Viz Lab Summer Grant 2007

2007 Reports 2006 Reports 2005 Reports 2004 Reports

Foraging Behavior of Age-0 Lake Trout (Salvelinus namaycush)

Beth Holbrook Graduate Student under Tom Hrabik, Biology

Introduction

Lake trout (Salvelinus namaycush) are the primary native predatory fish in the Laurentian Great Lakes. In the 1950s, lake trout populations crashed throughout the Great Lakes due to a combination of sea lamprey predation, overfishing, and reduction in spawning habitat. Populations became so low that natural reproduction ceased in all of the Great Lakes, except Lake Superior. The ecological importance of this species combined with its value to the commercial and sport fishing industries resulted in efforts to rehabilitate lake trout populations. These measures include controlling sea lamprey populations, stocking lake trout, eliminating commercial fishing, limiting sport fishing, and establishing refuges where fishing is prohibited. Despite over fifty years of management intervention, Lake Superior is the only Great Lake where natural reproduction occurs and lake trout populations have rebounded.

There is evidence that impediments to natural reproduction may be occurring the first several months after lake trout eggs hatch in the spring, yet little is known about this period of lake trout development. In order to determine whether competition, food availability, or bioenergetic constraints might be limiting survival, it is important to better understand the mechanisms by which age-0 lake trout forage. The purpose of this research was to use VDIL software to analyze video images of age-0 lake trout feeding trials to accurately determine parameters used in foraging models.

Overview of research

Foraging potential in fisheries is commonly approximated using mathematical models. One common model is an encounter rate model (Z) developed by Gerritsen and Strickler (1978):

where Rij is the reaction distance of the lake trout, Vj is the swimming speed of the lake trout, Vi is the swimming speed of the prey, and ui is the prey density. To estimate daily foraging potential, the encounter rate model is multiplied by the probability that a fish will locate, pursue, attack, and retain a prey item.

The purpose of our research was to accurately estimate these model parameters by conducting a series of age-0 lake trout foraging trials in the laboratory. We replicated Lake Superior temperatures by conducting all experiments in a temperature-controlled chamber at 8ºC. We also simulated natural light conditions by using cyan LED lights ranging from 450-550 nm, the predominant wavelengths found at depths greater than 45 m in Lake Superior. The illuminance of the LED lights was varied from 0 lux to 1960 lux to simulate a range of depths and nighttime conditions under which age-0 lake trout may forage.

Lab foraging experiments were conduced with two different prey species, amphipods (Hyallela spp.) and mysids (Mysis relicta), a freshwater shrimp commonly found in Lake Superior. Previous diet studies have found that mysids are the primary food source for age-0 lake trout; however, they are extremely difficult to culture in the laboratory. Amphipods were used in the experiments until mysids could be captured in the field. Experiments were conducted from March until July, as lake trout grew from approximately 30 mm to 70 mm. Feeding trials were videotaped overhead using a Sony DCR-TRV250 digital video camera recorder with built-in infrared light that was used to capture feeding trials conducted at 2 lux or less.

Application of Viz Lab software

Recorded video of the feeding trials were analyzed by importing digital clips into Quicktime 7 Pro. The clips were exported as *.avi files in black and white, and the frame number was reduced to 15 fps for ease of analysis. The *.avi files were imported into MATLAB 7.0 and digitalized using a customized MATLAB software program (Hedrick 2007). Digitalized coordinates were exported for the length of the lake trout, the length of the prey item, the search speed of the lake trout, the burst speed of the lake trout, the swimming speed of the prey item, and the vector of attack. The coordinates were imported into an Excel spreadseet and scaled to calculate distances (mm), speeds (mm/sec), and angles of attack. Quicktime 7 Pro was also used to view the video clips to analyze the probability that an age-0 lake trout would locate, pursue, attack, and retain a prey item during each feeding trial.

Results

Excluding 0 lux, age-0 lake trout reacted at a similar distance to prey at low light levels compared with high light levels. At 0 lux, lake trout fed unsuccessfully—during our trials we only had one successful capture that occurred at 0 mm (Figure 1). Lake trout searched for prey at a reduced swimming speed compared to their burst speed upon location of prey (Figure 2). Amphipods had a much slower swimming speed than lake trout, and were usually unsuccessful at avoiding capture (Figures 2, 3). The majority of attacks were located anterior to the age-0 lake trout (Figure 4), suggesting that age-0 lake trout relied heavily on visual detection and were not as reliant on other sensory systems such as their lateral line while foraging.

Figure 1 Reaction distance (mm) plotted against light levels (lux) for amphipods. Error bars are 2*SE.

Figure 2 Swimming speed (mm/sec) for lake trout while searching for prey, after detecting prey (“burst”), and swimming speed of amphipods. Error bars are 2*SE.

Figure 3 The probability of location, pursuit, attack, and retention. The overall probability for these separate events occurring was 0.69. Error bars are 2*SE.

Figure 4 Reaction angles and distances of attack. Each point represents a prey item that was attacked, with the age-0 lake trout oriented vertically (shown by the dark shape in the middle of the polar plot). The majority of attacks occurred between 90º and 280º.

References

Gerritsen, J., Strickler, J., 1978. Encounter probabilities and community structure in zooplankton: a mathematical model. J. Fish. Res. Board Can. 34, 73—82.

Hedrick, Tyson L. "DLT Data Viewer 2", Digitizing and DLT in MATLAB. 2007. http://www.unc.edu/~thedrick/digitzing/