Variability in the zooplankton assemblages in relation to environmental variables in the tidal creeks of Sundarban estuarine system, India

The present study illustrates a holistic account of zooplankton community dynamics in relation to physico − chemical variables in the tidal creeks of Indian Sundarbans estuarine system. Out of 11 water parameters, seven parameters (Temp., salinity, DO, turbidity, PO 4 − P, NO 3 − N and NO 2 − N) differed signicantly (p ≤ 0.05) among seasons. A total of 63 zooplankton taxa were recorded with the predominance of Copepoda, varying in ranges from 59.55 to 73.13% of the total zooplankton population. PERMANOVA design depicted the signicant variations of zooplankton population both spatially (F = 2.313; p = 0.001) and temporally (F = 6.107; p = 0.001). Out of 41 species of Copepoda recorded, 14 species (Paracalanous parvus, Parvocalanous dubia, Bestiolina similis, Acrocalanous gibber, A. gracilis, Acartia erythraea, A. spinicauda, Pseudodiaptomus serricaudatus, P. annandalei, P. aurivilli, Oithona brevicornis, O. similis, Longipedia weberi and Microsetella norvegica) indicated as ‘characterizing species’ in the creek environment, and highlighted the euryhaline nature as well as broad range of thermal tolerance of these species. β – diversity index (Index of Multivariate dispersion) reected higher values (β = >1) in the creeks (S4, S2 and S6), those are experienced with high anthropogenic pressure. On the whole, the calculated mean value of α − diversity (d (cid:0) =4.07; H'=2.31) indicated ‘good’ zooplankton diversity. Water parameters viz., Temp., salinity, DO, turbidity, PO 4 − P and NO 3 − N were found to have inuence on the distribution, abundance and diversity of zooplankton in the creeks. More specically, the linear model (DistLM) exhibited two variables viz., temperature and salinity were the primary controlling factors in shaping the zooplankton community compositions in the creek environment. (2−tailed) between water variables and zooplankton groups, and between the copepod families were also computed. ANOVA (post hoc) was performed to understand the signicant variations of copepod families in seasons. Further, we performed the PERMANOVA to test the differences (p ≤ 0.05) between the zooplankton samples (abundance) in terms of seasons and stations. Alpha diversity (α-diversity) based on species richness per sample (SRp), Shannon diversity (H (cid:0) ) (Shannon-Weiner 1949), richness (d (cid:0) ) (Margalef 1958), Simpson diversity index (1-λ (cid:0) ) (Simpson 1949) and Evenness index (Pielou 1977) were measured. Additionally, one-way ANOVA was performed to identify the signicant (p ≤ 0.05) differences of the richness of zooplankton taxa across stations. Beta (β) diversity was measured using two multivariate approach viz., index of multivariate dispersion (MVDISP; Warwick and Clarke 1993), and β−dissimilarity index based on Bray curtis dissimilarity (Bray Curtis 1957). The index of multivariate dispersion (IMD) was also applied as a multivariate stability index to evaluate the stability changes in the zooplankton community (Warwick and Clarke 1993). The graphical representation of k-dominance curve in both the year was employed to investigate the dominance pattern of zooplankton taxa across seasons. A Bray-Curtis similarity matrix was then calculated in the six stations sampled during the different phases. Cluster analysis (group averaged) and multi-dimensional scaling (MDS) were used to assess the similarity of community structure among samples; which were then tested using analysis of similarity (ANOSIM) to understand the signicant differences between stations with respect to zooplankton species composition. Similarity percentage (SIMPER) routine was performed to categorize species responsible for similarity (characterizing species) and dissimilarity (discriminating species) between the groups, and also the relative contribution of individual species to the total zooplankton community. RELATE test, a non-parametric type of Mantel test, was performed to comprehend the signicant pattern and With of (β−dissimilarity) in the different assemblages, these exhibited as key zooplankton species that differentiated between the assemblage, and their variability makes them discriminating species in the different zooplankton assemblages. (2.71) and minimum at S2 during monsoon (1.66). The stations S1 and S6 showed similar assemblage pattern with regard to the richness, evenness and diversity of zooplankton. Throughout the study period, the zooplankton assemblage was not uniformly distributed in the studied stations and uctuated greatly. The zooplankton species evenness (J (cid:0) ) depicted that, it was relatively higher at S4 during pre−monsoon (0.86) and lower at S5 during monsoon (0.55). Results of ANOVA (post hoc test) showed signicant variations (p ≤ 0.05) of diversity indices (d (cid:0) , H (cid:0) and 1−λ (cid:0) ) across seasons except J (cid:0) . The correlation matrix exhibited a signicant positive correlation between d (cid:0) and H (cid:0) (r=0.923; p ≤ 0.01), and also d (cid:0) with 1−λ (cid:0) (r = 0.83; p ≤ 0.05). Similarly, J (cid:0) had positive correlation with H (cid:0) (r = 0.42; p ≤ 0.05) and 1−λ (cid:0) (r = 0.61; p ≤ 0.01). All the correlation coecient values of the indices showed positive correlation between them, implying similar pattern of species interaction in the zooplankton assemblage throughout the study. variability (stability) during the study period. The β−dissimilarity index revealed the inter-stations differences in the species distribution patterns in the study area. The highest β−dissimilarity percentage was found to be between S2 and S6 with 56.81% and lowest between S3 and S5 (50.34%).


Introduction
Tropical and sub-tropical regions of the world are home to diverse assemblage of species (Cornils et al. 2007). Plankton are cosmopolitan; distributed across all ecosystems such as lakes, pools, reservoirs, hill streams, rivers, estuaries and the oceanic compartments (Cloern et al. 2014). Ecological studies emphasizing on phytoplankton and zooplankton communities has been exhaustively studied at recent times since there is a vivid environmental in uence on the planktonic communities (Thabet et al. 2018). Tropical estuarine ecosystems in this respect are fascinating for studying the dynamics of zooplankton, due to its frequently changing hydrological conditions and rigorous biological processes (Islam et al. 2006). Zooplankton communities play a pivotal role in the functioning of aquatic ecosystems as it provides crucial linkages in the aquatic food web (Capriulo et al. 2002;Sotton et al. 2014). The information regarding the abundance and composition, and its relationship with the environmental parameters is indispensable for understanding the ecological processes of a particular area (Sousa et al. 2008). Monitoring and understanding the alterations in the dynamics of the zooplankton community over time can also provide a deep insight about the ecosystem functioning, which forms the basis for developing predictive models based on natural and anthropogenic alterations in the environment, especially those induced by climate variations (Tommasi et al. 2013). Therefore, it is imperative to study spatial and temporal structure of the zooplankton community to comprehend the state of ecosystem health.
Tidal creeks are a pre-eminent part of the estuarine ecosystem. These are extensive, bountiful and home to enormous biodiversity (Mallin 2004). They are highly fecund coastal environments in terms of aquatic biodiversity (Wiegert and Freeman 1990), acting as a support system to complex food webs (Posey et al. 2002). They also serve as feeding and nursery grounds, providing food and habitat to numerous sh, and other species of commercial importance (Lawal-Are et al. 2010). Several estuaries across the world have numerous tidal creeks with diverse water quality characteristics which pilot the ecosystem functions (Lerberg et al. 2000). Off late, the increased efforts to develop coastal areas for various economic purposes have resulted in the tidal creeks losing their ecological value (Vernberg and Vernberg 2001). Anthropogenic e uents such as nutrients, pesticides, heavy metals and other chemical pollutants contaminate the creek waters to a dangerous level (DeLorenzo et al. 2001). The net result is that many of these productive and ecologically signi cant ecosystems are degraded to various extents. Thus, these ecosystems are highly susceptible to environmental changes and human induced changes; nevertheless, their ecological signi cance is underestimated which is re ected by the fact that signi cantly a smaller number of studies have been conducted on these systems as compared to other known estuarine systems (Mallin 2004). estuarine water. In the present study, the sampling sites were chosen based on the varying hydrographical conditions and anthropogenic interference, which includes shing, agriculture and other commercial activities etc. Six sampling sites viz., Sikarpur (SIK) , Sagar (SAG), Hatipeta (HAT), Chemaguri (CHE), Phooldubi (PHO) and Kachuberia (KAC) (hereafter referred to as S1, S2, S3, S4, S5 and S6, respectively) were selected on the six major creeks. Station S1 and S4 opens into Mooriganga river, and S3, S5 and S6 opens into river Hooghly, whereas S2 opens into the Bay of Bengal. The study period is considered as post−monsoon (November to February), characterized by low to negligible rainfall and low temperature; pre−monsoon (March to June) characterized by squalls and convectional rainfall and high temperature, and Monsoon (July to October) characterised by torrential rains with sultry weather conditions (Chaudhuri et al. 2012). The geographical locations of the sampling sites and study area are shown in Fig. 1.

Sampling methodology
Monthly sampling was carried out from six sampling stations between July 2016 and June 2018. For determination of water quality, sub-surface water samples (0.5 m depth) were collected using Niskin water sampler and transferred immediately to pre−rinsed polyethylene bottle (1.0 L). Water samples were brought to the laboratory in cold condition for further nutrients analysis. We obtained in situ measurements of water temperature (Temp.) by a mercury thermometer (P−601466), pH with a digital pH meter (pH 620, Eutech Instruments, Singapore) and turbidity (Nephelometric turbidity unit) was measured by a Turbidity meter (Model No. EI331E). The salinity, dissolved oxygen (DO) and total alkalinity (TA) were estimated on board by following titrimetric methods (APHA, 2012). The water samples collected for dissolved nutrient analysis were ltered through GF/F lter paper (mesh size-0.7 μm) for removing the particulate matter and the ltrate was stored at −20°C until further analysis. Nutrients such as nitrite (NO 2 −N), nitrate (NO 3 −N), phosphate (PO 4 −P) and silicate (SiO 4 −Si) were analysed following standard spectrophotometric procedures described in Strickland and Parsons (1972). For analysis of biochemical oxygen demand (BOD), water samples were collected separately in 500 ml polyethylene bottles (HDPE, Tarson). BOD was estimated by 5−day BOD test (APHA 2005). All the methods were standardized as per ambient conditions, and blank measurements were taken into consideration for the procedures.
Zooplankton samples were collected by ltering 100 L of water through 50 μm plankton net (mouth diameter: 75cm) from the sampling sites. To avoid the large-scale variation in zooplankton and dial vertical migration, all the samples were collected during morning hours between 6.00 am to 10.00 am. The collected samples were preserved in 4% buffered formalin and transported to the laboratory at the earliest for qualitative and quantitative analysis. A total of 144 zooplankton samples were analysed in the present study. The Sedgewick-Rafter counting cell method was applied for enumeration of zooplankton by employing trinocular light microscope (Axioster plus − Carl Zeiss). The taxonomic composition of the samples was analysed to the lowest possible taxa following standard taxonomic identi cation keys (Kasturirangan 1963;Conway et al. 2003;Al−Yamani et al. 2011;Shiel 1995). The abundance of zooplankton was expressed as number of individuals per cubic meter (ind.m -3 ).

Data analysis
The water quality parameters were normalized by transforming log(x+1) except pH prior to analyses. The water quality parameters in different stations were subjected to one-way analysis of variance (ANOVA), and Post hoc test (Duncan's multiple range tests) using SPSS v.21. The statistical signi cance of spatio-temporal differences of water variables were tested with a Permutational Multivariate Analysis of Variance (PERMANOVA), a non-parametric multivariate statistical test. Pearson's correlations (2−tailed) between water variables and zooplankton groups, and between the copepod families were also computed. ANOVA (post hoc) was performed to understand the signi cant variations of copepod families in seasons. Further, we performed the PERMANOVA to test the differences (p≤0.05) between the zooplankton samples (abundance) in terms of seasons and stations. Alpha diversity (α-diversity) based on species richness per sample (SRp), Shannon diversity (H ) (Shannon-Weiner 1949), richness (d ) (Margalef 1958), Simpson diversity index (1-λ ) (Simpson 1949) and Evenness index (Pielou 1977) were measured. Additionally, one-way ANOVA was performed to identify the signi cant (p≤0.05) differences of the richness of zooplankton taxa across stations. Beta (β) diversity was measured using two multivariate approach viz., index of multivariate dispersion (MVDISP; Warwick and Clarke 1993), and β−dissimilarity index based on Bray curtis dissimilarity (Bray Curtis 1957). The index of multivariate dispersion (IMD) was also applied as a multivariate stability index to evaluate the stability changes in the zooplankton community (Warwick and Clarke 1993). The graphical representation of k-dominance curve in both the year was employed to investigate the dominance pattern of zooplankton taxa across seasons. A Bray-Curtis similarity matrix was then calculated in the six stations sampled during the different phases. Cluster analysis (group averaged) and multi-dimensional scaling (MDS) were used to assess the similarity of community structure among samples; which were then tested using analysis of similarity (ANOSIM) to understand the signi cant differences between stations with respect to zooplankton species composition.
Similarity percentage (SIMPER) routine was performed to categorize species responsible for similarity (characterizing species) and dissimilarity (discriminating species) between the groups, and also the relative contribution of individual species to the total zooplankton community. RELATE test, a nonparametric type of Mantel test, was performed to comprehend the signi cant pattern of the change of the zooplankton species assemblage between the two years (2016−17 and 2017−18). Pre-treatment of all the biological data was done by square root transformation to achieve the normality of the zooplankton abundance data. To identify and quantify the environmental variables that in uenced the zooplankton community variability, BIO-ENV (Biota and/or Environment matching) and the distance − based linear model (DistLM) were conducted. BIO−ENV procedure re ected the best set of environmental variables that explained the patterns of zooplankton communities. A marginal test (non-parametric signi cance test), was used to determine the variation in biological data, and each environmental variable that can explain. Further, sequential test was done to examine whether any particular variable contributed signi cantly to the explained variation in biological (zooplankton) data (Clarke et al. 2014). The tted DistLM was visualized using the distance-based redundancy analysis (dbRDA). Then we obtained the model using the Akaike information criterion (AICc), and the stepwise selection procedure. The above stated analyses were performed by using PRIMER v 6.0 (Clarke and Gorley 2006).

Results
Variations of water quality parameters The variations of water quality variables in different seasons of the study sites are shown in Table 1. No signi cant spatial differences (p≥0.05) in water variables were found, while the temporal differences exhibited signi cant heterogeneity (ANOVA post hoc test, p≤0.05). Out of 11 water parameters, seven parameters (Temp., salinity, DO, turbidity, PO 4 −P, NO 3 −N and NO 2 −N) differed signi cantly (p≤0.05) among seasons. The creeks water remained alkaline throughout the study period. Lowest pH was observed at S1 (6. 51), and highest at S6 (8.2) throughout the sampling period. Seasonal mean surface water temperature ranged from 20.06±3.48 ºC to 31.0±2.41 ºC. The highest (32.70 ºC) and lowest (16.9 ºC) value was recorded at S1 in June and S6 in January, respectively. A wide variation of salinity was observed during the study period. The maximum salinity was recorded at S2 in June (27.2 ppt), and lowest at S5 in September (0.60 ppt). The salinity values varied at the different sampling stations according to their distance from the sea. Differences in dissolved oxygen concentration of the stations were found to be no signi cant. The higher magnitude of dissolved oxygen was obtained during post-monsoon (7.20±0.27 mgl -1 ), and it decreased from monsoon (6.35±0.15) to pre-monsoon (5.80±0.20). Turbidity was recorded in the expected level with its peak during monsoon and lowest in pre-monsoon. Mean seasonal turbidity varied from 23.35±14.48 to 88.63±16.25 NTU with its maximum (104.0 NTU) at S6 in October, and minimum (12.0 NTU) at S3 in January.
The seasonal trend of nutrients was not uniform in the studied stations. NO 3 −N contents were comparatively higher during monsoon, and its concentrations were recorded maximum at S6 and minimum at S1. A similar type of seasonal trend was noticed for dissolved PO 4 −P and NO 2 −N in the sampling sites. The mean concentrations of PO 4 −P and NO 2 −N were found to be decreasing from monsoon to post-monsoon, and again it started showing reverse trend during pre-monsoon. Dissolved SiO 4 −Si exhibited wide variation across seasons; however, post hoc analysis portrayed no signi cant difference (p≥0.05) among stations. Though ANOVA (post hoc test) portrayed no signi cant variations of water variables across the stations; PERMANOVA analysis re ected the signi cant spatial (F=1.459, p = 0.046) and temporal (F=7.923, p=0.001) differences in water variables in the studies creeks. Intra-relationship among various water quality parameters is described in Table 2. Water temperature exhibited signi cant positive correlation with turbidity (r = 0.682; p≤0.01), PO 4 −P (r = 0.535; p≤0.05) and NO 3 −N (r = 0.575; p≤0.05). Salinity in turn, was found to have a correlation with nutrient parameters viz., PO 4 −P (r = 0.443),

Zooplankton abundance, composition and distribution
The mean abundance of total zooplankton and copepod population in the studied stations are given in Table 3. A total of sixty three zooplankton taxa were recorded throughout the study period. Among zooplankton major groups, Copepoda was most prominent in terms of species richness and numerical abundance, varying in ranges from 59.55 to 73.13% of the total zooplankton. Among zooplankton Calanoida alone contributed 46% followed by Cyclopoida (15%) and Harpacticoida (6%). The other groups contributing signi cantly were larvae (13%), crustacean nauplii (9%) and Mysids (4%) (Fig. 2). The mean seasonal abundance was maximum during the pre-monsoon (50,861.11±23,702.63 ind.m -3 ) and lowest in the monsoon (29,805.56±17,571.72 ind.m -3 ). On the whole, the quantitative spectrum of total zooplankton ranged from 10,000 to 1,10,000 ind.m -3 with its maximum abundance at S1 and minimum at S5 during February and October, respectively. The station-wise average numerical abundance of zooplankton in different seasons is illustrated in Fig. 3. As a whole, holoplankters dominated throughout the study period in all the stations contributing 71.4−87.9% to the total zooplankton community with the highest in monsoon and lowest in pre−monsoon. The meroplankters contributed 11.2-34.62% to the total zooplankton density with the highest in pre−monsoon and lowest in monsoon. The abundance of holoplankters and meroplankters in the different stations are shown in Fig. 4
Albeit, Copepoda can be considered as the major group, which comprised of a large proportion of total zooplankton population in all the months, the occasional co−dominance of larvae (comprising of the polychaete, cirripede, isopod, amphipod, gastropod and sh larvae) and mysids were also documented. The mean abundance of larvae was maximum during pre−monsoon (6,222 ind.m -3 ) and minimum in monsoon (2,500 ind.m -3 ). Throughout the study period, the numerical abundance of larvae was maximum in October and minimum in July. Rotifera and Cladocera contributed 1% each to the total zooplankton throughout the study period. The temporal abundance of Rotifera exhibited two peaks (bi-modal pattern) in August and November. There was signi cant increase in average numerical abundance of Rotifera from pre−monsoon (55 ind.m -3 ) to post−monsoon (430 ind.m -3 ). Cladocera did not vary signi cantly across the seasons. Chaetognatha was represented by a single species, Zonosagitta bedoti with an average numerical abundance of 386 ind.m -3 , and exhibited its maximum (1,505 ind.m -3 ) in January. Its abundance exhibited unimodal pattern across the seasons. Among the others, Mysids contributed signi cantly to the total zooplankton density, whereas Hydromedusae were spotted sporadically during the study period. Results of ANOVA (post hoc test) showed, no signi cant spatial differences in zooplankton (groups), while temporal differences exhibited signi cant heterogeneity (p≤0.05) among the zooplankton groups such as Copepoda (F= 4.061; p = 0.039), Cladocera (F= 7.046; p = 0.007), Chaetognatha (F=3.902; p= 0.05), crustacean nauplii (F=13.442; p= 0.001) and larvae (F=7.902; p= 0.005). Other three major groups viz. Rotifera (F=1.214; p=0.324), Eggs (F=1.220; p=0.323) and Mysids (F=1.73; p=0.930) did not show statistically signi cant variations across seasons.

Species similarity
The Cluster analysis on the similarity of zooplankton species abundance in the studied stations are shown in Fig. 5. The species abundance resulted in three clusters. The maximum similarity (83.9%) was observed between the stations S1 and S2 during post−monsoon, and lowest similarity (73%) between S3 and S6 during pre−monsoon. On the whole, the species composition in the samples exhibited >65% similarity between the stations. The grouping of species composition was further substantiated by the NMDS, and the plot also gave similar compositional pattern among samples. The assemblage of pre−monsoon differed substantially from that of the other two seasons, with regard to the zooplankton composition (Fig. 6). Results of ANOSIM re ected the signi cant (Global R = 0.101; signi cance level 0.1%) spatial differences of zooplankton species composition. The zooplankton community also depicted the signi cant temporal variation (R= 0.565; P 0.1%). The k-dominance curve extracted for the period 2016−17 re ected that the species abundance and diversity was noticeably higher during the pre−monsoon as compared to the other seasons (Fig. 7a), whereas in the following year (2017−18) the curves overlapped indicating no differences in the species diversity, illustrating the similar dominance patterns (Fig.7b). RELATE routine analysis further substantiated the signi cant (ρ = 0.399; P 0.1%) inter annual differences of the zooplankton community compositions between 1 st and 2 nd year study period. Though, one-way ANOVA (post hoc) results portrayed no signi cant variation of the zooplankton groups spatially (p>0.05), PERMANOVA design depicted the signi cant variations of zooplankton population both spatially (F=2.313; p = 0.001) and seasonally (F = 6.107; p = 0.001). On the whole, the spatio−temporal variations of zooplankton were found to be signi cant (F = 2.0; p = 0.001) in the study area.

Diversity indices α−diversity
The occurrence of Copepoda species varied considerably in seasons. Maximum species richness (36 species) was recorded at the station S3 during pre−monsoon followed by S2 (34 species) and S5 (32 species) during post−monsoon and pre−monsoon, respectively. The average species richness of the Copepoda was 28±4 species across the stations. On the whole, the seasonal species richness was highest during pre−monsoon (41 species) followed by monsoon (37 species) and post−monsoon (31species) throughout the study period. In respect of total zooplankton, maximum richness was accounted for 53 taxa at S2 during post−monsoon, whereas it was minimum at S4 (34 taxa) during monsoon. Signi cant spatial variations in the zooplankton taxa during monsoon (F = 5.667; p = 0.000) and post−monsoon (F = 9.634; p = 0.000) was observed, whereas the variation was not signi cant during pre−monsoon (F = 1.474; p = 0.219).
The Shannon-Wiener diversity (H'), Margalef's species richness (d ) and Pielou's evenness index (J ) in the studied stations for both the year is shown in Fig.  8. As with total zooplankton population, the d was highest at S1 during post−monsoon (5.47) and lowest at S5 during monsoon (2.55). The H values also depicted similar trend with the maximum at S1 during post−monsoon (2.71) and minimum at S2 during monsoon (1.66). The stations S1 and S6 showed similar assemblage pattern with regard to the richness, evenness and diversity of zooplankton. Throughout the study period, the zooplankton assemblage was not uniformly distributed in the studied stations and uctuated greatly. The zooplankton species evenness (J ) depicted that, it was relatively higher at S4 during pre−monsoon (0.86) and lower at S5 during monsoon (0. The relative dispersion (inter stations β−dissimilarity) value calculated using the multivariate dispersion index. The stations S4 (=1.11), S2 (=1.13) and S6 (=1.16) showed higher values (>1) indicating high β−diversity compared to S3 (=0.72) and S5 (=0.77). The zooplankton species turnover was registered in the impacted stations (S4, S2 and S6) as the values reached β =1. However, the pair-wise comparison (stations compared) of IMD depicted lower values (<1.0) indicating community variability (stability) during the study period. The β−dissimilarity index revealed the inter-stations differences in the species distribution patterns in the study area. The highest β−dissimilarity percentage was found to be between S2 and S6 with 56.81% and lowest between S3 and S5 (50.34%).
For all the sampling stations, the correlation between zooplankton abundances and environmental variables were established by multivariate BIOENV analysis. The analysis con rms a set of environmental parameters that is related to spatial variations of zooplankton. BIO-ENV analysis (employing Spearman rank correlation method) re ected that the water quality parameters viz., pH, salinity, DO, turbidity, PO 4 −P, NO 3 −N were correlated with the zooplankton community (Table 6). The best set of correlation between zooplankton and the individual environmental variable obtained in the combination of salinity and PO 4 −P with the maximum coe cient of 0.210, whereas the minimum correlation coe cient (0.170) was recorded only with salinity. The correlation coe cient (ρ = 0.21) with the signi cance level 1.6% indicated that, two variables i. e. salinity and PO 4 −P were the controlling factors for the distribution of zooplankton in the studied stations. To quantify the additional explained variables, Marginal tests (distance based linear model, DistLM) was performed to obtain a signi cant correlation between the zooplankton and each of the environmental variables. The results showed that, the signi cance correlation (p≤0.05) has been observed, with the variables such as Temp., salinity, DO, turbidity, PO 4 −P and NO 3 −N. The Sequential tests also revealed similar results with the former having signi cant correlation (p≤0.05) with three variables viz., Temp., salinity and turbidity ( Table 7). The tted model (DistLM) explained that the two variables; Temp. and salinity were the deterministic parameters to explain the zooplankton community compositions in the studied stations (Table 8).

Discussion
Zooplankton community structure and its distribution pattern The density and taxonomic diversity of the organisms in samples is primarily dependent upon the mesh size of sampling nets (Turner and Dagg 1983). The abundance of zooplankton, particularly the copepod nauplii and small copepods (Oithona spp.) were undervalued by the use of large mesh nets for sampling. For example, a mesh size of >150 µm is fairly large to sample quantitatively the small copepods and meroplankters, that sometimes dominate the estuarine zooplankton community (Fulton 1984). In our study, overall, the copepod nauplii and juvenile stages of copepods accounted great majority in the samples. While pro ling the zooplankton taxa in the creeks of Sagar Island, Sundarbans, a total of 63 zooplankton taxa were documented. During the present study, Copepoda contribution to the total zooplankton population was ranging from 59.5 to 73.13%, which was in line with the ndings of Nandy and Mandal (2020) from Matlah river system, where the authors reported the dominance of Copepoda (59 to 87%) to the total zooplankton population with the monsoon maxima (87%). However, in the present context, a pre−monsoon maximum was evident, which was contrast to the observations made by Nandy and Mandal (2020), but in agreement with the other workers (Ramaiah et al. 1996;Bhattacharya et al. 2015). The recruitment of neritic species through tidal in uenced massive ingress of seawater into the creeks also could be one of the plausible causes for higher abundance of Copepoda in pre−monsoon (Mishra and Panigrahy 1996). Sarkar et al. (1986) found predominance of Copepoda (73−96%) in the total zooplankton population with highest values in pre−monsoon, and lowest during late monsoon period in the creeks of Sagar Island which supports the ndings of the present study. Calanoid copepods being the dominant group followed by Cyclopoids and Harpacticoids during pre−monsoon in the present study was in line with Bhattacharya et al. (2015). Sai Sastry and Chandramohan (1995) also reported similar kind of Copepods contribution viz., Calanoid > Cyclopoid > Harpacticoid during pre−monsoon with herbivorous and omnivorous copepods being the principal trophic components in the Cochin backwaters ). The abundance of Copepoda steadily increased from post−monsoon to pre−monsoon with rising trend of the salinity in the present study. The recorded low zooplankton abundance during monsoon could be attributed to heavy in ux of freshwater coupled with abrupt hydrological changes, which in  14 Copepod species has been documented as having perennial existence in the studied creeks, which may be due to their continuous breeding nature, high reproductive capacity coupled with suitable environmental conditions and food availability in the ecosystem (Ramaiah and Nair 1997; Santhanam and Perumal 2003). Similar type of species dominance has been reported from Mandovi and Zuari estuarine system and Godavari estuary by Padmavathi and Goswami (1996) and Sai Sastry and Chandramohan (1995), respectively. The dominance of Copepod families viz., Oithonidae, Paracalanidae, Acartiidae and Pseudodiaptomidae in the present study were at par with the ndings of Neumann-Leitao et al. (1992); McKinnon and Klumpp (1998). The higher abundance of species Acartia sewelli, Eucalanus crassus, Candacia bradyi, Acrocalanous longicornis during pre−monsoon, and A. plumosa, Centropages dorsipinnatus during post−monsoon, probably explains the limited period of existence, and highlighting the temporal shift in species abundance in the creeks of Sundarbans. The prevalence of high numbers of Oithona spp. among other copepods in the present results was similar to the reports of Gallienne and Robins (2001), which described them as the most abundant and ubiquitous copepod in the world. High abundance of Oithona spp. in the zooplankton population could be attributed to their smaller body size, omnivorous feeding habit (Kumar 2003), and high reproductive capacity (Santhanam and Perumal 2003). Almeda et al. (2010) suggested that the lower metabolic requirements of Oithonids compared to Calanoids might explain the high abundance of the former group, both in coastal eutrophic waters and in oligotrophic oceanic environment.

Formation of groups and co-existence
In the present study, the copepod families such as Paracalanidae, Candacidae, Oithonodae, Onceidae, Ectinosomatidae and Longipedidae showed positive correlation amongst themselves (Table 4) indicating that there is a close association between the families and species under these families can increase or decrease in conjunction with one another. Thus, it provides evidence that these families coexisted and combined to form their own group in the zooplankton assemblage. The coexistence of these copepods reveals their advantageous life history traits in the local environment (site-speci c) which made them a successful inhabitant (Dur et al. 2007) in these creeks habitat. The families Candacidae and Tortanidae may have formed their own separate group, as they were both negatively correlated with most of the other families. Similar nding was also reported by Bhattacharya et al. (2015) from Sundarban waters, where the authors speculated that there is congeneric association of ve and six species of family Acartiidae and Paracalanide, respectively, from Jambu Island, Indian Sundarbans. These coexisting species exhibited high adaptability and opportunistic nature, by means of shifting their feeding habit to adopt in a highly variable estuarine environment Mwaluma et al. 2003).
Holoplankters predominated throughout the year in all the stations with slightly higher numbers in monsoon months, whereas meroplankters were highest during post−monsoon, which indicates a key role of salinity in controlling the zooplankton community . The co-dominance of meroplankters (polychaete larvae, decapod larvae, crab zoea, bivalvia, gastropod veliger, isopod larvae and cirripede larvae) is suggestive of the fact that these organisms take an important role in the coupling of bentho-pelagic food webs (Rakesh et al. 2008; Nandy and Mandal 2020). Another characteristic feature of the present observation was the relatively large occurrence of copepod nauplii, which could be attributed to high density of older copepods (Uye et al. 2000) round the year. Chaetognatha, Zonosagitta bedoti was found to be maximum during post−monsoon which was in consistent with the ndings made by Nandy and Mandal (2020). Bhattacharya et al. (2015) reported higher percentage contribution of developing stages of Z. bedoti in their study from Sundarban mangrove wetland, and inferred that the species prefer estuarine environment for its development. In the present study, the mean abundance of Z. bedoti (386 ind.m 3 ) was higher than the former study (55 ind. m 3 ), which may be due to its salinity, geographic heterogeneity and sampling strategy. The gradual increase in the abundance of Chaetognaths from the post−monsoon to pre−monsoon season corresponding to the increase in salinity which was also reported by Sarkar et al. (1985) from Hooghly−Matlah estuarine system.
Signi cantly less contribution was made by Rotifera and Cladocera, which was often more speci c to low saline water environment (Godhantaraman 2001), where they prevailed only during monsoon and late monsoon period. The occurrence of genera Brachionus and Ceriodaphnia during monsoon months indicates they are least tolerant to higher salinity, and transported passively by river run-off. In agreement to our study, similar observations have been made earlier by Govindasamy and Kannan (1991), where low abundance of Rotifera and Cladocera in the zooplankton community in Pichavaram mangroves was observed. The accumulation of meroplanktonic larval forms indicated the availability of rich food supply (planktonic forms) in the mangrove dominated creeks, low predation pressure (Kimmerer 1991), high growth rate (Turner 2004), and a balanced food concentration. In addition, these larval forms are also ushed by the sea during the spring tide, which also partially explains the neretic larval supply in the creeks. According to (Morgan 1995), the avoidance of predators may also in uence the larval release and dispersal in the estuary, which could also be the reason for higher abundance of larvae, resting eggs and nauplii in the mangrove dominated creeks. The muddy substrate of the inner estuary supports a rich polychaete community, whose larvae were also more evident in the studied creeks. Correlation between zooplankton assemblage pattern and environmental traits The factors determining the seasonal and interannual variations of zooplankton assemblages are salinity, temperature, sediment type and origin of the fauna previously reported by Selifonovo (2008). The variations in the zooplankton community were signi cantly correlated to the physico-chemical factors (especially the water temperature, salinity, DO, turbidity and PO 4 −P) in the studied creeks. Abdul et al. (2016) ascribed that water variables (temperature, salinity, transparency and DO) signi cantly explain the principal variations in the zooplankton species composition in a Nigerian tropical coastal estuary, which was in line with the present observation. In estuarine environments, the salinity gradient has control over the overall species composition (Bollens et al. 2011). Therefore, a substantial positive correlation between salinity and various groups of zooplankton (Copepoda, Cladocera, Chaetognatha, crustacean nauplii, larvae, mysids and total zooplankton) were seen in the present study. This was at par with many other previous ndings that emphasized the in uence of salinity on zooplankton community (Paturej et al. 2009;Dube et al. 2010). However, the inversely proportional relationship between zooplankton density and salinity has been described from the Matlah system (Nandy and Mandal 2020), with a note that the predominance of low saline species of genus Paracalanus, Acartia and Acrocalanous are indicating the estuarine in uence in their studies. In aquatic ecosystems the rise in temperature has been associated with elevated abundance and zooplankton species diversity (Buyurgan et al. 2010). The maximum density and diversity of zooplankton were recorded during pre−monsoon in the present study, which supports the notion. Furthermore, previous studies (Salvador and Bersano 2017; Du et al. 2020) noted that the increasing temperature promoted the abundance of small bodied copepods (P. parvus and copepod nauplii in particular), which are frequently associated with the increase in temperature and eutrophic condition in the coastal environment. Our results also corroborated with the former ndings where there was maximum abundance of small-bodied copepods (especially P. parvus, B. similis, P. dubia and O. brevicornis) during pre−monsoon season. This probably hinted towards the a nity for temperature, and increase in numerical abundance of zooplankton that could be directly proportional to the increasing temperature in the creek habitats. The relationship between zooplankton taxa, temperature and high concentrations of nutrients (NO 3  Aquatic turbidity is also a causative factor in decreasing zooplankton and it accelerates copepod mortality, which was evident in Sundarban estuarine system (Nandy and Mandal 2020), and many other estuaries (Gordina et al. 2001;Park and Marshall 2000). In the present study too, turbidity negatively in uenced on the zooplankton community. The mechanisms driving zooplankton community compositions are di cult to distinguish, but it is clear from the above discussion that water variables such as temperature, salinity, DO, turbidity and PO 4 −P have profound in uence on the zooplankton community structure in the creeks water, which was evident from the BIOENV and marginal test. Furthermore, the tted model (DistLM) also depicted two variables viz., water temperature and salinity were the deterministic variables that could explain the distribution of zooplankton community. This is in agreement with previous studies which emphasized the impact of salinity and temperature on zooplankton distribution (Paturej et al. 2006;Gao et al. 2008;Sebastian et al. 2012). Additionally, the recruitment of zoobenthos populations in the creeks was largely ensured by the local populations, and their distributions and settlements are related to the tidal ow.

Conclusion
The present study reveals the spatio-temporal dynamics of zooplankton at various geographical locations stands unique, and provides a comprehensive information on the status of the ecological interaction with the zooplankton community in the creek environment. In the present investigation, the calculated mean value of d and H' (4.07 and 2.31, respectively) indicating 'good' zooplankton diversity in the creeks system. The Index of multivariate dispersion re ected higher values (β = >1) in the creeks (S4, S2 and S6) that implied the systems are in stress due to varying anthropogenic pressure. Strong seasonal and spatial variations were recorded in the zooplankton community in uenced by varying environmental variables. Temperature and salinity were positively correlated to the increase in zooplankton abundance. On the whole, water parameters such as temperature, salinity, DO, turbidity and PO 4 − P and NO 3 − N were found to have in uence on the distribution, abundance and diversity of zooplankton in the creeks. More speci cally, among these variables, our tted distance based linear model (DistLM) exhibited two variables viz., temperature and salinity were the principal factors in controlling the zooplankton community compositions in the creek environment. Overall, the study highlights the necessity of incorporating different approaches for a better understanding of zooplankton community structure in mangrove environment including Indian Sundarban ecoregion. The present work shall help in the ecological assessment of the creek ecosystem as well as providing the baseline information on tidal creek ecosystem.

Declarations
Ethical Statement: The authors declare that they have strictly followed all the rules and principles of ethical and professional conduct while completing the research work. No speci c permission was required to collect the zooplankton samples at the study sites.
Consent to participate: Not applicable Consent to publish: Not applicable Con ict of Interest: The authors declare that they have no con ict of interest.
Availability of data: Data will be provided on request to the Corresponding author.    The abundance of total zooplankton and total copepods are expressed in thousands

Stations
Major species contribution to similarity within the assemblage (Characterizing species) (Avg. sim: 69%) Oithona brevicornis (12.17%), P. parvus (11.40%), Crus. nauplii (9.66%), B. similis (9.01%), P. dubia (7.35%), P. serricaudatus (5.15%), Polychaete larvae (4.34%), O. similis (4.18%), A. spinicauda (4.10%), A. erythraea (3.8%), A. gibber (3.66%) (Avg. sim: 63%)    Figure 1 The geographical locations of the sampling sites and study area of Sagar Island, Sundarbans Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Non metric dimensional scaling map of the zooplankton community in the studied stations The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.