Length–weight relationships of fishes collected from the Korangi-Phitti Creek area (Indus delta, northern Arabian Sea)

Introduction

The Korang-Phitti Creek area is located on the northern side of the Indus delta in the northern Arabian Sea. This complex network of small creeks has vast mangrove-dominated areas that serve as nursery and spawning grounds for many important pelagic and demersal fish species. As a result of the area's productivity, it is exploited by small-scale fishermen; overfishing of certain fish stocks in the creeks have been a matter of great concern as the number of fishermen escalate each year (Meynell, 1995). Despite the importance of the fisheries within the region, no baseline biological data (i.e. length–weights) are available for many fish species from these waters. Such data are critical for comparing growth rates, examining age structure, or estimating production and biomass of the fish stocks (Ricker, 1975). In the present study, the parameters of the length–weight relationship are reported for 41 fish species (mangrove residents and important coastal fishes) that visit the Korangi-Phitti Creek area.

Materials and methods

Fish samples in the Korangi-Phitti Creek area (24°45′N, 67°20′E) were collected at 123 stations from February 1999 to December 2001. Monofilament gill nets of 8.9, 5.7, and 3.8 cm stretch mesh sizes were used. Multiple mesh sizes of gill nets were used in order to efficiently sample the wide range of size groups of fish encountered and to collect a range of immature to fully mature specimens. All fish were stored on ice in the field and brought to the laboratory where they were identified to the lowest possible taxon, measured to the nearest 0.1 cm total length (TL), and weighed to the nearest 0.01 g total weight (TW); where possible, the sex determined. Only combined values were reported when too few individuals were sexed to produce a significant relationship; total N represents the total number of fish in the catch and n represents the random subsample of fish measured after outliers were removed. Nomenclature for the families followed Nelson (2006).

Length–weight parameters were estimated for the entire sample of each species and by sex, as appropriate according to the following linear regression using PROC REG (SAS Institute, 1989): log TW = log a + b log TL, where TW = weight (g), TL = length (cm), a = y-intercept, and b = slope. To test for possible significant differences between sexes (P < 0.05) a Student’s t-test was used for the comparison of two slopes (Zar, 1984). Outliers were identified in SAS (PROC REG, RSTUDENT option) using tests for influence. Only high influence points (Studentized residuals beyond ±2.96) believed to be due to measurement, recording, or identification error were removed from the data set before final parameter estimation. Outliers in small datasets (N ≤ 25) were identified using Dixon’s test statistic for outliers (Sokal and Rohlf, 1981). A Student’s t-test was also performed to determine whether the estimated values of the slope (b) significantly deviated from isometric growth (b = 3).

Results and discussion

A total of 14 196 fish representing 41 species, 33 genera, and 20 families were collected during the sampling period. The most abundant taxa were clupeids, Anodontostoma chacunda, Nematalosa nasus, and Sardinella gibbosa. The number of fish collected varied by species, ranging from a low of 10 individuals each for Brachirus orientalis and Synaptura commersonnii, to 5413 for S. gibbosa. Length–weight relationships of all 41 species are summarized in Table 1, which provides sample size (n), minimum and maximum total length, the length–weight parameters (a, b, s_b), and the coefficient of determination (r²). Along with providing data on species from a previously understudied ecosystem, this study provides data on seven species in which there were no previous reports of length–weight relationships (Sardinella brachysoma, Arius arius, Trachinotus mookalee, Nuchequula blochii, Gerres limbatus, B. orientalis, and S. commersonnii).

The coefficient of determination (r²) of the length–weight relationships ranged from 0.723 for Liza carinata to 0.995 for S. commersonnii, all models were significant and r² values were > 0.90 for 35 species (sexes pooled). The estimated b values ranged from a minimum of 2.204 for L. carinata to a maximum of 3.535 for Pseudorhombus arsius. Fourteen species included enough sexed individuals to be able to conduct sex-specific length–weight regressions (Table 1). Two species (N. nasus and Scatophagus argus) were found to have differences in the sex-specific b-value. In both species, males had a significantly lower b value (2.609 and 2.764, respectively) than females. There were 16 species that showed isometric growth (b = 3) with estimated b-values not significantly (P > 0.05) different from three. Eight species showed positive allometric growth (b > 3) and 17 species showed negative allometric growth (b < 3).

The length–weight relationship was relatively ‘weak’ for several species (Thryssa dussimieri, H. kelee, N. nasus, S. gibbosa, Sardinella sindensis, and L. carinata), with r² ranging from 0.680 to 0.866 (Table 1). The low r² may be the result of several factors such as gut fullness or maturity stage (T. dussimieri, H. kelee, N. nasus, S. gibbosa, and L. carinata) and low sample size (S. sindensis). The influence of season, habitat, ontogenetic changes, sex, gonad maturity, diet, stomach fullness and health on length–weight relationships has been identified in many studies (Richter et al., 2000; Moutopoulos and Stergiou, 2002; Frota et al., 2004, among others). Although the estimations of the length–weight relationship did not take these factors into account, except for the sex when available, these data represent important baseline data needed for management of the fisheries in the region.