Temporal and spatial distribution of fish and shrimp assemblage in the Bakkhali river estuary of Bangladesh in relation to some water quality parameters

Abstract

Fish and shrimp species, together with water quality data, were collected from two different stations located inside the Bakkhali river estuary of Bangladesh during winter, premonsoon and monsoon periods. Significant temporal differences were observed for water temperature, salinity and dissolved oxygen. The average catch of fish and shrimps per net between stations varied between 1.89±0.36kg at station 1 and 7.54±4.39kg at station 2, while the average catch in winter, premonsoon and monsoon periods found to be 2.79±1.08kg/net, 6.31±1.03kg/net and 5.06±2.89kg/net, respectively, with a significant difference in catch per net between stations although no significant difference in catch per net was observed between seasons. A total of 18,467 individuals of fish (35 species) and shrimp (10 species) were found in the present study. Three species of shrimps were observed to be dominant (>10.0%) and these were Metapenaeus lysianassa (17.07%), Ambassis dussumieri (14.54%) and Macrobrachium villosimanus (12.13%). Clear differences in faunal abundances were observed between seasons and stations with higher mean abundances during winter (1747.83±421.99 individual/5kg) and at Station 1 (1444±866.74 individuals/5kg). Similarly, the diversity indices, both Shannon–Wiener and Margalef, showed significant differences between stations and seasons (except Shannon for stations). Analyses of similarity (ANOSIM) results confirm both spatial and temporal differences in species community structure with a highly diverse assemblage. Canonical Correspondence Analysis results indicated that salinity and transparency were the main variables influencing fish and shrimp distribution in the Bakkhali river estuary.

Keywords:

• Ambassis dussumieri
• analysis of similarity
• Bakkhali river estuary
• Metapenaeus lysianassa
• species assemblage

Published in collaboration with the University of Bergen and the Institute of Marine Research, Norway, and the Marine Biological Laboratory, University of Copenhagen, Denmark

Introduction

Estuaries are transition zones between sea and freshwater; they are occupied by a combination of freshwater and marine species as well as juveniles (Claridge et al. 1986). They serve important economic functions including transportation, industry and tourism, but also drainage of waste from domestic, industrial and agriculture activities (Heip & Herman 1995; Raz-Guzman & Huidobro 2002). Simultaneously, these ecosystems offer protection, not only for resident species, but also for a wide range of marine and freshwater species which migrate there at certain stages of their life cycle (Weinstein 1985; Weisberg et al. 1996; Cowley and Whitfield 2002; McLusky and Elliott 2004; Blaber 2000). Fish assemblage structure of estuaries is characterized by high diversity and high abundance, especially for juveniles (Whitfield 1999). An examination of the ecological factors is important in defining habitats for fishes and has been the main focus of many previous studies (Able 1999; Martino & Able 2003). Most estuaries are characterized by high biological productivity associated with relatively extreme and varying environmental conditions (Day et al. 1989; Kennish 1990; Whitfield 1999). As boundary systems between watersheds and the sea, estuaries exhibit environmental gradients that favour the recruitment of a variety of species with diverse physical and trophic structures (Sánchez & Raz-Guzman 1997; Harris et al. 2001; Kimmerer et al. 2001). Since estuaries serve as nurseries for many commercially important fish and crustaceans (Shenker & Dean 1979; Weinstein 1979; Rakocinski et al. 1996; Blaber 2000; Elliott & Hemingway 2002; Akin et al. 2003), it is necessary to examine the environmental factors that shape the species assemblage structure. Fishes play an important role in estuaries as they constitute permanent and temporary community components, with marine species visiting these habitats for feeding, reproduction, growth and protection (Raz-Guzman & Huidobro 2002). The distributions of fish within biologically and physically complex estuarine systems may be influenced by many mechanisms. Several estuarine ecologists have pointed out that biotic processes, such as competition and predation, may be important in driving the occurrence of spatial and temporal patterns of fish abundance and assemblage in estuaries (Holbrook & Schmitt 1989; Ogburn-Matthews & Allen 1993; Lankford & Targett 1994; Barry et al. 1996).

By nature, estuarine habitats are highly productive (Nixon et al. 1986; Day et al. 1989) and their role as nursery grounds for fishes is well documented for temperate (Powles et al. 1984; Elliott et al. 1990; Kennish 1990; Drake & Arias 1991; Szedlmayer & Able 1996; Whitfield 1999; Blaber 2000; Shackell & Frank 2000; Elliott & Hemingway 2002) and tropical regions (Raynie & Shaw 1994; Sanvicente-Añorve et al. 2000; Harris et al. 2001; Cowley & Whitfield 2002; Franco-Gordo et al. 2003). Several biological and abiotic factors affect the occurrence and habitat of fish and shrimp within estuaries. These factors include salinity, temperature, turbidity, dissolved oxygen (DO), freshwater inflow, structural attributes of habitat, depth, geographic distance from the estuary mouth, and hydrography (Gunter 1961; Blaber & Blaber 1980; Weinstein et al. 1980; Rogers et al. 1984; Zimmerman & Minello 1984; Thorman 1986; Peterson & Ross 1991; Sogard & Able 1991; Cyrus & Blaber 1992; Rakocinski et al. 1992; Cowen et al. 1993; Everett & Ruiz 1993; Szedlmayer & Able 1996; Fraser 1997; Maes et al. 1998; Marshall & Elliott 1998; Araújo et al. 1999; Wagner & Austin 1999; Whitfield 1999; Hagan & Able 2003; Jaureguizar et al. 2003; Martino & Able 2003).

The assemblages of fish in estuaries are variable both in terms of species composition and distribution patterns (Harris et al. 1999). Changes in species assemblage are continuous, according to reproductive seasons of the species and the environmental fluctuations (Whitfield 1994; Harris & Cyrus 1995; Hettler & Hare 1998; Garcia et al. 2003). However, through discussion with local fisherman from the Bay of Bengal, Bangladesh, during the present study, there seems to be a general tendency for estuarine fish larvae to peak in abundance during the monsoon in this region. Furthermore, Hossain et al. (2007) also reported a similar trend for juvenile fish species at Naaf river estuary. Fish and shrimp assemblage structure in the estuaries of Bangladesh has not been well studied; although there are some scattered works on different biological aspects of the coastal estuarine system of Bangladesh (Hossain et al. 2007), none of them examined the species assemblage structure.

The Bakkhali river estuary located at the southeastern part of Bangladesh is heavily supported by small scale and multigear fisheries. The coastal areas show a typical tropical multi-species fisheries ecosystem. There are about 490 species of fishes (Hossain 1971) and 19 species of shrimps/prawn (Chowdhury & Sanaullah 1991) available in this area. These fisheries are characterized by fishing households rather than commercial organizations and play a greater role in sustaining the livelihoods and ensuring the food security of large numbers of rural people throughout the developing world (Whitmarsh et al. 2003). The future of these potentially huge resources has not been well documented. However, for the sustainability of this fishery resource proper scientific study is an urgent task. Hence, the present study has been designed to provide an extensive report on the fish and shrimp assemblage structure of Bakkhali river estuary in relation to water quality parameters.

Materials and methods

Study area

The Bakkhali river estuary is located at the southeastern coast of the Bay of Bengal in Bangladesh (Figure 1). A number of small streams originating from the south-eastern hills of Mizoram (India) meets at the Naikhongchhari of Bandarban district and form the river Bakkhali. It flows through Naikhongchhari and Ramu of Cox's Bazar district and falls into the Moheshkhali channel of the Bay of Bengal. This river is relatively wide compared to other rivers of the Cox's Bazar district and has a length of about 67km. The Bakkhali river estuary has a semidiurnal tidal regime. Its hydrology is also heavily influenced by monsoon wind. The tidal range varied between 0.07m and 4.42m during neap and spring tide respectively (Hossain and Lin 2001). Salt intrusion extends up to 6km upstream where a rubber dam was constructed for irrigation purposes.

Figure 1.  Map of Bakkhali river estuary and location of sampling stations.

The bottom of this river consists mainly of mud and sand particles. The estuarine zone is also characterized by long intertidal mudflats where mangrove vegetation (mainly Avicennia sp.), natural ullo grass Imperata cylindrica, cord grass Spartina sp. and sea grass Halophila beccarii are present (Hena et al. 2007). The lower part of this estuary is heavily influenced by anthropogenic and industrial activities including fish harbours, fish processing plants and a large number of fish and shrimp farms. The large amount of organic and inorganic waste changes the chemical characteristics of the water body by producing toxic substances, which ultimately affect the biodiversity. Samp took place at two stations (Figure 1), one (St1) about 5km upstream from the estuary, which is protected from the sewage and anthropogenic intervention, and another (St2) at the lower stream near the mouth of the estuary, heavily influenced by domestic and industrial activities. Apart from these two stations, nets were not set on a regular basis in the other areas of the Bakkhali river estuary. Net setting and collection of samples was largely dependent on the local fishermen who have used these areas for generations. Therefore, these areas are allowed for fishing only by the local fishermen. Hence, through negotiation with the local fishermen these two stations were considered for the present study.

Sampling gear

The fish and shrimp samples were collected using barrier nets known locally as ‘Char jal’ (Figure 2). In Cox's Bazar region, Char jal are used to catch various aquatic species from river banks inundated during high tide. Net fencing is made from bamboo poles which are submerged during high tide. Samples are collected during low tide. The net frame is around 2.5m in height and 150m in length, forming part of a circle so that there is approximately 120m distance between two ends of the net. The nylon net has a mesh size of 0.8cm. Bamboo poles are secured on the shore of the river during low tide. During high tide the water is allowed to enter and after 2–2.5h of the high tide the fishermen secure the upper portion of the net and create a barrier. At low tide all the animals inside the fence become trapped and the fishermen harvest the fish, shrimp and crab. Finally, the net is released from the bamboo and is ready for the next high tide.

Figure 2.  A Char jal in the Bakkhali river estuary.

Sampling periodicity

Samples were collected each month between December 2007 and August 2008. Of the four seasons specified by Mahmood et al. (1994), three seasons were chosen – the winter (December, January and February), premonsoon (March, April and May) and monsoon (June, July and August) – to conduct the sampling. Sampling was done during the full moon and new moon, as during these periods higher abundance of fish and shrimps were reported by the fishermen. No samples were collected during the post monsoon period (September, October and November) as the fishermen become engaged in Hilsha fishery, which is the single largest commercial fishery of Bangladesh (Mazid 2002). Fishermen are engaged by Hilsha boat owners to leave this less-profitable Char jal fishing during the post monsoon season.

Sample collection

Sample catches from Char jal were taken directly from the nets. In the laboratory, samples were sorted and identified to species level (Fischer & Whitehead 1974; Shafi & Kuddus 1982a, 1982b; Talwar & Jhingran 1991; DeBruin et al. 1995). The total numbers of each species and their wet weight from each net were also recorded. During sampling, in situ water quality parameters were measured at each sampling site. The salinity, pH, temperature and dissolved oxygen were determined by using a refractrometer (NewS-100, TANAKA, Japan), a pen pH meter (s327535, HANNA Instruments), a thermometer in centigrade and a DO meter (HI 9142, HANNA Instruments), respectively. A Secchi disc (20cm diameter) was used to measure the water transparency.

Data analysis

Diversity of the species assemblage was expressed by the Shannon–Wiener index (H') (Shannon 1949; Shannon & Weaver 1963; Ramos et al. 2006) using the following formula:

where S is the total number of species and Pi is the relative cover of ith species.

Richness was measured by Margalef index (d) (Margalef 1968) using the following formula:

where S is total species and N is total individuals.

For environmental parameters (temperature, salinity, DO, pH and water transparency) one-way analysis of variance (ANOVA) was used to calculate if there is any difference between two stations. The same procedure was followed for seasons. Prior to ANOVA tests, all data were checked for normality using the Kolmogorov–Smirnov test and homogeneity of variances using Levene test (Sokal & Rohlf 1998). Furthermore, in the event of significance, a post hoc Tukey HSD test was used to determine which means were significantly different at a 0.05 level of probability (Spjotvoll & Stoline 1973). The Kruskal–Wallis test (Akin et al. 2005) was performed on data which did not satisfy the assumptions of normality and homogeneity, after performing diverse data transformations (Clarke & Warwick 1994). Except for salinity, all other water quality parameters met the criteria of normal distribution and homogeneity of variances. One-way analysis of similarity (ANOSIM) (Clarke & Warwick 1994) was used to conclude the significance of spatial and temporal variation in the fish and shrimp assemblage structure. This test is based on a Bray–Curtis rank similarity matrix and was calculated using log-transformed data. Similarity percentages analysis (SIMPER) (Clarke 1993) was used to observe the percentage contribution of each species to the average dissimilarity between samples of the various seasons and station-pair combinations. Hierarchical agglomerative clustering with group average linking and non-metric multi-dimensional scaling (nMDS) were performed to investigate similarities among stations and seasons (Clarke & Warwick 1994). This analysis was based on the Bray–Curtis similarity measure (Bray & Curtis 1957). Only species with more than 1% of the total species were included in the analysis to avoid any unusual effects of rare species. All the multivariate analyses were performed using the software PRIMER V6 (Plymouth Routines Multivariate Ecological Research) (Clarke & Warwick 1994). Associations between species and environmental variables were examined with the canonical correspondence analysis (CCA) using the ECOM 1.32 version (Environmental Community Analysis 2000) software. To reduce the effects of rare species, only species contributing >1% of the total based on all species and samples were included in CCA after log transformation (Log₁₀(x+1)). CCA was proposed to constrain the axes in classical Correspondence Analysis (CA) to be linear functions of a-priori defined or measured variables associated with species records. The ordination axes of CA are termed Eigenvectors. Each Eigenvector has a corresponding Eigenvalue, often denoted by λ. The Eigenvalue is actually equal to the (maximized) dispersion of the species scores on the ordination axis, and is thus a measure of importance of the ordination axis. The first ordination axis has the largest Eigenvalue (λ₁), the second axis the second largest Eigenvalue (λ₂), and so on. The Eigenvalues of CA all lie between 0 and 1. Values over 0.5 often denote a good separation of the species along axis (Jongman et al. 1995).

Results

Environmental parameters

The measured environmental parameters are summarized in Table I and illustrated in Figure 3. Water temperature ranged between 21.00°C (in winter; January 2008 at St1) and 31.00°C (in premonsoon and monsoon; March and July 2008, respectively, at St2) with a mean of 27.16±3.33°C. No significant difference was observed in temperature between stations (F_1,12=2.13, P=0.17). However, winter season showed a significant difference from monsoon and premonsoon (F_2,12=66.65, P=0.001) although there was no significant difference between premonsoon and monsoon.

Figure 3.  Temporal and spatial variations in mean environmental parameters at the study area (s, St, st, station).

Table I. Mean of the different environmental parameters in different seasons and stations during the study period.

Season	Sampling	Temperature (°C)	Salinity (ppt)	Transparency (cm)	pH	DO(mg l^−1)
Winter	St1W1	22.33±1.15	29.66±0.57	65.33±4.04	7.23±0.2082	4.15±0.48
	St2W1	23.33±0.57	30.66±0.57	58.66±2.30	6.86±0.5132	4.09±0.42
Premonsoon	St1P1	29.33±1.15	25.66±1.15	40±2.00	7.33±0.4726	4.85±0.14
	St2P1	30±1.00	26.33±0.57	35.66±3.78	7.03±0.05774	4.73±0.34
Monsoon	St1M1	28.66±1.15	2.33±0.57	29.33±2.51	7.3±0.3606	3.81±0.58
	St2M1	29.33±1.52	5.33±0.57	28±2.00	7.43±0.2082	3.65±0.49
Station 1		26.77±3.49	19.22±12.80	44.88±16.22	7.28±0.31	4.27±0.60
Station 2		27.55±3.32	20.77±11.74	40.77±14.03	7.11±0.37	4.16±0.59

Salinity values (mean 20.00±11.94) ranged from 2.00 ppt (during monsoon season; August 2008) to 31.00 ppt (during winter season; December 2007 and February 2008). No significant differences were found in salinity between the stations (H=0.788, P=0.375), while significant differences were observed among the seasons (H=15.316, P=0.001) with very low salinity during monsoon (Figure 3).

Oxygen concentration (mean=4.21±0.58) attained a maximum in March (5.02mg/l at St1) and a minimum in June (3.24mg/l at St1). No significant differences were found in oxygen concentration throughout stations (F_1,12=0.30, p=0.59); in contrast, significantly higher dissolved oxygen concentration was observed during premonsoon season (4.79±0.24mg/l) compared to winter (4.12±0.40mg/l) and monsoon (4.40±0.49mg/l) seasons (F_2,12=9.22, P=0.001).

Water transparency varied from 26cm (during monsoon; July 2008 at St2) to 70cm (during winter season; January 2008 at St1) with a mean of 42.83±14.86cm. Significant differences were observed in water transparency between stations (F_1,12=212.08, P=0.001). Similarly, water transparency exhibited a strong seasonal gradient (F_2,12=9.06, P=0.01). Mean value in winter season (62.00±4.69cm) was noticeably higher than the premonsoon (37.83±3.63cm) and monsoon season (28.66±2.16cm).

The highest pH value (7.7) was observed during the premonsoon season at St1, while the lowest pH value (6.3) was observed during the winter season, also at St1. Mean pH value was observed to be 7.20±0.34. No significant differences were found for pH between stations (F_1,12=1.20, P=0.29) and among the seasons (F_2,12=1.28, P=0.31).

Species community composition by weight

The catch per net at St1 ranged between 1.31kg (during winter season) to 2.52kg during the premonsoon period with an average of 1.89±0.36kg. The average catch per net at St1 during winter, premonsoon and monsoon seasons was found to be 1.64±0.29kg, 1.94±0.57kg and 2.10±0.07kg, respectively. Of the fish species, Liza tade was found to be the most abundant (23.6%) followed by Mystus gulio (11.96%), Gerres filamentosus (9.34%) and Terapon jarbua (8.76%). Among the shrimps, Macrobrachium villosimanus was found to be moderately higher (7.98%) in species composition by weight. Only 12 species contribute about 90% of the total catch. No significant difference was observed in average catch per net between winter, premonsoon and monsoon (F_2,6 =1.37, P=0.32).

The catch per net at St2 ranged between 2.71 and 14.77kg with an average of 7.54±4.39kg. The average catch per net St2 during winter, premonsoon and monsoon seasons was found to be 3.94±1.94kg, 10.67±2.12kg and 8.01±5.86kg, respectively. The highest percentage (27.40%) was found for the fish Mystus gulio followed by Acanthopagrus latus (10.25%), Cynoglossus cynoglossus (8.12%) and Macrobrachium villosimanus (7.47%). Only 13 species including those reported above contributed about 87.11% of the total catch (Table II). No significant difference was observed in average catch per net between winter, premonsoon and monsoon (F_2,6 =2.42, P=0.17).

Table II. Fish and shrimp species recorded in the Bakkhali river estuary from December 2007 to August 2008 showing relative contribution (%) to the total abundance by stations and seasons.

				Station		Season
Species name	Code	Total ind.	(%)	St1 (%)	St 2 (%)	Win (%)	Pre-mon (%)	Mon (%)
Metapenaeus lysianassa (De Man, 1888)	Mly	3152	17.07	15.27	21.35	27.57	4.45	0.00
Ambassis dussumieri Cuvier, 1828	Adu	2685	14.54	16.83	9.10	20.64	8.89	0.00
Macrobrachium villosimanus (Tiwari, 1949)	Mvi	2240	12.13	7.35	23.49	13.38	14.28	0.00
Terapon jarbua (Forsskål, 1775)	Tjr	1687	9.14	11.10	4.46	9.83	6.00	14.36
Gerres filamentosus Cuvier, 1829	Gft	1684	9.12	12.45	1.21	10.90	7.73	4.16
Liza tade (Forsskål, 1775)	Ltd	1130	6.12	6.96	4.13	4.57	7.52	9.92
Mystus gulio (Hamilton, 1822)	Mgl	861	4.66	2.77	9.16	0.10	3.10	31.55
Metapenaeus monoceros (Fabricius, 1798)	Mmc	810	4.39	3.85	5.67	0.22	4.50	24.70
Butis butis (Hamilton, 1822)	Bbt	769	4.16	5.56	0.84	0.10	12.74	0.57
Acanthopagrus latus (Houttuyn, 1782)	Alt	432	2.34	2.35	2.30	2.43	1.89	3.12
Eleutheronema tetradactylum (Shaw, 1804)	Etd	415	2.25	3.17	0.05	0.00	7.08	0.00
Glossogobius giuris (Hamilton, 1822)	Ggr	396	2.14	1.82	2.91	1.43	2.80	3.87
Valamugil speigleri (Bleeker, 1858-59)	Vsg	395	2.14	1.96	2.56	0.30	5.22	2.74
Cynoglossus cynoglossus (Hamilton, 1822)	Ccg	194	1.05	0.00	3.55	0.56	0.73	4.35
Pseudapocryptes elongatus (Cuvier 1816)	Plc	183	0.99	1.38	0.07	0.00	3.12	0.00
Penaeus semisulcatus de Haan, 1844	Pss	164	0.89	1.26	0.00	1.56	0.00	0.00
Macrobrachium rude (Heller, 1862)	Mrd	154	0.83	1.15	0.09	1.42	0.09	0.00
Mugil cephalus Linnaeus, 1758	Mcp	150	0.81	1.02	0.31	1.43	0.00	0.00
Johnius belangerii (Cuvier, 1830)	Jbg	147	0.80	0.00	2.69	0.00	2.51	0.00
Exopalaemon stylifera (H. Milne-Edwards, 1840)	Esf	130	0.70	1.00	0.00	1.24	0.00	0.00
Sardinella fimbriata (Valenciennes, 1847)	Sfb	124	0.67	0.79	0.38	0.00	2.11	0.00
Escualosa thoracata (Valenciennes, 1847)	Etc	99	0.54	0.76	0.00	0.00	1.69	0.00
Penaeus indicus H. Milne Edwards, 1837	Pic	65	0.35	0.50	0.00	0.00	1.11	0.00
Metapenaeus brevicornis (H. Milne Edwards, 1837)	Mbc	48	0.26	0.00	0.88	0.46	0.00	0.00
Portunus pelagicus (Linnaeus, 1758)	Ppg	47	0.25	0.00	0.86	0.32	0.22	0.00
Scylla spp.	Scy	37	0.20	0.00	0.68	0.13	0.39	0.00
Sillago sihama (Forsskål, 1775)	Ssh	36	0.19	0.22	0.15	0.34	0.00	0.00
Platycephalus indicus (Linnaeus, 1758)	Pid	31	0.17	0.09	0.35	0.09	0.38	0.00
Setipinna taty (Valenciennes, 1848)	Stt	31	0.17	0.06	0.42	0.00	0.39	0.38
Apocryptes bato (Hamilton, 1822)	Abt	24	0.13	0.09	0.22	0.11	0.20	0.00
Scatophagus argus (Linnaeus, 1766)	Sag	23	0.12	0.10	0.18	0.22	0.00	0.00
Cynoglossus lingua Hamilton, 1822	Clg	22	0.12	0.00	0.40	0.15	0.10	0.00
Ilisha megaloptera (Swainson, 1839)	Imp	21	0.11	0.10	0.15	0.20	0.00	0.00
Penaeus japonicus Bate, 1888	Pjp	18	0.10	0.00	0.33	0.00	0.31	0.00
Eleotris fusca (Forster, 1801)	Efc	15	0.08	0.00	0.27	0.14	0.00	0.00
Penaeus monodon Fabricius, 1798	Pmd	11	0.06	0.00	0.20	0.06	0.07	0.05
Lutjanus johnii (Bloch, 1792)	Ljn	10	0.05	0.00	0.18	0.00	0.17	0.00
Siganus javus (Linnaeus, 1766)	Sjv	8	0.04	0.03	0.07	0.08	0.00	0.00
Nuchequula blochii (Valenciennes, 1835)	Lbc	6	0.03	0.00	0.11	0.00	0.10	0.00
Coilia dussumieri Valenciennes, 1848	Cdm	5	0.03	0.00	0.09	0.00	0.09	0.00
Macrobrachium rosenbergii (De Man, 1879)	Mrb	3	0.02	0.00	0.05	0.02	0.00	0.05
Labeo rohita (Hamilton, 1822)	Lrh	2	0.01	0.00	0.04	0.00	0.00	0.09
Plotosus canius Hamilton, 1822	Pcn	1	0.01	0.00	0.02	0.00	0.02	0.00
Oreochromis niloticus niloticus (Linnaeus, 1758)	Tnt	1	0.01	0.00	0.02	0.00	0.00	0.05
Hyporhamphus limbatus (Valenciennes, 1847)	Hlb	1	0.01	0.00	0.02	0.00	0.00	0.05
Total number of taxa		45		28	41	30	32	16
Total number of individuals		18467		12996	5471	10487	5863	2117
Mean		1026		1444	607.89	1747.83	977.17	352.83
Stdev		803		866.75	477.25	421.99	207.77	22.81
Shannon				2.01	1.81	1.69	2.42	1.62
Stdev				0.33	0.56	0.24	0.13	0.23
Margalef				2.17	3.84	2.71	3.30	1.56
Stdev				0.13	0.66	0.24	1.23	0.40

However, a significant difference was obtained in average catch per net between St1 and St2 (F_1,16=14.74, P=0.001), although no significant difference was observed among the seasons (F_2,15=1.09, P=0.35).

Species community composition by number

During the study period a total 18,467 fish, shrimp and crab were collected from the Char jal with a mean abundance of 1026±803 ind/5kg of species (Table II). The maximum species abundance (2869 ind/5kg of species) was observed during the winter at St1 while the minimum (241 ind/5kg of species) was observed during the monsoon period at St2.

The species abundance per net in St1 ranged between 412 ind/5kg (during monsoon season) to 2869 ind/5kg during the winter period, with an average of 1444.00±886.74 ind/5kg. The average abundance during winter, premonsoon and monsoon seasons were found to be 2304.66±489.24 ind/5kg, 1585.66±341.70 ind/5kg and 441.66±25.73 ind/5kg, respectively. Ambassis dussumieri was found highest (16.83%) followed by Metapenaeus lysianassa (15.27%), Gerres filamentosus (12.45%) and Terapon jarbua (11.10%) in species abundance. Only 18 species contributed about 97.25% of the total catch. A significant difference was observed in average species abundance between winter, premonsoon and monsoon (F_2,6 =22.26, P=0.002).

The species abundance per net in St2 ranged between 241 ind/5kg to 1601 ind/5kg with an average of 607.88±477.25 ind/5kg. The average abundance during winter, premonsoon and monsoon seasons was found to be 1191.00±357.31 ind/5kg, 368.66±98.77 ind/5kg and 264±19.92 ind/5kg, respectively. The highest percentage (23.49%) was found for Macrobrachium villosimanus followed by Metapenaeus lysianassa (21.35%), Mystus gulio (9.16%) and Ambassis dussumieri (9.10%). Only 13 species contribute about 92.56% of the total abundance (Table II). A significant difference was observed between winter, premonsoon and monsoon (F_2,6=16.83, P=0.003). A significant difference in average species abundance was also observed between stations (F_1,16 =6.42, P=0.02) and seasons (F_2,15=8.58, P=0.003).

Species diversity

The Shannon–Wiener diversity index ranged between 0.95 (at St2 during monsoon) and 2.62 (at St1 during premonsoon) with a mean diversity value of 1.91±0.46 (Figure 4A). No significant difference was observed (F_1,16=0.915, P=0.353) in the mean Shannon–Wiener diversity values between the stations. However, this difference was found significant between the seasons (F_1,16=14.264, P<0.001) with higher mean diversity value (2.416±0.16) during the premonsoon period.

Figure 4.  Temporal and spatial variations of (A) Shannon–Wiener index and Abundance and (B) Margalef diversity index and Abundance of the Bakkhali fish and shrimp assemblage. St, station. St, station.

The minimum Margalef richness value (1.14) was observed at St1 during monsoon while the maximum value (4.50) was found in station St2 during premonsoon (Figure 4B) with a mean richness value of 2.52±1.08. The mean species richness values at St1 and St2 was found to be 1.85±0.49 and 3.20±1.09, respectively. In the case of seasons, the highest mean richness value (3.30±1.22) was observed in premonsoon while the lowest mean value (1.56±0.40) was observed during monsoon. Significant differences were found for Margalef's index for both stations (F_1,16=11.34, P=0.004) and seasons (F_1,16=6.757, P=0.008).

Species assemblage

The analysis of similarity (ANOSIM) showed significant difference in assemblage structure between stations (Global R=0.365; p=0.008) (Table III). Species assemblage at each station was found to be highly diverse and at St1 Liza tade (15.24%), Terapon jarbua (15.10%), Gerres filamentosus (13.45%) and Acanthopagrus latus (10.86%) were found to be the most dominant (>10%) species, while at St 2 Terapon jarbua (12.48%), Cynoglossus cynoglossus (11.57%) and Mystus gulio (11.16%) were found to be the dominant (>10%) species (Table IV). According to SIMPER results, other contributory species to the assemblage structure of the studied area were Glossogobius giurus, Metapenaeus monoceros, Metapenaeus lysianassa, Valamugil speigleri, Ambassis dussumieri and Butis butis (Table IV).

Table III. Result of one-way ANOSIM (R value and significant levels) and SIMPER analysis of fish and shrimp abundance between stations and different seasons.

Groups	ANOSIM		SIMPER
	R	P	Average dissimilarity (%)	Most discriminating species	Contribution (%)
St1 vs. St2	0.036	0.08	40.46	Metapenaeus lysianassa	10.23
	Analysis of similarity results among different seasons
	Global R=0.726	P=0.001
Winter vs. Premonsoon	0.491	0.04	34.70	Butis butis	10.62
Winter vs. Monsoon	1	0.02	48.29	Metapenaeus lysianassa	18.36
Premonsoon vs. Monsoon	0.594	0.02	39.95	Macrobrachium villosimanus	11.42

Table IV. Average similarity and discriminating fish and shrimp in each station using SIMPER analysis.

Average similarity (%)
Station 1 (69.63%)		Station 2 (69.53%)
Contributory species		Contributory species
Species	(%)	Species	(%)
Liza tade	15.24	Terapon jarbua	12.48
Terapon jarbua	15.10	Cynoglossus cynoglossus	11.57
Gerres filamentosus	13.45	Mystus gulio	11.16
Acanthopagrus latus	10.86	Liza tade	9.66
Glossogobius giurus	9.76	Glossogobius giurus	9.36
Valamugil speigleri	8.17	Metapenaeus monoceros	8.58
Metapenaeus lysianassa	5.69	Acanthopagrus latus	7.67
Mystus gulio	5.61	Valamugil speigleri	6.86
Ambassis dussumieri	5.47	Gerres filamentosus	5.65
Macrobrachium villosimanus	4.26	Butis butis	5.59
		Metapenaeus lysianassa	4.27

A highly diverse species assemblage was also observed among all seasons through SIMPER analysis (Table III). Significant difference were observed for temporal community structure of the studied area (Global R=0.726; P=0.001) with a clear separation of different seasons. The pair-wise comparison of seasons also showed distinct separation (Table V). Metapenaeus lysianassa (17.19%) and Ambassis dussumieri (13.43%) were found to be the most contributory species during winter, while during the premonsoon season it was Velamugil spaglari (11.23%) and Mystus gulio (10.38%). On the other hand, Mystas gulio (21.21%) and Terapon jarbua (17.12%) were found the most contributory species during the monsoon season (Table V).

Table V. Average similarity and discriminating fish and shrimp in each season using SIMPER analysis.

Average similarity (%)
Winter (79.07%)		Premonsoon (72.65%)		Monsoon (78.87%)
Contributory species		Contributory species		Contributory species
Species	(%)	Species	(%)	Species	(%)
Metapenaeus lysianassa	17.19	Valamugil speigleri	11.23	Mystus gulio	21.21
Ambassis dussumieri	13.43	Mystus gulio	10.38	Terapon jarbua	17.12
Liza tade	10.44	Terapon jarbua	9.76	Metapenaeus monoceros	13.68
Terapon jarbua	10.20	Liza tade	9.52	Liza tade	11.49
Macrobrachium villosimanus	9.68	Butis butis	8.5	Glossogobius giurus	9.65
Glossogobius giurus	9.49	Acanthopagrus latus	7.61	Valamugil speigleri	8.01
Acanthopagrus latus	9.32	Glossogobius giurus	7.21	Acanthopagrus latus	7.71
Gerres filamentosus	8.90	Metapenaeus lysianassa	7.16	Gerres filamentosus	7.16
Metapenaeus monoceros	4.33	Ambassis dussumieri	7.07
		Macrobrachium villosimanus	6.69
		Gerres filamentosus	5.97

At the similarity level of 65%, no marked separation, either for the stations or for the seasons, was observed by cluster analysis. Two clusters were identified – the first consists of St2 during monsoon and premonsoon period along with St1 during monsoon, and the second group consists of St1 during premonsoon and winter along with St2 during winter (Figure 5).

Figure 5.  Dendrogram (A) showing cluster based on Bray–Curtis similarity matrix of catch composition, and the ordination in 2D (B) using MDS on the same similarity matrix. St, Station; W, Winter; P, Premonsoon; M, Monsoon.

Canonical correspondence analysis

CCA eigenvalues of the first four axes were 0.36 (CCA1), 0.34 (CCA2), 0.11 (CCA3), and 0.10 (CCA4). Species–environment Pearson correlation coefficients for the first four axes were 0.96, 0.92, 0.82, and 0.79, respectively. The cumulative percentage variance of species for the first four axes (CCA 1–4) was 49.18. The first and second axes modelled 19.4 and 18.1% of species data, respectively. Therefore, the results obtained from the first two axes were plotted (Figure 6). However, the vector length of a given variable indicates the importance of that variable in CCA analysis. Salinity (0.92), which has the longest vector along the first axis, was significantly correlated with premonsoon at St1 (Figure 6). Furthermore, transparency was also significantly associated (0.81) with winter season of St1 and St2. As shown in CCA ordination (Figure 6), high values of salinity concentration are the most significant water parameters for Butis butis and Eleutheronema tetradactylum. High values of water transparency was associated with occurrence of Ambassis dussumieri and Gerres filamentosus. High values of pH are associated with the occurrence of Valamugil speigleri, Glossogobius giurus and Metapenaeus monoceros. However, Acanthopagrus latus showed the highest association with dissolved oxygen (DO), while no species was found to be closely associated with temperature.

Figure 6.  The CCA ordination of species abundance and environmental parameters (code for each species is given in Table II; St, station; p, premonsoon; m, monsoon w, winter).

Discussion

Environmental parameters

During the present study, no significant spatial variation was observed for temperature, salinity, dissolved oxygen and pH. This is most probably due to the presence of a rubber dam upstream of this river, which prevents freshwater influx in this area during winter and premonsoon and ultimately the whole area, i.e. both St1 and St2, is similarly governed by the brackish water entering through tidal influence. In the same way, during monsoon, the rubber dam remains open and huge amount of freshwater is discharged through the dam. As a consequence, the water parameters in both stations remain the same. However, a significant difference in water transparency was observed between St1 and St2, where lower transparency was observed in St2 compared to St1. This may be due to the higher water turbulence in downstream (St2) due to strong tidal fluctuation and anthropogenic causes such as the presence of the fish harbour, jetty, etc., in the mouth of the river (St2). Comparatively lower dissolved oxygen was observed in both stations. Hena & Khan (2009) also reported a lower level of DO in the same estuary (1.89–5.37mg/l). This may be due to the nearby domestic, agricultural and industrial waste water discharges which affect the water and sediment quality and lead to a hypoxic condition, as stated by Van Eck et al. (1991).

Fluctuations in water transparency influence the primary productivity which ultimately affects the fish distribution (Arthington & Welcome 1995; McAllister et al. 2001). Rozengurt & Hedgepeth (1989) reported changes in natural recruitment and species abundance in the Caspian Sea due to in increase in salinity. McAllister et al. (2001) also reported changes in species abundance due to a salinity increase. According to Maes et al. (2004), dissolved oxygen is one of the most important factors for fish abundance and distribution. Fish communities are highly affected by temperature within estuaries (Cyrus & Mclean 1996). A sudden increase or decrease in water temperature may cause fish mortality (Blaber 2000). Environmental parameters such as temperature, salinity, dissolved oxygen, water transparency and pH play an important role for species abundance and diversity (Whitfield 1999), especially for the tropical regions where the fluctuation of these parameters are frequently due to seasonal changes (Blaber 2000). The Bakkhali river estuary is no exception. Significant temporal differences were observed for temperature, salinity, water transparency and dissolved oxygen during the present study which may almost certainly affect the assemblage structure.

Species composition

The average catch per net at St2 (7.54±4.39kg) was found to be significantly higher compared to St1 (1.89±0.36kg). St2 is located in the river mouth and possesses a highly diverse type of larger-sized fish compared to St1. On the other hand, at St1 a small area of salt marsh is situated nearby, which acts as the nursery ground for small and juvenile fishes. As fish grew, their daily food requirement also increased and fish began to migrate to the river mouth/estuary in the search of food (Davies & Day et al. 1986). Hence, larger fish were caught and the total weight of the catch per net was higher were St2.

A total of 45 fish and shrimp species were recorded during the study. Among them are Metapenaeus lysianassa, Ambassis dussumieri, Macrobrachium villosimanus, Terapon jarbua, Gerres filamentosus, Liza tade, Mystus gulio, Metapenaeus monoceros, Butis butis, Acanthopagrus latus, Eleutheronema tetradactylum, Glossogobius giurus, Valamugil speigleri, and Cynoglossus cynoglossus, each contributing more than 1% of the composition. Islam et al. (1992) reported about 185 species from the coastal waters of Bangladesh collected from the estuarine sey bagnet. On the other hand Hossain et al. (2007) reported about 161 species collected by different types of net from Naaf river estuary located around 50km from the present study site. A smaller number of species observed in the present investigation is most probably due to the use of only one type of net, i.e. Char jal which catches only species living very near to the shore and closer to the bottom. Another reason is the controlled environment of the Bakkhali estuary by the rubber dam limiting the species abundance. Also, the dam-induced changes in water characteristics may have profound effects on species numbers in the river (McAllister et al. 2001).

During the study period, only one exclusively freshwater specimen fish (Labeo rohita) was collected. The rubber dam, constructed at the upper stream of the river, creates an unusual environment. During the monsoon season huge amounts of freshwater flow into the river and create a freshwater-influenced estuarine environment where salinity ranges were found to vary from 2.33±0.57 to 5.33±0.57. This supports the findings of the presence of a freshwater species in the investigated area. Freshwater fishes are usually incapable of osmoregulating in saltwater and consequently tend to be found in estuaries only when salinities decline to very low levels during periods of heavy freshwater discharge (Potter & Hyndes 1999).

Species abundance

The species abundance found in the Bakkhali river estuary is composed of small numbers of species with high contribution and a large number of species whose contributions are very negligible, a common feature of estuarine faunal populations (Gaughan et al. 1990; Harrison & Whitfield 1990; Drake & Arias 1991; Harris & Cyrus 1995; Whitfield 1999). The number of taxa in this study (45 species) was found to be lower than in the Naaf river estuary, another estuary close to the study area (Hossain et al. 2007). However, these types of judgements must be based on differences in sampling gear, sampling period and most importantly habitat characteristics. Moreover, each estuarine system may have a different abiotic environment (Blaber 1997), resulting from the tidal range, freshwater input, geomorphology and human pressure (Dyer 1997; McLusky & Elliott 2004) which also affects the species abundance. So a difference in species abundance is not likely to be the exception.

A remarkably lower number of species was observed in the upstream area (St1) of the Bakkhali river estuary, but the diversity index H′ was much higher. Out of total 45 species, 17 were relatively common in the upper stream which is characterized by juveniles of large-sized species and adults of small-sized species. The main reason for the high abundance of juvenile fishes is the presence of a salt marsh and seagrass bed at St1. These observations agree with the comments of Hena et al. (2007) – seagrass and salt marsh habitats are among the most productive ecosystems in the world in terms of the quantity of vegetation production closely linked to the high production rates of associated fisheries.

The lower stream (St2) has a wider zone and is characterized by strong tidal influence and mangroves. As a consequence, relatively higher numbers of species with brackish to marine origin were captured in this zone. Downstream, water transparency showed higher values (28.00–65.33cm) when compared to other estuaries of Bangladesh (Mahmood 1986). This is because freshwater flows are scarce here for a major part of the year leading to the presence of marine species which have been described as being related to clearer waters (Blaber et al. 1997). The exception occurred during monsoon months (June, July) when the rubber dam is opened causing flushing effect by freshwater flows from the upper, hilly areas.

An estuarine water body along with mangrove plants is the most productive region for zooplankton especially for shrimps and prawns (Hena & Khan 2009). The Bakkhali river estuary is influenced by mangrove plants including Avicennia alba, A. marina, and Acanthus ilicifolius, which have created a huge potential habitat for phytoplankton, zooplankton, shellfish and fish larvae. Plankton communities living in mangrove waters are well adapted to the water motion (Alongi 2009). Whether a source of food, shade, or refuge, mangrove forests are an important habitat for coastal organisms that either float or swim on the ebb and flood of the tide (Alongi 2009). Following this pattern, the Bakkhali river estuary also supports a huge abundance of shellfish of both marine and brackish origin. Although mangroves support fisheries by playing a significant role as nursery ground for shrimps including the giant tiger shrimp (Penaeus monodon) which is the major species of the industrial bottom trawl fishery of Bangladesh (Islam & Haque 2004), the present study did not show the same result for the penaeid group where P. monodon, P. indicus, P. semisulcatus and P. japonicus were found to be 0.06, 0.35, 0.89 and 0.10%, respectively. The same was encountered for Macrobrachium rosenbergii (0.02%), although it is the major species found in different estuarine studies of Bangladesh. On the other hand, Metapenaeus lysianassa (bird shrimp), Macrobrachium villosimanus and Metapenaeus monoceros contributed 17.07, 12.13 and 4.39% of total catch, respectively. This may be due to the use of Char jal, a different gear compared with the previous study, and a higher average salinity gradient for most of sampling period.

Species diversity

Seasonal variation in species diversity is a very common phenomenon in tropical estuaries and the estuary studied here is not different. However, no significant difference for the diversity values (both Shannon–Wiener and Margalef) between the stations indicated that the ecosystem for both tations was unique. However, the diversity values were lower than the value reported by Hossain et al. (2007) where the Shanon value was found to be 2.6. This difference may be due to the use of multigear fishing materials in the Naaf river estuary, the controlled environment of the Bakkhali estuary which limits the species interaction from the upstream communities and scarcity of marine species which normally come to the estuary for breeding.

Although several studies have reported the dominance of the resident species in the estuaries (Thompson 1966; Hotos & Vlahos 1998), in the case of the Bakkhali estuary no species was found to be dominant. Rather than a single species, five to six species were found dominant at different stations and different seasons. Blaber (2000) also stated that the estuarine resident species are a relatively insignificant proportion of the fish fauna available in an estuary and are generally all relatively small-sized fish.

Metapenaeus lysianassa was found most abundant overall (17.07%). However, the species showed its maximum abundance during the winter season (27.57%) and at St2 (21.35%). As M. lysianassa is a marine-dominant species, it only appears during the winter season at St2 where the salinity is very high. Ambassis dussumieri also showed the same trend except of higher abundance (16.83%) at St1. This is most probably due to the presence of salt marshes at St1. A similar trend was observed for Macrobrachium villosimanus except for the fact that they were dominant during premonsoon (14.28%) and winter (13.38%). As this species is a brackish species, it shows a higher abundance during premonsoon and winter seasons. Mystus gulio and M. monoceros were more abundant during the monsoon season at St2, while M. monoceros is a brackish to marine habitat species. However, their higher abundance at St2 during the monsoon period may be due to spawning. The same is also probably true for freshwater to brackish-water species such as Mystus gulio.

Species assemblage

Regarding spatial and temporal fish and shrimp assemblage structural, two major groups were indicated by cluster analysis in the Bakkhali river estuary. Group 1 comprises the sample of the monsoon season from St1 and St2 along with the premonsoon season from St2 within a 65% similarity level. The capture of abundant large-sized species in the premonsoon and monsoon seasons and the absence of the major contributing species like A. lysianassa, M. villosimanus and Aetapaeneus dussumieri this. Substantial The sample was taken from the lower stream during the premonsoon season at St2 included large numbers of adult fish species. The abundant presence of A. dussumieri, Terapon jarbua, Gerres filamentosus, Liza tade and Butis butis also supports.

Importatnt group 2 comprises the samples from St1 and St2 during the winter season and samples from St1 during the premonsoon period. In general, samples from St1 and St2 during winter showed the same trend on the basis of catchability.

Seasonality is the most important feature among different studied parameters affecting the fish and shrimp assemblage, similar to results from other estuaries (Whitfield 1989; Loneragan & Potter 1990; Drake & Arias 1991; Barletta-Bergan et al. 2002; Young & Potter 2003). In general, differences between seasons were observed to be more pronounced in this study. According to Lam (1983), the seasonal water variability in the spawning area has an important influence on the spawning activity, and on nursery areas (McErlean et al. 1973).

Canonical correspondence analysis

In CCA, species plotted closer to the vector, have stronger relationships with them. Species located near the origin either do not show a strong relationship to any of the variables or are found at average values of environmental variables (Marshall & Elliott, 1998). In this study, salinity, water transparency and pH were found to be three significant variables affecting species composition in the studied estuary. However, temperature and DO were not found to be significant. Half of the species in the estuary had average values in relation to environmental variables. Only two species, Eleutheronema tetradactylum and B. butis, indicated a strong response to the longitudinal salinity gradient. The five variables measured in this study explained species distributions well compared with most other estuarine studies where salinity and transparency were found to be the most influential factors for fish distribution patterns. For example, Marshall & Elliott (1998) found that five environmental variables accounted for 18.4% of the total species variation even though they included bottom, mid and surface values of each variable in CCA. Rakocinski et al. (1996) used 11 environmental variables that together explained only 21.9% of the total species variations in CCA. On the other hand, Martino & Able (2003) explained 29.9% of the total species variation in Mullica River Estuary, New Jersey, using five environmental variables that included salinity and geographic distance. However, during this study, environmental variables accounted for 49.18% of the total species variation. These results are also in agreement with Akin et al. (2005) for the Koycegiz Lagoon Estuary, Turkey.

Conclusion

In the Bakkhali river estuary, environmental influence was apparently more extensive during premonsoon and winter, when water transparency and salinity fluctuation leads to an increase in diversity. In this study, seasonality of the environmental conditions explained the major variations of the fish and shrimp assemblage. Seasonal variations occurred not only in total abundance and diversity, but also in the structure of the species assemblage of the Bakkhali river estuary. Besides seasonal variations, the assemblage also exhibited a defined spatial pattern. The migrating marine species Metapenaeus lysianassa was more abundant in the shallow salt marsh zones, while estuarine residents showed more or less equal distribution throughout the seasons except for Mystus gulio. The presence of this species seems to be more related to the spawning season during monsoons. In the Bakkhali river estuary, the four common fish species Ambassis dussumieri, Terapon jarbua, Gerres filamentosus, Liza tade were present for most of the sampling time, which is possibly due to their higher salinity tolerance. Favourable environmental conditions, mainly salinity, enable these fish species to spawn. Therefore, it can be said that these species use the Bakkhali estuary as a spawning ground.

Editorial responsibility: Geir Ottersen (St1, St2)

Notes

Published in collaboration with the University of Bergen and the Institute of Marine Research, Norway, and the Marine Biological Laboratory, University of Copenhagen, Denmark