Dominant species drive seasonal dynamics of the fish community in the Min estuary, China

The study was carried out in the brackish area (25°50’00”–26°20’00”N; 119°30’00”–120°00’00”E) of the Min estuary (Fig. 1). The climate of the area is characterized by a typical subtropical monsoon with seasonal variations. The mean annual temperature is 19.85°C with a range of 9.8–32.2°C (Hu et al. 2017). The mean annual discharge of the Min River is 1760 m³ s⁻¹, with a seasonally uneven distribution as a maximum value occurs in April–July (average 3200 m³ s⁻¹) and a minimum in October–March (average 620 m³ s⁻¹; Yang et al. 2007; Hu et al. 2014). The mean depth of the river is 3 m upstream and the maximum depth is 30 m downstream (Zhang et al. 2015). The tide is irregular and semi-diurnal, and salinity significantly increases when the runoff drastically decreases (Fang et al. 2017).

Figure 1

Fish surveys were performed seasonally (May in spring, August in summer, November in autumn and February in winter) at 11 sampling locations in 2015 (Table 1). Bottom trawling was used (horizontal aperture 7.5 m, vertical opening height 3 m, deploy distance 45 m, and mesh size 63 mm at the net opening and 25 mm at the cod end). The net was operated for half an hour at each sampling site at a towing speed of approximately 3.3–4.3 knots (corresponding to 6.02–7.85 km h⁻¹) to collect fish. Samples from each site were put into an ice container by site groupings for preservation and sent to the laboratory for further analysis. Species were taxonomically identified according to the monographs “Fishes of the Fujian Province” (Part I, II; Fishes of the Fujian Province Editorial Subcommittee, 1984; 1985), and scientific names were checked against www.fishbase.org At each site, the number of individual species was counted to determine their abundance and each species was weighted to determine the biomass.

Non-metric multi-dimensional scaling (nMDS) was performed on species biomass across sampling locations in the Min estuary. The resulting ordinations were examined for seasonal groupings that indicate potential structuring within the fish community. The non-parametric analysis ANOSIM was used to test the statistically significant (p < 0.05) differences between the sampling seasons (Clarke & Gorley 2006).

To identify different aspects of the species diversity, the demersal fish community was analyzed employing four types of indices representing the main components of the diversity (Table 2): 1) species richness, using two indices as the number of species S and Margalef’s species richness index D (Margalef 1958); 2) species evenness, which refers to how close in numbers each species in a habitat is, e.g. Pielou’s index J’ (Pielou 1966), the most commonly used evenness index, despite significant controversy over its performance (Heip et al. 1998), and Heip’s evenness index E_Heip (Heip 1974), mainly sensitive to the variation in rare species (Beisel et al. 2003). The less evenness in communities between the species, the higher J’ value is. Unlike J’, E_Heip is less sensitive to variation in the number of species (Smith & Wilson 1996). 3) heterogeneity, described by the Shannon– Wiener index H’ (Shannon & Weaver 1949) and the Simpson concentration index λ (Simpson 1949), which combine both the number of species and evenness components in a single value. The index H’ is assumed to be sensitive to changes in the abundance of rare species, while λ is strongly weighted toward dominant species (Peet 1974). It is assumed that a high value of H’ represents a high ecological quality status, while a high value of λ indicates a low ecological quality status. In our study, we chose an alternative one 1−λ to maintain a similar trend of variation as H’; and 4) taxonomy, describing taxonomic and phylogenetic characters of the fish community and helping to improve knowledge for conservation purposes. The first three sets of indices were determined based on the abundance data. To avoid the bias of abundance and biomass, presence/absence data as well as the default value 1 for the branch length of each taxonomic category were used to determine the taxonomic indices, including: 1) average taxonomic distinctness Δ⁺, an average distance tracing through the taxonomic tree between every pair of individuals in a sample; 2) variation in taxonomic distinctness ˄⁺, considering the evenness of taxa distribution across the hierarchical taxonomic tree (Warwick & Clarke 1995). Differences among all analyzed indices were examined using one-way ANOVA. Calculations of all indices and multivariate analysis were performed using the software PRIMER V6 (Clarke & Gorley 2006).

Dominant species were identified with the index of relative importance IRI (Pianka 1971) calculated as follows: IRI = (N% + W%)×F%, where N% and W% are the ratios of each fish species relative to the total species caught by number (N) and by weight (W) respectively, and F is the occurrence frequency of that fish species. In general, the criterion for defining dominant species varies. They were determined according to IRI values of the top species, e.g. in different areas; species with IRI >1000 (Zhu et al. 1996), or IRI >500 (Tan et al. 2012), or IRI >100 (Wang et al. 2011) were used in a discriminatory way to identify dominant species. In this study, species with IRI >1000 were grouped into dominant groups and species with values of 500–1000 into common groups.

The ecological niche index, describing the n-dimensional space associated with survival and reproduction of living organisms, has been frequently used to analyze the shift of dominant species through interspecific relationship, and can be calculated as follows:

where B_i refers to the ecological niche breadth, P_ij is the ratio of species i at sampling site j relative to the total number of fish at sampling site j, and r is the total number of sampling sites (Levins 1968). To explain the competition between two species, the overlap of the niche breadth was calculated according to the formula:

(Pianka 1973), where O_ik is the niche overlap between species i and k, with the value range of 0–100 expressed in percent; P_ij and P_kj are the ratios of the number of species i and k to the number of individuals at site j. Differences in ecological niches of the dominant species in different seasons were examined using one-way ANOVA.

Table 3 shows taxonomic characteristics of fish species, their abundance and biomass in different seasons. A total of 127 species belonging to 91 genera, 49 families and 14 orders were sampled. In total, 57 species were from Perciformes, accounting for about 45% of the total number of species, followed by about 10% from Clupeiformes and 9% from Pleuronectiformes. At the family level, both Sciaenidae and Gobiidae ranked first, each accounting for 8% of the total number of species, followed by Engraulidae (7%) and Tetraodontidae (7%). Cynoglossus and Takifugu were the dominant genera, each contributing 4% to the total number of species, followed by Pampus, Thryssa and Dasyatis with 3% respectively.

As far as the seasonal aspect is concerned, 64 species were sampled in spring and their number increased to 78 species in summer, then decreased to 49 species in autumn and 46 species in winter. In the non-metric multi-dimensional scaling analysis and the similarity test ANOSIM, the taxonomic composition of fish communities in different seasons could be effectively distinguished (p < 0.01), except between autumn and winter based on the abundance data, while considering the biomass data, fish assemblages in summer and winter showed significant differences compared to other seasons (Fig. 2).

Figure 2

Table 4 shows eight diversity indices for different seasons. The two species-richness indices show a significant correlation at 0.823. They were not correlated with other indices, except for D and H’ that showed correlation at 0.693. The evenness index E_Heip and J’ showed significant relevance at 0.9249. Two heterogeneous indices H’ and J’ showed a high correlation at 0.9434, both of which were to some extent related to E_Heip and J’. Interestingly, J’ showed higher relative values with H’ and 1 − λ than E_Heip. The average taxonomic distinctness Δ⁺ was negatively correlated with the variation in taxonomic distinctness ˄⁺ with relevance −0.6708, and both were independent of other indices (Fig. 3).

Figure 3

The dominant species in the Min estuary showed seasonal variability. The species Harpadon nehereus occurred as the dominant species in three seasons, except for spring. Both Cynoglossus abbreviatus and Polydactylus sextarius occurred as dominant species alternatively in two seasons; the former in spring and winter and the latter in summer and autumn. In addition, Trachurus japonicus and Secutor ruconius were the dominant species in spring and replaced by Pennahia argentata and Upeneus japonicus in summer, followed by Coilia mystus and Collichthys lucidus in winter (Fig. 4).

Figure 4

Table 5 shows all Pianka values of niche overlap among the dominant species in each season. In spring, there were four pairs showing a high niche overlap, including Harpadon nehereus and Pennahia argentata, and three pairs among Cynoglossus abbreviatus, Coilia mystus and Collichthys lucidus. In summer, Collichthys lucidus and Cynoglossus abbreviatus showed the highest value of 94.70%; other relatively high values were 67.94% for Harpadon nehereus and Pennahia argentata, 54.28% for Harpadon nehereus and Polydactylus sextarius, and 51.16% for Polydactylus sextarius and Pennahia argentata. There were no significantly high values of niche overlap in autumn, while values between Coilia mystus and Harpadon nehereus – 59.23% and between Coilia mystus and Cynoglossus abbreviatus – 53.11% were considerably high. In winter, in addition to the overlap between Collichthys lucidus and Cynoglossus abbreviatus at 80.08%, the species pair of Secutor ruconius and Upeneus japonicus showed an almost complete overlap of the ecological niche at 99.99%.

The Min estuary is an important fishing area with a density of 997.36 kg km⁻² of fish biomass, higher than that in coastal waters of the East China Sea (884.72 kg km⁻²) and the Bohai Sea (275.30 kg km⁻²), and lower than in the Yellow Sea (2323.57 kg km⁻²; Huang et al. 2010). The Min estuary also borders on the famous eastern Mindong Fishing Ground and the southern Minnan-Taiwan Bank Fishing Ground with higher productivity in China, providing an important place for migratory fish species, e.g. T. japonicus, H. nehereus and C. lucidus etc., to spawn, nurse or winter. Knowledge of the taxonomic composition of fish assemblages, even on a seasonal basis, would be beneficial in terms of knowing how fish use this estuary for their development.

In 2015, a total of 127 fish species were sampled in the Min estuary, which is more than 77 species sampled in the Yellow River estuary in 1959–2011 (Shan et al. 2013) and 62 species sampled in the Yangtze estuary in 2010–2011 (Shi et al. 2014). Unlike the temperate character of the Yangtze and Yellow River estuaries, the Min estuary is subtropical with a higher water temperature, which supports higher species richness. In terms of taxonomic composition, Engraulidae were the common dominant family in fish catches in the Yellow River estuary during all years of sampling (Shan et al. 2013), similar to the Min estuary. Furthermore, Sciaenidae comprising more subtropical species also dominated in the fish community from the Min estuary, e.g. Larimichthys crocea and Pennahia argentata were abundant in summer, and Collichthys lucidus prospered in spring, autumn and winter. In addition to natural differences in fish assemblages between all these estuaries resulting from different environments, changes in fishing methods, tools and regulations could also affect the taxonomic composition of fish, e.g. harvest regulations significantly contributed to fish recruitment failure and catch-per-unit-effort decline of saltwater bass Paralabrax spp. (McClanahan & Mangi 2001). The exclusion of closed areas could potentially increase catch rates, while the exclusion of beach seines could lead to an increase in other types of fishing gear but a reduction in the total catch (Jarvis et al. 2014).

Biotic factors also play an important role in the estuarine fish community. Large seasonal environmental differences in a subtropical estuary lead to changes in seasonal composition. The content of nitrogen and phosphorus in the Min estuary was high, adjusted by diluted water of the Min River. As a phosphorus-limited eutrophicated estuary, phosphorus showed a relatively higher value in autumn and winter than spring and summer (Zheng 2010). Chlorophyll a is usually associated with the distribution of zooplankton, where its presence plays an important role in controlling the distribution of some dominant species (Marques et al. 2007; Ensign 2014). It showed significant seasonal variation, with productivity ranging from high to low in summer, autumn, spring and winter (Xiao 2014). Meanwhile, migratory species began to arrive at the Min estuary in spring, which led to an increase in species richness. In summer, with the arrival of an increasing number of species as well as an increase in primary production, fish species richness significantly increased compared to other seasons. Furthermore, a four-month (March–June) fishing ban in the Min estuary was in force (while scientific research was officially permitted). Sampling in the restricted season and just after the season could certainly lead to a better catch, with higher diversity and biomass. On the other hand, due to the aftereffects of compensatory fishing intensity after the fishing ban, the autumn fieldwork resulted in a poorer catch than expected from the theoretical natural composition. As a geologically southern estuary with warm water in winter, the estuary still maintains a number of fish species during this season.

Species diversity is a multi-component concept to expound thoroughly the biological and ecological characters of fish communities (Purvis & Hector 2000). Our results show not only that a single diversity descriptor cannot provide a complete description of species diversity, but also that in some cases it cannot even encapsulate a complete description of a specific diversity component. In addition, some of the descriptors considered complementary according to theoretical works proved to be redundant.

Estimates of the number of species (S and D) in the Min estuary were not correlated with other indices considered in our study, as was the case with the Gulf of Lions (Mérigot et al. 2007). Species richness remains the most comprehensive index for nature conservation purposes, despite such drawbacks as high sensitivity to difficulties in accurately estimating the actual number of species at different sample sizes (Gaston & Spicer 1998; Margules & Pressey 2000).

In the Min estuary, E_Heip and J’ of the fish community showed the same pattern. Although these two indices were calculated based on H’, E_Heip demonstrated greater reliability and could prove more efficient. The two most popular heterogeneous indices, H’ and 1 – λ, were strongly related. The Simpson index is primarily a measure of dominance, especially of the first two or three species, whereas H’ is more strongly affected by species in the middle of the species rank sequence (Whittaker 1972). Although H’ and 1 − λ were significantly correlated with D and E_Heip, with a correlation coefficient of 0.50–0.70, combining the number of species and evenness into a single diversity index does not facilitate the description of a fish community (Bell 2000) as the number of species and evenness are related to different aspects of diversity. In fact, the number of species and species evenness are related to different responses of species to environmental factors (Ma 2005; Nyitrai et al. 2012).

The taxonomic diversity is expected to allow for taxonomic relationships between individuals and thus to provide additional information to classical species diversity indices. The loss of taxonomic diversity of fish can lead to a loss of ecological responsiveness to environmental fluctuations and a loss of ecological functions providing goods and services to ecosystems ( Miranda et al. 2005; Ramos-Miranda et al. 2005). To simply show the meaning of taxonomy and the evolution of fish in a sampling area, presence/absence data would be better to avoid any disturbance resulting from abundance and biomass. In our study, in addition to the negative correlation between the two taxonomic indices at −0.6708, they were independent of other indices and should be used in biodiversity conservation.

The IRI index is a good indicator to describe fish communities by integrating abundance and biomass into a single index. In the Min estuary, the dominant species showed seasonal turnover by reasonable use of resources. In spring, T. japonicus was the most dominant species, with the IRI value twice as high as in the case of the subsequent dominant species, i.e. C. abbreviatus and S. ruconius. It is considered that geological features and oceanic dynamics of the Min estuary provide higher habitat diversity and thus can provide a wider range of potential microhabitats for fish to coexist (Shi et al. 2014). On the basis of ecological traits, it appears that these three species showed specific habitat preferences and feeding habits to avoid competition for food and space. For example, T. japonicus is a typical migratory species moving back and forth between different zones as well as between the lower and upper layers of the sea. It is highly predatory, feeding mostly on planktic crustaceans and small fish. In spring, it migrates to the estuary area for feeding and shows the highest feeding intensity in the whole year (Zhang et al. 2016; Yan et al. 2018). The species C. abbreviatus is a medium-sized fish that feeds on benthic invertebrates in the bottom sediments of the coastal area (Ni 2003; He et al. 2018). The species S. ruconius, which inhabits the lower water layer, lives in large groups mainly in coastal seabed sand and mud and feeds on small plankton. In spring, it arrives at the estuary for spawning in June–July (Du et al. 2010; He et al. 2018). In this season, higher niche values were determined for C. abbreviatus, C. mystus and C. lucidus, as well as for H. nehereus and P. argentata, where only C. abbreviatus was the dominant species at that time. High values of niche overlap among common or rare species and lower values of niche overlap among all dominant species, suggesting a difference in their feeding habits or habitat requirements, could effectively reduce the species competition to maintain the ecological balance. Although low overlapping cannot be expected to be an indication of strongly interspecific competition (Losos 1996), it may imply that these species can segregate spatially or overlap extensively, depending on the spatial distribution of their resources (Hofer et al. 2004).

In summer, after finishing their persistent migration, T. japonicus and S. ruconius abandoned the estuary, leaving a rich ecological space for new dominant species, including P. sextarius, P. argentata, H. nehereus and U. japonicus. Species P. argentata, P. sextarius and U. japonicus are also migratory and increasingly come to the Min estuary in May and June for spawning and then quickly dominate the area due to their varied feeding habit, e.g. P. argentata mostly feeds on nekton (especially fish and crustacean), P. sextarius feeds on shrimps and U. japonicus feeds on benthic invertebrates (especially macrura and mollusk). The species H. nehereus is a local resident in the middle-lower water layer and gradually increases its population. It is omnivorous, feeding mainly on zoobenthos. In this season, only pairs of C. abbreviatus (common species) and C. lucidus (rare species) showed a high niche overlap at 94.70. Interestingly, several dominant species, such as H. nehereus, P. sextarius and P. argentata, showed a significant overlap at 50–70, which could be explained by the highest productivity in summer to satisfy the resource demand of these species. Alternatively, the co-occurring pattern can be expressed by a complementary niche, e.g. a high overlap in one niche dimension (spatial dimension) compensated by a low overlap in at least one of the other dimensions (feeding or temporal gradient; Schoener 1974; Pusineri et al. 2008; Nagelkerke & Rossberg 2014).

In autumn, half of the P. sextarius population gradually abandoned the sampling area, but still kept the dominant position. Meanwhile, H. nehereus rapidly advanced to an absolutely dominant status. The IRI values of the two dominant species – H. nehereus and P. sextarius – were 4211 and 3209 respectively, about 5–8 times higher compared to other common species, and the two species showed no ecological niche conflicts.

In winter, P. sextarius continued to migrate from the sampling area and its ecological space was quickly occupied by C. abbreviatus. Two commercial species, C. mystus and C. lucidus, common species in other seasons, preponderantly dominated the assemblage. These two species always co-occurred together and showed similar feeding habits, preying mostly on zooplankton (e.g. copepods, Mysidacea). The species Harpadon nehereus preserved its status. The dominant species pair of C. abbreviatus and C. lucidus showed a high niche overlap at 80.08 and the pair of C. mystus and C. lucidus showed a considerable spatial niche overlap at 54.67. If resources for species are not in short supply, two organisms can share the resources without detriment to each other, even if they show relations in niche overlap and the competitive effect (Pianka 1974). The simultaneous occurrence of these dominant species indicated that there are enough food resources and habitat space in the Min estuary, which could be attributed to the absence of many migratory species and the decrease in feeding intensity in winter, despite decreasing temperature and primary production.