Stock structure delineation of Kawakawa Euthynnus affinis (Cantor, 1849) from Malaysian Borneo using multivariate morphometric analysis

2.1 Specimen accumulation

Overall, 188 individuals of Kawakawa, E. affinis from eight localities of Malaysian Borneo were collected (Table 1, Fig. 1). The specimens were collected from the landing sites specified in the annual report by the Department of Fisheries Malaysia (DoF), which provided information on the availability of fish at these locations. Samples were collected from three main geographic areas: Celebes Sea (CS), Sulu Sea (SS), and South China Sea (SCS) (Fig. 1). Based on references discovered by Collette and Nauen (1983), all samples were morphologically recognized and validated. All specimens were transferred to the Molecular Ecology Laboratory at Universiti Sains Malaysia's School of Biological Sciences. The specimens were rinsed with running water and gently tapped before being placed on a flat surface with a black background for improved visibility. The samples were tagged meticulously and captured using a Nikon D90 digital camera, and morphometric variables were measured using a digital caliper with a 0.1 cm precision (Fig. 2).

Close

Fig. 1

Sampling locations of Euthynnus affinis specimens collected from the South China Sea, Sulu Sea, and Celebes Sea regions.

Export to PPT

Close

Fig. 2

Morphometric measurements of Euthynnus affinis (Kawakawa). TL: total length, FL: fork length, SL: standard length, HL: head length, BD: body depth, ED: eye diameter, SnL: snout length, DP: distance of pectoral fin, DD1: distance of the first dorsal fin, DD2: distance of the second dorsal fin, DV: distance of ventral fin, DA: distance of anal fin.

Export to PPT

2.2 Morphometric analyses

Using a digital calliper, twelve morphometric variables were measured, including head length (HL), snout length (SnL), fork length (FL), standard length (SL), total length (TL), body depth (BD), eye diameter (ED), distance of pectoral fin (DP), distance of the first dorsal fin (DD1), distance of anal fin (DA), distance of the second dorsal fin (DD2), and distance of ventral fin (DV) (Fig. 2). The study used multivariate discriminant function analysis to identify attribute combinations that optimize E. affinis population differentiation, aiming to estimate the proportion of accurately classified recycled characters, indicating discriminatory behavior in the classification process. SPSS version 25, a widely used statistical software in social sciences research, was used for discriminant function analysis. (Nie et al., 1975). The Principal Component Analysis (PCA) was performed using PAST software (version 4.03) (Hammer et al., 2001) to determine the maximum potential diversity of body shapes for population classification purposes. To analyze the relationships between the eight E. affinis populations, the (UPGMA) Unweighted Pair Group Method with Arithmetic Mean was utilized to perform a grouping assessment/cluster analysis (CA) with a hierarchical structure. This method relied on Mahalanobis distances (Hammer et al., 2001). Elliott et al. (1995) propose an equation for normalizing morphometric variables that accounts for size variation across populations: Madj = M(Ls/L0)^b, where M is the raw measurement, Madj is the size-adjusted measurement, Ls is the mean standard length across all samples, L0 is the standard length for each sample, and b is determined for each character from the observed data as the slope of the log M on log L0 regression using all fish across populations.

3.1 Sampling data

One hundred and eighty-eight specimens of Kawakawa, E. affinis were collected from eight landing sites in Malaysian Borneo, covering three main geographic areas: Sulu Sea, South China Sea, and Celebes Sea. Samples were collected from four major landing sites: Kota Kinabalu (Sabah), Bintawa, Pulau Bruit, and Mukah (Sarawak) representing the South China Sea, Sandakan and Kudat (Sabah) representing the Sulu Sea, and two more landing sites, Lahad Datu and Tawau (Sabah) representing the Celebs water regions (Table 1, Fig. 1).

3.2 Discriminant function analysis (DFA)

The discriminant function analysis identified five functions, 1, 2, 3, 4, and 5, with variances of 81.0 %, 9.5 %, 6.5 %, 2.8 %, and 0.2 %, indicating a significant link between population segregation and the value of 0.80 for the first function (see Table 2). The Wilk's lambda test assigns a probability of p = 0.000 to the analyses of functions 1 through function 5 (Wilk's Lambda = 0.244) and a probability of p = 0.000 to functions 2 through function 5 (Wilk's Lambda = 0.679) (Table 2). When the morphometric measures examined in this research were placed through DFA assessment based on character grading in function 1, the morphometric measurement with the highest score was the total length (TL, 0.916), followed by the standard length (SL, 0.898), body depth (BD, 0.890), fork length (FL, 0.885), head length (HL, 0.736), pectoral fin distance (DP, 0.711), eye diameter (ED, 0.706), and snout length (SnL, 0.696) (Table 3).

As determined by the DFA analysis, the percentage of eight E. affinis populations correctly identified is Sandakan, Sabah (50 %), Lahad Datu, Sabah (66.7 %), Tawau, Sabah (20.8 %), Kudat, Sabah (33.3 %), Kota Kinabalu, Sabah (33.3 %), Bintawa, Sarawak (20.8 %), Pulau Bruit, Sarawak (54.2 %), and Mukah, Sarawak (29.2 %) (Table 4). The DFA test used in this investigation revealed that the population from Lahad Datu, Sabah, had the greatest percentage of correctly classified individuals (66.7 %). In contrast, the populations of Tawau in Sabah and Bintawa in Sarawak both showed the highest percentage of misclassification at 79.2 % (Table 4). Following the findings of a canonical discriminant function assessment, the typical size of all Bornean groups was almost the same (Fig. 3). Population overlap may be the cause of this resemblance in size.

Close

Fig. 3

Canonical discriminant analysis based on morphometric measurements of eight populations of Euthynnus affinis.

Export to PPT

3.3 Principal component analysis (PCA)

With an estimated body shape variation of 82.562 % and a highly significant eigenvalue of 9.907, the first component of the principal component analysis exhibits the highest variance among the 12 components used to estimate variations in body shape [(it is considered significant if the eigenvalue is greater than 0.3 (Nimalathasan, 2009)]. The first four principal components (PC) explained 96.417 % of the common variance (82.562 %, 8.792 %, 3.179 %, and 1.884 % for PC1, PC2, PC3, and PC4, respectively) (Table 5). PCA analysis revealed that the overlapped patterning between PC1 and PC2 in the scatter plot shown in Fig. 4 strongly supports the restricted variation in the shape of E. affinis body, for which there are very high and significant eigenvalues that can be used to distinguish E. affinis according to typical body pattern. Fig. 4 demonstrates that the population is mixed with no visible distinction, indicating that physical homogeneity exists amongst Malaysian Borneo groups. In general, outcomes from the PCA assessment corroborate those from the DFA research. According to a Mahalanobis distance-based hierarchical cluster assessment, the eight E. affinis populations were separated into three distinct groups (Fig. 5). The first category includes Kota Kinabalu (Sabah), Mukah, Bintawa, and Pulau Bruit (Sarawak) populations. The second group comprises E. affinis populations from Lahad Datu and Kudat (Sabah), while the third group includes populations from Tawau and Sandakan (Sabah).

Close

Fig. 4

Scatterplot of first two principal components of Euthynnus affinis (n = 188) based on morphometric measurements.

Export to PPT

Close

Fig. 5

Dendrogram generated from cluster analyses of morphometric measurements using Mahalanobis distances among eight populations of Euthynnus affinis collected from Malaysian Borneo's waters.

Export to PPT

4.1 Discriminant function analysis (DFA)

The discriminant analysis helped identify data segregation functions, revealing the organization of species stock by identifying discrete clusters (Aminan et al., 2020). The experiment revealed that Function 1 in DFA demonstrated the highest efficacy in differentiating E. affinis, with a remarkable variability of 81.0 % and a Wilk's lambda score of 0.000 (Table 2). Total length (TL) was found to be the most important factor in determining E. affinis population membership, with the highest contribution score, followed by standard length (SL), body depth (BD), fork length (FL), head length (HL), pectoral fin distance (DP), eye diameter (ED), and snout length (SnL) in function 1 (Table 3). Interestingly, canonical discriminant analysis, on the other hand, reveals that the Malaysian Borneo's population is morphologically homogenous (Fig. 3).

The observation supports previous research by Aminan et al. (2020) and Gain et al. (2017), which demonstrated that the key feature (function 1) in DFA evaluation can differentiate between Rasbora spp. and Cirrhinus cirrhosus populations, respectively. Other research carried out in the Mediterranean region of Tunisia indicated that the standard length, the length of the pectoral fin, and the length of the fork are all parameters that have a role in classifying the population of bullet tuna, Auxis rochei (Allaya et al., 2016). Similarly, Karakulak et al. (2016) found a strong correlation between the pectoral fin length of Euthynnus alletteratus and Thunnus thynnus, as well as the length of the head, the first predorsal fin, and the height of the second dorsal fin. Anzeer et al. (2021) further substantiated this observation by employing discriminant function analysis to identify variations in body shape and size among larvae of Dascyllus carneus. The results demonstrate significant heterogeneity, particularly in the head, trunk, and tail. Different environmental niche utilization, food availability, and prey species may also contribute to differences in head morphology (Cronin-Fine et al., 2013). This investigation's findings align with those of El Mghazli et al. (2021), who achieved a 96 % accuracy rate in identifying Horse Mackerel Trachurus trachurus specimens from the Atlantic waters off Morocco's coast.

Even among closely related fish species, it is important to note that differences in body shape and size are essential for recording morphological differences, differences in sizes and shapes, and constructing phylogenetic correlations based on body shape changes. It is also vital to emphasize that body shape and size are the most significant instruments (Imtiaz and Md Naim, 2018). Aquatic organisms, like fish, must adapt to their habitat's conditions to maintain their biological systems' normal functioning. The interesting thing about the morphological analysis done in this study is that there is variation across populations of the same species, which may signal changes in the environment and behavior of the species (Jalili et al., 2015). The observed variations in morphometric characteristics can be attributed to both intrinsic and extrinsic factors, including independent and dependent elements, as well as exogenous variables like food availability and water quality (Kumar et al., 2019). To predict group membership data, the greater standard of resemblance of the investigated components is utilized (Kočišová and Misanková, 2014; Aminan et al., 2020). Fisheries management, biodiversity conservation, and natural resource management depend on population discrimination and species recognition. Based on our investigation, it is evident that DFA, PCA, and CA are instrumental in differentiating the E. affinis population across Malaysian Borneo. These analytical tools not only play a crucial role but also provide a substantial foundation for comprehending essential information about E. affinis. This understanding is pivotal for the efficient management and safeguarding of regional resources in Malaysian Borneo.

4.2 Principal component analysis (PCA)

The first four Principal Components (PCs) addressed 96.417 % of the diversity in body shape in total variables (Table 5). The discovery is consistent with the findings of Addis et al. (2014), who revealed that the first two PCs captured 56.97 % of the shape variance in Thunnus thynnus. Imtiaz and Md Naim (2018) found 80 % variations in body shape of Nemipterus genus and reported comparable findings. Based on these observations, we conclude that changes in E. affinis body shape exist in total body shape. Consequently, this result is analogous to that of Jalili et al. (2015), who found that morphological differences between Alburnus filippii were determined by snout tip and body depth variation. One noteworthy illustration of the prevalence of the results of this study on marine fish populations is the study done by Mounir et al. (2019). They discovered that the most relevant characteristics describing variations in Sardina pilchardus populations from Morocco's Atlantic coast were head length, dorsal fin, and body depth. Moreover, Nasri et al. (2019) showed that body depth, the length of the head, dorsal fin, and peduncle tail are significant in detecting variance among Cyprinion species in Persian waters using morphometric approaches. The PCA results generally complement the DFA findings, indicating that morphological homogeneity exists among E. affinis populations from Malaysian Borneo. Periodic movement in these regions might explain the multiple dispersion of these samples, implying that this population has a temporal and geographical residency in this location.

Intriguingly, the results of this study have confirmed previous research on the E. affinis genetic composition across many coastal environments, including Taiwan (Chiou and Lee, 2004), the Philippines and Peninsular Malaysia (Santos et al., 2010), the Indian Ocean North (Kumar et al., 2012), Malacca Strait (Masazurah et al., 2012), and Tanzanian coastal waters (Johnson et al., 2016), where the research shows that there is just one genetic pool of E. affinis. Based on the hierarchical cluster analysis results, the E. affinis population was classified into three categories (Fig. 5). The first cluster consists of individuals from Kota Kinabalu (Sabah), Mukah, Bintawa, and Pulau Bruit (Sarawak), the second comprises population from Lahad Datu and Kudat (Sabah), and the third group includes population from Tawau and Sandakan. The unique geographical conditions and environmental limits experienced by each population may account for the observed differences across these populations (Siddik et al., 2016). The ecology of the ocean, which includes many different biological zones and factors, including the quantity of light that reaches the ocean floor, the amount of food that is available, and the speed at which the water moves, has also been connected to such variations (Imtiaz and Md Naim, 2018). Populations of fish living in different locations may look different from one another because of environmental, genetic, and selective interactions that cause differences in morphology across species (Sajina et al., 2011). Based on cluster analysis, our findings are comparable to those of Siddik et al. (2016) and El Mghazli et al. (2021), who unearthed that populations of gangetic whiting (Sillaginopsis panijuswere) from Bangladesh and Atlantic horse mackerel (Trachurus trachurus) from Morocco were divided into two groups.

It is vital for the sound and efficient management of fisheries resources to better understand the population organization of the species being evaluated. This is because each stock has to be managed individually to maximize productivity (Tzeng, 2004). Therefore, ignoring the stock analysis approach can result in an array of issues, such as the setback of genetic variability, modifications in biological characteristics such as making fish smaller, overexploitation of populations that produce less, and erroneous projections of how mitigation strategies might indeed influence stocks. These problems can all contribute to a decline in overall stock abundance (Cronin-Fine et al., 2013). Morphometric differences across regional groupings can be related to different genetic structures or various environmental influences in each location (Tzeng, 2004). Fish exhibit greater diversity in phenotypic traits inter and intra-population compared to any other vertebrate, making them more susceptible to such changes, eventually modifying their morphological characteristics. Fish also have more physical variety amongst populations than other animals (Verma et al., 2014; Benali et al., 2021). Morphometric discrepancies amongst groups are anticipated to be related to geographic separation and differing evolutionary origins (Aminan et al., 2020).