180 
Fishery Bulletin 1 14(2) 
cause most survival data come from acoustic tagging 
studies, it is essential to improve our understanding 
of the underlying cause of mortality events (i.e. preda- 
tion, environmental, or other) from these types of in- 
struments. We developed a tool to address this research 
need by designing floating, baited, predation-event re- 
corders (PERs). These recorders allow estimation of 
relative predation rates in various environments, and 
reveal information about mortality produced by differ- 
ent species of fish predators. 
Our objectives were to investigate the feasibility of 
1) developing and constructing a passive, baited, GPS- 
enabled PER, 2) evaluating relative risk of predation 
mortality and 3) observing and identifying individual 
predators and associated predation events. We estimat- 
ed relative predation mortality and identified preda- 
tion hot spots upon juvenile Chinook salmon to com- 
pliment ongoing acoustic telemetry surveys. We were 
able to accurately identify the location of individual 
predation events, reliably identify predators, and the 
recorder system was easily deployed and retrieved by 
a boat-based crew of 2 people. 
Materials and methods 
PER construction 
PERs were constructed from an approximately 75 cm 
length of 76 mm diameter, schedule 40 polyvinyl chlo- 
ride (PVC) pipe. The bottom end was fitted with a PVC 
cap which was glued in place and the top end was fit- 
ted with a two part threaded, removable cap. Attached 
to the top cap was a GARMIN® TT' N T0 3 GPS transpon- 
der (available at website) set to update and record its 
position every five seconds. A predation-activated timer 
was attached to the bottom cap. The design of the timer, 
similar to that of Somerton et al. (1988), is connected 
to a baited line attached to a magnet, which is slotted 
inline into a receptacle on the timer, housing a mag- 
netic switch (Fig. 1). When the bait is pulled, the mag- 
net is removed, activating the timer that records the 
precise timing of the predation event. All GPS trackers 
were controlled and their tracks recorded with a GAR- 
MIN® Alpha 100® 1 handheld base-station unit plugged 
into a laptop computer located on board the boat. Up 
to 20 GPS trackers may be tracked simultaneously in 
real-time with one Alpha 100® handheld unit, as long 
as all trackers are within approximately 14 kilometers 
line-of-sight. More trackers can be tracked if multiple 
base stations are used. 
Attached to each predation timer was a 50 cm 
length of 3.6 kg breaking-strength fluorocarbon leader. 
A sub yearling fall-run juvenile Chinook salmon from 
the Mokelumne River fish hatchery was attached to 
the distal end of the fluorocarbon leader by means of 
3 Mention of trade names or commercial companies for iden- 
tification purposes only and does not imply endorsement by 
the National Marine Fisheries Service. 
a loop threaded through the mouth and operculum. A 
seven-gram split-shot style weight was placed approxi- 
mately 10 cm above the fish. Approximately 2.25-kg of 
lead shot was placed inside the bottom of each PER 
as ballast, which served to keep PERs upright while 
submerging all but the upper most 10 to 15 cm, where 
the GPS receiver was attached. GoPro 3 underwater 
cameras, with 64 gigabyte storage SanDisk 3 memory 
cards (available at website), were attached to a subset 
of 3 PERs opposite the predation timer and aimed di- 
rectly at the attached smolt (Fig. 1). PERs were spray 
painted in a green and brown camouflage pattern to 
reduce visibility and obtrusiveness in the upper water 
column, but the top 10 to 15 cm above the water line 
were painted a bright safety orange and marked with 
reflective tape for easier visual identification by pass- 
ing watercraft. Onset® HOBO® 3 pendant temperature 
and light data loggers (available at website) were at- 
tached to each PER (Fig. 1) so that we could relate 
environmental variables to predation events. 
Field trials 
All field trials were conducted within a 1-km study 
site (lat. 37.806°N, long. 121.317°W, lat. 37.799°N, long. 
121.313° W) on the lower San Joaquin River located 
approximately 1.5 km downstream from Mossdale, CA. 
The depth of the sampling site ranged from 3.65 m to 
0.6 m and had a mean depth of 1.98 m; the minimum 
effective depth of the PERs was 0.6 m, which repre- 
sented approximately 88% of the total wetted stream 
channel. Depth of the entire sampling site was mea- 
sured and mapped with boat-mounted sonar. River ve- 
locities ranged from 0.49 m/s to -0.32 m/s (mean veloc- 
ity: 0.27 m/s. Negative values denote a flood tide and 
reversal of flow going upstream. Channel width ranged 
from approximately 70 m to 90 m. The sinuosity index 
(SI) of the study reach was 1.21. SI is a measurement 
of a river or stream’s deviation from the shortest pos- 
sible downslope path. A value of 1.0 indicates a per- 
fectly straight channel, whereas increasing values of 
1 are representative of increased meandering (Muel- 
ler, 1968). Ten PERs were repeatedly deployed on 9 
separate trials, either one hour before sunrise or one 
hour before sunset by a 2-person boat crew, all track- 
ers remained within line of site of the boat while de- 
ployed. Each trial consisted of 2 separate deployments 
mid channel; if a tether became beached or fouled it 
was promptly retrieved and redeployed at mid-channel. 
Each re-deployment was considered a unique deploy- 
ment on its own. 
The procedure for deploying PERs was 1) activation 
of GPS transponders/GoPro cameras (30 sec/PER), 2) 
attachment of the salmon smolt to predation timers (1 
min/timer), 3) release of PERs (1 min/PER), 4) transit 
of PER through study site (45 min. to 1 h. depending 
upon river velocity), 5) retrieval of PERs and record- 
ing of timer data (20-30 min). Digital predation timers 
were immediately recorded upon retrieval. The cumu- 
lative time spent preparing, deploying, fishing, and re- 
