Characterization of Fish Community Structure And Development of An Estuarine Fish Community Index For Temporarily Closed Estuaries (TCEs) From India’s Western Coast

We present for the rst-time data on sh assemblage structure for ten temporarily closed estuaries (TCEs) along the western coast of India. Fish community structure differed signicantly between TCEs based on species presence/absence, estuarine use and feeding mode - differences based on estuarine surface area and geographical position (northern and southern locations). Marine migrant and freshwater stragglers dominated the ‘sh guilds’ in all estuaries, while piscivores and zoobenthivores were the major feeding guilds. We used the estuarine sh community index (EFCI) and a combined anthropogenic pressure index (CPI) to determine ecological quality of the TCEs. The application of functional guilds and EFCI could represent the use of TCEs by sh communities, and functional similarities existing between sh assemblages of these estuaries, despite considerable taxonomic, physical and chemical differences. Analysis of EFCI indicated that the ecological health of TCEs in south-western India ranges from “poor” to “good”. Through this index, we also established a signicant relationship between CPI, EFCI and ecological quality in various TCEs (EFCI = A + (b X CPI)). Thus, the EFCI provides an integrated measure of the ecological status of sh community for TCEs along Indian coast. 2) to develop and compare EFCI and CPI to investigate the health status of TCEs and 3) to establish a relationship between the EFCI and anthropogenic stressors identied in TCEs. Overall, our results would help improve the management of poorly-known TCEs from India, which are important not only for food security and livelihoods, but also for a range of ecosystem services.


Introduction
Temporarily closed estuaries (TCEs) are microtidal estuarine systems, which remain open to the sea for a short period of time are very common in tropical region (Whit eld 1992; Elliott et al. 2007). TCEs are more variable and complicated ecosystems than permanently open estuaries due to the frequent alteration between open and closed mouth phases : Human et al. 2016). The mouth state of TCEs is mainly driven by the interaction between sediment transport as a result of waves, tides or stream in ow (Harrison and Whit eld 2004;. As a result of frequent variation in physico-chemical parameters, TCEs are more sensitive to changes in ow conditions caused by anthropogenic activities such as sewage e uent discharges (Lawrie et al. 2010). Mouth closure events associated with high river-ow cuts off tidal exchanges (i.e. the scouring force of high river ow which helps keep the mouth open) with the ocean, resulting in extended periods of conditions under which salinity and temperature strati cation can develop, as well as in the depletion of oxygen and nutrients create considerable physiological demands of the shes that occupy these systems (Whit eld, 1999).
The predominance of estuarine and estuarine-associated marine sh species (which caters for both dependant and nondependent taxa) in TCEs is an indication of the important nursery function of these systems, as it serves as migration routes and refuge areas for a variety of sh species (Whit eld 1999; Elliott et al. 2002). Marine juvenile sh recruit into TCEs not only when the mouth opens, but also during marine overwash events when waves from the sea wash over the sand bar at the mouth (Cowley and Whit eld 2001; Whit eld and Bate 2007). The longer open mouth phase, together with the greater habitat variation in large TCEs, promotes high sh diversity within these systems (Vorwerk et al. 2003). Unfortunately, as a result of global change, TCEs are now undergoing rapid and critical changes that, in some cases, have already led to the loss, or the signi cant deterioration of their natural condition (Human et al. 2016). Land use and erosion in catchments, barriers, impoundments and uncontrolled water abstractions, mining operations, eutrophication, chemical pollution, overexploitation of both living and non-living resources, and alien invasive species are threatening the integrity of TCEs in the tropics (Human et al. 2016). Addressing the challenges facing sustainable management of TCEs is therefore critical, as in some cases the ecological integrity, biodiversity and nursery function of such estuarine systems have already been compromised (James et al. 2007).
The TCEs in India, particularly along the south west coast is poorly known with regard to their status, threats and sh community structure, except for a few preliminary studies (Bijukumar and Sushama 2000;Johny et al. 2016; and Kiranya et al. 2018), which are however restricted to a general description of spatial and temporal variations in species diversity, and mostly limited to a single estuary. No comprehensive studies have yet been undertaken to quantify sh assemblage structure, habitat quality and the in uence of anthropogenic stressors on sh communities in these TCEs. For many TCEs, the guild approach could represent a simple tool to de ne the extent of marine in uence and freshwater discharge, and the estuarine sh community index (EFCI), a simple and composite index de ned separately for permanently open, and temporarily closed estuaries can help determine the ecological quality of estuaries (Harrison and Whit eld 2004). In general, the guild approach is useful to determine functioning, hierarchical structure and connectivity and to simplify complex ecosystems (Dewailly 1995;Albaret et al. 2004;Garrison and Link 2000;Lobry et al. 2008;Franco et al. 2006). Guilds have a high value as a tool for understanding the functional structure of complex ecosystems such as estuaries (Garrison and Link 2000), especially about estuarine sh assemblage studies (e.g., Dewailly 1995). Such studies have been undertaken in estuaries of South Africa (Harrison and White eld 2006) and Europe (Dewailly 1995). Globally, the use of sh as an index for assessing the quality of transitional water bodies such as estuaries are well documented (Pérez-Dominguez et al. 2012). The sh-based multi-metric indices for monitoring the environmental quality was initially proposed by Karr (1981) Multi-metric indices are constructed from an array of sh ecological attributes and are therefore, considered to be superior to single metric assessments (Perez-Dominguez et al. 2010;Fonseca et al. 2013). In particular, these indices are considered to have greater relevance to different ecological regions (Fausch et al. 1990). The EFCI is a novel index for monitoring the status of ecological condition of the estuaries on the verge of deterioration from anthropogenic and natural pressures, as well it is also a tool to document the recovery of estuaries due to rehabilitation and restoration measures implemented by management authorities, (Roset 2007) due to the absence of rehabilitation or restoration activities in the selected TCEs, the EFCI could serve as a tool to evaluate the ecological condition of the same.
The aim of our study was to improve the knowledge of TCEs along the south west coast of India through ful lling the following objectives 1) to delineate similarities/dissimilarities between sh guild (estuarine use and feeding mode) structure of TCEs, 2) to develop and compare EFCI and CPI to investigate the health status of TCEs and 3) to establish a relationship between the EFCI and anthropogenic stressors identi ed in TCEs. Overall, our results would help improve the management of poorly-known TCEs from India, which are important not only for food security and livelihoods, but also for a range of ecosystem services.
Material And Methods 2.1. Study area Ten TCEs located along the south-western (Kerala) coast of India, Ponnani, Chettuva, Kodungallur-Azhikode (Northern region: between 10.1º and 10.7ºN) and Vembanad (Central region: between 9.5º and 9.7ºN), Veli-Akkulam, Kadinamkulam, Akathumuri, Anchuthengu, Poovar and Poonthura (Southern region: between 8.5º and 8.7ºN) ( Fig. 1) were selected for the study. The size, origin, depth, nature of the adjacent freshwater and marine habitat, ecological features of each estuaries were illustrated in Table 1. They have some form of narrow connection to the Arabian Sea, and also simultaneously receive freshwater input from adjacent river systems. In all these estuaries, tides are semi-diurnal and microtidal, with average tidal range between 1 to < 2m, except for the northern estuaries (Ponnani and Kodungallur-Azhikode), which are meso-tidal with an average tidal range between 2 to 3m. Proximity of these estuaries to urban settlements triggers various anthropogenic stressors in addition to large scale seasonal changes in their biotic and abiotic components (Coast et al. 2020). Many of these estuaries are ecologically important, and have international signi cance, especially the Vembanad estuary, which is designated as a Ramsar site (MoEFCC (2020). Despite the variation between sampling sites, all TCEs included in this study are, by virtue of their location, relatively shallow and small except the Veli-Akkulam estuary. Relevant physiographic, biological and ecological information for the ten TCEs were gathered and presented in Appendix A and B. A master checklist was compiled by combining all the species observed from the ten TCEs using presence-absence data. For small coastal ecosystems such as estuaries, using sh guilds is more e cient than using taxonomic identities (Elliott et al. 2007) to analyse standard features of ecosystems and community functioning. Accordingly, for comparing the ecological structures of sh assemblages between estuaries, sh species were grouped into guilds based on their estuarine use and feeding mode as de ned by Elliot et al. (2007), and Dewailly (1995). Guilds based on estuarine use comprised of estuarine resident species (ES), freshwater migrants (FM), marine migrants (MM), and marine stragglers (MS), anadromous (AN) and freshwater stragglers (FS) (see Appendix C for de nitions). Feeding mode functional groups consisted of zoobenthivores (ZB), herbivores (HV), omnivores (OV), zooplanktivores (ZP), detritivores (DV) and piscivores (PV). Cluster analysis, using the Bray-Curtis similarity measure, was used for estimating similarity between estuarine sh assemblages based on species presence/absence, estuarine use guild and diet categorization guild. Software package Primer 6 was used for determining the similarity measure ( (2004) de ned a total of 14 metrices to evaluate the estuarine integrity, which covers four broad sh community aspects, viz., (1) species diversity (ii) species abundance (iii) nursery function and (iv) trophic integrity. Description of these matrices is as follows: 1) species richness represents the number of sh fauna present in the ecosystem, 2) rare or threatened species is a measure of the number of rare or threatened species, 3) introduced species identi es the threat to native sh populations caused by non-native species (Harrison and Whit eld 2004), 4) species composition, a measure of the amount of similarity in the sh assemblage composition (Harrison and Whit eld 2004), 5) species relative abundance quanti es the relative proportion of species in a TCEs with a reference estuarine sh community, 6) number of species that make up 90% of the abundance is a measure of dominance, 7) number of estuarine resident species, 8) number of estuarine-dependent marine taxa quanti es how successfully an estuary ful lls its role as a nursery habitat, 9) relative abundance of estuarine resident species, 10) relative abundance of estuarine-dependent marine species (both 9 and 10 are complementary to quantitatively assess estuarine habitat quality and nursery function), 11) number of benthic invertebrate feeding sh taxa provides an indirect measure of the condition of benthic invertebrate fauna, 12) number of piscivorous taxa was included on the basis that diverse and abundant top carnivores are typically representative of the broader trophic network within estuaries (USEPA 1997), 13) percentage abundance of benthic invertebrate feeding shes and 14) percentage abundance of piscivorous shes (both 13 and 14 provides a quantitative, and complementary analysis of trophic integrity).

Reference conditions and metric calculation
Each metric varies in their measurement from numbers of taxa, to similarity values, and measures of abundance. To effectively combine these measurements into a single index, it is necessary to transform the metrics into standard scores. Each metric was allocated a discrete score between 1 and 5 according to degree of deviation from the reference (Appendix D). Metric scores and thresholds were established using a combination of a review of other metrics that vary in their measurement from numbers of taxa, to similarity values and measures of abundance. To effectively combine these various measurements into a single index, it is necessary to transform the metrics into standard scores. Each metric was therefore allocated a discrete score between 1 and 5 according to the degree of deviation from the reference. Metric scores and thresholds were established using literature from tropical estuaries (Lepage et al. 2016), and personal discussions with experts in estuarine research in the study region. Thresholds xed for evaluating the habitat quality were (1)  A simple composite pressure index (CPI) for an estuary was calculated as a summation of corresponding scores of various pressures. Scores were given based on the severity of anthropogenic pressure in each estuary. The following indicators were quanti ed: 1) habitat loss (including a percentage reduction in bathymetry and topography, 2) water pollution, including nutrient enrichment and eutrophication (in %), 3) destructive shing and overexploitation, 4) climate changeinduced pressures including extreme climatic events, 5) alien invasive species, 6) tourism and recreation, 7) barriers for sh migration, and 8) sand mining. The least score of 1 was given for those pressures that range from 10 to 30%, score 3 and 5 were allocated if these pressures were in the range of 30-60% and 60-90% respectively. Absence and presence of an individual pressure were scored 1 and 5, respectively. Thus, the anthropogenic pressures identi ed in different estuaries (a total of eight stressors) were scored on a scale ranging from 1 to 5 (pressure index). The CPI was estimated (statistically) to determine any signi cant relationship between EFCI and anthropogenic pressures. Final pressure index or the CPI was calculated for the entire TCEs as the sum of all indicator scores, and were within values of 0 (no disturbance) and 72 (very high disturbance). To determine the relationship between CPI and EFCI, Pearson correlation coe cient was used. A simple linear regression model was also tted following the least square method using the formula,

Taxonomic distribution
Total of 212 sh species within 64 families were recorded from the ten TCEs covered in this study (Appendix E). Of these, 126 species were reported from at least one of the ten estuaries, and 44 species were found exclusively in a single estuary (PON). Five species, Etroplus suratensis, Pseudetroplus maculatus, Oreochromis mossambicus, Mugil cephalus, and Megalops cyprinoides, were common in all the TCEs. Cichlidae, Ambassidae, Megalopidae, Mugilidae, and Carangidae were the common families recorded in all the estuaries, with the number of families ranging from 17 (in KAD) to 50 (in PON).

Taxonomic similarity
Similarity in taxonomic composition between estuaries was determined using binary (presence/absence) and quantitative (percentage composition) data. The NMDS plot revealed the presence of four distinct clusters (Fig. 2). Similarly, the analysis of species' presence/absence data based on cluster analysis showed a clear division into four separate groups of estuaries with a high level of dissimilarity between the groups (Fig. 3). First group comprised of two estuaries, PON and CHT (35% similarity in taxonomic composition), second cluster comprised of two estuaries (KOD and VEM), which were more similar in their species composition (55%), the third cluster comprised of three estuaries (AKA, POV and VEL (with 50% similarity) The fourth cluster included three estuaries, ANC, PT and KAD, where KAD and PT found more similar in species composition (60%), while ANC showed 45% similarity in taxonomic contribution.

Analysis of functional guilds
Estuarine use guilds Cluster analysis of estuarine use guilds demonstrated a clear division of estuaries into four groups with low dissimilarity (Fig. 6). First group comprised of three estuaries; KOD, CHT and VEM, all of which were relatively dissimilar within the group, the second cluster also included three estuaries, AKA, ANC and PT, which were similar in their ecological guild composition, the third cluster comprised of three estuaries; KAD, VEL and POV. A single estuary, PON formed the nal cluster, which was dissimilar in guild composition compared to the other three clusters.
Feeding mode guilds Relative proportions of feeding mode guilds varied among the estuaries (Fig. 4.B) -major guilds being PV (31.5%), ZB (24.6%), OV (16.8%), DV (11.4%), HV (9.5%) and ZP (6.3%). NMDS plot for feeding mode guilds revealed the presence of four clusters (Fig. 7). Estuaries POV, VEL, PT, ANC, AKA and KAD formed the rst cluster and KOD and VEM comprised the second cluster. The third and fourth clusters consisted of single estuary each -CHT and PON, respectively. Cluster analysis of feeding mode guilds showed a clear division into four separate groups with high dissimilarity (Fig. 8). The rst group consisted of four estuaries, CHT, KOD, VEM and AKA of which two viz., VEM and AKA were relatively similar within the group. Second cluster comprised of three estuaries, VEL, ANC and PT, which were identical in their feeding mode functional group distributions. Third cluster comprised of only two estuaries KAD and POV, while PON was the only estuary that was classi ed into the fourth cluster.
Analysis of habitat quality EFCI was measured for de ning the ecological quality of the ten TCEs considered in this study. Through this index, we established signi cant relationships between anthropogenic stressors and EFCI. In addition, scoring thresholds were also identi ed for each metric per estuary, representing signi cant variations with increasing stressors over time. Major anthropogenic factors and extent of each stressor, which in uences the habitat quality of the TCEs of Kerala were also analyzed. In VEM, the major anthropogenic stressor was habitat loss (60-80%), in PON, destructive shing and overexploitation contributed to 30-60% of the anthropogenic stressors, and in VEL, both habitat loss and water pollution were the major stressors (up to 60%). Barriers for sh migration was detected in CHT, KOD and VEM. Stressors such as tourism and recreation activities, and alien invasive species were observed in all TCEs. The pressure, from alien species invasion was highest in VEL. The composite pressure index (CPI) yielded a minimum score between 16 (KAD, ANC, POV, PT and AKA) and 24 (VEM) ( Table 2). Estuaries along the northern coast had greater values of CPI than those along the southern coast. Simple regression analysis revealed that EFCI has a signi cant relationship with the pressure index (R 2 = 0.31) (Fig. 9). Based on EFCI, it was realized that three estuaries, PT, AKA and POV were in "moderate" state; Four estuaries, VEM, PON, CHT and KOD were under 'good" condition, while remaining three estuaries, ANC, VEL and KAD were in a "poor" state (Table 3). Table 2 Description and values used to score anthropogenic pressures in the ten TCEs covered in the present study (codes as per

Discussion
Studies on sh guild structure and ecological quality of TCEs estuaries are scarce in the tropics, and therefore the present study, helps bridge this knowledge gap. A major constraint of this study was data comparability, i.e., the underlying data use for the analysis were collected over different time scales by different authors and involved varying objectives, methodologies and analysis, all of which might act as potential sources of variability between the datasets. Therefore, to overcome the bias in data quality and quantity, we attempted to present the sh assemblage structure of TCEs by following a guild approach using estuarine use, and feeding mode, as indicators.

Species diversity in TCEs
Total Kok and Whit eld 1986). Thus, it can be assumed that the low sh diversity observed in most TCEs studies might be due to one or a combination of the above reasons.
Mugilidae was the most speciose family (S = 32), likely due to its successful ability to recruit during over-wash conditions when the estuarine mouth is closed (Cowley and Whit eld 2001). Dominance, as well as numerical abundance of shortlived estuarine resident species (E. suratensis and P. maculatus), were evident in all TCEs of Kerala, a phenomenon reported earlier in other parts of the world (Potter et al. 1990), that of estuarine resident species adapting and dominating estuarine sh communities, numerically. Oreochromis mossambicus was the only freshwater migrant and exotic species recorded in all TCEs of Kerala, being numerically abundant at all sites, and posing a great risk to native species (Regi and Biju Kumar 2018).
The invasion of Clarius gariepienus was reported in the Veli-Akkulam estuary by Regi and Bijukumar (2012). Though C. gariepienus cannot tolerate salinities above 10 psu (Whit eld 1996), reduced salinities (< 15 ppt) associated with high river ow allows this species to enter and colonize estuarine habitats (Harrison and Whit eld 2004). In this study, the presence of elasmobranchs such as Carcharhinus limbatus and Scoliodon laticaudus was reported only in a single estuary (Ponnani), and could be likely due to the high salinity ingress into the estuarine mouth. Several individual estuaries covered in this study such as Ponnani, Anchuthengu, Veli-Akkulam, Akathumuri and Poovar also contained a large percentage of juveniles of estuarine residents as well as marine migrants. It has been suggested by Cowley and Whit eld (2001) that, recruitment of juveniles into estuaries take place due to barrier over-wash resulting from tidal currents. In terms of taxonomic composition, Poonthura, Akathumuri, Kodungallur-Azhikode and Vembanad had higher similarity compared to other estuaries, as illustrated by the results of the cluster analysis and NMDS. Separation of these two gradients might have a result of by coastal deposition and anthropogenic pressures, as has been observed elsewhere by Cardoso et al. (2011).

Analysis of functional guilds
In the cluster analysis, our results showed that Veli-Akkulam, Kodungallur-Azhikode, Chettuva, Ponnani, Kadinamkulam, and Akathumuri estuaries were relatively dissimilar in their estuarine-use guild composition within the group, while Anchuthengu and Poonthura were similar. Despite the dissimilarity between the sh assemblages of the above-mentioned estuaries, sh diversity was higher in the Ponnani estuary, which is larger area in comparison to the others. This is consistent with the fact that the estuarine surface can directly in uence estuarine sh assemblages ( Estuarine residents were represented in all TCEs investigated in our study, dominating the sh assemblage in terms of numbers, but with relatively low number of species, similar to previous observations (Selleslagh et al. 2009;Franco et al. 2006). On the other hand, freshwater migrants contributed very low proportions to assemblage structure and were absent in Poonthura, Anchuthengu, Kadinamkulam and Ponnani estuaries. This observation is similar to those made in South African TCEs (Harrison and Whit eld 2008), and could be attributed to the increase in salinity associated with estuarine mouth opening limiting the appearance of freshwater species (Marshall and Elliott 1998;Dolbeth et al. 2013). Oreochromis mossambicus, Ompok bimaculatus, Xenentodon cancila and Aplocheilus lineatus were most common freshwater migrants (FM) recorded in all the TCEs.
Among the TCEs of Kerala, anadromous species (AN) were found to occur in Poonthura, Vembanad, Ponnani, Chettuva and Poovar estuaries. Elliott et al. (2007) stated that marine migrant opportunists and marine stragglers inhabit estuarine areas where salinity is high, and marine stragglers (MS) occur inside the estuary in lower densities, due to a small area and mouth opening (Dolbeth et al. 2013). It is thus presumed that the occurrence of marine stragglers in various TCEs of Kerala is decided by these factors. The feeding guild system was designed to allow the aggregation of sh species that utilize similar food resources (Elliott et al. 2007). With respect to feeding mode functional groups, an overall dominance of piscivorous taxa (81 species) followed by zoobenthivores (53 species) was observed in the TCEs investigated in the current study. Some of the dominant piscivore guilds were represented by members of the family Carangidae, Ariidae, Sciaenidae, Epinephelidae, Clupeidae and Elopidae.
The guild, ZP was the least represented guild (15 species) and occurred in only two estuaries viz., Anchuthengu and Poovar (single species in each TCEs), a contrasting pattern to those in the Forth estuary (UK) (Mathieson et al. 2000), where, planktivorous species (PS) dominated the sh assemblage. Franco et al. (2006) demonstrated that European estuaries were dominated by invertebrate feeders (IV), with invertebrates and sh feeders dominating most sh assemblage in Portuguese estuaries (Cardoso et al. 2011). A smaller representation of the ZP guild, in contrast to the high number of piscivores (PV) (81) reveals the presence of juveniles of marine migrants and other guilds in all TCEs of Kerala due to marine connectivity and reduced freshwater in ow (Bijoy Nandan et al. 2012). The strong marine in uence that characterizes tropical estuaries probably provides reasonable access for marine piscivores to establish in the estuarine system. In this study, we found that the estuarine ichthyofauna of TCEs was dominated by piscivores and the presence of apex predators were absent in all the estuaries except Ponnani. Whit eld (1999) stated that the individuals that enter estuaries are usually juveniles whose diet is mainly composed of sh. Many Indian authors reported the abundance of piscivorous taxa in TCEs associated with the breaching of estuarine mouth (Grace, 2014: Kiranya et al. 2018), and so it can be assumed that the frequent breaching of the TCEs (both naturally and arti cially) would lead to the abundance these taxa in almost all estuaries. Although the estuarine ecosystem mainly supports detritus food chain, many piscivores shes depend on the estuaries for completing their larval or juvenile stages ((Morson et al. 2019).

Analysis of habitat quality
A sh-based index, EFCI was determined for de ning the habitat quality of TCEs of Kerala. The use of sh community analyses as indicators of ecological health through the implementation of the EFCI has proven to be extremely useful in this study. The EFCI approach described in this study also represents a simpli ed method where values are set and evaluated directly according to the level of anthropogenic pressure using regression analysis, thus exhibiting a signi cant relationship with anthropogenic pressures. The EFCI scores in this study ranged from 34 to 56, indicating that the ecological health of the TCEs ranges from "Poor" to "Good" state. Harrison and Whit eld (2004) determined the EFCI for the Sezela estuary (South Africa) and reported the range of EFCI value between 30 and 46. In this study, only four estuaries, Vembanad, Kodungallur-Azhikode, Ponnani and Chettuva were in 'Good' state. Estuaries of least concern such as Veli-Akkulam, Kadinamkulam and Anchuthengu were in 'Poor' state. Estuaries such as Poonthura, Akathumuri and Poovar were assessed to be under 'Moderate' state but are likely to move to 'Poor' state in the absence of management measures or policies to protect these fragile ecosystems. Several studies have suggested the alarming decline in Kerala's estuarine biodiversity primarily due to unsuitable and unethical shing practices (Regi and Bijukumar 2012; Kiranya et al. 2018).
Large scale decline in health of estuaries in Kerala and reduction of sh stocks have occurred as a result of habitat degradation, water pollution, including nutrient enrichment and eutrophication, destructive shing and overexploitation, climate change-induced pressures including extreme climatic events, invasive alien species, tourism and recreation, barriers for sh migration and sand mining (Personal communication). In TCEs investigated in the present study, pressures such as water pollution, tourism and recreation activities, destructive shing and overexploitation, and invasive alien species showed signi cant impacts, contributing to 10-60% of the total stressors. The pressure index, CPI also revealed variations between the TCEs. We also tried to estimate a simple linear regression between EFCI and CPI to see the impact of anthropogenic pressures on EFCI. Lepage et al. (2016) also used this correlation to establish the impact of pressures on sh community structure and ecological quality. The relationship yielded satisfactory results, and EFCI and CPI were positively correlated. Though the regression t was below reasonable estimates (R 2 = 0.3, might be a result of the limited number of estuaries analyzed here), the impact of anthropogenic pressures on EFCI was evident in the regression plot.
The estuary most affected by invasive alien species was Veli-Akkulam, which is also subjected to large-scale habitat degradation (Regi and Bijukumar 2018). The analysis of EFCI also con rmed the "poor state of the estuary. In Vembanad Lake, habitat loss is the primary anthropogenic stress factor, negatively affecting 60-80% of the ecosystem; however the estuary was revealed to be in a 'good state" according to the EFCI score. Compared to all TCEs, Ponnani estuary is the only one that suffers from destructive shing activities, but the EFCI value indicated that the estuary is in a 'good state" mostly due to the remarkable resiliency of estuarine species. Among the ten estuaries, barriers of sh migration were noticed in Chettuva, Kodungallur-Azhikode and Vembanad, while the remaining seven estuaries were free from any barriers to sh migration.

Conclusion
Our work highlighted the critical role played by temporarily closed estuaries (TCEs) in maintaining sh assemblages. Ichthyofaunal community structure in small (e.g., Poonthura, Poovar, Kadinamkulam etc.) and large (e.g., Vembanad and Ponnani) TCEs differed signi cantly. Variations in functional guilds were found in all of the approaches used to analyze sh assemblages such as species presence/absence, species composition, diversity, evenness, family level composition, and structure of sh guilds (estuarine use and feeding mode). Though the analysis presented here is preliminary, the EFCI and CPI classi ed TCEs of Kerala as poor, moderate and good. Proper pollution abatement measures and enforcement of environmental protection measures should be planned and implemented to restore/improve the habitat quality of the moderate and poor TCEs, besides long-term monitoring of sh faunal composition as indicators of ecosystem health.