Identification Application Of Geometric Morphometrics In Ommastrephide Economic Zone

Ommastrephide bartramii and Dosidicus gigas is the largest and one of the most abundant of the family Ommastrephide and the subfamily Ommastrephide. These spices play the significant role in marine ecosystems as both predators and prey, and are regarded as indicator species for changes inthe marine environment. Ommastrephide bartramii and Dosidicus gigas has the surface layer cluster behavior, the characterisitics with fast growth and complex population structure. Masterig the population structure of oceanic squids have many important significances for protecting and managing these species. According to the statolith and beak ecological samples of Ommastrephide bartramii caught in North Pacific ocean and Dosidicus gigas caught in equatorial waters off Peru, Costa Rica and Chile. These beak and statolith were analyzed based on traditional measurement, Fourier measurement and landmark measurement to identify species and populations structure. The main results are as follows:(1) According to statolith samples about Ommastrephes bartramii caught in North Pacific ocean and Dosidicus gigas caught in equatorial waters off Peru from May to October in 2013, their statolith morphology were analyzed by traditional measurement and Fourier analysis for species identification. Nine of variables were measured on the statolith; digitalized photographs of statolith were transformed to Ellipse Fourier descriptors(EFDs). Finally, nine statolith variables and 77 standardized EFDs were analyzed by principal component analysis(PCA) and stepwise discriminant analysis(SDA). The correction of classification was also compared between two methods. The results showed that the characteristics of morphology were narrow body, long and narrow rostrum and large dorsal tallied with the Ommastrephe bartramii statolith; the statolith of Dosidicus gigas statolith had large wing and narrow rostrum. All the morphological variables had significant difference between two species(P<0.01). The highest variables were Total Length(TSL/ML) and Maximum width(MW/ML) respectively with traditional measurement. Result of PCA showed that first thirteen components accounted for 70% of total contribution of mutation, and the scatter-plot reveals relative overlapped within two species. SDA showed that dorsal lateral length(DLL/ML), lateral dome length(LDL/ML), wing length(WL/ML), Rostrum width(RW/ML) entered as coefficients in the function, total correct classification was 94%. 12 harmonics of 77 EFDs were entered in SDA with Fourier analysis with total correct classification of 99%.(2) According to beak samples about Ommastrephes bartramii caught in North Pacific ocean from August to October in 2013 and Dosidicus gigas caught in equatorial waters of Equator from April to July in 2014. Three upper beak and lower beak were analyzed based on traditional measurement and Fourier analysis for species identification. Twelve of variables were measured on the upper beak and lower beak, digitalized photographs of beak were transformed to Ellipse Fourier descriptors(EFDs). Finally, 77 standardized EFDs were analyzed by principal component analysis(PCA) and stepwise discriminant analysis(SDA). The result showed that the Upper Hood length(UHL), Upper Crest length( UCL), Upper Rostrum length( URL), Upper Rostrum width( URW), Upper Lateral Wall Length(ULWL), Lower Lateral Length( LCL), Lower rostrum Length( LRL), Lower rostrum width( LRW), Lower crost Wall Length(LLWL), Lower Wing Length( LWL) had significant difference between two species(P<0.01), Upper Wall Length(UWL), Lower Hood Length(LHL)had no significant difference between two species(P>0.05). Result of PCA showed that the highest variadbes were Lower rostrum Length( LRL) and Lower rostrum width( LRW), first sixteen components accounted for 83.58% of total contribution of mutation, and the scatter-plot reveals relative overlapped within two species. Three are 16 principal components with 83.58% cumulative contribution of upper beak and 73.76% cumulative contribution of lower beak. SDA showed that Dorsal Lateral Length(DLL/ML), Lateral Dome Length(LDL/ML), Wing Length(WL/ML), Rostrum Width(RW/ML) entered as coefficients in the function, total correct classification was 95%. 18 harmonics of 77 EFDs were entered in SDA with Fourier analysis with total correct classification of 97.6% and 85.7% respectively.According to beak samples about Ommastrephes bartramii caught in North Pacific ocean from January to May in 2014, and Dosidicus gigas caught in equatorial waters off Peru from December in 2013 to February in 2014. Upper beak and lower beak were analyzed by processing image using relative warp as principal component, sixteen RW values were used stepwise discriminant analysis for species identification, and grad distortion via thin plate spline. The results showed that the 1st, 2nd and 3rd of upper beak RWs accounted for 74.63% of the total variation. Among which TypeⅠlandmarks 2 and 3, TypeⅡlandmarks 7 and 8,Type Ⅲ landmarks 6 and 10 are important contributors, and could explain the most variation. Total correct classification could reach 84.15%, or 82.05% by cross validation. Relative warp showed that the 1st, 2nd and 3rd of lower beak accounted for 79.14% of the total variation, which TypeⅠlandmarks 5 and 6, which TypeⅡlandmarks 9 and 10, which Type Ⅲ landmarks 6 and 8 could explain the most variation. Successing in identifying the lower beak could reach 95.75%, which similar to cross validation.(3) According to the statolith morphology of three stocks of Dosidicus gigas caught in Costa Rica, Peru and Chile from May in 2008 to October in 2013. The differences of three stocks of statolith were analyzed by processing image using relative warp as principal component, twenty four RW values were used stepwise discriminant analysis for species identification, and grad distortion via thin plate spline. The results showed that TypeⅠlandmarks 2 and 13, TypeⅡlandmarks 3, 4, 5 and 6,Type Ⅲ landmarks 14 are important contributors, and could explain the most variation. The results showed that the 1st, 2nd and 3rd of CR statolith RWs accounted for 74.76% of the total variation; PE statolith RWs accounted for 52.25% of the total variation; CH statolith RWs accounted for 56.84% of the total variation. Successing in identifying of CR statolith could reach 100%, with similar to cross validation; which PE statolith could reach80.3%, cross validation could reach 77%; CH statolith could reach 78.6%, with similar to cross validation.