Boggs Estimating capture depths of longline-caught pelagic fish 



655 



et al. (1990) weighed ~10-12kg, whereas longhne- 

 caught bigeye tuna in the present study averaged 

 >30kg. 



Results of the present study apply predominantly to 

 daytime habitat depths, but an important difference ap- 

 parently exists between the daytime and nighttime 

 depth distributions of bigeye tuna (Holland et al. 1990). 

 At night, tracked bigeye tuna move upward to ~70- 

 90m at temperatures of 23°-25°C. Confirmation of 

 this nocturnal behavior comes from a new nighttime 

 longline swordfish fishery that has recently developed 

 in Hawaii using chemical light sticks. Although this 

 fishery deploys very shallow (generally <90m) gear, 

 the bycatch of bigeye tuna is surprisingly high (S. 

 Pooley, NMFS Honolulu Lab., pers. commun., April 

 1991), indicating that bigeye tuna have a shallow night- 

 time depth distribution. 



The small number of yellowfin tuna caught in this 

 study makes estimated habitat depth (40-200 m) less 

 certain, but it does not differ much from the 90-230 m 

 depth found in Suzuki and Kume (1982) and Yang and 

 Gong (1988). Tracking studies (Carey and Olson 1982, 

 Holland et al. 1990) show yellowfin tuna spend most 

 of their time at depths <100m. Depths of the highest 

 longline CPUEd for yellowfin tuna in the present 

 study (40-80 m; Fig. 7) are similar to the depths (30- 

 80 m) at which tracked yellowfin tuna in Hawaii spend 

 over 50% of their time during the day (Holland et al. 

 1990), tending to confirm that yellowfin tuna habitat 

 is mostly in the mixed layer. 



Methods for estimating habitat depths in the present 

 study could be improved by increasing the number of 

 TDRs deployed or by developing a model, calibrated 

 with TDRs, to predict gear depth based on wind and 

 current measurements, divergence or convergence of 

 floats, and stops and starts in deployment and retrieval. 

 Procedures to estimate the capture depths of fish 

 caught while hooks are sinking or rising could also be 

 developed, but would depend on very accurate time- 

 keeping, since the gear rises rapidly during retrieval 

 (Fig. 3). 



Catch by moving hooks 



The catch of shallow-swimming species on deep hooks 

 moving through shallower depths could reduce the 

 selectivity of gear designed to catch deep-swimming 

 species. The results show that moving longline hooks 

 are more effective (per unit time) than settled hooks 

 at catching billfish, mahimahi, some sharks, and most 

 other non-tuna species. However, the majority of these 

 fish are caught on settled hooks, because of the longer 

 time that hooks are settled (Fig. 5). The relative 

 amount of time hooks are moving vs. settled is the only 

 aspect of the commercial daytime tuna longline opera- 



tions that differs much from the fishing method used 

 in this study. The gear is left in the water longer and 

 then retrieved more rapidly during commercial fishing, 

 so hooks spend less time moving and more time settled. 

 This may result in greater proportions of fish being 

 caught on settled hooks by commercial fishermen than 

 in the present study. 



Eliminating shallow-settled hooks should greatly 

 reduce the catch of shallow-swimming species. For non- 

 tuna species, deploying and retrieving the gear less 

 often (as in commercial operations) should decrease the 

 CPUT (catch-per-unit-time), but would increase the 

 CPUE because the latter increases with set duration. 

 In contrast, bigeye tuna CPUT and CPUE should in- 

 crease with less frequent deployment and retrieval, 

 because CPUT is highest for settled hooks. 



The mechanism for increased CPUT on moving hooks 

 for non-tuna species is unclear. Moving bait may be 

 more attractive than settled bait, but the low number 

 caught on sinking hooks (Fig. 5) suggests that gear 

 motion alone is not responsible for increased catch rate. 

 Perhaps a gradual aggregation of fish around the gear 

 (or the vessel) while the gear is settled contributes to 

 the catch rate by rising hooks. 



Although hook timer data provide a reliable way to 

 confirm when fish are caught on settled hooks, such 

 data may be less reliable as a measure of fish caught 

 on moving hooks, because of the uncertainty regarding 

 fish with timers triggered at recovery (Fig. 5). These 

 fish are not included in the number captured on mov- 

 ing hooks (Table 4); their timer readings cannot be 

 distinguished from ones triggered after being brought 

 aboard. Therefore, the estimates of fish caught on mov- 

 ing hooks (Table 4) may be too low. Alternatively, if 

 these readings indicate a tendency for some fish to not 

 activate timers until they struggle during recovery, 

 then the estimates of fish caught on moving hooks may 

 be too high. In either case, only inferences regarding 

 CPUT on moving and non-moving hooks, and the 

 estimated proportions of fish caught on moving hooks, 

 are affected by this uncertainty. The estimates of 

 catches on non-moving hooks are conservative, and 

 confirmed capture depths are not affected. 



The higher proportion of fish caught on moving hooks 

 in 1989 compared with 1990 (Table 4) could have been 

 caused by moving hooks being less visible in 1990, since 

 branch lines were more often recovered after dark 

 (Table 1). Sets also lasted longer in 1990 (Table 1); this 

 may have increased the relative proportion of catches 

 on settled vs. moving hooks. The CPUT in relation to 

 sinking, settled, and rising gear, and to the time of day, 

 should be explored further using the techniques devel- 

 oped in the present study. 



A TDR attached to vertical and regular rope longline 

 gear sometimes records abrupt depth changes as a fish 



