Seasonal Changes in Benthic Fish Population Influenced by Salinity and Sediment Morphology in a Tropical Bay

Approximately 98% of all marine species are supposed to belong to the benthos [1]. Benthic organisms link the primary producers, such as phytoplankton with the higher trophic levels, such as finfish, by consuming phytoplankton and then being consumed by larger organisms. Many benthic fish species rely on other bottom prey as food sources for part of, or in some cases, all of their life history [2] and inturn become the rich source of food for fishes in bathypelagic and mesopelagic zones. Hence, they play a very important role in the marine food web. As benthic fishes are a good source of food for mesopelagic and benthopelagic finfishes, it is very important to know their status in an ecosystem for a better understanding of the fishery management.


Introduction
Approximately 98% of all marine species are supposed to belong to the benthos [1]. Benthic organisms link the primary producers, such as phytoplankton with the higher trophic levels, such as finfish, by consuming phytoplankton and then being consumed by larger organisms. Many benthic fish species rely on other bottom prey as food sources for part of, or in some cases, all of their life history [2] and inturn become the rich source of food for fishes in bathypelagic and mesopelagic zones. Hence, they play a very important role in the marine food web. As benthic fishes are a good source of food for mesopelagic and benthopelagic finfishes, it is very important to know their status in an ecosystem for a better understanding of the fishery management.
Communities of fishes on the continental shelf have rarely been studied beyond the compilation of species list for given areas. Information on benthic fish assemblages is particularly scarce in the Bay of Bengal where demersal fish are heavily exploited as principal targets or as by-catch [3]. Most studies have examined mega faunal assemblages, while, only few studies of fish community structure have focused on patterns of spatial and temporal variation in composition, abundance and distribution of demersal fish assemblages of the continental shelf and slope at higher latitudes [4][5][6][7][8][9][10]. Even though on a world wide scale there are probably many areas of shelf and slope that have not been explored, they were evaluated to a little extent in tropical waters [11][12][13][14]. Yet, few benthic studies edging the Pisces have to be conducted in the tropics compared with higher latitudes. In the Indian context, demersal fish assemblages have been examined by few authors over a few decades in Arabian Sea (west coast of India) [15][16][17][18][19]. In the Bay of Bengal (east coast of India), comparatively less work was done on the demersal fish component [20][21][22]. Even these are mainly directed towards the whole demersal fish (collected with huge trawl nets) community with cursory accounts of ecology and lists of fauna. But studies focusing exclusively on benthic fish (collected with small Naturalist's dredge) component are scarce in both temperate and tropical zones.
The present study area, Nizampatnam Bay (Lat. 15º 28N' to 15º 48' N and Long 80º 17' to 80º 47' E) is a shallow tropical Bay located in the southern province of Andhra Pradesh, Bay of Bengal, East coast of India. This Bay has a coastline of 122km occupying an area of 1825Sq. Km with a maximum depth of 34m. The discharges from the river, saltpans, aquaculture ponds and mangroves built along the coastline are influencing the ecology of the bay fauna.
Hence; a study was conducted to evaluate the seasonal variability in the diversity and faunal associations of macrobenthic population and the effect of different physicochemical parameters on the same. From the available macrobenthos data, benthic fish component was segregated and evaluated for seasonal variability. This paper is the first available data set on the benthic fish species density, diversity, assemblages and influence of environmental factors on benthic fishes from Bay of Bengal, East coast of India.

Materials and Methods
During this investigation, four cruises (FKKD-fishing trawler) were conducted through two successive Post-monsoon seasons Observations on physicochemical characteristics of the seawater temperature, dissolved oxygen and salinity were made using an onboard WTW probe. Sediment samples were collected using a van Veen grab (Hydrobios, Kiel, Germany) having a mouth area of 0.1m 2 . These samples were oven dried at 60 0 C and stored. The sediment texture was estimated using a particle size analyzer Master sizer 2000, Malvin Instruments (USA) for soft sediments and Sieve Shaker, Retsch instrument (Germany) for hard sediments. Organic matter (%) content was determined by the Wet Oxidation method of Walkley-Black using rapid titration procedure but as modified by [21]. Altogether 128 biological (dredge) samples were collected with a Naturalist's dredge. On board, all the organisms were washed with seawater in a large metal tray and after separation the animals were carefully transferred into polythene containers, labeled and preserved in 7% formaldehyde/methylated alcohol for later study. In the laboratory, benthic fishes were segregated, counted (ind. haul-1) and identified up to species level using standard identification keys [23][24][25][26]. Debatable identifications were mostly referred to taxonomists at the Zoological Survey of India, Kolkata.
The data was subjected to univariate analyses carried out in PRIMER 6.1 [27]. Shannon-Wiener diversity (H' Loge), Margalef richness (d) and Pielou's evenness (J') were determined according to routines implemented in PRIMER. Multivariate analysis consisted of estimating Bray Curtis similarity after suitable transformation of sample abundance data. The similarity matrix was subjected to both clustering (hierarchical agglomerative method using groupaverage linking) and ordination (non-metric multidimensional scaling, MDS) using PRIMER 6. Significance tests of sample groupings were made using the ANOSIM (1-way) randomization test. The contribution of each species to groupings noticed in the clustering and ordination analysis was examined by SIMPER (similarity percentages) implemented in PRIMER [28] to quantify percentage contribution of each species to similarity within each group (i.e. characteristic) of samples and to dissimilarity between different groups. Other routines (e.g. BVSTEP), namely stepwise searches of combinations of species considered to be ultimately responsible for the observed pattern in the biotic assemblages, were also carried out using PRIMER. Canonical Correspondence Analysis (CCA) [29] was performed to possible correlations between environmental variables, benthic fish species and variance in site pattern, using a form of stepwise regression. A Monte Carlo permutation test (unrestricted) was used to determine the significance of species environment relations. Simple multiple regressions of environmental parameters against biological indices (H', J and d) were performed in SPSS 14 to explain the extent of variance in the distribution of biological indices influence by a set of environmental parameters.

Sediment characteristics
The predominant sediment texture in the Nizampatnam Bay was clayey silt (11 stations) followed by sandy (5 stations), silty (2 stations) and mixed type (more or less in equal proportions of sand, silt and clay-2 stations) (Figure 3). At stations that lie nearer to the shoreline, the sediments consisted of relatively soft (mud) substrata (sts. 1, 2, 3, 4, 6, 7, 8, 11, 14, 15 and 20).  Mean particle diameter (MPD) of sediment varied between 4.83μm and 888.0μm (153.7±30.9) showing an inclination with depth. From the above observations, it can be concluded that organic matter, % silt and clay declined with inclination in depth, whereas MPD and % sand increased with depth in the Nizampatnam Bay. It is generally believed that sediments with coarse particles are deficient in organic matter whereas fine grain sediments show organic enrichment.

Examines Mar Biol Oceanogr
Copyright © Annapurna C

Volume 1 -Issue -3
Canonical Correspondence Analysis (CCA) implemented in CANOCO was used to ascertain the influence of environmental factors on benthic fish species distribution at the study sites. We found that the distributions were almost the same whether the entire set of species or just the discriminating species identified through BVSTEP analysis were used. The best subset of benthic fish species found by the BVSTEP procedure seemed to be comprised by the same eight species resulted in SIMPER from the list of 30 species encountered from the study area. Monte Carlo permutation tests (with forward selection) were used to identify which environmental variables (out of 9) explained the variance significantly (p<0.05 level) in benthic fish distribution and species pattern. From the CCA, it was found that four ordination axes cumulatively explained 98.6% of the species variance with the first two axes being important explaining 87.9 to 92.0% of the species variance ( Table 3). The eigenvalues for the first two canonical axes were 0.613 and 0.028 respectively. The environmental axis 1 (87.9%), strongly associated with depth (r =-0.716), temperature (r = 0.8411), salinity (r = -0.4249) and dissolved oxygen (r = 0.5043). While axis 2 showed significant correlations with depth (r = 6843) and dissolved oxygen (r = 0.5012), Axis 4 was strongly associated with % sand (r = -0.6642), % silt (r = 0.6705) and MPD (r = -0.6837). From the triplot, it was clearly evident that the benthic fish community of this Bay was segregated into two groups. Group I typified the stations monitored in post-monsoon season constituted the Cociella crocodilus assemblage and Group II contains the stations monitored during pre-monsoon represented by Pisodonophis boro association.

Discussion
Spatial and temporal heterogeneity among macro benthos in Bay environments under tropical conditions has been assigned to events taking place in sediment properties, ambient salinity and such other environmental factors [30]. Several factors like location, depth, and distance from shore, river proximity and oceanographic environmental parameters (bottom water temperature, salinity, dissolved oxygen, sediment organic matter, mean particle diameter and texture) appeared to be important for determining benthic fish distribution patterns in this study. In the Bay water body at Nizampatnam, terrigenous inputs through river discharge, materials flowing out of salt pans, estuaries, mangroves and transport through neritic inflow appeared to determine the hydrography and sediment granulometry that in turn could have influenced distribution of benthic fishes [31].
Altogether 30 species of benthic fish species representing 20 genera, 12 families, 7 orders and 1 class were encountered in this study. The average values of Shannon-Weiner diversity index H' recorded during post-monsoon was 1.3±0.4 and in pre-monsoon was 1.2±0.3. The benthic fish diversity of Nizampatnam Bay can be considered as poor as the mean H' recorded in this study was 1.3 and 1.2. Buchanan [32], Longhurst [33], Desai & Kutty [34] have also indicated that species diversity was poor or moderate in tropical benthic communities they studied.
Using CCA routine implemented in CANOCO, benthic fish communities were linked with environmental variables (depth, bottom water temperature, salinity, dissolved oxygen, sediment texture, mean particle diameter and organic matter). The first axis of the CCA had an eigenvalues of 0.613, implying a large percentage (87.9) of explained variance [28,39]. Together, the four ordination axes cumulatively explained 98.6% of the species variance with the first two axes being important explaining 87.9 to 92.0% of the species variance ( Table 3). The noteworthy feature, however, is the high correlation (weighted correlation coefficient >0.4) between faunal abundance and environmental variables on all 4 CCA axes. From the triplot in CCA ordination ( Figure 5), it is clearly evident that the distribution of Pisodonophis boro, Trichonotus arabicus, Acentrogobius cyanosmos and Pseudorhombus elevatus (Group II, pre-monsoon) assemblage seemed to prefer high saline waters and deeper areas with coarse sediments. The Nizampatnam Bay's salinity varies widely from season to season and from year to year, depending on the amount of fresh water flowing from its rivers. During drier months (pre-monsoon), the Bay is usually saltier. Salinity also increases with depth. Fresh water remains at the surface because it is less dense than salt water. The water on the Bay's eastern shore tends to be saltier than water on the western side. This is due to two factors: Most fresh water enters the Bay from its northern and western tributaries, through the Krishna River. So, saltier water moving up the Bay veers toward the eastern parts of the Bay. These high saline waters available in the deeper zones of this Bay favoured the survival of Pisodonophis boro, Trichonotus arabicus, Acentrogobius cyanosmos and Pseudorhombus elevatus association which is sensitive to salinity fluctuations. These four fishes that occur in the pre-monsoon are predatory as they are surviving in sediments with high sand proportions dominated by epifauna benthos [40].
In contrast, Cociella crocodilus, Cynoglossus lida, Cynoglossus punticeps and Astroscopus zephyreus (Group I, post-monsoon) association has a negative relation with salinity and preferred the locations with soft (high mud clay and silt proportions) substratum. During the post monsoon, the fluctuations in the bottom water parameters (salinity, temperature and dissolved oxygen) were high due to different mixing and circulation patterns resulted by the influence of freshwater influxes from terrestrial systems. It is obvious that, the sediments of the stations that lie nearer to the coast were clayey silt, whereas stations that lie away from the shore were dominated by sand. This point reflects the influence of river water inflows into the Bay during post-monsoon. Consequently, the Bay substratum exhibited clayey-silt and organic matter rich with sediments in post-monsoon and in the stations that lie nearer to shoreline. Cociella crocodiles, Cynoglossus lida and Cynoglossus punticeps was mainly adapted to a bottom habitat feeds on polychaetes and detritus, amphipods, copepods and small molluscans [41]. In Nizampatnbam Bay, the abundance of infaunal benthic organisms was higher during post-monsoon than pre monsoon. These infaunal benthic organisms of Nizampatnam Bay were dominated by amphipods and polychaetes [40]. The terrestrial organic matter by land drainage and river runoff could contribute significantly to the overall load of organic matter [42][43][44]. Several workers documented that the relationship between the grain size and sediment organic content showed a negative correlation [45][46][47]. These organic matter enriched sediments favoured the survival of infaunal which inturn favoured the survival of Cociella crocodilus, Cynoglossus lida, Cynoglossus punticeps and Astroscopus zephyreus during post-monsoon. These four fishes can also tolerate wide ranges of fluctuations in bottom hydrography as all these four fishes were habituated to survive in marine and brackish water [48]. Hence, from the above findings, it is clearly evident that differences in the distribution of benthic fish assemblages in Nizampatnam Bay were influenced by monsoonal changes in the salinity of bottom water and sediment morphology. Monte Carlo permutation tests confirmed the significant association (p<0.05) between environmental variables and the benthic fish species distribution off Nizampatnam Bay, Bay of Bengal.