Orbesen et ai.: Mortality rates of Thunnus thynnus associated with pelagic longline gear in the Gulf of Mexico 
1 7 
(vers. 9.4; SAS Institute, Inc. Cary, NC). The following 
model was applied: 
Mortality = H j + D j + T ^ + C\ + »S m + L n , (1) 
where mortality = the probability of a fish being dead 
at-vessel; 
H = the i th type of hook used (circle hook, 
J-hook, weak hook); 
D = they th maximum hook depth (in the 
water column, meters); 
T = the k th target species (swordfish, 
tuna, mixed); 
C - the I th sea-surface temperature (SST, 
in degrees Celcius, measured by the 
vessel during gear deployment); 
S = the ?n th soak duration (time from 
last hook deployed to first hook re¬ 
trieved); and 
L = the n th fish length (straight fork 
length, in centimeters). 
All factors were modeled as continuous variables with 
the exception of hook type and target species. To test 
for an effect of hook type on mortality, least square 
means were generated as estimates of mortality for 
each hook type. 
Tagging 
From 2010 through 2015, we deployed 41 PAT-MklO 
pop-up satellite archival tags (PSATs; Wildlife Comput¬ 
ers, Inc., Redmond, WA) on bluefin tuna captured on 
PLL vessels fishing in the GOM (Fig. 1). These PSATs 
were programmed to archive pressure (depth), ambient 
temperature (in degrees Celsius), and light intensity 
every 10 s. Each PSAT was equipped with a corrodible 
burn pin that detaches the tag on a preprogrammed 
date (90-365 d) or when the PSAT has been at a con¬ 
stant depth (±5 m) for a 24-h period, indicating that ei¬ 
ther the tag is no longer attached to the animal or the 
animal has died. Upon detachment of a tag, profiles of 
depth and temperature and proportions of time spent 
in 14 user-defined depth (time at depth) and tempera¬ 
ture (time at temperature) bins were summarized into 
1-h (4 PSATs) or 4-h (37 PSATs) periods and transmit¬ 
ted through the Advanced Research and Global Obser¬ 
vation Satellite system. For those cases in which the 
tag was physically recovered (11 PSATs), the full archi¬ 
val data set was obtained and analyzed. 
All PSATs were equipped with a surgical-grade, ny¬ 
lon toggle anchor and an RD-1800 (Wildlife Comput¬ 
ers, Inc.), a device designed to sever the PSAT link 
before hydrostatic pressure damages the tag (typically 
at a depth of approximately 1800 m). The tether rig¬ 
ging had 3 variations over the course of this study. 
The tags deployed in 2010 and 2011 (5 PSATs) were 
equipped with double-crimped monofilament teth¬ 
ers. In a concurrent study on yellowfin tuna, several 
tags were recovered with visible teeth marks on the 
PSAT and float, raising a concern that monofilament 
might result in an increase in attachment failures (C. 
Brown, unpubl. data); therefore, in 2012 tethers were 
constructed from stainless steel cable (22 PSATs). 
However, during a failed tagging attempt, a PSAT fell 
overboard. This PSAT, which was still equipped with 
a stainless steel tether, was observed to be negatively 
buoyant. Subsequent buoyancy testing indicated that 
rigged PSATs were very sensitive to small changes in 
weight, and the previous design with a double-crimped 
tether also resulted in tags being negatively buoyant. 
All subsequently deployed tags (14 PSATs) were rigged 
with a single-crimped monofilament tether, to allow the 
PSAT to float with the tether attached, and each of 
these PSATs was checked for positive buoyancy prior 
to deployment. 
All PSATs were deployed by NMFS-trained observ¬ 
ers aboard commercial fishing vessels that targeted 
yellowfin tuna with PLL gear in the GOM between the 
months of February and May of each year (2010-2015). 
Observers were given strict guidance to tag any live 
bluefin tuna, regardless of condition. When a fish was 
released by using methods similar to those of a nor¬ 
mally operating PLL vessel, care was taken so that the 
fish remained in the water for tagging, and the hook, 
for the most part, was not removed. 
Operational changes 
A study of PLL gear that employed a design with al¬ 
ternating hook types found that a new 16/0 weak hook 
could reduce catches of bluefin tuna in the GOM PLL 
fishery by an estimated 56.5% (Foster and Bergmann 1 ) 
from levels observed when a typical circle hook was 
used. In that study, hook timers, devices that measure 
the total time a fish spends hooked on a line, were at¬ 
tached to a portion of the gangion. The results of that 
study led the NMFS to mandate the use of these weak 
hooks in the GOM PLL fishery, and this regulation 
went into effect in 2011 (Federal Register, 2011). Prior 
to this rule being enacted, fish tagged in 2010 (4 fish) 
were captured by using regular-strength 16/0 circle 
hooks. In 2012, tags were deployed in conjunction with 
a continuation of the weak hook study (30 fish). For 
fish captured on a leader with an accompanying hook 
timer, total time on the line was obtained. All remain¬ 
ing fish were deployed on PLL sets by using the 16/0 
weak hook (7 fish). 
Determination of mortality 
Postrelease mortality mostly has been estimated to oc¬ 
cur shortly after release because of acute injury (Mu- 
oneke and Childress, 1994; Stokesbury et al., 2004; 
Wilson et al., 2005). However, delayed mortality due 
to loss of ability to feed or infection can occur sever¬ 
al days to weeks postrelease (Burns and Froeschke, 
2012). Although increasing the duration of monitor¬ 
ing beyond several days allows the incorporation of 
delayed mortalities, there is a risk of confounding re¬ 
sults with mortality unassociated with the initial cap- 
