Harding et al. : Regional and seasonal patterns of epipelagic fish assemblages from the central California Current 
263 
Reyes (38°00'N) (Fig. 1). Point Arena defined the 39 
northern border of NC, and Pt. San Pedro the 
southern border of GF. The boundaries of the 
GF are generally well recognized, and similar 
borders have been reported elsewhere (Ainley, 
1990; MacFarlane et al., 2005). We worked from 
chartered commercial trawlers and the NOAA 
ship David Starr Jordan. Samples were collected 
over 5-13 consecutive days in June or July (here- 
after “summer” cruises) and for 5-11 consecutive 
days in September or October (hereafter “fall” 
cruises) for six years, from 2000 to 2005. All jF 
sampling was conducted during daylight hours, -§ 
between sunrise and sunset. There were 23 and §, 
11 fixed trawling stations in the NC and GF, re- ™ 
spectively, and both regions were sampled during 
every cruise, although not every station was vis- 
ited each time (Table 1). Effort was not equally 
distributed among years, seasons, or regions 
because of operational constraints, weather, and 
other project objectives. 
We used a Sea-Bird SBE 19 CTD (conductivity, 
temperature, and depth) profiler with added sen- 
sors for hydrographic sampling conducted imme- 
diately before or after trawling at each station. 
Five environmental variables were measured 
in 1-m depth bins from the surface to the bot- 
tom: water temperature, °C (TMP); salinity, ppt 
(SAL); water density, kg/m 3 (DEN); photosyn- 
thetically available radiation, (pE/sec)/m 2 (PAR); 
and chlorophyll-a concentration, pg/L (CHL). 
Values from 1 to 15 m depth were averaged for 
each cast, and therefore represented about the 
same range as the vertical spread of the trawl 
net. Owing to occasional instrument failures, not 
all of the CTD sensor data were collected on ev- 
ery cast (Table 1). Bottom depth in m (DEP) and 
distance offshore (DIS) in km were measured for 
every haul. 
Trawling and sample processing 
We used a 264 Nordic rope trawl and 3.0-m 2 foam-filled 
pelagic doors to collect epipelagic fish and inverte- 
brates. Net dimensions were approximately 14 m (ver- 
tical at mouth) by 27 m (horizontal at mouth) by 194 
m (length). Vertical spread was measured in the field 
with depth recorders attached to the head and footrope, 
and the measurement of horizontal spread had been 
estimated and provided previously by the manufacturer 
(NET Systems, Bainbridge Island, Washington). Effec- 
tive mouth area was assumed constant at 380 m 2 . The 
codend liner was constructed of 6x10 mm knotless nylon 
and did not usually retain fish <40 mm total length 
or small invertebrates such as krill. Floats on the 
headrope and bridles helped maintain the net near the 
surface (usually within 1.0-1. 5 m) continuously during 
tows. Sets were made at depths of >37 m, except at four 
shallower stations where the bottom was thought to be 
free of snags. Tow duration was 6-40 min (mean=22 
124 
123 
Longitude °W 
Figure 1 
122 
Location of trawl stations off central and northern California 
where epipelagic fish species were collected during 2000-05. 
Symbols show stations of the two regional sampling groups: 
north coast (NC) and Gulf of the Farallones (GF). Also shown are 
station labels for selected eastern GF (23, 24, 25) and western 
GF stations (FS = Fanny Shoal, NFI=North Farallon Island). 
Three National Marine Sanctuaries (NMS) are indicated. 
min) and inversely proportional to jellyfish ( Chrysaora 
fuscescens and Aurelia spp.) density. When abundant, 
these large jellyfish reduce sampling efficiency and 
can damage nets. Tow speed varied from 5.0 to 8.0 
km/h (mean=6.5 km/h), depending on sea conditions 
and vessel. Tow distance was measured either with 
a mechanical flow meter pulled alongside the boat or 
calculated with GPS; tow distance varied from 0.5 to 
4.8 km (mean=3.0 km). 
Most fishes retained in the codend were identified 
to species, or occasionally to higher taxonomic levels, 
and counted. Larval and other small fish <40 mm 
were seldom retained. However, postlarval osmerids, 
scorpaenids, and flatfishes were occasionally captured 
and these were identified to family or sometimes to 
broader groups (e.g., “flatfish larvae”). Very large hauls 
were subsampled by volume or weight, and total spe- 
cies abundance was estimated from the composition of 
subsamples. Size class distinctions were made for only 
one species: juvenile Chinook salmon (<250 mm fork 
length) were counted separately from larger individuals 
(>250 mm fork length, hereafter called “adults”). Adult 
salmon and other highly mobile species are thought to 
