Demetras et al.: Use of underwater recorders to quantify predation of |uvenile Oncorhynchus tshawytscha 
183 
o 
CL 
0.80 - 
0 10 20 30 40 
Time (min) 
Figure 3 
Estimated survival function for the Cox regression of 
time to predation on water velocity and median depth. 
Broken lines indicate a point-wise 95% confidence in- 
terval about the survival function. 
smolt was missing, it was a confirmed predation event. 
Video analysis also indicated that the tethered smolts 
remained in an upright and active swimming position 
during deployments. 
Discussion 
Predation upon juvenile Chinook salmon in the Sac- 
ramento-San Joaquin Delta and resulting effects on 
population level has been a topic of debate. The pre- 
sumption that predation may play a significant role in 
survival was investigated with the use of statistical 
models on winter-run Chinook salmon by Lindley and 
Mohr (2003) and Hendrix. 4 Neither analysis implied a 
substantial link between striped bass predation and 
Chinook salmon survival. However, the quality of the 
data used in statistical analyses is a major determi- 
nant of the strength of the results, and diet data for 
many piscine predators in the delta is lacking (Gross- 
man et al. 2 ). The addition of robust data from new 
methods for quantifying predation may help fill this 
crucial gap for future modeling efforts. Our results, 
that predation was greatest at maximum positive wa- 
ter velocities (outgoing tide) and lowest at more nega- 
tive water velocities (incoming tide), are in contrast 
with those of Anderson et al. (2005), who found that 
4 Hendrix, N. 2008. A statistical model of Central Valley 
Chinook incorporating uncertainty: description of Oncorhyn- 
chus Bayesian analysis (OBAN) for winter run Chinook, 18 
p. R2 Resource Consultants Inc., Redmond, WA. [Available 
at website, accessed November 2014.] 
0.70 - 
0 10 20 30 40 
Time (min) 
Figure 4 
Estimated survival function for the Cox regression of 
time to predation over a series of water velocities from 
the minimum (negative) to the maximum (positive) wa- 
ter velocities by increments of 0.1 m/s. 
survival of juvenile salmon was influenced more by 
travel distance than travel time or velocity. However, 
there are some important differences between stud- 
ies. First, we conducted our study on a much smaller 
spatial and temporal scale that quantified individual 
predation events and therefore characterized more 
proximate, short-term processes. Secondly, our study 
system was strongly tidally influenced, to the extent 
that tidal water movements may have substantially 
affected predator behavior. 
Acoustic tag technology for basin-scale studies has 
become the standard for assessing movement and sur- 
vival of fish, particularly in salmonids (Perry et al., 
2010; Michel et al., 2013). However, these studies are 
often expensive and do not reveal the mechanisms or 
locations of mortality. Although researchers are de- 
veloping acoustic tags that report predation events 
through a change in tag ID code, it may be hours to 
days before the digestion-based processes trigger the 
predation event to be detected by a receiver (Afen- 
toulis and Schultz, 2014). Furthermore, acoustic tags 
designed to report predation events do not identify 
predator species or distinguish the difference between 
predation events and the scavenging of tagged fish af- 
ter some other cause of mortality. As such, alternative 
methods and instrumentation, such as PERs are need- 
ed to complement acoustic tagging studies to evaluate 
predation mortality. 
Fishery and water managers often call for investiga- 
tions of predator control along with predator habitat 
manipulation as a management tool, and some preda- 
tor-reduction studies have been implemented (e.g., Por- 
ter, 2011). Evaluations of predator control and preda- 
