450 



Fishery Bulletin 100(3) 



dual channels enabling sequential transmission of pres- 

 sure and temperature. Seven pressure tags and three 

 pressure-temperature tags were deployed. All, except one, 

 were inserted into the stomach. To retard regurgitation, 

 two small hooks were tied to each tag with VICRYL dis- 

 solving sutures before being placed in the stomach. One 

 tuna struck a tag with hooks during a calibration experi- 

 ment, resulting in the tag attaching without the tuna leav- 

 ing the water This fish was tracked for five hours until the 

 tag detached. 



Standard tracking techniques were used (see Holland 

 et al., 1985). The tracking equipment, including the re- 

 mote-controlled hydrophone rotator, is described in Pep- 

 perell and Davis ( 1999). We attempted to track the fish at 

 a distance of about 400 m, which we judged from signal 

 strength, using distance and signal-strength calibrations 

 carried out before tracking. The decoded data from single- 

 channel tags were logged together with global positioning 

 system (GPS) latitude and longitude, nominally at 1-sec- 

 ond intervals, on a computer with VEMCO VSCAN/GPS 

 software (Vemco, 1992). Temperature and depth data from 

 dual-channel transmitters were logged at about 3-second 

 intervals. 



Conductivity-temperature-depth profiles were taken at 

 the start and finish, and at convenient times during each 

 track, with a digital data loggers international profiler 

 Additional information collected during each track in- 

 cluded bottom depth, air temperature, wind speed, cloud 

 cover ( 1-8 sectors covered), brightness ( 1-4), and presence 

 of surface and subsurface schools of SBT In 1994. we also 

 recorded the depth distributions of SBT schools detected 

 on the echosounder under the tracking vessel. 



The tracking data were processed using a filtering pro- 

 gram to delete spurious data, interpolate missing data, 

 and generate data files at nominated time intervals. The 

 program compared each depth with the seven previous 

 data points (which were nominally recorded at 1-second 

 and 3-second intervals for single- and dual-channel tags, 

 respectively) and then deleted spurious records based on 

 expected maximum and minimum depths and maximum 

 depth-change rates. Similarly, a temperature filter de- 

 leted spurious values based on expected maximum and 

 minimum temperatures and maximum temperature- 

 change rates in data from the dual-channel transmitters. 

 Complete vertical tracks were plotted against time with 

 20-second interval data. The position and sustained swim- 

 ming speeds of tracked tuna were assumed to be the same 

 as that of the tracking vessel. By using GPS data, speed 

 was calculated over the ground from the straight-line 

 distance travelled between points in 10 minutes. Verti- 

 cal movements and horizontal movements in relation to 

 the vessel were not considered. An attempt was made to 

 match both the heading and speed of the tracked tuna so 

 that the vessel reflected the movements of the fish. 



Depth distribution of tuna was examined from 20-sec- 

 ond interval data aggregated in 5-m depth bins stratified 

 by day and night. The number of observations in each 

 depth bin was then expressed as a proportion of the total 

 number of observations in each day-night stratum. The 

 speed distribution of tuna was examined from 10-minute 



interval data aggregated in 0.1 m/s bins stratified by day 

 and night. 



Results 



The tracking system was developed for use on a variety 

 of commercial vessels. In April-May 1992, five fish were 

 released and two were tracked successfully (Table 1). 

 Modifications were then made to the system and different 

 transmission frequencies were selected to reduce inter- 

 ference. The improved system was used on two cruises 

 in January and March 1993. Four fish were tracked for 

 periods of up to 49 h (Table 1). 



In 1994, tracking was carried out farther west because 

 fish schools were not sighted farther east (Fig. 1 ), probably 

 due to warm water remaining in the northwest section of 

 the study area (Fig. 2). The whole period (8 January-3 

 February) was characterized by windy conditions that 

 made catching and tracking fish difficult. Ten SBT were 

 released with tags (Table 1). The longest track was 31 h, 

 but some tracks were short because the fish regurgitated 

 their tags. Other tracks were terminated voluntarily after 

 sunset to maximize the number of daytime observations 

 on different fish. 



Tuna 1 was released SW of Liguana Island on 29 April 

 1992 at 08.55 h (Figs. 1 and 2A). It headed WSW towards 

 the shelf edge until tracking was terminated after 12 

 hours. Acoustic interference from the tracking vessel 

 affected calculation of depth at certain vessel speeds, re- 

 sulting in some gaps in the vertical data (Fig. 3A). During 

 the day the fish spent most of its time at the surface, but 

 frequently dived to near the bottom. Tuna schools spotted 

 at the surface appeared to break up and then re-appear 

 sometime later, more or less at the same rate that the 

 tracked SBT moved in and out of the surface waters. The 

 fish dived shortly after sunset and then returned to the 

 mixed layer and oscillated around 30 m until tracking was 

 terminated at 2108 h. 



Tuna 2 was released SW of Liguana Island on 30 April 

 1992 at 1330 h (Fig. 1). The fish moved west throughout 

 most of the track but slowed and changed direction many 

 times from 0400 h. These direction changes occurred when 

 the tuna reached cooler surface waters (Fig. 2A). It re- 

 mained in the 17°C water, moving south along the bound- 

 ary with the 16°C isotherm. Its vertical behavior can be 

 seen in Figure 3B. Like tuna 1, this fish made a brief dive 

 just after sunset. 



Tuna 3 was released at The Lump, 28 km west of Ward 

 Island on 27 January 1993 at 15:22 h and was tracked for 

 49 h (Figs. 1 and 2B). It moved into waters near the shelf 

 edge, a movement matched by a conventionally tagged 

 tuna that was released at 33°15'S, 133°49'E on 26 Janu- 

 ary 1993 and that was recovered five days later at 34°S, 

 132°17'E by a commercial fishing operation. The vertical 

 behavior of tuna 3 is shown in Figure 4A. It spent con- 

 siderable time near the surface in association with other 

 tuna during the late afternoon on the first day It seldom 

 surfaced on the second day; neither did the other tuna. 

 A commercial pole-and-line boat fishing in the same area 



