120 
Fishery Bulletin 110(1) 
require simulations of various biomass levels, possible 
distributions, and sampling strategies (Simmonds and 
Fryer, 1996). In the meantime, to reduce the sampling 
variance and improve the description of species demog- 
raphy, acoustic and trawl sampling may be increased in 
areas of higher expected abundance, which is seasonally 
dependent (Clark and Janssen Jr., 1945; Zwolinski et 
al., 2011; Demer et al., 2012). Moreover, target trawling 
may increase the number of fish samples to better de- 
fine the demography of the least abundant species. The 
historical distribution of the dominant CPS and, more 
recently, model prediction of potential sardine habitat 
indicate that sampling off Oregon and Washington dur- 
ing the spring is likely useless for the assessment of 
sardine, jack mackerel, and Pacific mackerel but may 
be relevant for potentially emerging anchovy and her- 
ring populations. 
Sardine, mackerels, and anchovy have been shown 
to alternate dominance in the CCE in response to low- 
frequency variability in the oceanographic conditions of 
the North Pacific (MacCall, 1996). While those changes 
occur naturally and cyclically, they can be exacerbated 
by extreme fishing pressure. It follows logically that 
timely management actions may optimize the exploita- 
tion of the populations and aid their recovery (Petitgas 
et al., 2010; Radovich, 1982). Therefore, monitoring 
by synoptic, periodic, fisheries-independent multispe- 
cies surveys is indispensable for a successful transition 
to ecosystem-based fisheries management (Ecosystem 
Principles Advisory Panel, 1999; Rice et al., 2005). 
Presently, acoustic-trawl surveys uniquely provide 
synoptic multi-species information over large oceanic 
areas. Unlike the DEPM, acoustic-trawl surveys do not 
depend on the timing of spawning. In contrast to trawl 
surveys, echosounders sample much larger water vol- 
umes and ranges of organism sizes. There are, however, 
some significant challenges when conducting acous- 
tic-trawl surveys. The principal challenge is accurate 
classification of the echoes. With the fre- 
quency-dependent backscatter informa- 
tion, it is possible to objectively separate 
fluid-like scatterers (e.g., zooplankton, 
bladderless fish) from gas bearing organ- 
isms (e.g., fish with gas-filled swimblad- 
ders), and, to some degree, classes of 
organisms within those groups. How- 
ever, CPS with similar morphological 
features and sizes exhibit similar spec- 
tral responses, which require, to date, 
disambiguation by physical sampling. 
In the short term, acoustic-trawl sur- 
veys for monitoring CPS should include 
more trawl sampling, both directed and 
random, and the use of towed optical 
devices. Data from these will serve to 
reduce the uncertainty associated with 
species identification, especially for the 
less abundant species. Classification al- 
gorithms may be refined and the results 
validated by using data from nonlethal 
sampling devices, e.g., towed stereo 
cameras. 
The survey design could be optimized 
for improved acoustic and trawl sam- 
pling of species with low or patchily dis- 
tributed abundances in inaccessible in- 
shore regions (e.g., anchovy and herring) 
or outside the survey area (e.g., Pacific 
mackerel). Additionally, the combination 
of multi-frequency echosounder systems 
with multibeam (Cutter and Demer, 
2008) and omnidirectional sonars may 
serve to quantify potential biases due to 
the surface blind zone (Scalabrin et al., 
2009) and nearsurface avoidance of fish 
(De Robertis et al., 2010). 
In the medium term, species-specific 
habitat models, similar to that for sar- 
dine (Zwolinski et al., 2011) should be 
5900 7600 
10 15 20 25 30 
U5 
10 15 20 25 30 
Standard length (cm) 
Figuire 7 
Estimated biomass-weighted length distributions for the northern stock 
of sardine ( Sardinops sagax) in the California Current Ecosystem (CCE) 
during spring 2006, 2008, and 2010 surveys. The dashed lines repre- 
sent the estimated length composition from the assessment (Hill et al., 
2010). There is no indication of significant recruitment. Consequently, 
the sardine population is aging and declining. The instantaneous net 
mortality rate was estimate to be 0.56 by fitting an exponential-decay 
function to abundances derived from acoustic— trawl data. 
