Edwards and Pernn Annual dolphin mortality 



629 



The U.S. data set for 1987 provided a unique oppor- 

 tunity to investigate these effects. That year, for the 

 first time, observers accompanied nearly 100% of trips 

 by U.S. -registered purse-seiners to the ETP. By sub- 

 sampling the 1987 data set, various observer levels 

 and fleet sizes could be simulated and their effects on 

 dolphin mortality estimates studied. In the absence of 

 any other available data, these data were taken by 

 NMFS as roughly representative of the international 

 fleet. The simulation results guided subsequent legis- 

 lation regarding mortality comparability criteria for 

 non-U.S. fleets in the ETP. 



Total annual mortality estimates are presented here 

 for three dolphin groups (all stocks combined): spotted 

 dolphin (Stenella attenuata); whitebelly spinner dol- 

 phin (Stenella longirostris longirostris); and common 

 dolphin (Delphinus delphis). These three groups rep- 

 resent the range of data types (good, intermediate, 

 and poor in quality) in the 1987 data set, and are 

 generally representative of data types collected every 

 year that may be affected by variable fleet sizes and 

 levels of observer coverage in future fishing years. 



Accuracy and precision of the mortality estimates 

 are presented in terms of relative bias and coefficient 

 of variation. Frequency distributions and analytic con- 

 fidence intervals are examined for underlying patterns 

 in mortality data responsible for generating observed 

 patterns in accuracy and precision. 



Methods 



Data 



During 1987, 34 U.S. -registered purse-seine 

 vessels carrying certificates that permitted 

 fishing "on dolphins" made a total of 134 

 fishing trips to the eastern tropical Pacific 

 Ocean (ETP). Of these vessels, 33 made 124 

 trips that were observed by either a National 

 Marine Fisheries (NMFS) or an Inter- Ameri- 

 can Tropical Tuna Commission (IATTC) tech- 

 nician. These 124 observed trips comprised 

 the entire population of trips for this study. 

 They accounted for 93% of all trips and 96% 

 of all days spent at sea in the ETP by Un- 

 registered tuna purse-seiners during 1987. 

 Trips made by individual vessels within fleets 

 were considered replicates and observed trips 

 were assumed to represent unobserved trips 

 accurately. Data consisted of records of date, 

 time, trip number, set number, and kill per set. 

 Between 4 and 20 vessels carrying observ- 

 ers departed each month during 1987 (Fig. 1). 

 Approximately the same number (although 



not necessarily the same vessels) of observed vessels 

 (about 30) fished in the ETP during any given month 

 (Fig. 2). Just over half the trips (65/120) lasted about 

 three months. Trip duration ranged from one to six 

 months (Fig. 2). Number of trips incurring mortality 

 of each dolphin group type ranged from 93 (offshore 

 spotted dolphins; Fig. 3) to 12 (common dolphins). 



Dolphin mortality data were not stratified geographi- 

 cally, despite the demonstrated efficiency of this strati- 

 fication scheme (e.g.. Hall and Boyer, 1986) because it 

 can be applied only when data sets provide reasonable 

 numbers of reasonably homogeneous data per stra- 

 tum. Data tend to be so variable for the less frequently 

 killed groups, such as common dolphins, that mortal- 

 ity estimates tend to be very imprecise (and infrequent) 

 even without stratifying data. This is likely to be a 

 problem for all dolphin types for small fleets in every 

 year 2 . 



Data quality differed between group types in both 

 frequency and extent of kill. Data for offshore spotted 

 dolphins were the "best" in terms of being most plenti- 

 ful and most evenly distributed between trips collect- 

 ing data. The frequency distribution of kill per day per 

 trip (i.e., number of dolphins killed per number of days 

 at sea, for a given trip) is relatively smooth (Fig. 3). 

 Most trips killed about 0.5 dolphins per day, a few 



-At the time this study was conducted (1988-1989), there were no 

 legal requirements for data stratification. Subsequent to this study, 

 the U.S. Congress enacted legislation (Federal Register Notice Vol. 

 54, #43, Tuesday March 7, 1989; 54 FR9438-9451 ) requiring stratifi- 

 cation of data for annual mortality estimates into three geographic 

 strata and two dolphin groupings (common dolphins, and all other 

 dolphins). 



"> Oi. 



UJ £■* 



tn 



0C 20 



O 



Li. 



O ,6 



DC 



HI 



JAN FEB MAR APRIL MAY JUNE JULY AUG SEPT OCT NOV DEC 

 I I Cruises at sen MONTH ! ] Month departed 



Figure 1 



Number of observed U.S. -registered tuna purse-seine vessels at sea, and 

 number of observed vessels leaving port each month during 1987. 



