 periods, owing to the anthropogenic effects and habitat degradation (Xu, Wang, Zhao,Yang, & Zhuang, 2019). Despite zooplankton distribution patterns are closely related to the abundance of fishery resources, however, researches on zooplankton ecology in Yangtze River Estuary are still extremely limited. Wang, Chen, Tang, and Han (2016) systematically expounded the relationship between zooplankton community characteristics and water environmental factors in the Yangtze Estuary and its adjacent waters after the Three Gorges Dam impoundment. Chen, Dai,Xu, Li, and Mei (2015) also studied the impact of human activities on the zooplankton in this area. Consequently, more extensive and in-depth researches on vertical migration and spatial distribution are essentially needed in this important ecotone.

In present study, given the scenario of dominant runoff of the south branch of the Yangtze Estuary, substantial suspended solid and intense predation pressure of estuaries, we hypothesized that planktonic crustaceans residing in various water strata adopt different migratory behavior against the variations of estuarine environmental factors and to maintain their optimal niche in the Yangtze Estuary. Our aim is to test the hypothesis by analyzing the fine-scale (three-hourly) temporal variability of planktonic crustaceans density by water depth, concerning diel light, water temperature, total dissolved solids (TDS) from June 2017 to May 2018; and the effects of fish predation were also discussed.

2.Materials and methods

2.1.Sampling area

The hydrological basin of the Yangtze River, with an area of 18 ×10km, is the largest river flowing into the East China Sea, which provides an average annual runoff of 9.248 ×10m. Its terminal part is divided into two branches (south and north), separated by the Chongming Island, aggregating again near the mouth. Here, the Yangtze Estuary, the influence of tidal mixing both river flow and neritic waters and interactions among various drainage patterns dominate overwhelmingly, overall resulting in the complex and dynamic environment(Fan & Song, 2014). According to its geographical location and tide characteristics, two parallel sampling sites (site 1:N3135.500,E12120.004; site 2:N3133.500, E12120.000) (Fig. 1) were set in the south branch of the Yangtze Estuary from June 2017 to May 2018, and samples were collected during the spring tide period of each month(Appendix Table A1).

Fig. 1.Location of sampling sites in the Yangtze estuary.

2.2.Field sampling and laboratory analysis

Zooplankton sampling was carried out in accordance with Marine Monitoring Specifications of China (Zhang, Hong, Xu, Chen, & Jin,2007, chap. 7). Water samples from surface (0.5-m from the surface),middle (half the depth), and bottom layers (0.5-m from the sediment bottom) were monthly collected at two sites (St.1 and St.2) at 3-h intervals for 24 h monthly. Three-hourly samples were taken for 24 h,through four flood tides (10:00, 16:00, 22:00, and 4:00) and four low tides (7:00, 13:00, 19:00, and 1:00). Planktonic crustacean’s samples were collected at three depth layers according to the water depth of each station. Ten liters’ surface water was collected for quantitative samples with a 5-L Plexiglass water sampler, quantitative samples then passed through 13# zooplankton nets (mesh diameter:112 μm) that filter and concentrate samples into 50-mL plastic bottles. Samples were fixed and preserved with 5% formalin solution and transported to the laboratory for analysis. The middle layer and bottom layer sampling methods were similar to those used for surface sampling. At each sampling interval,hydrological parameters were measured in-situ with relevant instruments and a 1-L sub-sample was collected for the analysis of chemical parameters.

The samples with high numbers of organisms were sub-sampled by aspirating supernatants, and then 1-mL subsample was examined and identified to the lowest possible taxonomic level under an Olympus CX21 100 ×microscope. In all samples individuals were quantified, with densities calculated and expressed as ind./L. Taxonomic identification of zooplankton was performed mainly following the relevant taxonomic pieces of literature (Chen, Lin, Xu, & Han, 2011; Cheng, Wang, Li, & Sun,2014; Sun, Li, Cheng, Jin, & Yang, 2015; Xiang, Yu, & Chen, 2015;Zhang, Zhao, & Tao, 2010).

Hydrological parameters encompassing water temperature (Wt,expressed inC), pH, dissolved oxygen (DO, expressed in mg/L), and total dissolved solids (TDS, expressed in mg/L) were measured at three depth layers with an HQ40d series portable instrument (Hach) at 3 h intervals for 24 h monthly. Depth (D, expressed in m) and transparency(SD, expressed in cm) were measured with sounder and Secchi disk,respectively. Chemical parameters encompassing the permanganate index (COD, expressed in mg/L), total nitrogen (TN, expressed in mg/L) and total phosphorus (TP, expressed in mg/L) were measured without vertical stratification at 3 h intervals for 24 h monthly with a Hach DR3900 spectrophotometer in the laboratory.

2.3.Data calculation and statistical analysis

(1) Weighted mean depth (WMD): reflects mean depth distribution of the collected crustacean zooplankton (Thackeray, George, Jones,& Winfield, 2006), this index is described by the following formula:

WMD

=

Σ(nidi)

/

Σni

;(2) Morisita index (MI): an established single-sample index of dispersion, which reflects the patterns of dispersion of the collected organisms in various water layers (Elliott, 1977). This index was calculated as follows:

MI

=

n

×

(

∑

X

−∑

X)

/

[(

X)

−∑

X]

;

(3) The Berger–Parker index (D): reflects the dominance of the collected organisms (Lin, Duan, Hu, & Han, 2003), which was calculated as follows:

where

ni

is the density (individuals per m) of sample

i

,

di

is its depth considered as the midpoint of the layer,

x

is the density of collected organisms in a given sampling unit,

n

is the number of sampling units,

N

is the total density of the sample, and

fi

is the frequency of the sample. A species is considered dominant if

D

is greater than 0.02. MI is equal to 1 for a random distribution, less than this for a regular distribution and greater than 1 for an aggregated distribution.

Tables and figures were performed using Origin 2017 software(Origin Lab, USA). SPSS 18.0 and CANOCO 4.5 software were carried for statistical analysis. The two-way repeated measure ANOVA was performed with water layers as main factor and sampling period (spring,summer, autumn and winter) and diel time (24 h) as sub-factors to determine the variations in planktonic crustaceans and water variables among factors above-mentioned. All variables were analyzed through two-way repeated measure ANOVA except TN, TP, COD, salinity and SD, which were determined the variations among seasons using one-way repeated measure ANOVA. Before ANOVA, parametric test was performed and all conformed to the requirements of normality and homogeneity of variance. If the ANOVA was significant, Bonferroni test was further applied for multiple comparisons of the means. Pearson correlation analysis and Redundancy Analysis (RDA) were used to calculate the association between species with the environmental factors (Ter Braak, & Smilauer, 2002). Zooplankton density and the environmental data were subject to log(x+1) conversion processing (Muylaert,Sabbe, & Vyverman, 2000) and species with less than 30% occurrence were not included in the RDA to improve clarity of the ordination triplots.

3.Results

3.1.Environmental variables

The ANOVA results indicated that the interaction effects of water layers and diel time was not significant (Table 1,

P

>0.05). But, Wt, DO and pH in different water depths and seasons showed significant interaction effects (Fig. 2, F =8.802, 6.268 and 4.462, respectively;

P

<0.05). The mean value of Wt fluctuated considerably, showing at least three-fold higher during summertime (27.6 ±0.8C) than during the winter period (9.2 ±0.9C). The higher value of Wt was recorded at the surface as compared to the middle during spring and winter period (F =7.578, 11.175;

P

=0.028, 0.012, respectively). Considering DO, the surface value was higher significantly than other depths in the summer(F =12.609,

P

=0.007), but no significant vertical difference was observed during other sampling periods. Concerning pH, the mean value varied between 8.20 ±0.21, in summer (surface), and 7.37 ±0.14, in winter (bottom). Mean pH values were generally higher at the bottom as compared to the middle during spring (F =2.209,

P

=0.024), whereas higher at the surface than deeper layers (F =9.298,

P

=0.005). TDS was,however, almost vertically homogeneous during sampling periods.Water depth varied with the tide fluctuation, ranged from 11.1-20.0 m (average 15.6 ±2.1 m) during our study period. The variation of SD was dramatic, being very high during the summer period but very low during the winter period and significant seasonal differences were identified (F =5.949,

P

<0.001). There was a significant seasonal variation in concentration of TP, with higher mean values recorded in autumn than during the spring and summer period (F =10.791,

P

=0.005). The TN concentration peaked in spring, significantly higher than in other seasons (F =8.839,

P

<0.001). Salinity fluctuated slightly between 0.10 and 0.16 during four seasons. There were no significant differences in salinity among sampling periods except for the lowest value recorded during the summertime (F =2.489,

P

=0.158).

3.2.Dominant species of planktonic crustaceans

Fig. 2.Wt, DO, pH and TDS in different vertical layers and seasons. ＊ denotes significantly difference (ANOVA, P＜0.05).

Fig. 3.Diel variations of Morisita’s index for the dominant zooplankton species.

A total of 51 species of planktonic crustaceans belonged to 26 genera and 16 families were observed, including 25 cladocerans taxa and 26 copepods taxa. The 10 dominant species of cladocerans include

Bosmina

longirostris

,

B

.

fatalis

,

Daphnia galeata

,

Moina affinis

,

D

.

carinata

,

D

.

cucullata

,

M

.

macrocopa

,

Diaphanosoma leuchtenbergianum

,

B

.

coregoni

,and

Chydorus sphaericus

(D =0.54, 0.03, 0.11, 0.02, 0.07, 0.08, 0.09,0.08, 0.05, and 0.05, respectively). Seven dominant copepods species were

Sinocalanus dorrii

,

Schmackeria inopinus

,

S

.

forbesi

,

Mesocyclops leuckarti

,

Eucyclops serrulatus

,

E

.

denticalatus

, and

Cyclops vicinus

(D =0.63, 0.14, 0.23, 0.05, 0.13, 0.04, and 0.05, respectively).

3.3.Patterns of dispersion of dominant planktonic crustaceans

All specie showed an aggregated pattern of dispersion on each sampling time, except that the MI of

S. dorrii

at 16:00 and

S. inopibus

at 19: 00 was <1 (Fig. 3). The dominant planktonic crustaceans in the Yangtze Estuary are concentrated and distributed in different water strata at different times. However, the temporal variations of MI indicated considerable variations in the degree of aggregation in each species over the sampling period. The most prominent aggregation occurred in

B. longirostris

at 13:00 and 19:00 (MI =7.91, 7.25, respectively). The degree of aggregation during flood tide was higher than that during ebb tide. Pearson correlation analysis showed that the diurnal variations of MI were positively correlated with vertical difference of Wt and pH but were contrary to TDS, which indicated that vertical distribution of Wt,pH and TDS affected the aggregation and DVM of species (Table 2).

3.4.DVM characteristics and seasonal variations

The WMDs of total zooplankton ranged from 3.6 m to 12.2 m during the sampling period. The vertical migration amplitude of zooplankton was 4.3 m in spring, 3.8 m in summer, and 2.6 m in autumn and winter.We found a significant difference in the diel variation of WMD during the sampling period (F =6.901,

P

=0.034), which showed that different migratory patterns occurred in different seasons (Fig. 4).In spring, the significant differences in WMDs were detected over diel time; and the clear DVM patterns were observed (F =10.353,

P

=0.035), that the crustaceans descended to the deeper layer in the morning and ascended to the upper layer in the afternoon and at night.The WMD was deeper at 13:00 p.m. and shallower in 7:00 and 10:00 a.m., exhibiting perceptible upward migration with WMD ranging from 4.6 m to 8.9 m. In contrast, the migratory pattern was reversed during nighttime and afternoon (13:00 to 22:00 p.m.) by an ascent of organisms.In summer, the significant differences in WMDs were observedduring nighttime (F =5.203,

P

=0.021), which showed that the organisms ascended to the upper layer in the twilight and middle of the night, whereas descended to the deeper layer at night. A perceptible migratory behavior occurred at 19:00 p.m. by the ascent of the twilight period (16:00–19:00 p.m.) and the decent during nighttime(19:00–22:00 p.m.). Then, the organisms reminded in the shallower layer at 4:00 a.m. with significant upward migration. The above phenomenon was not significantly perceptible during the daytime.

Table 1 Results of ANOVAs and post-hoc on water phy-chemical parameters with respect to sampling period, diel time, water layer and their interactions.

Table 2 Pearson correlation coefficient between Morisita’s index and environmental factors between surface and bottom.

In autumn and winter, unlike former seasons, it was not possible to detect the significant DVM, which showed that these organisms were distributed around the mid-depth both at night and day except for the ascent in the middle of the night (1:00–4:00 a.m.) in autumn and the descent in the morning (7:00–10:00 a.m.) in winter.

3.5.DVM types

The DVM patterns of zooplankton are directly related to the migratory behaviors of the dominant species. Therefore, it is necessary to investigate the DVM of the dominant planktonic crustaceans in the Yangtze Estuary. The effects of the water strata and its interaction with sampling period or diel time on planktonic crustaceans are presented in Table 3, which indicated that the interaction of water strata and seasons was not significant (

P

>0.05). The significant interaction effect, hence,between diel time and water depth on several dominant planktonic crustaceans is showed in Fig. 5.A detailed analysis of the

S. dorrii

abundance emphasized the variations among water strata, with significantly higher values at 7:00 a.m.(F =9.320,

P

=0.033) and 1:00 a.m. (F =6.910,

P

=0.015) at surface and lower values at 13:00 p.m. (F =7.971,

P

=0.034) at surface, suggesting the occurrence of normal vertical migration, that is, descending to the deeper strata during daytime and ascending to the upper strata at night. Similarly, for

S. inopibus,

the organisms generally remained close to the bottom at 13:00 p.m. (F =6.961,

P

=0.026) and ascended into the water column during nighttime. In contrast, an inverse diel vertical distribution pattern was observed for

B. fatalis,

which showed their upper peaks of abundance during daylight period while appeared at higher bottom densities during nighttime (

P

<0.05).In term of

M. leuckarti

, at bottom, species peaked in abundance in the afternoon (13:00 p.m.: F =9.224,

P

=0.041) and were few in numbers at 7:00 a.m. (F =4.537,

P

=0.020) and 16:00 p.m. (F =8.684,

P

=0.037), exhibiting a perceptible upward and downward migration occurred in the morning and afternoon, respectively. Concerning

B. longirostris

, the organisms mainly resided at the bottom at 1:00 a.m.(F =11.131,

P

=0.043) and then migrated upwardly at 4:00 a.m. (F =6.559,

P

=0.021). While vertical heterogeneity of zooplankton almost did not occur during the dark period for

M. leuckarti

and daylight period for

B. longirostris

.Regarding

S. forbisi

, unlike former species migrating along the water column, they were frequently sampled throughout the water column without vertical stratification and with the absence of vertical migration(

P

<0.05).

Fig. 4.Seasonal variation in the WMD of total planktonic crustacean in the Yangtze estuary. Vertical bars denote standard deviation. The WMDs having the same letter are not significantly different from each other (P＞0.05, Bonferroni). Shaded area indicates nighttime.

3.6.Dominant species in relation to environmental parameters

Based on zooplankton relative abundance (≥0.02) and the frequency of occurrence (≥30%), 6 types of zooplankton were selected for DCA analysis of surface, middle and bottom layers, which showed that the maximum lengths of the gradient were all less than 3 (0.673, 0.625 and 0.614, respectively). Hence, the linear model redundancy analysis(RDA) was selected to examine the relationship between zooplankton and environmental factors. The first two ordination axes explained 75.7% of the variance in species and 89.9% species-environment relation at surface, 88.9% and 98.4% at middle and 60.0% and 88.1% of these relationships at the bottom, respectively (Appendix Table A. 2).Environmental factors that have significant impact on zooplanktons have been selected with Monte Carlo-permutation tests during RDA analysis. At the surface layer, the dominant species were mostly influenced by Wt, TDS, DO, and pH. The first RDA axis was positively correlated with Wt, whereas it was negatively correlated with

S. dorrii

,

S. inopibus

and

S. forbesi

, which exhibited a significant negativecorrelation between Wt and above species. The second RDA axis was positively correlated with pH and

B. longirostris

, whereas negatively correlated with TDS and DO. In the case of the middle layer, the dominant species were mostly influenced by Wt, TDS, salinity and TN.The first axis was primarily in relation to most species encompassing

S. dorrii

,

S. inopibus

,

S. forbesi

and

B. fatalis

in the positive direction and highly associated with TDS in the negative direction, which exhibited significant negative correlation between TDS and above species. The second RDA axis was highly associated with Wt,

M. leuckarti

and

B. longirostris

in a positive direction, whereas highly associated with salinity and TN in the negative direction. In the case of the bottom layer,the dominant species were mostly influenced by Wt, TDS, and pH. The first axis was in relation to all species, Wt and pH in the positive direction, and TDS in the negative direction, which exhibited that all species were negatively correlated with TDS but positively correlated Wt and pH (Fig. 6).

Table 3 Results of ANOVAs and Bonferroni tests on dominant species of planktonic crustaceans with respect to sampling period, diel time, water layer and their interactions.

Fig. 5.The abundance dynamics of six dominant species at the surface, middle and bottom water layers during a whole diel period. Shaded area indicates nighttime.

4.Discussion

4.1.Implication of DVM pattern

The planktonic crustaceans existing in Yangtze estuary performed different vertical distribution and DVM patterns, which varied between species, interpreting species-specific migratory patterns (Fig. 5). The typical behavior of normal DVM, nocturnal upward and diurnal downward migration, occurred in species

S. dorrii

and

S. inopibus.

This phenomenon was somewhat in concordance with the old stages copepods living in Japanese coastal waters (Ueda, 1987), which exhibited nocturnal migration into the upper layer from the deeper layer where they resided during daylight period. However, nocturnal upward migration was not perceptible in the Haughton River estuary (McKinnon& Klumpp, 1998).

B. fatalis

, in our study, showed higher densities at surface and bottom during daylight and dark period, respectively, with migration expressed as the reverse DVM. The reverse DVM of cladocerans have been observed by Marcos, Michelli, and Paulo (2015) and Ghidini and Santos-Silva (2011) in Xambrˆe Lagoon and Amazonian lakes, respectively. However, the reverse DVM of cladocerans rarely occurred in estuaries excepting for the observation of marine copepods nauplii in the Matang estuary (Chew, Chong, Ooi, & Sasekumar, 2015),which generally remained close to the bottom at night and ascended into the water column in the day. One possible explanation to the reverse DVM is those small-sized cladocerans, such as

Bosmina

, and copepod nauplii are less sensitive to visual predators or to evade non-visual,nocturnal predators (Ohman, 1990). Moreover, the adults of copepod species in the Matang estuary preferred shallow layers during nighttime and dispersed in the day, occupying both the surface and the bottom(Chew et al., 2015). Similar migratory behavior was not recorded for

M. leuckarti

, but for

B. longirostris

which were generally sampled throughout the water column during the day, but they were found dominated in the upper layer at 4:00 a.m. The

M. leuckarti

, however,descended to deeper layers in the morning and ascended to upper layers in the afternoon, performing homogeneous vertical distribution at night.

Fig. 6.RDA ordination triplots of environmental parameters and dominant crustaceans in the Yangtze estuary (S.d =S. dorrii, S.i =S. inopibus, S.f =S. forbesi, M.l =M. leuckarti, B.l =B. longirostris, B.f =B. fatalis).

4.2.The response of DVM pattern to diel light

In the presence of higher light intensity during daytime, zooplankton was consistently more abundant in deeper layer in the morning (7: 00 a.m.-13: 00 p.m.) except

B. fatalis

(Fig. 5). These results were somewhat in concordance with copepods residing at the Mondego estuary, which performed a negatively phototactic reaction resulting in a downward migration with the increases of light intensity especially in summer(Gonçalves et al., 2012). All before-mentioned indicated that intense migratory behavior of crustaceans was associated with high light intensity in the day, that is, all species showed obvious and uniform photopathy (Cohen & Forward, 2009). During daylight period, the higher light intensity not only results in the descent of zooplankton performing photopathy, causes simultaneously increases in the primary productivity of phytoplankton, which in turn, facilitates the grazing of zooplankton and phytophagous fish in the Yangtze Estuary. The WMD of planktonic crustaceans in the Yangtze Estuary varied seasonally, with greater migratory amplitudes in the day (7: 00 a.m.-19: 00 p.m.) than at night (22: 00 p.m.-4: 00 a.m.) during spring and winter periods (Fig. 4).Meanwhile, the degree of aggregation (MI values) in each species was more prominent in the day than at night (Table 2). The above-mentioned response to light stimuli agrees with Xue, who stated that the movement gradient of cladocera existing in the Xiangxi river estuary was influenced by light intensity for some reason (Xue, Han, Cai, & Liu, 2006).Laurent (2011) suggests that zooplankton tend to move in the direction of the original or higher light intensity when the light intensity changes,which was following the superficial decreases in the abundance of most dominant crustaceans at midnight (22: 00 p.m. to 1:00 a.m.) (Fig. 5).Because present light intensity was below the sensory limit of crustaceans, which became inactive and descended to deeper water layers. In our study,

S. dorrii

and

S. inopibus

, which exhibited a species-specific behavioral response to diel light with spending the nighttime primarily in the upper water layer and migrating down at dawn to spend the daytime in deeper layers. The diurnal light may lead vertical stratification of Wt with higher temperature at the surface, which was positively correlated with the degree of aggregation of species (r=0.812,

P

=0.014, N =8) (Table 2), and species

S. dorrii

and

S. inopibus

living in the surface layer were negatively correlated with Wt (Fig. 6). The DVM pattern of them, hence, was similarly connected with the avoidance of damage from higher Wt. Similar migratory behavior was observed by Simoncelli, Thackeray, and Wain (2019), who mentioned that

Daphnia

always sank during the observational period with higher temperatures.Orcutt and Porter (1983) proposed that a decrease in water temperature causes an increase in the period of a reproductive cycle. Therefore,migrating zooplankton residing in the deeper and cooler layers during the daytime may have a longer period of the reproductive cycle, and thus facilitates the maintenance of the zooplankton community in the Yangtze Estuary.

4.3.The response of DVM pattern to TDS

It was worth noting, in our study, that all dominant species were negatively correlated with TDS in the whole water column (Fig. 6).Similar findings have been reported by Shao, Liu, Zhang, Sun, and Lin(2017), who mentioned the negative correlation between zooplankton abundance in the Yangtze Estuary and concentration of suspending matter at both surface and bottom during spring and autumn period.One possible explanation is that dominant runoff of the south branch of the Yangtze Estuary results in increasing concentration of TDS, which adverse to the increases in the primary conductivity of phytoplankton,resulting in the decreases in the abundance of zooplankton (Hoover,Miller, Landry, DeCarlo, & Mackenzie, 2008; Yang et al., 2014). Moreover, the negative effect of suspending sediment on the migratory behavior of planktonic copepods has been confirmed by a turbid estuarine study (Castel & Veiga, 1990), which stated that copepods residing in the Gironde estuary had a higher probability of being washed out of estuary by benthic sediment when the organisms reached the bottom layer. These observations are somewhat in accordance with the negative correlation between vertical stratification of TDS and the degree of population aggregation (r=−0.740,

P

=0.036, N =8) (Table 2). In this case, zooplankton may also increase simultaneously the risk of being washed out of the estuary by benthic sediment as they descend to the bottom layer to avoid predators. Therefore, there are both benefits and risks for migrating zooplanktons and it is essential for them to weigh between advantages and disadvantages to maintain their optimal niche in the Yangtze Estuary.

4.4.The response of DVM pattern to fish predation

One of the most compelling evidence for the ultimate role of predation in driving DVM is the survey on the Yangtze Estuary (Jiang,Zhang, & Zhong, 2015) that

Coilia nasus

had a higher feeding coefficient to

Sinocalanus

and

Mesmackeria

(

ε

=0.608 and 0.477, respectively).And among the feeding composition of

Coilia nasus

in several stages, the feeding density of the early and older stage for

M. leuckarti

and

S. dorrii

accounted for 69 and 75% of total feeding density, respectively. It can be speculated, therefore, that

S. dorrii

performed normal migration patterns to escape their fish predators and increase their reproductive and growth fitness, is not only influenced by light and Wt. In particular, between 2017 and 2018 in the Yangtze Estuary, fish larvae and juveniles that occurred in very large numbers at the surface during daytime (Li et al.,2020) fed on zooplankton during the day, which may induce the higher densities of larger prey organisms encompassing

M. leuckarti

and

S. dorrii

to reside at the bottom where there is lower light intensity and thus, less risk of predation by visual fish predators. However, there is little grazing food and increasing risk of being washed out of the estuary by benthic sediment in the deeper layer. Therefore, the trade-off among above factors is necessary for migrating zooplanktons, which, in turn, perform various migration. Compared with larger zooplanktons, small-sized

Bosmina

will lower risk of fish predation, resulting in high reproductive rate and abundance in the estuary and reverse DVM. This may be one of reasons for miniaturization of zooplanktons. Meanwhile, the reserve DVM also may affect spatial distribution of predator performing normal DVM and fishery resources in the Yangtze Estuary, that is, fish and other predators concentrate at depth to feed on zooplankton during nighttime, whereas they migrate to the upper water in the day with response to the presence of food (Olivier, Stelly, & Laurent, 2018).Moreover, there has been a decrease in the total density of planktonic crustaceans comparing to 2010 in the south branch of Yangtze Estuary(Yin & Xu, 2016), which could result in the decrease in the fishery resource.

5.Conclusion

Our study suggested that the vertical aggregates pattern of dominant planktonic crustaceans in the Yangtze Estuary showed a negative correlation with the TDS and a positive correlation with Wt. The DVM patterns of crustaceans were possibly influenced by vertical light density, Wt, and TDS in the water column, as well as the fish predation effect; whereas, the nutrient, pH, DO, and CODwere not supposed to be key factors. Therefore, planktonic crustaceans residing in the Yangtze Estuary adopted various migratory patterns encompassing normal migration (

S. dorrii

and

S. inopibus

), reverse migration (

B. fatalis

) and homogeneous vertical distribution (

S. forbisi

) against the variations of diel light, TDS and fish predation, which proves well the hypothesis.Although migration in response to light intensity change is the proximate aspect underlying DVM, other factors encompassing predatorspecific chemicals and vertical gradients of temperature may modify this response. The effect of light stimuli and its interaction with predation and thermocline on the DVM of estuarine zooplankton deserves a valuable topic to be further studied. Our results will provide basic data support for the protection of fishery resources in Yangtze Estuary.

CRediT authorship contribution statement

Tianli Li: Conceptualization, Methodology, Software, Validation,Data curation, Writing - original draft, Visualization, Writing - review &editing. Tonglin Liu: Investigation. Ruhan Liu: Investigation. Lijing Chen: Resources, Supervision, Project administration, Funding acquisition.

Acknowledgments

This publication was funded by the Science and Technology Commission of Shanghai Municipality (19050501900).

Appendix A.Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.aaf.2020.05.008.