Marin-Enriquez and Muhlia-Melo: Environmental and spatial preferences of Coryphaena spp 
11 
camber 2013. All environmental data were downscaled 
to l°xl° in order to match the spatial resolution of the 
fishery database. 
Analysis of incidental catch 
Kolmogorov-Smirnov D test and Bartlett’s B test were 
conducted to assess normality and homoscedasticity of 
the data. Depending on the results, parametric or non- 
parametric (Kruskal-Wallis H test) analyses of vari¬ 
ance were conducted to evaluate significant seasonal 
and interannual differences in dolphinfish ICPUE. 
Quadrants with catches higher than the lower limit 
of the upper quartile of ICPUE were considered high 
catch quadrants (Andrade and Garcia, 1999). Results 
from these analyses were used to evaluate possible dol¬ 
phinfish migration patterns. 
Relation of dolphinfish incidental catch to environmental 
and spatial variables 
To assess the relationship of ICPUE to environmental 
variables, and because of their apparent nonlinear re¬ 
lationship, a 3 rd degree polynomial linear model was 
fitted for each one of the variables in relation to the 
log-transformed ICPUE. We fitted a model for each 
different size class, in order to highlight differences 
in environmental preferences at different dolphinfish 
life stages. Standard validation techniques (residuals 
vs. fitted values, histogram of residuals) were used to 
assess the relevance of the fitted models to the data 
(Zuur et ah, 2009, p. 23). F-tests were then conducted 
to evaluate the significance of model fit. 
With the purpose of evaluating possible segregation 
due to size, we created histogram-like figures for the 
estimated number of fish caught per size class of fish 
and each spatial and environmental variable at the 
l°xl° resolution. Spatial distribution maps were cre¬ 
ated with R, vers. 3.0.1 (R Core Team, 2013), and its 
maptools package, vers. 0.8-34. 
The monthly latitude of the 25°C isotherm at 120°W 
was extracted from the monthly mean SST data from 
the satellite imagery by using the contourLines and 
convCP functions of the R PBSmapping package, vers. 
2.68.68. Possible relation between mean latitude of high 
catch quadrants and mean latitude of the isotherm was 
explored by using the Pearson correlation coefficient (r). 
All calculations were performed with R and with a 
significance of 0.05 for all statistical tests. 
Results 
Spatiotemporal variation of incidental catch 
A total of 627 positive sets were reported in the 10-year 
fishery database, which resulted in a total of 55,406 
fish caught, and a per-quadrant maximum, minimum, 
mean, and standard deviation (SD) of 3527, 1, 88.36, 
and 269.76 fish, respectively. Maximum and minimum 
ICPUE values were 3527 fish/set, with a mean of 55.32 
fish/set (SD 189.47). The monthly maximum and mean 
number of positive sets per quadrant was 46 sets and 
1.81 sets (SD 3.09). 
Data of ICPUE was nonnormally distributed 
(D^O.38, P<0.05) and nonhomoscedastic for monthly 
(B,n,6271=817.41, P<0.05) and year-to-year variation (B, 9 
627) =706.81, P<0.05). With the exception of December, all 
months had at least one set with 1 fish/set. The high¬ 
est value of 3527 fish/set was found during July 2004. 
Significant differences were found in monthly ICPUE 
(if (11 62 7)=58.85, P<0.05): in general, maximum values 
occurred during June and October, and minimum values 
occurred during January and November (Fig. 1A). 
Yearly, maximum values of ICPUE were found during 
2004, and minimum values were reported for 2008 (Fig. 
IB). Year-to-year variation was also statistically signifi¬ 
cant (H { 9 6 27)=26.72, P<0.05). 
Two areas with high dolphinfish concentration (>300 
fish/set) were found, one in the open ocean around, 10- 
20°N and 115-125°W, and a second one near the Baja 
California Peninsula (BCP), around 20-27°N and 113- 
116°W. Catches around 100-300 fish/set were found 
in the northern coastal part of the study area (~33°N, 
117°W) and in the entrance of the Gulf of California, 
around 23°N and 107°W. In the southern part of the 
study area, scattered quadrants with high estimated 
ICPUE were also present, especially west of 110°W 
(Fig. 2). 
Analysis of size classes 
A total of 5306 dolphinfish (9.50% of total fish) of the 
small-size class were caught; 21,387 (38.60%) belonged 
to the medium-size class, and 28,713 (51.82%) belonged 
to the large-size class. Number of fish caught by size 
class was nonnormal (D^O.41, P< 0.05) and nonho¬ 
moscedastic (-8,2 , i88d= 347.83, P<0.05). Significant dif¬ 
ferences were found between fish caught for different 
size classes (H i2 , i88i)=449.25, P<0.05). An average of 
8.50 fish (SD 74.30) in the small-size class (min. 0, 
max. 1706) were caught per quadrant per month. Mean 
for the medium- and large-size classes were 34.10 fish 
(SD 151.03; min. 0, max. 2185) and 45.80 fish (SD 
153.44; min.0, max. 1689) per quadrant per month, 
respectively. 
No apparent well-defined spatial pattern due to size 
class was observed. Dolphinfish of all size classes were 
caught both on coastal and offshore waters throughout 
the study area (results not shown). 
For the small-size class, most fish were caught dur¬ 
ing July (45% of total fish for that size class, 2428 fish), 
and 0 fish were caught during December. July was also 
the month when most incidental catch of fish in the 
mid-size class occurred (30.50% of total fish [6521 fish 
for the size class]). Only 0.03% (8 fish) of total fish of 
the medium-size class were caught during December. 
For the large-size class, most fish were caught during 
October (32.22% of total fish for the size class, 9522 
fish), and a monthly minimum of 0.08% of total fish 
