Domeier et al.: Tracking Thunnus thynnus onentalis with the aid of an automated algorithm 



299 



End 



Tstart > t(1), t{1) > t(2), 



T(2) > t (3), t(3) > end 



Figure 4 



Diagram to show how the best track is calculated by summing the cost of arcs for 

 all possible paths and then choosing the track of least cost. 



merit patterns and habitat use of Pacific bluefin tuna 

 in the eastern Pacific. Monthly data were combined 

 within each tag data set prior to performing utilization 

 distribution analyses with the Home Range Exten- 

 sion for ArcView (version. 1.1c, BlueSky Telemetry, 

 Aberfeldy, Scotland) that employs the fixed kernel 

 method (Rodgers and Carr, 1998). Results were dis- 

 played as volume contours displaying the main centers 

 of activity for each fish during a given time period. 

 Initial analyses allowed us to combine data so that 

 figures could be minimized. For the archival tag data, 

 consecutive months with similar spatial distribution 

 were combined and individual fish with very similar 

 tracks were combined. All data from fish that were 

 PSAT tagged were combined by month because of the 

 relatively sparse data compared with the data from 

 the archival tags. PSAT tag data provided a glimpse at 

 year-to-year variations in bluefin distribution (August 

 2000 through October 2002), whereas the archival tag 

 data were for a single year and allowed for a monthly 

 comparison within one year (August 2002 to Septem- 

 ber 2003). 



The near daily position data provided through the 

 PSAT Tracker analyses allowed us to calculate the 

 swimming speed of each fish. This was done by simply 

 dividing the horizontal distance between consecutive 

 data records by the time between consecutive data re- 

 cords ( 1-4 days). 



Results 



Tag recoveries 



Fifteen of the PSAT tags transmitted data after remain- 

 ing on the fish from 2 to 191 days (Table 2). Unfortu- 

 nately some of these tags did not transmit usable data. 

 Fourteen of them provided a pop-up location and eleven 

 of them transmitted enough data for some level of analy- 

 ses of behavioral and movement patterns. The Microwave 

 Telemetry PSAT tag provided an archival data set with 

 a one-hour sampling schedule. The Wildlife Computer 

 PSATs transmitted data summaries that included a 

 daily water column profile of temperature (obtained from 

 the deepest dive) and the percent time each fish spent 

 within predetermined temperature and depth bins. 



Four archival tags were recovered after a period at 

 liberty of 16 hours to 385 days (Table 2). The 16-hour ar- 

 chival tag recovery was made from a recreational angler 

 very near the point of release; this tag was not used for 

 any analyses. The three tag recoveries made after 300 

 days came from a purse-seine vessel. Two of these three 

 recaptured fish spent several weeks in a grow-out pen 

 before the tags were discovered; the dates the fish were 

 in the pen were not used for any analyses. The light 

 stalks of tags 441 and 159 were damaged during recov- 

 ery. For these tags, the internal temperature and pres- 

 sure sensors were verified by Lotek data, but external 



