230 
Fishery Bulletin 116(3-4) 
o 
5= 
Q) 
0.14 
0.08 
0.06 
0.04 
0 . 02 - 
0 
r : 
n 
J \ 
1 \ 
-Q1 
Q2 
A 
Q3 
1 \ 
i / \ 
-Q4 
J \ 
J 
• A \ 
r/ \ v 
10°S 
r^j W 
/ /' J 
/ / / - 
\\\ 
10°N 
20°N 
30°N 
40°N 
I 
100 
150 
200 
250 
300 
350 
- 400 
450 
500 
170°W 160°W 150°W 140°W 130°W 120°W 
O 
a> 
T> 
3 
Figure 1 
(A) The meridional (top) and zonal (bottom) distributions of quarterly ef¬ 
fort of the Hawaii-based longline fishery, averaged across the full time 
series (1995-2015). Dashed black vertical lines are drawn at 20°N and 
26°N (top) and at 150°W (bottom) and delineate the divides between 
the SE and NE regions and the SW, CW, and NW regions, and dashed 
gray lines are drawn at ±1° for comparison of alternate options (top). (B) 
Map depicting the 5 regions by which the Hawaii-based longline fishery 
was examined overlaid on the climatological (1995-2015) median depth 
of preferred thermal habitat of bigeye tuna (Thunnus obesus) (8-14°C) 
obtained from Global Ocean Data Assimilation System (shaded) and the 
depth of the 1.0 mL/L oxygen-concentration threshold from World Ocean 
Atlas 2013 data (contoured every 100 m from 100 to 500 m, with stippling 
where the depth is less than 100 m). The white line in panel B encom¬ 
passes grid cells with fishing effort from at least 3 vessels over the full 
time series. Effort in panel A is from logbook data. 
ging data from Howell et al. (2010) 
indicate that when bigeye tuna are 
at depth during the daytime, which 
is when the fishery for bigeye tuna 
operates, they are primarily in wa¬ 
ters within this thermal range. The 
GODAS data were used to determine 
the minimum, maximum, and median 
depths of preferred thermal habitat 
of bigeye tuna in 2 ways: 1) these 
depths were determined for all grid 
cells (each l°xl°) with fishing effort 
at any point in the time series; 2) 
quarterly GODAS data were weighted 
by the number of hooks set in each 
grid cell during each quarter. We 
then used standard linear regression 
to evaluate whether there were sig¬ 
nificant (P<0.05) linear trends in both 
unweighted and weighted median 
depths of preferred thermal habitat. 
Where significant trends were found, 
we used linear regression to deter¬ 
mine how the depth of preferred ther¬ 
mal habitat changed over the years 
studied. 
Results 
Fishing effort 
Seasonal variability The Hawaii-based 
longline fishery exhibited strong sea¬ 
sonal movement during the period 
studied, 1995-2015. Figure 1A shows 
the temporally averaged meridional 
and zonal distribution of effort (num¬ 
ber of hooks set) each quarter. On 
the basis of this distribution, as well 
as the 150°W boundary between the 
2 fishing convention areas of the Ha¬ 
waii-based fishing grounds, we divided 
the fishery into the 5 regions shown in 
Figure IB: northeast (NE); northwest 
(NW); central west (CW); southwest 
(SW); and southeast (SE). Together, 
Figures 1A and 2 show the movement 
of the fishery by quarter throughout 
the year. In the first quarter of the 
year, most of the effort took place in 
the SW region north of 10°N and in 
the CW region. During the second 
quarter, effort was concentrated in the 
SW and NW regions. The fishery then 
underwent a large geographic shift in 
the third quarter, and most of its effort 
was directed within the NE region. Ef¬ 
fort occurred closest to Hawaii in the 
CW region during the fourth quarter. 
