Cooper et al.: Spatiotemporal catch patterns and population distributions of Lampris megalopsis and L. incognitus 143 
The spatial range of the deep-set fishery also changed 
over time. In 2014-2018, the fishery expanded eastward 
to 125°W (compared with a limit of 135°W in 1996--2013), 
and less area was fished south of 10°N (Fig. 1). Even after 
the eastward expansion in 2014, the majority of deep-set 
fishery operations were still centered around the main 
Hawaiian Islands, with 74% of the fishery operating west 
of 150°W during 2014-2018 in comparison with 91% 
during 1996-2013. 
As with effort, the spatial range of the shallow-set fishery 
also decreased in 2014-2018 compared with the range in 
1996-2013 (Fig. 2). As with the deep-set fishery, effort for 
the shallow-set fishery was centralized around the Hawai- 
ian Islands in 1996-2013 but transitioned to more of a 
bimodal distribution, with one peak around the Hawaiian 
Islands and another around 140°W, 30°N in 2014-2018. 
The CPUE for opah was highest in the northeast region 
of both the deep-set and shallow-set fishery, roughly east 
of 140°W and north of 25°N, with a noticeable gradient 
increasing from west to east (Figs. 1 and 2). The results 
of comparing data between the gear types indicate that 
CPUE was highest in the deep-set fishery. From 2014 
through 2018, CPUE of deep-set longlines in 5°-by-5° 
blocks ranged from 3.0 to 6.0 opah/1000 hooks in the north- 
east extent of the fishery and from 0.5 to 2.0 opah/1000 
hooks west of 140°W. The CPUE of shallow-set longlines 
peaked at about 2 opah/1000 hooks in the northeast area 
and was generally <1 opah/1000 hooks throughout the 
rest of the fishery range. 
Spatial distributions of species 
For species identification in this study, an additional 
679 samples from 37 trips were added to the 415 samples 
from 102 trips used in the study conducted by Hyde et al. 
(2014), to examine distribution in the region that overlaps 
with the pelagic longline fisheries. The region and time 
frame over which the samples were collected differed for the 
2 programs (Table 1). The samples from Hyde et al. (2014) 
came from west of 139°42’W, whereas the more recent sam- 
ples came from east of 139°6’W. Both programs sampled 
across months, although no samples from November were 
available to Hyde et al. (2014). Bigeye Pacific opah were 
identified in all 5°-by-5° blocks, although percent occur- 
rence varied spatially. In the region around Hawaii and to 
the south, only bigeye Pacific opah were identified in the 
blocks. North of the Hawaiian Islands (>25°N), smalleye 
Pacific opah were identified in all blocks with the propor- 
tions ranging from 0.1 through 0.8 and with the highest 
proportional abundance in the blocks closest to the U.S. 
West Coast. (Fig. 3). 
The Akaike information criteria for GAMs with k of 4, 
6, and 8 were —2669.9, —2726.1, and —2742.0, respectively. 
The effective degrees of freedom for these GAMs were 2.98, 
4.88, and 6.75, respectively. The effective degrees of freedom 
of 2.98 for the GAM with a k of 4 is very close to 3, which 
is 1 less than k, indicating that this GAM may be over- 
smoothed. The 3 GAMs explained 56.2%, 59.1%, and 60.0% 
of the deviance in the genetic sampling data, respectively. 
Table 1 
Overview of sampling of bigeye Pacific opah (Lampris 
megalopsis) and smalleye Pacific opah (L. incognitus) from 
which genetic data used in this study were obtained. Data 
were compiled from Hyde et al. (2014) and from genetic 
analyses conducted since 2014 of samples from the eastern 
North Pacific Ocean. 
Period 
2009-2010 
Data type (from Hyde et al.) 2017-2018 
Effort (no. of trips) 102 37 
Sample size 415 679 
Latitude mode 30.00°N 22.45°N 
Latitude range 10.68-31.72°N 19.83-33.39°N 
Longitude mode 217.56°E 227.53°E 
Longitude range 190.52-220.27°E 9 220.93-232.89°E 
Model fits appeared adequate, and residuals were approxi- 
mately normally distributed in each case. 
All 3 GAMs predicted that bigeye Pacific opah would 
dominate (with proportions of bigeye Pacific opah 20.7) 
in waters west of 140°W, particularly in the southwest 
extent of the study area, and that smalleye Pacific opah 
would dominate (with proportions of bigeye Pacific opah 
<0.3) in waters east of 130°W (Fig. 4). The 2 species were 
predicted to co-occur in roughly equal proportions between 
140°W and 130°W. Predictions from the 3 GAMs varied 
spatially, particularly in the area north of 25°N and east of 
150°W, where fishing effort was generally lower. However, 
estimates from all models indicate a strong west-to-east 
gradient of high-to-low probability of occurrence of bigeye 
Pacific opah. 
The results from jackknifing indicate low levels of vari- 
ance over the study area, with a maximum coefficient of 
variability of ~4% (Fig. 5). The area with the highest coef- 
ficient of variability (155—140°W, 30-35°N) was the area 
with the least data, and it encompasses waters predicted 
to have areas dominated by bigeye Pacific opah and mixed 
areas. Generally, waters predicted to be dominated by 
one species or the other had lower variability. 
Catch per unit of effort by species dominance 
The results from the GAMs and the maps of CPUE were 
combined to examine CPUE in the 3 types of areas clas- 
sified by species dominance: dominated by bigeye Pacific 
opah, dominated by smalleye Pacific opah, and mixed 
areas. The lowest CPUE was in areas predicted to be dom- 
inated by bigeye Pacific opah (west of 140°W), and the 
highest CPUE was in the eastern portion of the mixed 
area and the area predicted to be dominated by smalleye 
Pacific opah (east of 130°W) (Table 2, Figs. 1 and 2). 
The results from comparing data for the 2 periods indi- 
cate that CPUE declined slightly in areas dominated by 
