344 



Fishery Bulletin 105(3) 



roughly 1800 nesting females in 1995 

 (Spotila et al., 1996); however, a more 

 comprehensive evaluation indicates that 

 the total western Pacific metapopulation 

 may contain 2700-4500 breeding females 

 (Button et al., 2007). Satellite telemetry 

 studies have linked leatherback turtles 

 foraging along the U.S. West Coast with 

 one of the two largest remaining nest- 

 ing beaches, Jamursba Medi (Papua, In- 

 donesia) (Benson et al., 2007a), which 

 experiences peak nesting activity dur- 

 ing the austral winter. No links to the 

 U.S. West Coast have been identified for 

 animals nesting during the austral sum- 

 mer at nearby Wermon, Papua, Indonesia 

 (S. Benson and P. Button, unpubl. data) 

 or in Papua New Guinea (Benson et al., 

 2007b). 



In a recent analysis of nest counts at 

 Jamursba Medi, an average of about 750 

 females were estimated to nest annually 

 between 1993 and 2004 (Hitipeuw et al., 

 2007). Efforts are underway to deter- 

 mine the relationship between the num- 

 ber of females nesting annually and the 

 total number of females in the Jamurs- 

 ba-Medi nesting population; however, it 

 is thought that this population currently has at least 

 1000-2000 nesting females (Spotila et al., 1996; But- 

 ton et al., 2007). Capture studies off central California 

 during 2000-2005 documented that about 67.5% (27 of 

 40) of foraging leatherback turtles were female (S. Ben- 

 son and P. Button, unpubl. data). Our average annual 

 estimate of 178 leatherback turtles along the California 

 coast, therefore, should correspond to approximately 

 120 females. It is difficult, however, to evaluate this 

 number in relation to the total Jamursba-Medi nest- 

 ing population, because insufficient data are available 

 on migration intervals between nesting beaches and 

 foraging grounds. If each nesting year corresponds to 

 one year at the California foraging grounds, then an 

 average of about 16% of Jamursba-Medi females (120 

 divided by 750) potentially use the California foraging 

 area; however there is evidence that leatherback turtles 

 do not alternate nesting and foraging at such regular 

 intervals. Remigration intervals to nesting beaches can 

 range up to seven years (Price et al., 2004; Button et 

 al., 2005), and some turtles have returned to forage off 

 the U.S. West Coast during consecutive years without 

 nesting (S. Benson and P. Button, unpubl. data). Fur- 

 ther studies of remigration patterns and foraging site 

 fidelity of western Pacific leatherback turtles will be 

 required to resolve the proportion of these turtles that 

 forage off California. 



Estimates of foraging abundance in this study vary 

 markedly among years (Fig. 3) and have a number of 

 known sources of downward bias. First, the California 

 study area includes only the nearshore environment 

 to a water depth of about 92 m (50 fm), but leather- 



-3 -2 -I () I 2 



12 month average NOI, Jan-Dec 



Figure 5 



Regression of estimated leatherback turtle CDermochelys coriacea) 

 abundance versus 12 month average Northern Oscillation Index (NOI; 

 Schwing et al., 2002) for the period January-December of each survey 

 year (indicated next to the points). Filled symbols represent years in 

 which surveys were conducted coastwide; open symbols represent years 

 in which estimates of coastwide abundance were based on central Cali- 

 fornia abundance (see "Materials and methods" section). 



back turtles also have been captured incidentally in 

 drift gill nets set in deeper waters (Julian and Beeson, 

 1998; Carretta et al., 2005). Second, the estimate cal- 

 culated in this article represents an average snapshot 

 abundance, and will be an underestimate of the true 

 number of individuals using the area to the extent that 

 residence times in the study area are less than our 

 three-month study season. For example, if leatherback 

 turtles forage within the study area for two months, 

 then turtles observed in August would likely be differ- 

 ent individuals from those observed during October or 

 November, but average line-transect densities would 

 only reflect the presence of a single turtle. Similarly, 

 if leatherback turtle density is not constant through- 

 out the three-month study period, the pooled estimate 

 presented in this study will be lower than the peak 

 seasonal abundance. Lastly, no estimate of perception 

 bias was available for leatherback turtles in this study; 

 however, small groups of small dolphins and porpoises 

 are missed about 33% of the time (Forney et al., 1995). 

 Therefore, it is likely that the detection of available 

 turtles along the transect line is less than 100%. 



The estimate of g(0) developed in this study is the 

 first to be based on dive records of leatherback turtles 

 within a foraging region. Although the three turtles 

 exhibited remarkably similar proportions of time within 

 near-surface waters (Table 2), there is some uncertainty 

 in the estimate because it is based on limited afternoon 

 deployments in = 'i) within a single year (2005). Further- 

 more, the depth to which turtles are visible to aerial 

 observers may vary in space and time, as turbidity 

 changes. The estimated g(0) =0.471 should, therefore, 



