Application of otolith shape analysis for stock discrimination and species identification of five goby species(Perciformes: Gobiidae) in the northern Chinese coastal waters

We tested the use of otolith shape analysis to discriminate between species and stocks of five goby species( Ctenotrypauchen chinensis, Odontamblyopus lacepedii, Amblychaeturichthys hexanema, Chaeturichthys stigmatias, and Acanthogobius hasta) found in northern Chinese coastal waters. The five species were well differentiated with high overall classification success using shape indices(83.7%), elliptic Fourier coefficients(98.6%), or the combination of both methods(94.9%). However, shape analysis alone was only moderately successful at discriminating among the four stocks(Liaodong Bay, LD; Bohai Bay, BH; Huanghe(Yellow) River estuary HRE, and Jiaozhou Bay, JZ stocks) of A. hasta(50%–54%) and C. stigmatias(65.7%–75.8%). For these two species, shape analysis was moderately successful at discriminating the HRE or JZ stocks from other stocks, but failed to effectively identify the LD and BH stocks. A large number of otoliths were misclassified between the HRE and JZ stocks, which are geographically well separated. The classification success for stock discrimination was higher using elliptic Fourier coefficients alone(70.2%) or in combination with shape indices(75.8%) than using only shape indices(65.7%) in C. stigmatias whereas there was little difference among the three methods for A. hasta. Our results supported the common belief that otolith shape analysis is generally more effective for interspecific identification than intraspecific discrimination. Moreover, compared with shape indices analysis, Fourier analysis improves classification success during inter- and intra-species discrimination by otolith shape analysis, although this did not necessarily always occur in all fish species.