Abstract
Length-weight relationships (LWR) were described for 53 species of shallow-water fish caught with bottom trawls in a depth range between 7.3 and 108.1 m during September 2013 in the Colombian Caribbean Sea (fishing area 31 of the FAO). A linear regression was performed using the logarithmically transformed data to calculate a and b coefficients and their 95% confidence interval of the LWR for 53 fish species corresponding to 28 families and 44 genera. Six fish species showed a maximum total length greater than that reported in FishBase: Astrapogon alutus (102.00 mm), Eucinostomus harengulus (162.00 mm), Haemulopsis corvinaeformis (293.00 mm), Cyclopsetta chittendeni (390.00 mm), Etropus crossotus (224.00 mm), and Bairdiella ronchus (415.00 mm). A total of 24 species (45.3%) exhibited isometric growth, 21 species (39.6%) negative allometric, and 8 species (15.1%) positive allometric. This study shows the first estimates of LWR for 25 species of shallow-water fish in the Colombian Caribbean Sea. We found interdependence of growth parameters as a function of fish body shape. The analysis of the relative condition factor indicated that 21 fish species (39.6%) showed poor growth conditions. The results obtained from this study contribute to fill information gaps on shallow-water fish populations and also help fisheries scientists in future population assessment studies in the Colombian Caribbean Sea.
1. Introduction
Shallow-water fish represent about 90% of the catch in artisanal fisheries, which has a very important role in the local economy and food security in the Colombian Caribbean Sea [1–4]. Additionally, fish species are also an important fraction of the bycatch fauna in the industrial shrimp trawl fishery both at the local level [5–8] and in other regions around the world [9–11].
The length-weight relationships (LWR) provide information on the type of growth, the state of the species, habitat conditions, and the morphometric characteristics of the species, mainly in the species subjected to fishing exploitation [12–17]. LWR parameters are obtained from length-frequency data, which are very useful for estimating biomass and comparing the life history of species between regions [12, 15, 18]. However, LWR parameters may vary between habitats and regions, so accurate estimation of local parameters is essential for comparative studies in fish stock assessment [19, 20]. Additionally, the condition factor based on LWR data is relevant for examining the welfare of fish populations [15, 21, 22]. However, despite their importance, information on LWRs and condition factors is only available for a limited number of fish species [15, 23] and is very scarce in data-poor fisheries from the Colombian Caribbean Sea.
In this way, the goal of this study was to determine the LWR and the relative condition factor of 53 species of shallow-water fish in the Colombian Caribbean Sea with the purpose of contributing to the knowledge of the biology of shallow-water fish, located within the United Nations Food and Agriculture Organization fishing area 31 (FAO).
2. Materials and Methods
2.1. Sampling
The study area was located between Punta Gallinas (12°10′N, 71°14′W) and Urabá Gulf (9°03′N, 76°53′W) in the Colombian Caribbean Sea, fishing area 31 from FAO (Figure 1). Samples of shallow-water fish species were collected on a research survey using trawl sampling at depths between 7.3 and 108.1 m in September 2013. The sampling period corresponds to the rainy season, which is characterized by heavy rains and weak winds [24]. A total of 5094 fish were captured, to which the total length (TL in mm) of each individual was recorded from the tip of the mouth to the extended tip of the caudal fin using an ichthyometer with a precision of 1 mm, and the total body weight (W in g) was recorded with a precision of 1 g using an electronic scale. The care and use of experimental animals complied with Autoridad Nacional de Licencias Ambientales de Colombia (ANLA), animal welfare laws, guidelines, and policies as approved by the University of Magdalena reference number 1293–2013. The information of the body shape and the parameters of the length-weight relationships (LWR) for each one of the fish species was discussed in the FishBase (https://www.fishbase.se/) 02/2023 database [25].

2.2. Statistical Analyses
The LWR parameters of the fish species were determined by applying the following allometric equation [15, 26, 27]:where Wi is the total body weight (g), Li is the total length (mm), a (intercept) and y b (slope) are the estimated parameters applying the linear regression model with the log-transformed data according to the following equation:
The corrected back-transformed predicted value of the response variable was calculated by multiplying the back-transformed predicted value by the correction factor (cf), where RSE is the residual standard error and loge(10) is used to adjust for the base of the logarithm used [28]:
To evaluate the type of isometric growth if b = 3.0, negative allometric if b < 3.0, and positive allometric if b > 3.0, a t-student test was used to determine significant differences from the estimated value of b and its 95% confidence interval (C.I.) [29].
To evaluate the influence of body morphology on the growth parameters a (intercept) and b (slope) of the LWR of the fish species, a robust multiple regression model was applied with the data grouped according to body shape (eel-like, elongated, fusiform, short and/or deep), indicating the negative allometric, isometric allometric, and positive allometric growth zones [15, 23].
The relative condition factor (Krel) of the evaluated fish was determined according to the following equation [15, 27]:where W is the observed body weight (g) of an individual and aLb is the estimated weight from the length-weight relationships for that individual’s length. A good growth state of the species was identified when the Krel value ≥1.0, while the species was in poor growth conditions when the Krel value <1.0 [22, 27]. A one-sample t-test was used to verify significant differences between the Krel and the expected value of Krel = 1.0 [29]. All statistical and graphical analyses were performed in the R 4.2.3 language [30], using the modelr, FSAmisc, moments, and ggplot2 packages [31–34].
3. Results
A total of 53 fish species belonging to 28 families were analyzed, of which the Sciaenidae and Haemulidae families were the most representative with 7 and 6 species, respectively (Table 1). Regarding body shape, 29 species showed a fusiform body shape, 13 short and/or deep, 9 elongated, and 2 eel-like. The most abundant species were Eucinostomus gula (Quoy & Gaimard, 1824) with 618 specimens, followed by Menticirrhus americanus (Linnaeus, 1758), Larimus breviceps (Cuvier, 1830), and Etropus crossotus (Jordan & Gilbert, 1882) with 467, 453 and 414 specimens, respectively (Table 1). Total lengths (TL) for all species ranged from 24.00 mm to 940.00 mm (Table 1). Species with a maximum total length greater than that reported in FishBase were Astrapogon alutus (Jordan & Gilbert, 1882) (102.00 mm), Eucinostomus harengulus (Goode & Bean, 1879) (162.00 mm), Haemulopsis corvinaeformis (Steindachner, 1868) (293.00 mm), Cyclopsetta chittendeni (Bean, 1895) (390.00 mm), E. crossotus (224.00 mm), and Bairdiella ronchus (Cuvier, 1830) (415.00 mm) (Table 1).
Linear regressions were significant for all species (), with coefficients of determination (r2) between 0.81 and 0.99, except for E. gula with the lowest value of 0.77 (Table 2). The intercept of the linear regression (a) showed a range of values between 2.1120E − 08 for Gymnothorax ocellatus Agassiz, 1831 and 1.1419E − 02 for Balistes capriscus Gmelin, 1789, while the slope parameters (b) were between 1.79 for B. capriscus and 3.69 for G. ocellatus (Table 2). About 24 species (corresponding to 45.3%) showed isometric growth (b = 3.0), 21 species showed (39.6%) negative allometric growth (b < 3.0), and 8 species showed (15.1%) positive allometric growth (b > 3.0) (Table 2, t-student test and the C.I. analysis of b). It is important to highlight that 9 species evaluated do not have LWR values in FishBase (https://www.fishbase.se/), and the first LWR report for 25 species of shallow-water fish in the Colombian Caribbean Sea is shown in Table 2.
The parameters of the LWR linear regression, the intercept (a) and slope (b), are highly dependent on the body shape of the fish species (Figure 2). Thus, eel-like species tend to be positive allometric, while elongated species tend to be isometric or positive allometric. However, the fusiform and short and/or deep species tend to be negative allometric but some are isometric or positive allometric. Robust multiple regression of log(a) as a function of b and body shape explained 88% of the variance, intercept (a) = 3.4203 ± 0.4217, slope (b) = −2.8190 ± 0.1451, and r2 = 0.881. The values of the relative condition factor (Krel) varied between 0.94 for Rhomboplites aurorubens (Cuvier, 1829) and 1.03 for G. ocellatus (Table 2). A total of 21 species (39.6%) showed a Krel value <1.0 although without significant statistical differences compared to an expected mean value Krel = 1.0 (, more detail of t-test in Table 2).

4. Discussion
92.5% of the fish species evaluated presented a range of parameter b between 2.5 and 3.5 [15], indicating normal growth dimensions [35, 36], except for B. capriscus, Diodon holocanthus Linnaeus, 1758, R. aurorubens with values of b < 2.5, and G. ocellatus with b > 3.5, which showed a narrow length range with respect to the maximum total length reported in FishBase (https://www.fishbase.se/) [25], common in values of b < 2.5 or >3.5 [15, 37].
Compared with other studies in the Colombian Caribbean Sea and the LWR information reported in FishBase (https://www.fishbase.se/), some variations of the b parameter and the type of growth of the species were observed, which may be related to factors such as ontogeny, feeding (amount, quality and size), sex, maturity stage, health, seasonality, habitat, length range, and sample size [26, 27, 38–41]. The differences found in growth can also be attributed to the different sampling methodologies and fishing gear used. In this study, data were collected from a research cruise (independent of the fishery), and trawls were used, which is a nonselective fishing gear, which allowed sampling with a wide spectrum of lengths of fish species. Additionally, the study area presents a high variety of habitats such as coral reefs, seagrass meadows, soft bottoms, among others [42–44], which are important for the life cycle of the species and depending on their state of conservation can be determinant in the health condition and the morphometric characteristics of the sampled species.
We demonstrate the interdependence of the parameters (intercept and slope) of the LWR depending on the body shapes of the fish, similar to what was reported in other studies [15, 45]. Regarding the relative condition factor, we found that around 39.6% of the species evaluated are probably in poor growth conditions, which may be related to high fishing pressure derived from industrial and artisanal fisheries that operate in the study area. All the fish species analyzed are part of the target catch and/or bycatch of the fisheries, and the fishermen use fishing gear with low catch selectivity such as trawl nets, seine nets, and gillnets that could affect the welfare of fish species and cause negative impacts on habitat [2, 6, 7, 46–48]. Although the results of this study were obtained from mixed sexes, they are of great importance for fisheries management, since there is no specific fishing gear for catching each sex and the regulations derived from fisheries apply to the entire population [17].
LWR information is reported for 9 assessed species that currently have no data in FishBase (https://www.fishbase.se/) [25]. In this study, the first estimates of LWR for 25 species of shallow-water fish species in the Colombian Caribbean Sea are presented. This work contributes with biological information on the LWR and the relative condition factor of demersal fish from shallow waters that can be very useful for future research on the assessment of marine populations, which is important for the management and conservation of fish species exploited as target catch and/or bycatch by industrial and artisanal fisheries in the Colombian Caribbean Sea and FAO fishing area 31.
5. Conclusions
In this study, the first estimates of LWR for 25 species of shallow-water fish species in the Colombian Caribbean Sea are presented. We demonstrate the interdependence of the parameters (intercept and slope) of the LWR depending on the body shapes of the fish. 39.6% of the species evaluated are probably in poor growth conditions, which may be related to high fishing pressure derived from industrial and artisanal fisheries. This work contributes with biological information on the LWR and the relative condition factor of demersal fish from shallow waters that can be very useful for the management and conservation of fish species exploited.
Data Availability
The data used to support the findings of this study are available upon request to jparamo@unimagdalena.edu.co or arodriguezj@unimagdalena.edu.co.
Disclosure
This study was part of Alfredo Rodriguez’s PhD thesis.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This study is a contribution of the Tropical Fisheries Science and Technology Research Group (CITEPT) at the University of Magdalena in Colombia. We thank the researchers of the CITEPT Research Group, who collected the data on board the vessel “ADRIATIC” and analyzed the fish samples in the laboratory. Special thanks to the researcher Jorge Viaña from the Institute of Marine and Coastal Research (INVEMAR), who helped with the assignment of the commercial categories of the fish species. Alfredo Rodriguez was sponsored by the Fondo de Ciencia, Tecnología e Innovación (FCTeI) del Sistema General de Regalías (SGR) and the Doctoral Excellence Scholarship Program Bicentenario del Ministerio de Ciencia, Tecnología e Innovación (Minciencias). The scientific fishery sampling was funded by Autoridad Nacional de Acuicultura y Pesca (AUNAP) and University of Magdalena.