28 
Fishery Bulletin 1 15(1) 
(synonymous with at-vessel, hooking, and initial mor- 
tality; Dapp et ah, 2016a). The immediate mortality of 
incidentally caught sharks can be substantial (Dapp 
et ah, 2016a), and incidental capture has been identi- 
fied as a major driver of declines in shark populations 
(Dulvy et ah, 2014). 
Immediate mortality rates of sharks during longline 
capture can be influenced by a multitude of factors, 
such as, but not limited to, capture duration, hook type, 
hook size, species-specific biology, and crew-specific 
handling practices (Campana et ah, 2009; Morgan and 
Carlson, 2010; Dapp et ah, 2016a, 2016b; Gilman et ah, 
2016). Factors contributing to the immediate mortal- 
ity of sharks caught by the U.S. Atlantic pelagic long- 
line fishery have previously been examined by using 
the data set of the National Marine Fisheries Service 
(NMFS) Pelagic Observer Program (POP) (Serafy et ah, 
2012; Gallagher et ah, 2014). A recent assessment by 
Gallagher et al. (2014), who used the POP data set, re- 
vealed that hook depth, soak time, sea-surface temper- 
ature (SST), target catch (assumed to be either sword- 
fish or species of tunas), and fork length had varying 
effects on the likelihood of immediate mortality for 12 
species of sharks. 
We expanded upon earlier research by investigating 
how additional factors (geographic zone and manage- 
ment regulations) contribute to the likelihood of im- 
mediate mortality of longline-caught sharks, using the 
U.S. pelagic fishery logbook (USPL) data set^. Temporal 
analysis of immediate mortality rates with this data 
set can provide valuable information on the efficacy of 
regulatory measures designed to decrease immediate 
mortality of sharks. For example, Serafy et al. (2012) 
attributed temporal differences in immediate mortality 
of several species to management regulations requiring 
the use of circle hooks during longline fishing. Simi- 
larly, zone-specific rates of immediate mortality can 
provide information on how operational differences be- 
tween geographic zones and fleets cause variations in 
immediate mortality. 
The USPL data set is the largest U.S. fishery data 
source available in terms of number of animals caught. 
The large sample size of this data set makes it more 
suitable than experimentally derived data sets for ana- 
lyzing spatiotemporal influences on the mortality rates 
of sharks. The USPL data set comprises mandatory 
data recorded by commercial fishermen and has been 
a valuable resource in previous studies of changes in 
shark relative abundance or catch rates (Baum et ah, 
2003; Baum and Myers, 2004; Mandelman et ah, 2008; 
Baum and Blanchard, 2010). However, issues with spe- 
cies identification and misreporting can limit the value 
of logbook data (Burgess et al., 2005). Accordingly, the 
accuracy of commercial logbook data should be scru- 
tinized and carefully compared with data collected by 
trained observers. 
‘ NMFS (National Marine Fisheries Service). Southeast Fish- 
eries Science Center, Miami Laboratory, 75 Virginia Beach 
Drive, Miami, FL 33149. 
Despite these concerns, Mandelman et al. (2008) and 
Baum and Blanchard (2010) concluded that catch rates 
for several easily identifiable, commercially valuable, 
or commonly captured species of sharks in the USPL 
data set were not significantly different from catch 
rates recorded in the POP data set. However, several 
congeneric species that have similar names or are mor- 
phologically similar (e.g., silky shark [Carcharhinus 
falciformis], dusky shark [C. obscurus], and night shark 
[C. signatus]', longfin mako [Isurus paucus] and short- 
fin mako [/. oxyrinchusY, and common thresher shark 
[Alopias vulpinus] and pelagic thresher [A. pelagicus]) 
were consistently misidentified in the USPL data set, 
even by trained observers (Reardon et al., 2006; Baum 
and Blanchard, 2010). 
We sought to expand upon earlier studies of logbook 
accuracy by comparing the immediate mortality rates 
of sharks with accurately recorded catch rates of sharks 
recorded in the USPL data set to the immediate mor- 
tality rates of these species recorded in the POP data 
set. This comparison allowed us to determine whether 
records of dead and live discards were accurately re- 
corded in the USPL data set. It is important to appreci- 
ate that the USPL was not designed for this purpose; 
it was implemented as a marine resource management 
tool to monitor fishery changes in catch per unit of ef- 
fort (additional information is available from the web- 
site). However, if shark mortality is accurately recorded 
within the USPL data set, the large number of records 
in the data set will make it a powerful and statistically 
robust tool for identifying factors that contribute to im- 
mediate mortality during longline capture. 
For our analysis, we summarized the number of 
sharks retained, discarded dead, and discarded alive by 
the U.S. pelagic longline fleet from 1992 through 2008 
by using the USPL data set. In addition, we evaluated 
the accuracy of mortality recordings by commercial fish- 
ermen in the USPL data set by comparing our results 
with the findings from previous immediate mortality 
studies in which similar species, fishing techniques, 
and locations were used. If accuracy was validated, we 
went on to assess factors contributing to the likelihood 
of shark bycatch being discarded dead to propose new 
variables that can be included when recording logbook 
data to enhance data collection. 
Materials and methods 
Overview of the data set and study methods 
To achieve our goals, we first screened the USPL data 
set for the period of 1982-2008 to remove inaccurate 
records. Quantitative species-specific comparisons of 
immediate mortality between this data set and other 
published sources were made to assess the accuracy of 
the USPL data set. If accuracy was verified, a statisti- 
cal analysis of shark mortality was then conducted to 
determine factors correlated with immediate mortality 
within the USPL data set. 
