Zamora-Garcia et al.: Catch rate, length, and sex ratio of Merluccius productus in the Gulf of California 367 
113°W 
Pacific Ocean 
Catch rate (kg/tow) 
= <1000 
2 1000-2000 
= 2000-3000 
m@— >3000 
A Localities 
=F Delfin Basin 
@ Islands 
=} Vaquita refuge 
t_3 Upper Gulf of 
California and 
Colorado River Delta 
Biosphere Reserve 
Figure 2 
Map of the study area showing the fishing grounds for and catch per tow of Pacific hake (Merluc- 
cius productus) recorded in the northern Gulf of California (NGC) in Mexico during the fishing 
season from January through March in 2015-2019. The fleet that targeted Pacific hake included 
ships from Puerto Penasco, Sonora, and San Felipe, Baja California, among other localities. Trawl 
tows followed the slope of the Delfin Basin south of the Upper Gulf of California and Colorado 
River Delta Biosphere Reserve and a refuge for vaquitas (Phocoena sinus). The black solid lines 
around the Delfin Basin indicate 100-m isobaths. 
The NC was divided by the duration of the trawl tow (from 
the end of net launch until the start of net collection), to 
obtain the catch rate, or catch per unit of effort (CPUE), in 
kilograms per hour. 
Observers recorded biometrics, including standard 
length (SL) and total fresh weight, from ~60% of tows per 
trip: 98 of 131 trips in 2015, 69 of 118 trips in 2016, 93 of 
181 trips in 2017, 105 of 170 trips in 2018, and 108 of 217 
trips in 2019. They selected random samples of fish (max- 
imum of 100 fish per set) and identified them to species 
following Lloris et al. (2003). Standard length of fish was 
measured from the tip of the mouth to the caudal peduncle 
to the nearest 0.1 cm. Total fresh weight was measured 
by using a digital balance with an accuracy of 1 g. The 
Specimens were dissected to determine sex and maturity 
stages. We used a 5-stage morpho-chromatic key to deter- 
mine the maturity of Pacific hake, according to Holden 
and Raitt (1975): immature (stage 1), developing (stage 2), 
mature (stage 3), spawning (stage 4), and postspawning or 
spent (stage 5). To ease comparisons, we considered fish 
that had signs of recent, current, or imminent spawning 
(stages 3-5) to be adults. Fish assigned to stages 1 and 2 
were considered juveniles. 
Data analysis 
Catch rate standardization From all the trawl tows com- 
pleted (number of tows [n]=817), we excluded 3 tows 
because of gear malfunction, according to Maunder and 
Punt (2004), and 9 tows completed in April 2019 because 
of incomparability. A log-transformed index of catch per 
unit of effort (logCPUE) of the positive values (number 
of values=805) was used for hypothesis testing because 
these values followed a log-normal-like frequency distri- 
bution. The geometric mean was used as the index for 
mean CPUE. 
First, we developed generalized linear models (GLMs) 
to test the effect of the selected explanatory variables. 
However, we found several nonlinear effects, a result 
that violates the main assumption of GLMs. Then, 
we used generalized additive models (GAMs), which are 
semiparametric extensions of GLMs with an additive 
predictor (Guisan et al., 2002), to identify significant 
explanatory variables (speed of trawl, mesh size in the 
codend, depth, and the hour of the day, month, and year) 
and the nature of the relationships with the logCPUE. 
We considered a normal probability distribution of 
