Environmental drivers and ecological implications of a ctenophore swarm: a case study from the Bay of Bengal, northern Indian Ocean

Gelatinous zooplankton belongs to the phyla Cnidaria, Ctenophora, Chaetognatha, Appendicularia, and Chordata (Class: Thaliacea) (Madin & Harbinson 2001), constituting the top-end predators of the marine planktonic food web. By virtue of their environmental tolerance and increased reproduction rates over other zooplankton taxa (Dawson & Hamner 2009; Purcell 2012), gelatinous plankton make swarms in many coastal systems worldwide (Boero et al. 2008; Purcell 2012; Purushothaman et al. 2020). Seasonal and regional fluctuations in climate help voracious carnivorous plankton-like ctenophores proliferate, leading to a decline in larval fish and juveniles, fish catch, and collapse of fisheries (Lynam et al. 2006; Purcell et al. 2007).

In oligotrophic marine systems like the Bay of Bengal (BoB), the phytoplankton community is dominated by picoplankton that is not ingested by copepods due to their small size, and hence most primary production gets diverted through the microbial loop (Sommer et al. 2002). Nutrients brought by rivers stimulate blooms of opportunistic (mostly mixotrophic) phytoplankton. The positive feedback from the bacterial decomposition of these plankton stimulate the bacterial stock, which get utilized by microzooplankton (MCZ) either directly or through heterotrophic nanoflagellates and subsequently by copepods. Since ctenophores mainly feed on copepods (Bigelow 1915), their grazing modifies the plankton food web characteristics, whereby the importance of the microbial pathway increases as the principal conduit of carbon transfer in the plankton food web. Because of the low nutritional value of gelatinous plankton (predominantly composed of water and low protein content), ctenophore biomass also gets redirected from fish towards remineralizing in the water column. The organic matter of dead and sinking ctenophores supports bacterial growth and nutrient remineralization (Lancelot et al. 2002). Thus, a positive feedback cycle exists between the microbial pathway and gelatinous carnivorous plankton-like ctenophores. Therefore, from the fisheries perspective, ctenophores are theoretically considered trophic dead-ends in the pelagic food web. However, there are reports on Pleurobrachia spp. being preyed upon by several fish species like Squalus acanthias, Cyclopterus lumpus, mackerel (Mortensen 1912; Scott 1914, 1924), and larger gelatinous carnivores like Cosmetira pilosella, Aequorea spp., Chrysaora isosceles, and Beroe gracilis (Fraser 1970; Greve 1971). Moreover, predatory benthic arthropods like Pagurus spp., Carcinus spp., and Crangon spp. prey on Pleurobrachia when they sink to the bottom during winter to conserve energy (Greve 1972). Thus, Pleurobrachia spp. also act as a crucial link in benthic–pelagic coupling.

Pleurobrachia pileus is an immigrant species that infiltrates the estuarine waters from the sea (van der Veer & Sadée 1984). P. pileus swarms follow the onset of phytoplankton bloom, causing depletion of copepods favouring phytoplankton growth. Its predator, B. gracilis, checks the growing population of P. pileus and resumes the phytoplankton–copepod link in the plankton food web (Kuipers et al. 1990). Annual P. pileus outbreaks have been previously associated with phytoplankton blooms and pulses in copepod and larval plankton densities (Greve 1971; Deason & Smayda 1982). Temperature is the primary physical factor responsible for increased Pleurobrachia globosa abundance (Wang et al. 2020). Swarms of P. globosa were reported from the coastal waters of Goa, India (Goswami 1982).

The environmental causes behind the rising swarms of ctenophores in coastal waters, which are densely populated by humans, can be attributed to natural climatic variations accompanied by the effects of global warming and other climate-induced changes. Major studies have been conducted on the swarm-forming invasive ctenophore species, Mnemiopsis leidyi. Warm water temperature associated with North Atlantic Oscillation Indices has been related to the prolonged occurrence of M. leidyi in great abundance in Narragansett Bay, Rhode Island (Purcell 2005). Moreover, several anthropogenic changes like pollution, a decrease in freshwater influx due to the construction of dams on rivers, eutrophication, overfishing, the introduction of invasive species, habitat modification, and associated variation in the zooplankton community lead to the fall in the zooplanktivorous fish population, enabling the ctenophores to invade and exploit the zooplankton (Bilio & Niermann 2004). High precipitation and riverine inputs lead to high nutrient loading, stratification, and increased primary production, which triggers zooplankton growth. Thus, variation in zooplankton abundances may benefit the ctenophore population. An investigation by Somchuea et al. (2022) also revealed that when there is a reduction in human activities and a decrease in anthropogenic pressure, there is a significant rise in live coral cover as well as fish species abundance and richness. It was proposed by Purcell (2005) that tropical species which live in warm temperatures may shift towards cooler waters with higher salinity and have shorter active seasons, and show opposite trends to temperate species that were associated with warmer waters and low salinity. In the North Sea, ctenophore P. pileus was reported during warming of the Sea but in South Africa, they were reported during the coldest part of the year (Purcell 2012). Since ctenophores consume ichthyoplankton and zooplankton, they act as predators and competitors of fish. They interfere with fisheries by clogging fishing nets, thereby causing detrimental impacts on human enterprises if present in swarming numbers. A swarm of M. leidyi was reported from Chesapeake Bay by Condon & Steinberg (2008), where large quantities of carbon fixed by primary and secondary producers got converted into a gelatinous mass which cannot be utilized by most higher trophic organisms. Thus, the shunting of carbon transfer negatively impacts the planktonic food web, especially fisheries production. van Walraven et al. (2017) studied the impact of the invasion of M. leidyi on the species composition and community structure of gelatinous zooplankton in the Dutch Wadden Sea. The gelatinous zooplankton composition remained largely the same before and after the swarm of M. leidyi. However, before the swarm, predation pressure by cnidarians and ctenophores on zooplankton and ichthyoplankton was low, due to their low abundance and temporal mismatch between gelatinous zooplankton and their prey. With the annual swarming of M. leidyi, though there was a reduction in the abundance of scyphozoans, there was an overall increase in the predation pressure on zooplankton, which was mostly contributed by M. leidyi. M. leidyi appeared to fill the vacant niches in the pelagic food web rather than compete with the local species during its swarming season. Vereshchaka et al. (2022) investigated the seasonal shift in the peak abundance of two invasive ctenophore populations of M. leidyi and Beroe ovata in the Black Sea. The research provided adaptive strategies of both ctenophores, which expand their ecological niches both spatially and temporally.

Although ctenophores constitute one of the main components of gelatinous zooplankton in the BoB, no concerted studies on the environmental factors that trigger their swarm and sustain the swarm event are available. Moreover, information on their ecological implications is also lacking. The present study aims to delineate the environmental factors that trigger a ctenophore swarm in the estuarine waters of Sundarbans and how the ctenophore swarm alters the plankton trophic pathways and fish community composition in comparison with the non-swarm areas. We lay two hypotheses for this study: first, ctenophore swarm occurrence is positively correlated with high salinity and ammonium levels and negatively correlated with MCZ density, resulting in higher MSP density (dominated by small copepods) and higher microbial loop activity in the swarm areas than in the non-swarm areas. Second, the ctenophore swarm reduces the abundance and diversity of commercially important fish species by competing with or preying on their larvae and juveniles. The main objectives of the study corresponding to the research question and hypotheses are as follows: (i) to investigate the hydrographical and biological characteristics of ctenophore swarm and non-swarm areas of the estuarine and coastal waters of Sundarbans. (ii) To compare the plankton community structure, biomass, diversity, and trophic pathways, as well as the fish community composition and diversity in the estuarine and coastal waters of Sundarbans. Therefore, the present study intends to offer the first account of the relationship between environmental factors and zooplankton, and fish community attributes during a ctenophore swarm in the northern BoB of the Sundarbans mangrove–estuarine complex. The study also details the effect of the ctenophore swarm on the pelagic food web.

The Sundarbans is the world's most extensive natural mangrove system formed by the Ganges–Brahmaputra delta. Mangroves provide coastline protection by dampening waves and reducing the impact of wave action, torrential storms, and tsunamis (Kamil et al. 2021). The Indian Sundarbans occupy the low-lying coastal plains of one of the highest fish-producing states – West Bengal on the northeast coast of India. A substantial mixing of fresh and saline water renders the coastal water brackish with varying salinity gradients in different seasons.

The BoB, though rich in silicate and phosphate, is poor in nitrate. Coastal upwelling helps increase nitrate availability and enhances productivity in the system (Muraleedharan et al. 2007). In winter, the weak northeast trade winds initiate a cold, dry atmosphere over the BoB, with slight rainfall along the coast. The resulting increased river discharge decreases the sea surface temperature (SST), leading to surface layer thermal inversion and reduced mixed layer depth in the northwestern BoB (Shetye 1993). Thus, moderate rainfall, low evaporation rate, upwelling, and availability of nutrients in the upper water column keep the nutrient levels high in coastal surface waters during winter (Gattuso et al. 1998). Consequently, the seasonal high nutrient levels and high solar radiation favour increased primary production during winter (Madhupratap et al. 2003).

The most favourable environmental conditions in the Sundarbans (concerning physicochemical and biological variables) make it the most preferred nursery ground for nearly 90% of the aquatic species on the east coast of India. Many commercially important fishes grow to maturity in the estuary and make up a large part of the near-shore fishery of the northern BoB. A sizable coastal population is dependent on the local fishery, and capture fisheries are treated as the backbone of Sundarbans' economy. Nevertheless, fisheries in Sundarbans face problems that impact the biodiversity, sustainability, and livelihood of fish resources and fisherfolk (Chandra & Sagar 2003).

Temperature and salinity profiles were recorded through a Conductivity–Temperature–Depth (CTD) profiler (SBE Seabird 19 plus). Dissolved oxygen (DO) was estimated by modifying Winkler's method (Grasshoff et al. 1999). For analysis of dissolved nutrients (nitrate, nitrite, ammonium, phosphate, and silicate) and chlorophyll a (Chl a) concentrations, water samples (1 L) were collected using a Niskin sampler (5 L, GENERAL OCEANICS, Model No. 1010) and brought in an ice box to the laboratory for spectrophotometric analysis (Parsons et al. 1984; Grasshoff et al. 1999).

MSP samples were collected from near-bottom to surface (vertical sampling) and sub-surface waters (2 m depth) using a Working Party (WP) plankton net (Hydro BIOS, diameter: 60 cm; mesh size: 200 μm) equipped with a digital flow meter (GENERAL OCEANICS, Model No. 26069). For horizontal sub-surface sampling, the net was towed for 10 min at 1–2 knots. Photographs of live gelatinous plankton were obtained and identified up to order level. Samples were collected in duplicate from each station. The MSP samples were immediately preserved in 4% buffered formaldehyde (Lenz et al. 2000); and biomass was estimated using wet weight (mg/m³) within 48 h (Wiebe et al. 1975; Kidwai & Amjad 2000a, 2000b). MSP samples were sorted into major taxonomic groups, and abundances were estimated (ind./m³) by considering the volume of water filtered through the net (Goswami 2004). The specimens were counted from sub-samples (typically 25–50% of the sample) and identified up to species level using a Leica stereo microscope (Leica M125 C) and standard taxonomic literature (Annandale & Kemp 1915; Kasturirangan 1963; Conway 2005).

Water samples for MCZ were collected using a Niskin sampler, filtered onto 20-μm nets, preserved in 1–2% acidic Lugol's Iodine solution and brought to the laboratory for taxonomic enumeration. MCZ samples were analyzed under an inverted microscope (Leica MC-120) following the protocols of Godhantaraman (2001) at 100 × –200× magnification and identified using a standard literature (Corliss 1979; Steidinger & Tangen 1996; Munir & Sun 2018). Fish samples, from the cast nets by fisherfolk were photographed and preserved in 5% formaldehyde. Their taxonomic identification was done using the standard literature (Dutta et al. 2013; Kottelat 2013).

Data generated on hydrographical variables, Chl a, MSP, MCZ, and fish were compiled and subsequently processed for multivariate statistical analyses. A new set of environmental variables explaining the variability among sampling locations were derived through Principal Component Analysis (PCA) of log-transformed and normalized data sets of environmental variables. Agglomerative Hierarchical Cluster Analysis (AHCA) and Canonical Analysis of Principal (CAP) coordinates were performed on the Brady–Curtis similarity matrix of the fourth-root transformed MSP data. AHCA and Non-metric Multi-Dimensional Scaling (nMDS) powered with Similarity Profile test, SIMPROF (Pi statistics), on the fourth-root transformed fish abundance data were used to check for the response of fish populations to the ctenophore swarm. SIMPROF analysis helps search for new divisions until the probability level associated with the Pi value exceeds the set significance level (P: 0.05). As per the data distribution patterns, the MSP species abundance was fourth-root transformed, and fish species abundance was log(X + 1) transformed to construct the Mondrian plots (inverse heat map) that show the relationship between samples and species association patterns. All the non-metric procedures were performed using the ecological statistical software PRIMER v.7.0.13 + PERMANOVA from PRIMER-E Ltd (Plymouth, UK).

The CTD data indicated that the hydrographical characteristics (surface (S) and column (C)) at the estuarine non-swarm (ctenophore) locations (WBS 1 and WBS 2) were similar to the coastal non-swarm locations (WBS 7 and WBS 8). On the contrary, the hydrographical properties in estuarine swarm locations (WBS 3, WBS 4, WBS 5, and WBS 6) differed from that of the non-swarm locations (Table 1). Global change factors like water temperature, salinity, DO, nutrient concentrations, precipitation, and river discharge influenced ctenophore aggregation in the study area. The water temperature at the swarm locations (S – ca. 19.22 °C; C – ca. 18.72 °C) was lower than the non-swarm estuarine (S – ca. 20.57 °C; C – ca. 20.37 °C) and coastal locations (S – ca. 20.34 °C; C – ca. 20.16 °C). However, salinity showed an opposite trend with higher salinity at swarm locations (S – ca. 24.64; C – ca. 25.01) and relatively lower salinity at the estuarine (S – ca. 22.46; C – ca. 22.57) and coastal non-swarm locations (S – ca. 19.40; C – ca. 19.76). The DO, nutrients (except for NH₄⁺), and Chl a concentrations were lower in the swarm area than in non-swarm estuarine and coastal areas.

Among various phytoplankton groups, diatoms were the most widespread in the study area. Centric diatoms were common in the estuarine swarm locations. MCZ abundance was the highest in the coastal non-swarm area (S: ca. 120,400 ind./m³; C: ca. 106,730 ind./m³) than in the swarm locations (S: ca. 38,400 ind./m³, C: ca. 30,560 ind./m³). The MCZ community, both in terms of abundance and diversity, was constituted mainly of ciliates and mixotrophic dinoflagellates. Foraminifera, rotifers, and crustacean nauplii contributed sparsely. Ciliate and dinoflagellate abundance was significantly higher in estuarine non-swarm locations (ca. 54,590 and 36,233 ind./m^3, respectively) than in the estuarine swarm locations (ca. 12,768 and 10,743 ind./m³, respectively).

Analysis of the MSP samples revealed the presence of 70 taxa in the study area, of which 49% (34 taxa) were gelatinous taxa from 22 genera and 13 families. Among the 12 species of hydromedusae recorded, Eirene brevigona, Clytia hemisphaerica, Clytia lomae, Clytia discoida, and Phialella quadrata were the most abundant. The dominant siphonophore species were Chelophyes appendiculata, Bassia bassensis, Diphyes chamissonis, and Lensia multicristata. The ctenophore population consisted of four species: P. pileus, P. globosa, B. gracilis, and B. ovata. P. pileus and P. globosa formed the swarm. Chaetognatha was also an essential component of the gelatinous plankton. Pterosagitta draco was the most dominant Chaetognatha in the surface waters, whereas Zonosagitta pulchra was most abundant in the column waters. The appendicularian, Oikopleura (Vexillaria) dioica, had maximum abundance in the surface waters.

The non-gelatinous taxa represented by 36 species belonging to 24 genera and 15 families were primarily composed of copepods, amphipods, and decapods. Copepods in the swarm locations were mainly constituted by the calanoids Acartia (Odontacartia) erythraea, Acartia (Acartia) danae, Parvocalanus elegans, and the cyclopoid Ditrichocorycaeus asiaticus. In the non-swarm locations, copepods were represented mainly by the calanoids Subeucalanus pileatus, Centropages orsinii, and Labidocera bengalensis. The sergestid Belzebub hanseni was the dominant planktic decapod, while Tullbergella cuspidatus and Lestrigonus bengalensis were the dominant amphipod species in the study area.

The MSP abundance (S: ca. 6,766 ind/m³, CV: 0.26; C: ca. 5,960 ind/m³, CV: 0.18) and biomass (S: ca. 3,903 mg/m³, CV: 0.23; C: ca. 3,455 mg/m³, CV: 0.22) in the estuarine swarm locations were significantly higher than that reported from the estuarine non-swarm area (abundance – S: ca. 2,900 ind/m³, CV: 0.03; C: ca. 2,693 ind/m³, CV: 0.33; biomass – S: ca. 1,210 mg/m³, CV: 0.33; C: ca. 1,194 mg/m³, CV: 0.42) and coastal non-swarm area (abundance – S: ca. 2,147 ind/m³, CV: 0.05; C: ca. 2,097 ind/m³, CV: 0.06; biomass – S: ca. 1,494 mg/m³, CV: 0.14; C: ca. 1,372 mg/m³, CV: 0.17). The higher MSP wet mass in the swarm area was due to the higher density of ctenophores and other gelatinous plankton. Ctenophore abundance in the swarm locations was 2–8 fold higher (S: ca. 1,150 ind/m³, CV: 1.99; C: ca. 1,940 ind/m³, CV: 0.44) than in the non-swarm area (S: ca. 571 ind/m³, CV: 1.20; C: ca. 232 ind/m³, CV: 1.96). Copepod abundance in the swarm stations was also higher (S: ca. 4,644 ind/m³, CV: 0.18; C: ca. 3,040 ind/m³, CV: 0.32) when compared to the non-swarm area (S: ca. 1,797 ind/m³, CV: 0.40; C: ca. 1,791 ind/m³, CV: 0.52). Copepoda formed the dominant group in terms of abundance and species diversity in both areas.

The AHCA of fourth-root transformed MSP species by stations data matrix revealed 29.07% dissimilarity in the MSP community between the swarm and non-swarm locations (SIMPROF Pi: 3.61, P = 0.1%) (Figure 2(b)). The CAP analysis highlighted the relationship of physicochemical variables, Chl a, and MCZ with the MSP community (Pearson correlation >0.2). While CAP1 defined the intra-group variability in the MSP community, CAP2 separated the swarm locations from the non-swarm locations. Correlation vectors of biological and environmental variables revealed that at the swarm locations, the MSP community characteristics such as total biomass and abundance and abundances of ctenophores, calanoids, and other non-gelatinous taxa (such as decapods, amphipods, polychaete larvae, gastropod, and bivalve veligers) correlated positively with ammonium and salinity. Conversely, water temperature, DO, nitrate, nitrite, phosphate, and silicate with negative loadings on CAP2 positively correlated with the plankton characteristics typifying the non-swarm locations (Chl a, dinoflagellates, ciliates, and copepodites) (Figure 2(c) and 2(d)).

The present study during the winter monsoon of 2018 revealed the principal role of seasonal hydrological features in forming gelatinous aggregations. An oceanfront that separates offshore waters from tidal mixed inshore waters promotes gelatinous diversity and abundance (Pagès et al. 1992), which is consistent with our observation of ctenophore swarms in the coastal waters of the Sundarbans mangrove–estuarine complex in the northern BoB. Unravelling the cause and effect of the gelatinous plankton swarming is challenging, and to our knowledge, this is the first study linking environmental changes to ctenophore swarm and its impact on plankton and fish communities in Indian waters.

The study revealed how the two distinct MSP assemblages identified in the sampling area corroborated the existing hydrographical settings. The MSP assemblage with high ctenophore density (swarm) characterized the estuarine locations of Sundarbans, where cooler high-saline waters prevailed. In contrast, the second MSP assemblage with low ctenophore density (non-swarm) was identified from the estuarine–coastal waters with warmer low-saline waters. The multivariate statistical approaches (PCA, AHCA, and CAP) used to decipher the environmental patterns suggested a significant difference between ctenophore swarm and non-swarm locations and how ctenophore swarm impacted different plankton communities. Water temperature, DO, and Chl a were lower in the swarm locations than in the non-swarm area, whereas salinity showed the reverse trend.

In the winter, intense rainfall under the NEM leads to substantial discharges from rivers (Nandy & Bandyopadhyay 2011) and nutrient enrichment of the estuarine-coastal waters. This, in turn, would stimulate plankton productivity in the otherwise oligotrophic surface waters of the northern BoB (Prasanna Kumar et al. 2010), leading to the migration of gelatinous organisms into the estuarine–mangrove waterways of Sundarbans, as observed in this study. Pollution of coastal waters by wastewater leads to the enrichment of organic and inorganic nutrients thereby degrading the water quality and increasing nutrient load, which stimulates phytoplankton blooms and ctenophore swarms (Sakib 2022). Although most of the dissolved macronutrients (nitrate, nitrite, phosphate, and silicate) were higher in the estuarine non-swarm area, ammonium was higher in the swarm locations. Biggs (1977) reports that gelatinous plankton such as ctenophores and hydrozoans turn over large amounts of ammonium, indirectly changing the plankton trophic structure. Thus, the observed higher ammonium levels in swarm locations comply well with Biggs' findings (1977). Increased ammonium concentration could also be associated with eutrophication in the estuarine waters from industrial and agricultural runoff brought down by rivers (Rice et al. 2016). However, if it was not for the gelatinous plankton, the elevated nutrient levels would be comparable between the swarm and non-swarm locations, at least within the estuary. Pitt et al. (2013) addressed the issue of nutrient depletion by gelatinous zooplankton, which is later released as dissolved organic matter (DOM) and mucus on the seabed and utilized by phytoplankton and microbes. Thus, in the event of a ctenophore swarm, a major portion of the nutrients are unavailable at higher trophic levels (Condon et al. 2011).

The CAP analysis helped to identify a positive correlation between SSS, ammonium, and MSP density with ctenophore abundance; and a negative correlation between MCZ abundance and ctenophore abundance. Similar results have been observed by Huang et al. (2021) during hydromedusa B. virginica bloom in the Nanhu Lake of China, where salinity, ammonium concentration and MSP abundance were recorded to be high, while the biomass of MCZ was low during the peak abundance of B. virginica. More than two-fold decrease in the size of MCZ was observed during jellyfish bloom. The study indicated the regeneration (by release/excretion) of ammonium by B. virginica and high grazing pressure on large MCZ assemblages by B. virginica and MSP. Moreover, Wang & Xu (2013) also showed a significant negative correlation between the biomass of jellyfish and the biomass of ciliates.

The Mondrian plot for MSP revealed five groups characterizing the study area. Group I, characterizing the non-swarm area, was composed of copepods with mixed trophic modes (carnivores, herbivores, and omnivores). Gelatinous zooplankton (ctenophores, hydrozoans, chaetognaths, appendicularians), and herbivores and omnivores copepods formed Group II, predominant in the swarm area. Group III comprised decapods and ichthyoplankton. Groups IV–V consisted of copepods, amphipods, some hydrozoans, and ctenophores present exclusively in the swarm area. Thus, apart from the hydrographical variability, the ctenophore swarming affected the MSP community, especially the copepod community structure, in the study area. While the majority of copepods in the swarm locations were small herbivores like Paracalanus indicus, Paracalanus parvus parvus, and A. danae, and omnivores like D. asiaticus, Oithona nana, and Oithona similis, the non-swarm station witnessed a greater abundance of large carnivore copepods like Euchaeta concinna, La. bengalensis, and Sapphirina nigromaculata, omnivores like C. orsinii and Calanopia sp., and herbivores like Canthocalanus pauper and S. pileatus. The disparity in the average size–structure of copepod community between swarm (dominated by small copepods) and non-swarm locations (dominated by large copepods) suggests that ctenophores selectively fed on large copepods and caused a decrease in the average size–structure of copepod community at the swarm locations. The grazing of small herbivorous and omnivorous copepods, in turn, decreased the MCZ (ciliates) abundance, phytoplankton density, and Chl a levels. Here, ciliates suffered both grazing loss (from omnivorous copepods) and growth attenuation (due to scarcity of phytoplankton prey). In contrast, the lower abundance of MSP and ctenophores in non-swarm regions led to reduced grazing pressure and, consequently, higher MCZ density (Figure 5).

High nutrient concentration in the non-swarm locations helped the large herbivorous copepods C. pauper, Cosmocalanus darwinii, and Undinula vulgaris attain high density in the phytoplankton rich waters, channelling organic carbon and nutrients through the conventional food chain constituted by micro phytoplankton–copepods. These calanoids are preyed upon by pelagic fish larvae and juvenile fish (Costalago et al. 2015), which correlates well with the high abundance of commercially important fish species obtained from the non-swarm area. On the other hand, copepods belonging to the genera Oithona are omnivorous ambush feeders and rely on a wide variety of food sources like phytoplankton, MCZ (ciliates and dinoflagellates), and copepod nauplii (Nakamura & Turner 1997). Their feeding habit is strongly coupled with the microbial loop (Turner 2004). Oithona spp. were the predominant copepods in the swarm locations where MCZ had a lower cell density, indicating the predominance of the microbial food chain in channelling the organic carbon and nutrients to higher trophic levels (Figure 5).

Among the species of ctenophores reported in this study, B. gracilis is a rare species in the BoB and is an invasive species in the Sundarban waterways. The prey–predator interaction between P. pileus and B. gracilis explains the co-occurrence of these species in the swarm area (Greve & Reiners 1988). As reported by Kuipers et al. (1990), the arrival of B. gracilis was noticed at the end of the P. pileus outburst. It is logical to infer that the P. pileus swarm initially caused a decrease in the copepod population as ctenophores are voracious feeders of copepods and started to collapse when B. gracilis predation on the swarm-forming species progressed. It was noteworthy that the MSP community disrupted by one ctenophore is restored by another, thereby shifting the balance back to copepods and phytoplankton (i.e., conventional food chain) (Figure 5). Thus, the present study shows that the NEM-sponsored environmental heterogeneity in the northern BoB can lead to the swarming of gelatinous plankton with tremendous ecological implications since they show extreme tolerance to anomalous environmental conditions (Purcell 2012).

The multivariate analysis of the ichthyofaunal data also suggested two fish assemblages, each characterizing the ctenophore swarm and non-swarm locations. Commercially important fish like Gudusia chapra, O. ruber, H. limbatus, and C. nebulosus were mainly recorded from the non-swarm locations. In contrast, ornamental fish like pufferfish and flatfish with low or no commercial value (Cynoglossus macrostomus, Butis butis, and T. oblongus) were obtained from the swarm area. Mystus gulio, a euryhaline omnivorous fish inhabiting the low saline waters (Gupta 2014), was observed in high abundance in the non-swarm stations, where salinity was low. Besides being used as food, this species is also exported as ornamental fish. Several studies report that gelatinous swarms dominated by ctenophores cause an increase in fish eggs and larvae mortality, leading to collapse of the fisheries (Kideys et al. 2005; Kawahara et al. 2006). Interactions between ctenophores and fish are complex and dynamic in nature. Ctenophores feed on ichthyoplankton, compete with planktivorous fish, and get preyed upon by mature fish species (Stoltenberg et al. 2021). A considerable vertical distributional overlap was reported between ctenophore M. leidyi and cod eggs and sprat eggs. Additionally, gut content analysis proved that M. leidyi fed on fish eggs (Haslob et al. 2007). Moreover, sinking dead ctenophores may cause anoxia in the water column, thereby declining fish stock (Stoltenberg et al. 2021). Thus, it can be deduced that the ctenophore swarm had a deleterious impact on the fish community through competition for food sources and predation.

Given the ecological and socio-economical implications of the ctenophore swarm, future studies in Indian waters should focus on ctenophore diversity, seasonality, and distribution. Seasonal studies will help to understand the impact of environmental factors on ctenophores and the ecological implications of ctenophores swarms on the plankton food web and fisheries.

This study markedly advances our knowledge about the environmental causes of the ctenophore swarm and its effect on the planktic community and pelagic fisheries. The study found that ctenophore swarm was driven by high salinity and ammonium levels and low MCZ density in the estuarine swarm areas. The ctenophore swarm altered the plankton food web structure by increasing MSP density (dominated by small copepods) and favouring the microbial loop. Ctenophore swarm structured the MSP community by modifying carbon and nutrient flow conduits in the plankton food web. The ctenophore swarm also negatively affected the fish community composition by reducing the abundance and diversity of commercially important species. Thus, the findings of the study answered the research questions and supported the hypotheses. The swarm of the ctenophore Pleurobrachia spp. was ephemeral and collapsed with the arrival of its predator, another ctenophore B. gracilis. At the time of sampling, the ctenophore population was migrating from the coastal waters towards the estuarine area. Thus the moderate abundance of P. pileus and P. globosa were also recorded in the coastal and estuarine non-swarm stations. Apart from its indirect negative effect on phytoplankton and MCZ, the Pleurobrachia spp. swarm directly impacted the MSP community, fish composition, and landings. The physical damages like clogging and damaging the fish nets by ctenophores (and other gelatinous plankton) and increasing the fishing time per vessel (and thereby the fuel consumption) also cause a decline in the quality and quantity of fish landings resulting in a substantial economic loss to the local fisher community. From an ecological and socio-economic point of view, it would be worth studying how the ctenophore-steered decrease in MCZ density led to trophic cascades in the plankton food web and influenced the pelagic fishery. Moreover, since our study was based on a single sampling event during the winter season when estuarine outflow was high, it did not capture the temporal or spatial variability of ctenophore swarm occurrence. Further studies are required to monitor ctenophore swarm dynamics throughout different seasons and locations in Sundarbans estuarine and coastal waters.

The study contributes to the understanding of ctenophore swarm ecology in tropical estuarine and coastal waters and its impact on plankton food web functioning and fisheries. The study also provides baseline information for future assessment of ctenophore swarm trends under changing environmental conditions. The study suggests that ctenophore swarms pose a potential threat to Sundarbans fishery resources and livelihoods and calls for effective management strategies to mitigate its negative effects. A dedicated long-term study on zooplankton in response to environmental vicissitudes is needed to understand the link between gelatinous ‘swarm’ events and their impact on ecosystem dynamics. This information will be valuable for ecosystem-based management of marine resources in the face of increased ocean acidification and climate change scenarios. It will contribute to realizing the targets in the United Nations Sustainable Development Goal-14 (SDG-14, Life underwater – to be achieved by 2030).

The authors (A.S., A.B., and J.P.) are grateful to the Director, Zoological Survey of India, for providing facilities to carry out this work and to the Department of Science and Technology-Science and Engineering Research Board, India (DST-SERB) (Project: ECR/2007/000087 and Project: CRG/2020/005212) for funding this study. Council of Scientific and Industrial Research, India (CSIR) is acknowledged for supporting A.S. with Senior Research Fellowship [sanction letter number 09/1181(0003)/2017-EMR-I]. R.M. gratefully acknowledges the financial (salary component) and logistics support from the Environment and Life Sciences Research Center, Kuwait Institute for Scientific Research (KISR), Kuwait. All authors thank the reviewers for their constructive feedback, valuable suggestions, and insightful comments, which helped us to improve the quality of the manuscript. The authors appreciate the editors of Journal of Water and Climate Change for timely handling of the manuscript.

A.S. and J.P. conceptualized the research objective. A.S. and A.B. curated the data sets. A.S., A.B., and R.M. performed the research methodology. A.S. and R.M. analyzed the data, applied software, wrote, and edited the manuscript. J.P. acquired the funding and administered the project. J.P. and R.M. supervised the preparation of the manuscript. D.B. provided resources and facilities to conduct this work.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.