Abstract.- During 1986, the Na- 

 tional Marine Fisheries Service be- 

 gan conducting long-term research 

 vessel surveys to determine trends 

 in population size of dolphin stocks 

 taken incidentally by tuna piu-se sein- 

 ers in the eastern tropical Pacific. 

 Line transect methodology was used 

 by observers aboard two vessels for 

 120 days each. We assumed the vari- 

 ability associated with the abundance 

 estimates would be relatively con- 

 stant during the sampling period and 

 investigated (1) annual changes in 

 population size of the northern off- 

 shore spotted stock that could be 

 detected within a 5-year (six survey) 

 sampling period, and (2) the number 

 of years required to detect a 10% an- 

 nual decline with a and /? error levels 

 of 10%. The abundance estimate of 

 the northern offshore spotted dol- 

 phin stock using the first year's data 

 was 929,000 animals. After 5 years, 

 a minimum 17.6% annual decline 

 could be detected during which 62% 

 of the stock would decrease. A 10% 

 annual decline can be detected in a 

 minimum of 8 years. Data from sub- 

 sequent surveys will be investigated 

 that may improve our ability to de- 

 tect smaller annual declines. 



Monitoring Trends in Dolphin 

 Abundance \n the Eastern Tropical 

 Pacific Using Research Vessels Over 

 a Long Sampling Period: Analyses 

 of 1986 Data, The First Year 



Rennie S. Holt 

 Stephanie N. Sexton 



Southwest Fisheries Center 



National Marine Fisheries Service, NOAA 



La Jolla, California 92038 



Manusci-i|it accepted 21 Septeml)er 1989. 

 Fishery Bulletin, U.S. 88:105-111. 



The National Marine Fisiieries Ser- 

 vice (NMFS) is responsible for assess- 

 ing the status of dolphins taken inci- 

 dentally by tuna purse seiners in the 

 eastern tropical Pacific (ETP) (Rich- 

 ey 1976). The status of spotted dol- 

 phins Stenella attenuata is of special 

 concern because it is the major spe- 

 cies taken by the fishery (Smith 1979). 

 Of the three stocks of spotted dol- 

 phins, the northern offshore stock 

 has been fished more frequently than 

 any other stock. The four stocks of 

 spinner dolphins S. longirostris, and 

 the four stocks of common dolphin 

 Delphinufi delphis, are also taken. 

 The four stocks of striped dolphin 5. 

 coeruleoalha and the Eraser's dolphin 

 Lagciiodelplns hosei are occasionally 

 caught (Holt and Powers 1982). These 

 five species are herein grouped and 

 termed target species. 



NMFS conducted assessments of 

 target populations in 1976 (SWFC 

 1976) and again in 1979 (Smith 1979) 

 using absolute stock abundance. The 

 validity of the absolute estimates de- 

 pended on several assumptions (i.e.. 

 all schools located directly on the 

 trackline were detected, schools did 

 not respond to the ship before being 

 detected, etc.). Unfortunately, not all 

 of these underlying assumptions were 

 met (Holt and Cologne 1987, Holt 

 1987). Therefore, in 1984 Congress 

 amended the Marine Mammal Pro- 



tection Act and mandated that an 

 alternative approach for assessing 

 stocks be used that was less sensitive 

 to biases. By using consecutive popu- 

 lation estimates and establishing the 

 first estimate as a base to which all 

 subsequent estimates would be rela- 

 tive, we can detect trends in stock 

 sizes over a long sampling period. 

 These relative estimates can provide 

 an assessment of stock condition if the 

 biases in the abundance estimates are 

 consistent over the sampling period. 

 In 1986, NMFS initiated a research 

 program to monitor dolphin popula- 

 tions in the ETP which would utilize 

 two research vessels for at least 5 

 years during which six sui-veys would 

 be conducted. The research design 

 (Holt et al. 1987) indicated that a 10% 

 annual rate of decrease in northern 

 offshore spotted dolphins could be 

 detected (a total 41%) decrease over 

 six surveys). Herein, we present the 

 population estimates for the first 

 year's survey data. We also discuss 

 effects of several factors on these 

 base estimates, and, assuming data 

 for subsequent surveys will have the 

 same level of precision (coefficient of 

 variation levels) as the first year, ex- 

 amine changes in population sizes 

 that can be detected in 5 years or, 

 conversely, the number of years re- 

 cjuired to detect various levels of 

 change. 



105 



