THE SEARCH BEHAVIOR OF SEA LAMPREY DURING THEIR NON-HOMING
REPRODUCTIVE MIGRATION IN THE GREAT LAKES
By
Trevor D. Meckley

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Fisheries and Wildlife – Doctor of Philosophy
2015

ABSTRACT
THE SEARCH BEHAVIOR OF SEA LAMPREY DURING THEIR NON-HOMING
REPRODUCTIVE MIGRATION IN THE GREAT LAKES
By
Trevor D. Meckley

Non-homing organisms are thought to rely on stable geophysical features (e.g., landslope), and encounter with distinct attributes of the environment (e.g., landmarks) in a predicable
sequence to locate reproductive habitat (e.g., rivers always occur on a coastline). However little
is known about how non-homing fishes complete a large-scale migration, as most research has
focused on homing fishes that rely on geomagnetic cues to return to a natal spawning site. The
invasive sea lamprey offers insight into non-homing migration, as individuals complete a single
non-homing migration to rivers following translocation by host fishes in the Great Lakes.
In this dissertation, Chapters 2 and 3, are devoted to developing a framework for
describing animal behavior from telemetered observations and development of a standardized
approach for assessing and filtering VEMCO Positioning System (VPS) data based on an
estimate of horizontal position precision (HPE). This methodology was imperative for
exploration of the sea lamprey migration with the underwater VPS telemetry technology.
In Chapter 3, we described how sea lamprey orient to a coast when in a lake and
hypothesized that sea lamprey navigate to the nearest coast by (1) orienting to the local
bathymetric gradient and (2) maintain straight movements counter to the local slope to move
towards shallow water. Three-dimensional (3-D) paths of migrating female sea lamprey were
obtained by an acoustic array with 3 km2 of coverage, centered 3.3 km from the coast in Lake
Huron. The findings of this chapter indicate that sea lamprey sampled an area of lake-bottom to

assess absolute hydrostatic pressure and to select a heading towards reducing pressure (shallower
water). In contrast to natal homing migrations, the sea lamprey appears consistent with nonhoming orientation to a general region with a simple set of rules based on local topography.
Chapter 4 focused on the sea lamprey migration along a coastline and near a river mouth.
Upon reaching a coastline, sea lampreys move parallel to shore. Prior studies indicated the
presence of larval odor in river water increased the likelihood that a migrant entered a river.
However, it was not known whether larval odor played a role in navigation (guiding the migrant
to the river mouth) or mediated habitat selection by labeling the suitability of a river for
spawning. In a two-year study using a 2 km2 acoustic array, the 3-D paths of sea lamprey were
documented as they approached, entered, or bypassed the Ocqueoc River in northern Michigan,
under one of two conditions: (1) low larval odor; and, (2) higher larval odor, created by increased
larval abundance plus the addition of the synthetic larval odor components, petromyzonamine
disulfate (PADS), petromyzosterol disulfate (PSDS), and petromyzonol sulfate (PZS) to a 1 x 10
-12 M concentration. A coupled hydrodynamic and dye concentration model predicted the
hydraulic conditions experienced by each sea lamprey by estimating water conditions (velocity,
temperature, etc.) at each fish position and allowed for assignment of whether a position was
inside or outside of the Ocqueoc River plume. Encounter with river water appears to trigger
localized search, regardless of larval odor content. However, when larval odor was abundant, the
migrant was more likely to enter the river. Whether a migrant enters a river, is modulated by the
presence of detectable larval odor, manipulation of river selection by invasive sea lamprey for
management is viable in rivers with high encounter rates. Finally, Chapter 5 covers the
implications for how altering migration routes in the sea lamprey via the application of
synthesized pheromones contained in larval odor could be profitable for management.

To my wife for five great years in Michigan…

iv

AKNOWLEDGEMENTS

I would like to thank my advisor, Dr. C. Michael Wagner immensely for his contributions
to my dissertation. His mentoring on research design, interpretation of results, and presentation
of findings has made me a much improved scientist. The members of my graduate committee all
supported different aspects of my dissertation and included Dr. James Miller, Dr. Daniel Hayes,
Dr. Charles Krueger, and Dr. Weiming Li. I thank my committee, Dr. Eliezer Gurarie, Dr. Tom
Binder, Dr. Christopher Holbrook, Dr. Frank Smith, and Jessica Barber for reviewing chapters. I
would like to specifically thank Dr. Eliezer Gurarie and Dr. Christopher Holbrook for their
collaboration and help on my dissertation, including teaching me the ways of R. Dr. Tuan
Nguyen and Dr. Mantha S. Phanikumar are responsible for the hydrodynamic models provided
in Chapter 4. I thank all personnel at the United States Geological Survey Hammond Bay
Biological Station, Millersburg, Michigan, and the United States Fish and Wildlife Service
Marquette Biological Station, Marquette, Michigan, for their assistance in animal capture,
housing, maintenance, and technical support during these projects. In addition, I would like to
thank specific individuals at the stations who have worked directly with me on my projects in
some capacity including: Dr. Michael Hansen, Dr. Roger Bergstedt, Erick Larson, Karen Slaght,
Jessica Barber, and Dr. Nicholas Johnson. I also received additional support from Glen Black,
Tyler Buchinger, Cory Olaf Brant, Dr. Michael Siefkes, Dr. Michael Twohey, Dr. Darryl
Horndorp, Captain Joseph Bergan, Dr. Robert Goodwin, Dr. John Hume, and Dr. Thomas
Luhring. I would like to give special thanks to field technicians for their hard work, dedication,
and patience during these projects: Eric Willman, Brett Diffin, Amber Masters, Greg Byford,
Skye Fissette, Carrie Kozel, Sarah Ptasznik, and Jeff Yaklin.

v

Lastly, I thank my entire family. My mom, dad, Janet and Joel have provided incredible
support for me during my Ph.D. degree. Most importantly I thank my wife, for all of her love and
support and for moving to Michigan, even though I went away every year to chase “ugly” fish.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................................ x
LIST OF FIGURES ................................................................................................................... xiii
CHAPTER 1 .................................................................................................................................. 1
AN ADAPTIVE WORKFLOW FOR STUDYING BEHAVIOR WITH TELEMETERED
OBSERVATIONS. .................................................................................................................. 1
ABSTRACT ............................................................................................................................... 1
INTRODUCTION ..................................................................................................................... 1
GENERAL WORKFLOW ....................................................................................................... 4
Question Development ............................................................................................................. 4
Data Collection........................................................................................................................ 5
Data processing ....................................................................................................................... 6
Analyses ................................................................................................................................... 7
APPROACH AND IMPLEMENTATION ............................................................................. 9
Step 1: Question Development .............................................................................................. 11
Step 2: Data Collection ......................................................................................................... 12
Step 3: Data Processing ........................................................................................................ 13
Step 4: Analyses ..................................................................................................................... 14
CONCLUSION ........................................................................................................................ 14
APPENDIX .............................................................................................................................. 16
REFERENCES ........................................................................................................................ 20
CHAPTER 2 ................................................................................................................................ 25
A PROCESS FOR FILTERING HYPERBOLICALLY POSITIONED UNDERWATER
TELMETRY DATA WITH HORIZONATAL POSITIONG ERROR (HPE) ..................... 25
ABSTRACT ............................................................................................................................. 25
INTRODUCTION ................................................................................................................... 26
Background ............................................................................................................................ 26
The VPS system ...................................................................................................................... 27
Filtering spatial data with HPE ............................................................................................ 28
METHODS .............................................................................................................................. 30
Error evaluation based on test data ...................................................................................... 31
HPE evaluation with test data ............................................................................................... 32
Filtering fish data .................................................................................................................. 34
RESULTS ................................................................................................................................. 34
Step 1: Establishing data quality objectives .......................................................................... 34
Step 2: Estimating the baseline position confidence in the array ......................................... 35
vii

Step 3a: Evaluating HPE filters consistent with the data quality objectives and guides ...... 36
Criterion 1 ............................................................................................................................. 37
Criterion 2 ............................................................................................................................. 37
Criterion 3 ............................................................................................................................. 37
Criterion 4 ............................................................................................................................. 38
Step 3b: Selecting an HPE cutoff .......................................................................................... 38
Step 3c: Evaluating the potential for introduced bias through application of HPE cutoff ... 39
DISCUSSION .......................................................................................................................... 39
ACKNOWLEDGEMENTS .................................................................................................... 45
APPENDIX .............................................................................................................................. 46
REFERENCES ........................................................................................................................ 60
CHAPTER 3 ................................................................................................................................ 64
HOW DO NON-HOMING FISHES FIND THE SHORE? EVIDENCE FOR
BATHYMETRIC ORIENTATION IN MIGRATING SEA LAMPREY ............................ 64
ABSTRACT ............................................................................................................................. 64
INTRODUCTION ................................................................................................................... 65
METHODS .............................................................................................................................. 68
RESULTS ................................................................................................................................. 80
DISCUSSION .......................................................................................................................... 83
ACKNOWLEDGEMENTS .................................................................................................... 90
APPENDIX .............................................................................................................................. 91
REFERENCES ...................................................................................................................... 149
CHAPTER 4 .............................................................................................................................. 158
DOES LARVAL SEA LAMPREY ODOR AID NAVIGATION OR GUIDE HABITAT
SELECTION DECISIONS BY ADULTS .......................................................................... 158
ABSTRACT ........................................................................................................................... 158
INTRODUCTION ................................................................................................................. 159
METHODS ............................................................................................................................ 162
General Methods ................................................................................................................. 162
Study Site ............................................................................................................................. 163
Experimental Subjects ......................................................................................................... 163
Pheromone application........................................................................................................ 164
VPS Array Performance and data treatment ....................................................................... 165
Hydrological Data ............................................................................................................... 166
Identifying active sea lamprey ............................................................................................. 169
Characterizing river plume encounter and entry ................................................................ 170
Movement Patterns .............................................................................................................. 172
Effect of Synthetic Odor ....................................................................................................... 174
RESULTS ............................................................................................................................... 174
Sea Lamprey Activity ........................................................................................................... 174
Characterizing river plume encounter and river entry ....................................................... 175
viii

Movement Patterns .............................................................................................................. 177
Effect of synthetic odor ........................................................................................................ 179
DISCUSSION ........................................................................................................................ 179
Movement outside of the river plume .................................................................................. 181
River plume encounter ......................................................................................................... 182
Movement in the river plume ............................................................................................... 184
The role of larval odor......................................................................................................... 185
River entry ........................................................................................................................... 187
Odor mediated manipulation of sea lamprey habitat selection ........................................... 188
Summary .............................................................................................................................. 189
ACKNOWLEDGEMENTS .................................................................................................. 190
APPENDIX ............................................................................................................................ 191
REFERENCES ...................................................................................................................... 291
CHAPTER 5 ............................................................................................................................. 298
ISOLATING OPPORTUNITIES FOR MANIPULATING SPAWNING HABITAT
SELECTION OF INVASIVE SEA LAMPREY (PETROMYZON MARINUS) IN THE
LAURENTIAN GREAT LAKES ....................................................................................... 298
ABSTRACT ........................................................................................................................... 298
INTRODUCTION ................................................................................................................. 298
RIVER ENTRY FRAMEWORK ........................................................................................ 300
River Plume Encounter ........................................................................................................ 300
River Entrance, Upstream Movement, and Tributary Selection ......................................... 301
River Retention .................................................................................................................... 304
MANIPULATING THE MIGRATION .............................................................................. 304
River Plume Encounter ........................................................................................................ 305
Cohort-Size Maintenance .................................................................................................... 305
Recruitment to habitat deficient for larval survival ............................................................ 306
Chemical barrier and Push-Pull ......................................................................................... 307
Larval Odor Removal (TFM) .............................................................................................. 308
CONCLUSION ...................................................................................................................... 309
APPENDIX ............................................................................................................................ 310
REFERENCES ...................................................................................................................... 317

ix

LIST OF TABLES

Table 1.1: Studies of animal behavior that focus on telemetry data should follow this general
framework and report on each component: question development data collection, data
processing, and analysis. Using the framework ensures that at the start of the study there is clear
consideration of what data is necessary to properly answer the question and describe the behavior
in an exploratory, explanatory or predictive context. ...................................................................17
Table 1.2: The components of Ethology. ......................................................................................18
Table 2.1: Filtering Objectives. The four specific criteria adopted to establish the data quality
objectives for the project ................................................................................................................47
Table 2.2: Criteria for selection of an HPE filter cutoff ..............................................................48
Table 3.1: Sea lamprey were primarily detected on the nearest coast within 72 hours of release.
This included 78 % of individuals that were observed stopping in the array and moving on the
first night (n=22) and 61% of all individuals released (n=67). Note that Time to reach the
receiver mean and min/max are only listed for the 22 individuals. Only 4 of 67 sea lamprey’s
first detections occurred outside of the nocturnal movement period, with 3 of 4 occurring on the
first day of release with 2 at Forty Mile point (East), 1 at the Ocquoec river mouth (South) and 1
near Cheobygran river mouth (West). ..........................................................................................92
Table 3.2: Individual t-tests for each sea lamprey of whether sea lamprey exhibited a persistence
in turning (theta; -pi to pi), different from 0, revealed that more sea lamprey had a bias of left
(negative) or right (positive) turns during phase 1 (P1), but not during phase 2 (P2). Most
persistent turns were counterclockwise. The magnitude is also greater for those that are
significant in phase 1 than phase 2. A pairwise t-test of the concentration in turning (Rho) for
each individual between phase 1 and phase 2 revealed that individuals tended to go straighter
during phase 2. Bold values are significant. .................................................................................93
Table 3.3: Pairwise t-tests to determine if there was a difference in the mean depth, standard
deviation in depth, mean depth when maintaining vertical depth (Hz Mean Depth), or ground
speed during phase 1(P1) and before the transition to phase 2 (P2). The tests support our
observations that these two phases are different. Sea lamprey had greater mean depths and less
variation during phase 1, and when moving at a particular depth in the water column, sea
lamprey spent time on the bottom during phase 1 and closer to the surface during phase 2. The
ground speed (GS) was also significantly lower during phase 1 than phase 2. The only
individuals with less variation in depth during phase two were those individuals swimming
mostly at the surface with occasional vertical excursions. Individuals T42-T56 did not have
pressure sensitive tags. ..................................................................................................................94

x

Table 4.1: Three specific criteria were adopted to establish the data quality objectives based on
an extensive review of data quality (Meckley et al. 2014a). ......................................................192
Table 4.2: Criteria for selection of an HPE filter cutoff. In the data values listed for each
criterion the HPE cutoffs are listed acceptable HPE cutoffs are listed for 2010 above and 2011
below. ..........................................................................................................................................193
Table 4.3: Akaike’s information criterion (AIC) values for linear mixed effects models using
river entry (Em) as a binary response variable to various combinations of fixed effects including
year (Y), iterative number of nights the sea lamprey encountered river water (N), a year by night
number interaction (YxN), average nightly discharge of the Ocqueoc River (DC: m3·S-1),
average lake temperature encountered (LT: C ̊), average nightly river temperature (RT; C ̊ ), the
difference in experienced lake temperature and river temperature (DFT; C ̊ ), minimum distanced
reached from the river mouth (D: km), and average water velocity experienced in the lake (V). ...
......................................................................................................................................................194
Table 4.4: Akaike’s information criterion (AIC) values for a nonlinear mixed-effects model in
the formulation described by Lindstrom and Bates (1990), using the function “nmle” in R (R
Development Core Team 2015). Ground speed (Vm) as a response variable to various
combinations of fixed effects including year (Y), the log of the cumulative distance traveled (D),
whether encounter with river water occurred yet in the path (E), or if the point was currently in
the plume (P), temperature at fish position (T), and an interaction effect between water current
magnitude and direction by fish heading interaction (CxF); and finally the random effect of
individuals (animal ID). ..............................................................................................................195
Table 4.5: Akaike’s information criterion (AIC) values for a nonlinear mixed-effects model in
the formulation described by Lindstrom and Bates (1990), using the function “nmle” in R (R
Development Core Team 2015). Path straightness (Sm) as a response variable to various
combinations of fixed effects including year (Y), the log of the cumulative distance traveled (D),
if whether encounter with river water occurred yet in the path (E), or the point was currently in
the plume (P); and finally the random effect of individuals (animal ID). Model 6 and 9 had the
lowest AIC scores. ......................................................................................................................196
Table 4.6: To evaluate the effect of a host of variables on river entry (Em), a mixed effects
logistic regression with a binary response variable (river entry: 1 or 0) was implemented in R
using glmer (lme4 package, Bates et al. 2014). The best fitting models, as determined by AIC,
for river entry included year and minimum distance reached from the river and either the lake
temperature experienced (Model 10) or the difference in lake temperature experienced and the
river temperature (Model 11). The fixed effects included year (Y), iterative number of nights
encountering river water (N), a year by night number interaction(YxN), average nightly
discharge of the Ocqueoc River (DC: m3·S-1), average lake temperature encountered (LT: C ̊),
average nightly river temperature (RT; C ̊ ), the difference in experienced lake temperature and
river temperature(DFT; C ̊ ), minimum distanced reached from the river mouth (D: km), and
average water velocity experienced in the lake (V). Here is Model 11 output. ..........................197

xi

Table 4.7: Output from the summary of the best fit nonlinear mixed-effects model for the
continuous response variable ground speed (body lengths per second), Model 9. The model
included the temperature experienced (T), whether encounter with the river plume had occurred
in the path (E), whether the sea lamprey was in the river plume (P), and an interaction between
water current and if the fish was moving towards or away from the current direction. The
correlation statement accounts for first order autocorrelation in the residuals that occurred from
analyzing multiple simultaneous steps relating to the same individual. The random effect of
animal ID is included. .................................................................................................................199
Table 4.8: Output from the summary of the best fit nonlinear mixed-effects model for path
straightness (0-1), Model 9. The model included the effect of the interaction between cumulative
distance traveled (D) and Encounter with the river plume and the fixed effect of being in the
river plume. The correlation statement accounts for first order autocorrelation in the residuals
that occurred from analyzing multiple simultaneous steps relating to the same individual. The
random effect of animal ID is included. The distance is divided by 1000 to put the variable in km
rather than m for easier interpretation. ........................................................................................201

Table 5.1: Six general strategies for manipulating sea lamprey migration behavior are presented
including the stimulus affected, how the stimulus is affected (action), the goal of the
manipulation and the type of manipulation. Only option one does not stand alone as a
management strategy as it manipulates behavior but not a decision to select habitat. ...............311

xii

LIST OF FIGURES

Figure 1.1: Analysis objectives aim to identify what is the animal doing (exploratory), why the
animal is doing what it is doing (explanatory), and can we anticipate how it will change what it
will do under different conditions (predictive)? The amount known about the animal often plays
a role in where researchers aim study objectives initially and we identify the usual process taken,
but the best behavioral studies aim to reach explanatory and predictive answers regardless of the
initial understanding of behavior. .................................................................................................19
Figure 2.1: Acoustic telemetry activities at the Hammond Bay field site. A schematic of the VPS
array that was located in Lake Huron around the mouth of the Ocqueoc River (blue line).
Triangles represent receiver (VR2W) positions. VPS array testing in 2010 included two
stationary tag tests (Gray dots, with median point as a black dot) and three mobile test transects
(black dots forming lines). The schematic is oriented with north up and the black line running
from left to right (east to west) represents the coast. .....................................................................49
Figure 2.2: The stationary test schematics and 2DRMS plots of all stationary test positions. Two
schematics depict all VPS positions during two stationary tag tests, including (a) 29,355
positions between the dates (6/17/2010 to 7/01/2010), and (b) 16,400 positions between the dates
(7/01/2010 to 7/08/2010), allowing us to evaluate array performance at two locations through an
extended period of time. The white dot in the center of the clusters is the median location. The
HPE versus measured error to the median point is shown for each estimated position during test
one (c) and test two (d). The white circles with black outline and red x represent twice the
distance root mean square error of x and y components of error within an HPE bin of one; 95%
of tag detections have an error less than this point within each bin. Note there is a minimum HPE
of 2.7 and 2.5 within the data. The line running between these points represents the 2DRMS and
the equation and fit for this line are shown in the top left corner of (c) and (d), respectively. Data
points above the 15 m bin, which can be seen in (a) and (b) are not shown in (c) or (d), because
they are outside of the zone of interest. .........................................................................................50
Figure 2.3: Proportion of test positions with measured error from 1 to 10 m. Percent of positions
with accuracy equal to or less than each measured value for the mobile tag test (circle),
stationary tag at location one (triangle), and the stationary tag at location two (square) depicted
an array with most positions having accuracy better than 6 m. The average error of the unfiltered
data for the stationary tag at location one (1.98 m), location two (1.11 m), and the mobile test
(6.83 m) are marked by representative symbols along the x-axis. ................................................52
Figure 2.4: Resultant data quality for HPE cutoffs of 3 to 15. The mean and maximum measured
error is below 1.77 m for all HPE thresholds, sufficient to meet criteria 1 for both stationary test
one (black) and test two (red). The maximum error exceeds 15 m at an HPE of 7 for test one and
8 for test two (violating criteria 2). The number of positions violating each test is located at the
top of the figure for HPE cutoffs of 3 to 15. .................................................................................53

xiii

Figure 2.5: Data loss versus error retention for HPE cutoffs of 3 to 15. The relationship of the
percent of incorrectly rejected positions of all acceptable positions and percent of incorrectly
retained positions of all retained positions suggested that HPE cutoffs of 8 to 10 for stationary
test one and 3 to 15 for stationary test two, met criteria 4. ...........................................................54
Figure 2.6: An evaluation of the performance of the VPS array with a mobile tag. (a) A
schematic depicting receiver positions (+) and the coast (black line) during the 2010 research
season. A mobile tag test was completed on 7/6/2010. The small dots represent the VPS
estimated positions during the mobile test. There were a total of 126 correctly retained positions
(black , <6 m error, <HPE 8), 5 correctly rejected positions (blue), 3 incorrectly rejected
positions (orange), and 5 incorrectly retained positions (red). The incorrectly rejected (n = 3)
positions occurred consecutively and were located at the furthest distance from the array in the
left transect. (b) A graph scaled to cover positions with an HPE between 0 and 20 depicts
correctly rejected, correctly retained, incorrectly rejected, and incorrectly retained positions. All
values with an HPE greater than 20 were correctly rejected (n = 3, not shown). Euclidean
distance is our best estimate of measured error. ...........................................................................55
Figure 2.7: HPE values for all fish positions. The schematic depicts the HPE value for each fish
position during active monitoring of study subjects. The coast is represented by the black line,
receivers are represented by black triangles, and all sea lamprey positions are color coded by
their associated HPE estimate and it is clear that HPE greatly increases when the sea lamprey
position is outside of the array proximity, other than for a few positions inside the array. ..........57
Figure 2.8: Fish positions filter by an HPE cutoff of 8. A depiction of how many of all fish
locations in 2010 would be rejected by an HPE filter of 8.0. Fish locations are represented by
small black dots, receiver positions are shown by red triangles and the coast is depicted by a gray
line. These are cumulative graphs with all sea lamprey tags shown in (a), fish positions with an
HPE below or equal to 8.0 are displayed in (b), and fish positions with an HPE greater than 8.0
are shown in (c). ............................................................................................................................58
Figure 3.1: Adult female sea lamprey were obtained from 5 rivers via barrier traps (Manistee
River (2a), Cheboygan River (1), Manistique River (3), Betsie River (2b), Ocquoec River
(“OCQ”), and parasites were caught attached to fish in the lake. A telemetry system (VPS,
VEMCO) composed of 41 receivers was used to position sea lamprey (“array”). There were
individual receivers that provided detection data along multiple locations on the coast and in
front of as well as in the Black Mallard (BM), Ocquoec (OCQ) and, St. Mary’s Rivers (SM).
Receiver strings were treated as individual detection points and were located near the Cheboygan
River mouth (CHB), 40 mile point (FMP), and Presque Isle (PRS). ............................................96
Figure 3.2: Each sea lamprey was released near the bottom in the middle (yellow star) of a 43
receiver array (orange dots) with 3 km2 coverage. A receiver in the center of the array at the
release point and on the east side of the array were not recovered, resulting in an irregular design.
The black arrow demonstrates the direction to the nearest coastline 172o (3 radians). ................97
Figure 3.3: The track of sea lamprey T04 is depicted as it exits the array (a), and below is the
relationship between backward path length and backward beeline distance of the track (b). An
xiv

arrow is drawn from the breakpoint to the end of the phase two to demonstrate the direction
headed during Phase 2 (a). In the lower graph the yellow line is a one to one relationship from 0.
A broken stick model fit is shown with the red lines representing the standard error and the black
line is the line fit (b). The break point separates phase one (orange) from phase two (blue). The
22 tracks of Group A are shown Figure 3.10. ...............................................................................98
Figure 3.4: Dotted lines connect phase 1 (orange) and phase 2 (blue), data points for each sea
lamprey subject, separated by a transition characterized by backward path length versus
backward beeline distance sinuosity. Both the mean value is shown (dot) as well as the standard
deviation in each axis (+). There is an increase in ground speed and a transition from a turning
bias in one direction to no bias in turning direction (a), as turning persistence overlaps 0 during
phase 2. Also apparent was a change in the mean depth from mostly movement on the bottom in
phase one with brief vertical excursions to the surface followed by movement on the surface
during phase two and brief vertical excursions to the bottom (b). ................................................99
Figure 3.5: The overall heading of 22 sea lamprey during phase two is depicted as a black dot
within one of 60 bins separating the circle and shows an average heading of 239o (gray arrow)
with a 95% confidence interval of 201-276o (shaded area) (a). The heading of sea lamprey with
respect to the ideal heading if following local bathymetry at the exit of the array is depicted by
the gray arrow with most sea lamprey clustered near 0 (mean= -7, 95% CI: 307-49o), repsenting
a movement towards shallower water, while a few individuals were observed moving opposite of
the ideal bathyemtric controur. The phase two headings were significantly oriented to local
bathymetry in the region where they exited the array (r = 0.32, p = 0.017; Rayleigh Test,
specified mean: 0o). .....................................................................................................................100
Figure 3.6: Sea lamprey (n=22) settled in the array during the day and began moving at night
(yellow), transitioning (green circle) between phase one (red arrow) and phase two (black arrow)
before they exited the array. Sea lamprey that moved north (b), east (c), south (d), and west (e)
are shown in individual frames with symbols indicating where each individual was first detected
on the coast. The array is color coded by depth 20-39(m).......................................................... 101
Figure 3.7: The track of sea lamprey T04 is depicted as it exited the array (a), and below is the
water column depth (b) and ground speed during the track (c). The graphs are color coded by 15
minute intervals. In inset “a” small circles represent receiver positions and the small colored dots
represent fish positions with a line showing the path. In inset “b” the red line represents the total
water column depth and the squares show the depth of the fish with respect to the total water
column depth through time. All 22 tracks are shown in Figure 3.10. .........................................102
Figure 3.8: Overall most observations occurred near the bottom (All observations; left panel).
During phase 1 most sea lamprey moved on the bottom and made occasional excursions to the
surface that returned to the bottom (center panel). Phase 2 consisted of more surface movements
with occasional excursions to the bottom, although a few individuals did move vertically and
only moved on the bottom (right panel). ....................................................................................103
Figure 3.9: Near shore sand waves in shallow water of Lake Huron (a) versus offshore detritus
and algae covered bottom without clear features associated to a common direction (b). ..........104
xv

Figure 3.10: There are 2 pages dedicated to each subject and only the data is shown for the
points when fish began to move following settlement, with the exception of transmitter 21 that
was shown moving immediately after release. The two pages correspond to the pathways of 22
subjects. Page one, “Movement Summary”, depicts the path of each sea lamprey through the
array when active (a) and is color coded by 15 minutes intervals corresponding to the lower
graph (b,c). Time is in Eastern Standard Time (UTC-4) and shown in a 24 hour time scale. The
lower graph depicts the vertical movements of sea lamprey for those individuals that had
transmitters with pressure sensors corresponding to the left y axis (b) and ground speed
corresponding to the right y axis for all individuals (c). The red line is the total water column
depth at each fish position. Note only T21 is shown moving during the day as it was notable
because it appears to orient during the day and then repeated the process at night but the night
time track was too short to consider in the movement classification. This suggests that this is a
daily orientation process. Page 2, “Movement Classification”, the lower graph depicts the
backward path length versus the backward beeline distance for each track. The point at (0, 0)
represents the point that the subject left the array. The yellow line is a one to one relationship
from 0. A perfectly straight path from the exit would move along this yellow line. The broken
stick model fit is shown with the black line and the red lines represent the standard error. The
break point is the point we use to separate phase one from phase two. Phase one is highlighted in
the in orange and phase two is depicted in blue. In the upper graph, the path of each individual is
shown and an arrow is drawn from the breakpoint to the end of phase two when the fish exits the
array coverage to demonstrate the heading of phase two that was used to test against the direction
to shallower water and the direction to the nearest coast using Raleigh tests. ...........................105
Figure 4.1: The study site was located in Hammond Bay of Lake Huron (A, B). Sea lamprey
were released at three locations (open circles). The study site was covered by a two kilometer
wide array that reached a kilometer from shore (C). The receiver positions in 2010 are
represented by triangles, and in 2011 they are represented by circles. In each year nine sync tags
were distributed in two rows across the array. During peak performance a transmitter could be
positioned every 15-45 s. Two of the six river receivers used for monitoring passage are
represented by black squares. The receivers in the river could not detect tags in the lake. ........203
Figure 4.2: The hydrodynamic model makes predictions of the current speed and direction and
lake temperature at each node (N=4346) of an unstructured grid fit to Hammond Bay (FVCOM
grid).The size of the triangular grid cells decreases near the river mouth to provide a higher
precision near our fish tracking array (minimum grid size: 9.15 m2). The hydrodynamic model
provides predictions of current speed and direction and water temperature. The model is three
dimensional and provides predictions in 2010 and 2011 at 20 slices through the water column
water column, independent of depth (m), which is color coded. ................................................204
Figure 4.3: Comparisons between vertical averaged currents form ADCP observations and
simulations taken in 2011. ..........................................................................................................205
Figure 4.4: Comparisons between observed temperature recorded by the ADCP and
hydrodynamic model simulations in 2011. .................................................................................206

xvi

Figure 4.5: Sea lamprey mostly began moving between sunset (orange vertical lines) and
nautical twilight NT (gray box) in most cases (A). Activity continued until between NT and
sunrise, although activity continued in lighter conditions than under which it commenced (B).
The double vertical lines represents sunrise and sunset as the time from NT changes with night
length. Sea lamprey entered the river throughout the night but the earliest they were observed
entering was just before NT (C) and despite activity being observed at sunrise in the lake, river
entry was never observed outside of NT in the lake (D). Two graphs show river entry as night
length changes with day length, this explains the dark gray areas (NT, all year), and light gray
areas (NT during some parts of the year). Focus on the 0 side of insets C and D. ....................207
Figure 4.6: The entry rate based on the total number of different nights that sea lamprey were
observed encountering river water in 2010 and 2011. The nights do not have to be consecutive
and don’t have to include the first night after release. The entry rate on the first night “1st night”
refers to the entry rate of sea lamprey that encountered river water on the first night of release
after settlement. The number of sea lamprey that fall into each category for each year are found
at the top of the graph. ................................................................................................................208
Figure 4.7: The 33 tracks of sea lamprey that did not encounter river water on the first night are
color coded and the arrow of the heading from the 10th to last point to final point are shown (A).
The circular plot in A (top) shows that most fish moved west out of the array. The exit arrow
(thick arrow) is shown for each sea lamprey and the average current experienced is shown by the
thin arrow, where the arrow points the current direction and the length represents current strength
(max: 27 cm·s-1,average: 4.9 cm·s-1) (B). The circle plots show the current directions observed
on the 33 nights (“Current”), the direction sea lamprey moved when the current was from the
East (“Fish-East Current”), the direction the sea lamprey moved when the current was from the
west (“Fish-West Current”). .......................................................................................................209
Figure 4.8: Each transect (set of two points) for sea lamprey that never encountered river water
were binned by 100 m increments from the coast based on their approach ratio (distance moved
towards the coast/total distance traversed). When sea lamprey were 500-1000 m from the coast
movement was directed towards the coast (1) before turning and moving parallel to the coast (0500 m). An increase in movement away from the coast occurs between 100-300 m from the coast
resulting in an average parallel movement along the coast of 303 m ± 221 m when exiting the
array (centered at an approach ratio of 0). ..................................................................................211
Figure 4.9: Ground speed (Body lengths per second) is shown for each step over the cumulative
path length for all sea lamprey (N=33) that did not encounter river water on the first night. Each
line represents a different individual. The pattern for before river plume encounter (black circles)
and after river plume encounter (red squares) is shown. Observations are restricted to a
cumulative path length of 2500 meters. A ramp up in speed to one body length per second is
evident in the first 500 meters of activity. ..................................................................................212
Figure 4.10: The google earth image depicts the distinct nearshore wave zone which includes a
comparatively steep series of banks and troughs that run horizontal to the coast compared the
shallow bathymetric gradient in Hammond Bay. The image captures the coastline in Hammond

xvii

Bay in front of the Ocquoec River Mouth, located in the bottom right of the image. The edge of
the banked zone extends from 204-308 m from shore in this area of the array. .........................213
Figure 4.11: A smoothed utilization distrution in 50 m cells of all active sea lamrpey points
following river plume encounter as defined by encounter with a dye concentraiton of greater than
5 ppm in 2010 and 2011. The white line represents the Ocquoec River. Movement extended
farther east and west along the coast in both cases but was not captured by the array. ..............214
Figure 4.12: The river plume occupied three general forms in Hammond Bay as represented by
the output of the dye concentration model. The image to the left represents a less common river
plume extending to the west. The central schematic depicts the river plume that occurs when the
plume switches from either east to west or occasionally when currents were moving in northerly
or southerly directions and pushed the plume out from shore. The most common river plume
extended to the east and extended out of Hammond bay. The river water was often trapped to the
east and west due to the shape of the coast. ................................................................................215
Figure 4.13: The temperature by Julian date for 2010 (Blue Line) and 2011 (Orange Line) is
shown. A symbol represents each fish if it was in the array on a given night in 2010 (Blue) and
2011 (Orange) at the maximum temperature encountered by the fish. Triangles represent fish that
entered the river on that night and circles representing fish that do not enter that night. An
individual fish is represented multiple times if it returns on multiple nights. ............................216
Figure 4.14: The percentage of the entire population of adult sea lamprey in Lake Huron that
return to the Ocqueoc River to spawn from 1995 to 2015 based on mark recapture (black circles)
and the change in entry between years (blue circles). A change of entry rate of 2 represents a
. doubling in the rate of entry. Vertical lines represent the times of TFM treatment to the Ocquoec
River. The change in entry rate from one year after TFM treatment to the second year shows that
the increase in entry rate between 2010 and 2011 (D) fits within the historical increase in rate (A:
1.8, B: 1.3, C: 3.1, D: 2.1). ..........................................................................................................217
Figure 4.15: The 40 sea lamprey that passed through the river plume based on the estimates of
the dye concentration at each fish position are shown in a schematic of Hammond bay, where the
black line represents the coast and the red dot represents the river mouth. Each page is labeled by
the year (10 or 11) and animal ID. The upper graph shows the sea lamprey movement on the first
night from where the sea lamprey settled in the array after release (orange dot). In the schematic
of movement, the gray arrow points the heading during the last 5 points of the track and the track
is color coded by 15 minute intervals of activity. The label of “Entered” indicates that sea
lamprey eventually entered the Ocqueoc River and the label of “Did Not Enter” indicates that the
sea lamprey never entered the Ocqueoc River. The number of bouts follows the label indicating
the number of different nights the sea lamprey was observed in the array. If the sea lamprey
entered and the bout is one, the individual entered at the end of the track shown. In some cases
entry is not captured well by the array (e.g. 10-17). The lower graph depicts the depth of the sea
lamprey during movement through time and the color of the points corresponds to the schematic
in the upper graph. The red line represents the estimated total water column depth. When the fish
is below the line it can be assumed the fish is on the bottom. ....................................................218

xviii

Figure 4.16: The 33 sea lamprey that did not encounter the river plume on the first night based
on the estimates of the dye concentration at each fish position are shown in a schematic of
Hammond bay, where the black line represents the coast and the red dot represents the river
mouth. Each page is labeled by the year (10 or 11) and animal ID. The upper graph shows the
sea lamprey movement on the first night from where the sea lamprey settled in the array after
release (orange dot). In the schematic of movement, the gray arrow points the heading during the
last 5 points of the track and the track is color coded by 15 minute intervals of activity. The label
of “Entered” indicates that sea lamprey eventually entered the Ocqueoc River and the label of
“Did Not Enter” indicates that the sea lamprey never entered the Ocqueoc River. The number of
bouts follows the label indicating the number of different nights the sea lamprey was observed in
the array. If the sea lamprey entered and the bout is one, the individual entered at the end of the
track shown. In some cases entry is not captured well by the array (e.g. 10-17). The lower graph
depicts the depth of the sea lamprey during movement through time and the color of the points
corresponds to the schematic in the upper graph. The red line represents the estimated total water
column depth. When the fish is below the line it can be assumed the fish is on the bottom. .....258
Figure 5.1: The steps of the spawning migration are analogous to the steps involved in trapping
sea lamprey (Bravener and McLaughlin 2013), although in this case the river takes the place of
the trap and the model depicts the transitions that occur when a sea lamprey enters a river in the
great lakes. The arrows refer to the probabilities of progressing from the lake to spawning habitat
or reversing through any prior step. The P refers to the probability of the event. .....................312
Figure 5.2: The distance sea lamprey travel from shore (a, dotted arrow), the distance the river
plume extends from the coast (b), the presence of other rivers on the coast and the number of
rivers encountered prior to reaching a river (c), are all important to a sea lamprey encountering a
particular river plume (Region of Coast). Following encounter with a river plume (d) sea
lamprey become available for manipulation of their habitat selection decisions including river
entry (e), or tributary entry (f). Five general manipulation scenarios are shown including the
manipulation of river plume encounter (event) (ii) or habitat selection (decision) (iii-vi). The
operator (+, - , =), indicates how the manipulation of adding an attractant (“attract”), a repellant,
or removing an attractant (“repel”), influences encounter rates at nearby rivers and entry of those
rivers. The manipulation of decisions through the addition of odorants (iii-vi), uses general terms
based on the effect desired and not specific odorants that will cause the desired effect. Encounter
is shown to change due to the order of sequential river plumes. Encounter at a river can change
due to a manipulation at a prior river but not the current river manipulated (decrease, iii;
increase, iv). It is unclear how past experience might influence entry of future rivers (iii, “?”). ......
......................................................................................................................................................313
Figure 5.3: Many river mouths have preexisting structures that may extend the distance the river
plume would otherwise reach from shore (i.e., as seen west of river mouths in Ludington, MI
(left) and Sagatuck, MI (right)). Structures of this type designed for boat traffic could present
unique solutions for influencing the sea lamprey migration and may be the only scenario that
would create a river plume extending a greater distance from shore than would otherwise occur
for that river. In each image the river plume can be faintly seen due to darker turbidity. ..........315

xix

Figure 5.4: The percentage of the entire population of adult sea lamprey in Lake Huron that
return to the Ocqueoc River to spawn from 1995 to 2015 based on mark recapture (black circles)
and the change in entry between years (blue circles). A change of entry rate of 2 represents a
doubling in the rate of entry. Vertical lines represent the times of TFM treatment to the Ocquoec
River. The change in entry rate from one year after TFM treatment to the second year shows that
the increase in entry rate between 2010 and 2011 (D) fits within the historical increase in rate (A:
1.8, B: 1.3, C: 3.1, D: 2.1). ..........................................................................................................316

xx

CHAPTER 1
AN ADAPTIVE WORKFLOW FOR STUDYING BEHAVIOR WITH TELEMETERED
OBSERVATIONS.

ABSTRACT
Ethology is now dominated by remotely sensed observations rather than visual
observations. The steps to answer ethological questions with telemetry should include question
development, data collection, data processing, and finally analysis, regardless of whether it is an
exploratory study of a rarely studied animal or a model organism. Following this structure and
reporting the steps clearly is important for ensuring high quality science and for sharing
increasingly complex analyses with other practitioners. Herein, we detail an approach for
answering proximate and ultimate questions of animal behavior with explanatory, exploratory,
and predictive analysis objectives.
INTRODUCTION

Ethology is the study of animal behavior. Mathews and Mathews (2010) defined behavior
as the ways in which an animal adjusts to and interacts with its environment (Matthews and
Matthews 2010). We extend this definition to define behavior as the regulation of bodily status
via spatial adjustments involving neuro-muscular systems operated in response to environmental
information to satisfy an internal state (or purpose for acting) (James Miller (MSU Entomology),
personal communication). This means that a proper description of a behavior would contain the
animal’s internal state (A), the local body adjustments (B), the spatial movement pattern (C), and
environmental conditions (D). This information makes it possible to sufficiently describe

1

behavior in terms of the underlying proximate mechanisms (causation and developmental
questions), and ultimate mechanisms (functional and evolutionary questions) of the behavior.
This does not mean that all four pieces aspects of the behavioral description must be known or
are equally important to answer behavioral questions, but it does mean that a complete
understanding of behavior includes all four parts. Ethology has transitioned from analyzing
visual observations of organisms to reliance on sequential observations of marked individuals, a
transition that has usually traded observation precision for the duration of time an individual is
observed (Sanderson 1966); but technology has reached a stage where telemetered datasets can
provide highly precise data over long time periods, even in aquatic environments (Cooke et al.
2013).
The technological advances have resulted in extensive datasets of spatial and bodily
movement. Although, the current proliferation of technology for capturing data and approaches
for analyzing data have muddied the water on what is actually being studied. Raising the
question of whether ethology should focus on animal displacements or if the focus should remain
on the classical study of animal behavior (ethology), with spatial movement representing a single
component. Spatial movement is only one aspect of behavior and only becomes interesting when
put into the context of the animal’s purpose for moving (objective) and the environment in which
it moves. Although disagreements occur over the semantics (Bateson 2012, Burghardt 2012), it is
usually agreed that observed movements should always be tied to their relevance in behavior.
Studies of the causes, mechanisms, and patterns of spatial movement have formally been
coalesced into a heuristic model termed movement ecology (Nathan 2008). Movement ecology
focuses on understanding behavior based on the movement signature, and for that reason it fits as
an aspect of ethology as Tinbergen (1963) describes ethology, “the biological study of behavior”

2

(Tinbergen 1963). The movement ecology model encourages consideration of the components of
behavior from a fresh perspective, but is not as useful for matching a question to a study design.
For this task I prefer to consider the three classifications of study objectives: (1) exploratory, (2)
explanatory, and (3) predictive objectives.
An exact and repeatable method to statistically classify movements into behaviors using
movement data is unrealistic, as different organism’s move at different scales and in different
ways and the measures used and data types available to provide the information vary with
technological advances (Gurarie et al. 2015, Calenge et al. Unpublished). The general work flow
however, is consistent for addressing behavior with movement data and remains independent of
the study animal or question, as behavioral questions are hierarchical in nature and can be
addressed in the order of exploratory, explanatory and predictive research objectives (Gurarie et
al. 2015, Calenge et al. Unpublished). From this hierarchy we can identify what the animal is
doing (exploratory), why the animal is doing what is it is doing (explanatory), and how an animal
will change what it is doing under different conditions (predictive). Identifying the general
process of the study and then reporting it in an organized framework can be a powerful approach
for sharing and relating methodology (Shamoun-Baranes et al. 2012, Table 1.1).
I recommend a general work flow that leads from question development to analysis
selection (Table 1.1). Regardless of whether the objectives fall within any or all exploratory,
explanatory, or predictive classifications; a strong plan will prevent failure of a practitioner to
capture the necessary information. I also highlight how this organization addresses the types of
questions falling under Tinbergen’s four pillars of ethology (Table 1.2).

3

GENERAL WORKFLOW

The study of animal behavior with movement comprises four steps including question
development, data collection, data processing, and finally analysis (Table 1.1). Analyses can be
structured into three types of objectives including exploratory, explanatory, and predictive
analyses (Gurarie et al. 2015). I use this structure to carry out research and I will expand on how
these steps should be completed.
Question Development
The way animal behavior has been recorded and defined in ethology has changed; but the
questions addressed have remained similar (Table 1.2). Tinbergen (1963) identified four main
questions to studying behavior: (1) what is the animal doing (causation), (2) what benefit does
what it is doing provide to the animal (functional), (3) how did behavior develop or change over
the lifetime of the individual (developmental), and (4) how did the behavior arise over the
history of the species (evolutionary) (Tinbergen 1963). These four questions of behavior still
resonate as a strong heuristic model with which to segregate types of questions (Bateson and
Laland 2013).
Tinbergen’s four questions can be answered via an exploratory, explanatory, and
predictive analysis hierarchy and this hierarchy informs us of an effective study design.
Causation questions fall within exploratory analysis objectives that use movement in the context
of the environment to characterize behavior. Anytime an analysis classifies a behavior based on a
movement pattern without an explicit understanding of the environment or animal objective, the
analysis is relying on major assumptions, which may be reasonable depending on the baseline
understanding of the species (Able 1991). There is no substitute to understanding the ecological
needs and general activities of the study animal; when little is known about the animal’s habits, it
4

becomes much harder to characterize patterns (Dodge et al. 2008) and to break behaviors into
sub behaviors (e.g. behavior: body care, sub behaviors: preening, washing; Shamoun-Baranes et
al. 2012). Developmental questions are usually just an extension of causal questions that seek to
understand how age and experience influence behavior (Jorge et al. 2011), seeking to understand
how the value of different behaviors fits into functional questions.
While relying on effective baseline exploratory evaluations, functional questions fall
into the explanatory analysis objectives in which the value of a defined behavior is the target.
Functional questions are often only answered from sensory manipulations in the field that are
able to alter the sensory apparatus of the animal (Barbin et al. 1998, Vrieze et al. 2010) or in the
lab where the sensory environment can be controlled instead of monitored (James et al. 2008).
Functional questions are difficult to answer absent a strong movement and environmental dataset
that can be combined to reveal movement mechanisms that rely on fixed (e.g. depth) and plastic
variables (e.g. water current). Lastly evolutionary questions fall into the explanatory and
predictive objective realm that ranges from simple testing to full model fitting. Evolutionary
questions can only be addressed after behavior is defined; the proximate value of the behavior is
elicited and can now be considered across other more complex variables (many generations).
Thus, explanatory and predictive analyses can only be reached after exploratory analyses have
been implemented; motivating our proposed hierarchy (Table 1.2).

Data Collection
Once questions are identified, the data required to answer the questions and collection
methods can be determined. Challenges to gathering information depend on the study system
and research questions. The scale of the question, size of the organism, type of environment, and
5

level of precision needed to answer the research questions, limits the types of questions that can
be address and the way questions are asked (Espinoza et al. 2011, Cagnacci et al. 2010, Cooke et
al. 2013, Meckley et al. 2014b). The greatest risk of advancing technology comes at the risk of
misusing it and ignoring major underlying assumptions about the behavior that are not addressed
by the collected data (Hebblewhite and Haydon 2010).

Data processing
The data processing stage involves: evaluating the quality of the collected data, filtering
data to match data quality with the needs of the research question, merging datasets, and
formatting data into a query-able form for use in analyses and summary statistics (ShamounBaranes et al. 2012, Meckley et al. 2014a, Dissertation Chapter 2). Remotely sensed
environmental and movement data is unlike bench-tested measurement tools (e.g., thermometer),
in that data quality varies in space and time, and the tool must be evaluated specifically in each
study until its performance is well understood. Data should be filtered if it is necessary to meet
the precision needs of the research questions (e.g., Dissertation Chapter 2, Meckley et al. 2014a).
After data are evaluated they must be merged into a single database that allows the connection to
be made between static attribute information of an individual, to dynamic trajectory data and
environmental conditions, a potentially challenging activity. Movement datasets require major
formatting efforts and careful annotation of the data is of paramount importance. When building
databases, careful consideration should be taken of the current and potential future uses, to
ensure data are accessible.

6

Analyses
Each classification of research objective has entire works dedicated to the objectives of
exploratory, (Gurarie et al. 2015), explanatory, and predictive analyses (Smouse et al. 2010) and
each category can contain a diverse number of analyses. An explosion in analysis diversity has
occurred to match the increasing mountain of diverse data (Cagnacci et al. 2010, Hebblewhite
and Haydon 2010, Smouse et al. 2010), though they all perform the task of either revealing
patterns in the movement, testing patterns in the data, or making predictions about how the
patterns might vary in different circumstances (Figure 1.1).
Exploratory approaches are designed to define behavior and specifically ask what is the
animal doing and how is the animal doing it, often including metric based, classification,
phenomenological time series, and mechanistic movement modeling (Gurarie et al. 2015). Each
approach (data type, collection method, and analysis) has different assumptions (Hebblewhite
and Haydon 2010), challenges, and value which must be evaluated prior to selection, though all
are designed to separate spatial or bodily movement patterns into distinct phases (Dodge et al.
2008, Gurarie et al. 2015). It is important to be aware that classifications vary in their specificity
to behavior depending on the level of understanding of the internal state, bodily movement, and
spatial movement (Dodge et al. 2008). For example a pattern may be classified very descriptively
based on spatial movement as spatially stopped or spatially active (Meckley et al. 2014a), but if
bodily movement is known it could be determined that the animal is swimming or not swimming
(Wilson et al. 2006); if the environment is also understood, it could be determined that the
animal is swimming with current, if the internal state is understood the pattern could be revealed
as selective tidal stream transport (McCleave and Kleckner 1982, Barbin et al. 1998), or if the
animal condition is monitored actively, movement could be related to condition in questions of

7

energetics or respiration (Wilson et al. 2006). When studies describe behavior at the highest level
of specificity without all the pieces of information they are making assumptions that could be
unsupported or potentially supported from other research. Exploratory analyses separate the
bodily and/or spatial movement signatures into patterns, which can be fed into explanatory and
predictive analyses to address higher level questions to describe behavior. Exploratory studies
also include mechanistic studies of how the behavior is performed in terms of mechanical ability,
energetics, and physiology, described by Nathan et al. (2008), as the biomechanical paradigm.
Explanatory analyses address the question of why the animal is doing what it is doing,
while predictive questions ask, how I anticipate the system would change if there is a change in a
covariate (Bowlin et al. 2010, Gurarie et al. 2015). In a movement context the purpose of
explanatory analyses is to understand how a set of behaviors relates to a set of environmental
covariates or a measure of success; while, predictive analyses determine how variables interact,
and how behaviors might be anticipated to change given a change in one of the covariates.
Explanatory and predictive analyses could address both functional and evolutionary questions
but they differ in the way they implement analyses and ask questions.
Explanatory analyses include correlative approaches and modeling approaches.
Predictive studies require full model fitting and most predictive analysis address questions that
take the form of: how will changes in a specific environmental variable influence the ecology (ie.
fitness) of a species given the behaviors that comprise the species life cycle as they are currently
described (e.g., Bowlin et al. 2010). Approaches are now capable of both explanatory and
predictive outputs from the same complex modeling framework (Goodwin et al. 2006). Classical
model fitting of variations of a random walk model as a null hypothesis are still used to explain
how organisms reach targets (Goodwin et al. 2006; Gurarie and Ovaskainen 2011), referred to as

8

the “random paradigm” by Nathan et al. (2008). Explanatory and predictive objectives are also
often used to understand animal decision making and determine why animals transition between
discrete behaviors, described as the cognitive paradigm (Nathan et al. 2008). Predictive models
are the most common for describing behavior in light of evolutionary contexts, although studies
of this nature are still somewhat rare and fall into the optimality paradigm as Nathan et al. (2008)
described it.
APPROACH AND IMPLEMENTATION

This framework is effective for matching data collection, data processing, and analyses to
research questions and helps avoid pitfalls of mismatching any of these components to a research
question. It also is valuable for sharing a reproducible method as was done by Shamoun-Baranes
et al. (2012). The current proliferation of perspectives on behavior are valuable (e.g.
psychology, movement ecology, etc.); however, it is important that work done in these realms is
tied back into the overarching behavior. Some of the measures used to describe behavior are
actually an output of behavior (e.g., speed or tuning), rather than a description of what the animal
is doing, where it is doing it, and why it is doing it, an important argument that has been recently
presented (Bateson 2012, Burghardt 2012), and should garner additional attention. Below I
describe how this framework has been used to answer pertinent fundamental biological questions
and management questions regarding the sea lamprey migration in the Great Lakes.
The sea lamprey lake migration in the Great Lakes has only begun to be studied in the
last 10 years (i.e., since 2004, e.g. Vrieze et al. 2011), initially involving both mark-recapture
studies and trajectory data from individuals monitored by manual underwater acoustic telemetry
(Vrieze et al. 2010, Vrieze et al. 2011). The research objectives have focused on how sea
lamprey encounter and enter rivers on the coast, and what role larval odor, a known migration
9

signal, has on the ability of lamprey to find a river or the decision to enter the river for the
overarching purpose of improving management of invasive sea lamprey in the upper great lakes
I have applied the research framework found in Table 1.1 to continue to answer important
questions detailing the way sea lamprey return from a large lake to enter streams to spawn.
Previous research has represented primarily exploratory studies with small field experimental
manipulation experiments (Vrieze et al. 2010, Meckley et al. 2014b). The two manual tracking
studies reported that sea lamprey transition from extensive movements typified by faster
movement and lower turning rates while vertically casting through the water column from
surface to the bottom when between rivers and intensive search typified by frequent stopping and
turning with in the region of the river plume (Vrieze et al. 2011, Meckley et al. 2014b). In
addition it was indicated that sea lamprey may rely on the physical edge of the coast as sea
lamprey rebounded along the coast horizontally casting and probing the shoreline, proposed as a
river mouth localization behavior in the presence of a dynamic river plume that may provide
poor directional information (Meckley et al. 2014b). Rebounding is a pattern of movement that
includes progress along the coast, while switching between periods of angling towards the coast
until reaching very shallow water followed by periods of angling away from the coast. Sea
lamprey clearly rely heavily on olfaction as Vrieze at al. (2010) observed few nasally occluded
sea lamprey successfully entering a local river and Meckley et al. (2014) observed an increase in
the amount of time sea lamprey spent in the presence of a partial larval odor, but no apparent
increase in habitat selection (river entry).
Two manual tracking studies have provided a baseline for making predictions about how
sea lamprey migration, though they each come with some major complications due to the use of
manual telemetry methods. One study did not have a reliable estimate of accuracy as they relied
10

on the bearing of the transmitter implanted fish from the boat and used the intersection of
multiple measurements to estimate position, accuracy in this case is influenced by fish ground
speed and their testing method was based on a fixed tag position (Vrieze et al. 2010). Bearing
based positioning is best with two boats taking simultaneous measurements (Holland et al.
1992). The second study did have a reportable accuracy but the accuracy was poor and prevented
some trajectory type data (e.g., turn angles) to be evaluated (mean error 71.2 m, Meckley et al.
2014). In both cases, motor boats were used to follow the fish and could have influenced
behavior of a vertically active fish as has been observed in other fishes (Draštík and Kubečka
2005).
Step 1: Question Development
The exploratory manual tracking studies provided key insights that led to question
development. I identified eight overarching questions. The invasive sea lamprey attaches to and
parasitizes large fishes that may displace sea lamprey outside of their control, leading us to
suspect that they are capable of orientation and migration to the coast following displacement
and to ask, (1) how do sea lamprey orient and return to the coast when displaced offshore
(Chapter 3)? Once arriving at the coast it is known that sea lamprey move parallel to the coast
but I wanted to evaluate this behavior more finely leading us to ask, (2) what aspects of their
extensive search influences what rivers they are likely to encounter (Chapter 4)? Once
encountering river water I wanted to know, (3) where does the decision to enter the river occur,
or maybe more specifically at what point does the search and assessment process for entry
become irreversible, (4) is larval sea lamprey odor important to navigation or only the decision to
enter a river, and (5) how do sea lamprey navigate within a plume, what senses and what
environmental information sources are likely used (Chapter 4)? These questions address both
11

fundamental biological questions and build to answer invasive sea lamprey management
questions such as, (6) why do some rivers garner stable reproduction of sea lamprey requiring
regular pesticide treatment, (7) what recommendations would I provide for manipulating sea
lamprey with a synthetic larval odor if it were available, and (8) what recommendations do I
suggest regarding the current way pesticide treatment is applied, which effectively turns down
the larval odor signal (Chapter 5)?
Step 2: Data Collection
I used an underwater acoustic position system from VEMCO, which is a new underwater
fine-scale passive monitoring system. The system provides two-dimensional positions derived
from raw detection data using proprietary hyperbolic positioning algorithms that estimate
positions from the time-difference-of-arrival (TDOA) of a coded signal at three or more
receivers in the VPS array (Smith 2013). Each detection record by a receiver consists of a
transmission code that is unique to a tag, the time of signal reception, and, depending on the tag
configuration, an environmental variable measured by the tag (e.g., pressure, temperature). As
with all TDOA-based processes, time synchronization of the autonomous receivers is critical.
The VPS system makes use of synchronization transmitters (sync tags) placed at stationary
locations throughout the array to account for time drift in individual receiver clocks (Andrews et
al. 2011, Smith 2013, Chapter 2) . Like other multi-receiver positioning systems (White and
Garrott 1990), the VPS provides a weighted average location from every combination of three
receivers that detected each tag transmission (hereafter referred to as a position ‘fix’), and an
associated PPE termed horizontal position error (HPE) (Espinoza et al. 2011, Smith 2013).

12

Step 3: Data Processing
The first data processing step entails putting data into a raw database that can be easily
accessed in the future. Animal data should be paired with environmental data at this time if any
variables are known. This step was tedious in the case of our sea lamprey data as a
hydrodynamic model was developed for several years of data (Chapter 4), so that at each fish
position I know the water temperature, velocity fields, and whether the sea lamprey was
encountering river water. In addition the animal and environmental datasets had to be matched in
MATLAB, the program used to develop the hydrodynamic models, and then transferred for
further analysis in a different program, R. At this time all raw data is stored in stable “.robj”
files, in which variables such as time are pre-formatted and the risks of storing time data in
“.csv” files accessed by Excel, can be avoided.
The second step entails defining data quality objects based on the research question and
future analysis methods followed by an assessment of the unfiltered data quality before finally
the data should be filtered to remove unacceptably erroneous positions. I developed a four-step
filtering approach which incorporates data-filtering principles developed for GPS tracking of
terrestrial animals. HPE was evaluated for its effectiveness to remove uncertain fish positions
acquired from the VPS system. The use of position precision estimates that reflect the confidence
in the positioning process should be considered prior to the use of biological filters that rely on a
priori expectations of the subject’s movement capacities and tendencies. Position confidence
goals should be determined based upon the needs of the research questions and analysis
requirements versus arbitrary selection, in which filters of previous studies are adopted. Data
filtering with this approach ensures that data quality is sufficient for the selected analyses and
presents the opportunity to adjust or identify a different analysis in the event that the requisite

13

precision was not attained. Ignoring these steps puts a practitioner at risk of reporting errant
findings (Meckley et al. 2014a, Chapter 2).Once data is prepared for a specific analysis it should
again be put into its own stable database for question and analysis specific analyses. Annotation
of these datasets is very important to avoid improper use by future practitioners.
Step 4: Analyses
At this point our research focused on exploratory and explanatory analyses. First, I use
exploratory approaches that characterize patterns in the movement of sea lamprey including
changes in sinuosity, heading, and ground speed (Chapter 3, Chapter 4). Analyses focused on
explaining patterns before and after changes in behavior with simple tests (Chapter 3) and more
complex mixed modeling approaches that take into account multiple observations of the same
individual and how environmental features associate with movement patterns (Chapter 4).
Through these approaches I can characterized the patterns and began to determine the behaviors.
Once different periods of the migration were characterized I moved past the exploratory analyses
to characterize what the discrete segments represent in the migration and how they functionally
influence how sea lamprey reach key points of the migration. In future work, I can apply the
exploratory and explanatory findings to develop predictive models that can predict things like
what aspects of rivers make them likely to be encountered and entered by sea lamprey.
CONCLUSION

This hierarchical approach to studying the sea lamprey migration through the monitoring
of the movement of sea lamprey and the underlying environmental context under which they
were moving has allowed us to progress from inaccurate preliminary studies that revealed very
general aspects of the sea lamprey migration to now being able to discuss fine scale behavior and

14

expand as well as correct our understanding of how sea lamprey migrate in the Great Lakes. I
have just scratched the surface of these collected datasets and in the future I expect analyses with
a predictive nature to become important.

15

APPENDIX

16

Table 1.1: Studies of animal behavior that focus on telemetry data should follow this general
framework and report on each component: question development data collection, data
processing, and analysis. Using the framework ensures that at the start of the study there is clear
consideration of what data is necessary to properly answer the question and describe the behavior
in an exploratory, explanatory or predictive context.
Question
Development
Identify
potential
questions
and
determine
the data
needed.

Data
Collection

Data
Processing

Exploratory
Analysis

Explanatory
Analysis
Look for
patterns
between
movement
and
environmental
data and/or
between body
and spatial
movement.

Predictive
Analysis
If the
environment
is modified
in a
particular
way, what
change in
behaviors
would we
anticipate?
Full Model
Fitting

Collect
environmental,
Assess data
body
quality needs
movement,
and actual data
and spatial
quality.
movement
data.

Classify and
merge spatial
patterns
and/or body
movement
patterns into
behaviors
(e.g., decision
tree)

Merge animal
data with
environmental
data.

Characterize
differences in
classified
phases

Model testing
and model
fitting

Build
Database

See Gurarie et
al. 2015 for
complete
description

Pattern
Testing

Filter data.

17

Table 1.2: The components of Ethology.
Ethology: The biological study of behavior.
Behavior: The regulation of bodily status via spatial adjustments involving neuromuscular systems, operated in response to environmental information to satisfy an internal
state (or purpose for acting).
Components of behavioral description:
Bodily Movement: The local adjustments of an organism’s neuro-muscular systems that
are part of a specific behavior but may or may not result in spatial movement. It
represents the mechanical component of behavior.
Spatial movement: The whole body adjustment as the result of bodily movement or
environmental propulsion. The spatial position as result of behavior.
Internal State: The underlying reason for the behavior or why the animal is doing what it
is doing. The underlying function and internal component of behavior.
Environmental conditions: The external information that the internal state uses to update
and subsequently inform the appropriate bodily movement.
Analysis

Aspects of Behavioral Study (Tinbergen’s 4 Questions)
Proximate: Questions that define behaviors.
Causation – “What is the animal doing,” is often addressed through
exploratory analyses and usually requires bodily or spatial movement
and at least limited environmental data.

Explanatory

Predictive

Exploratory

Developmental – “How does what the animal is doing change with
maturity, experience and environmental conditions?” is often
addressed with similar data and analyses as causation, though include
some explanatory and rarely predictive approaches.
Ultimate: Questions that aim to understand the impact of behaviors.
Functional (adaptive significance) – “What benefit do classified
behaviors provide to the organism,” requires an understanding of
causation, and usually requires spatial movement, bodily movement
and environmental information. These questions are addressed
through explanatory and predictive analyses.
Evolutionary – “How did the behavior form in a species, in light of the
environment and under intraspecific and interspecific interaction
pressures” is the most complex type of questions that require a strong
understanding of causation, developmental, and functional aspects of
behavior, and require explanatory and predictive analyses.
18

Initial Information: No knowledge of behavior

Analysis Objectives: Exploratory

Analysis Form:

Deep understanding of behavior

Explanatory

Descriptive Statistics

Tests

Predictive

Modeling

Figure 1.1: Analysis objectives aim to identify what is the animal doing (exploratory), why the
animal is doing what it is doing (explanatory), and can we anticipate how it will change what it
will do under different conditions (predictive)? The amount known about the animal often plays
a role in where researchers aim study objectives initially and we identify the usual process taken,
but the best behavioral studies aim to reach explanatory and predictive answers regardless of the
initial understanding of behavior.

19

REFERENCES

20

REFERENCES

Able, K. P. (1991). "Common Themes and Variations in Animal Orientation Systems."
American Zoologist 31(1): 157-167.
Andrews, K., N. Tolimieri, G. Williams, J. Samhouri, C. Harvey and P. Levin (2011).
"Comparison of fine-scale acoustic monitoring systems using home range size of a demersal
fish." Mar Biol 158(10): 2377 - 2387.
Barbin, G. P., S. J. Parker and J. D. McCleave (1998). "Olfactory clues play a critical role in the
estuarine migration of silver-phase American eels." Environmental Biology of Fishes 53(3): 283291.
Bateson, P. (2012). "Behavioural Biology: The Past and a Future." Ethology 118(3): 216-221.
Bateson, P. and K. N. Laland (2013). "Tinbergen's four questions: an appreciation and an
update." Trends in Ecology & Evolution 28(12): 712-718.
Bowlin, M. S., I.-A. Bisson, J. Shamoun-Baranes, J. D. Reichard, N. Sapir, P. P. Marra, T. H.
Kunz, D. S. Wilcove, A. Hedenström, C. G. Guglielmo, S. Åkesson, M. Ramenofsky and M.
Wikelski (2010). "Grand Challenges in Migration Biology." Integrative and Comparative
Biology.
Burghardt, G. M. (2012). "A Behavioral Biology for the Future." Ethology 118(3): 222-225.
Cagnacci, F., L. Boitani, R. Powell and M. Boyce (2010). "Animal ecology meets GPS-based
radiotelemetry: a perfect storm of opportunities and challenges." Philos Trans R Soc London
[Biol] 365(1550): 2157 - 2162.
Calenge, C., M. Basille, M. S. Boyce, S. Lele, M. Panzacchi, O. Strand and B. V. Moorter
(Unpublished). "How to identify preferred habitats? Using factor analyses and modelling
approaches for exploration, confirmation and prediction in habitat selection studies at the
population scale. ." Journal of Animal Ecology Special Issue.
Cooke, S., J. Midwood, J. Thiem, P. Klimley, M. Lucas, E. Thorstad, J. Eiler, C. Holbrook and
B. Ebner (2013). "Tracking animals in freshwater with electronic tags: past, present and future."
Anim Biotelem 1: 5.

21

Dodge, S., R. Weibel and A.-K. Lautenschütz (2008). "Towards a taxonomy of movement
patterns." Information visualization 7(3-4): 240-252.
Draštík, V. and J. Kubečka (2005). "Fish avoidance of acoustic survey boat in shallow waters."
Fisheries Research 72(2–3): 219-228.
Espinoza, M., T. Farrugia, D. Webber, F. Smith and C. Lowe (2011). "Testing a new acoustic
telemetry technique to quantify long-term, fine-scale movements of aquatic animals." Fish Res
108(2): 364 - 371.
Goodwin, R. A., J. M. Nestler, J. J. Anderson, L. J. Weber and D. P. Loucks (2006).
"Forecasting 3-D fish movement behavior using a Eulerian–Lagrangian–agent method (ELAM)."
Ecological Modelling 192(1): 197-223.
Gurarie, E., C. Bracis, M. Delgado, T. D. Meckley, I. Kojola and C. M. Wagner (Unpublished).
"A comparison of methods and practical guide to the behavioral analysis of animal movements."
Journal of Animal Ecology Special Issue.
Gurarie, E. and O. Ovaskainen (2011). "Characteristic spatial and temporal scales unify models
of animal movement." The American Naturalist 178(1): 113-123.
Hebblewhite, M. and D. T. Haydon (2010). "Distinguishing technology from biology: a critical
review of the use of GPS telemetry data in ecology." Philosophical Transactions of the Royal
Society of London B: Biological Sciences 365(1550): 2303-2312.
Holland, K., C. Lowe, J. Peterson and A. Gill (1992). "Tracking coastal sharks with small boats:
hammerhead shark pups as a case study." Marine and Freshwater Research 43(1): 61-66.
James, N. C., P. D. Cowley, A. K. Whitfield and H. Kaiser (2008). "Choice chamber
experiments to test the attraction of postflexion Rhabdosargus holubi larvae to water of estuarine
and riverine origin." Estuarine, Coastal and Shelf Science 77(1): 143-149.
Jorge, P. E., D. Sowter and P. A. Marques (2011). "Differential annual movement patterns in a
migratory species: effects of experience and sexual maturation." PloS one 6(7): e22433.
Matthews, R. W. and J. R. Matthews (2010). Insect Behavior. Netherlands, Springer
Netherlands.

22

McCleave, J. D. and R. C. Kleckner (1982). "Selective tidal stream transport in the estuarine
migration of glass eels of the American eel (Anguilla rostrata)." Journal du Conseil 40(3): 262271.
Meckley, T., C. Holbrook, C. Wagner and T. Binder (2014a). "An approach for filtering
hyperbolically positioned underwater acoustic telemetry data with position precision estimates."
Animal Biotelemetry 2(1): 7.
Meckley, T., C. Wagner and E. Gurarie (2014b). "Coastal movements of migrating sea lamprey
(Petromyzon marinus) in response to a partial pheromone added to river water: implications for
management of invasive populations." Can J Fish Aquat Sci 71(4): 533 - 544.
Nathan, R., W. M. Getz, E. Revilla, M. Holyoak, R. Kadmon, D. Saltz and P. E. Smouse (2008).
"A movement ecology paradigm for unifying organismal movement research." Proceedings of
the National Academy of Sciences 105(49): 19052-19059.
Sanderson, G. C. (1966). "The Study of Mammal Movements: A Review." The Journal of
Wildlife Management 30(1): 215-235.
Shamoun-Baranes, J., R. Bom, E. E. van Loon, B. J. Ens, K. Oosterbeek and W. Bouten (2012).
"From sensor data to animal behaviour: an oystercatcher example." PloS one 7(5): e37997.
Smith, F. (2013). "Understanding HPE in the VEMCO positioning system (VPS)." Halifax, NS:
VEMCO.
Smouse, P. E., S. Focardi, P. R. Moorcroft, J. G. Kie, J. D. Forester and J. M. Morales (2010).
"Stochastic modelling of animal movement." Philosophical Transactions of the Royal Society B:
Biological Sciences 365(1550): 2201-2211.
Tinbergen, N. (1963). "On aims and methods of Ethology." Zeitschrift für Tierpsychologie
20(4): 410-433.
Vrieze, L., R. Bergstedt and P. Sorensen (2011). "Olfactory-mediated stream-finding behavior of
migratory adult sea lamprey (Petromyzon marinues)." Can J Fish Aquat Sci 68: 523 - 533.
Vrieze, L. A., R. Bjerselius and P. W. Sorensen (2010). "Importance of the olfactory sense to
migratory sea lampreys Petromyzon marinus seeking riverine spawning habitat." Journal of Fish
Biology 76(4): 949-964.

23

White, G. and R. Garrott (1990). Waltham, MA: Academic Press Inc.
Wilson, R. P., C. R. White, F. Quintana, L. G. Halsey, N. Liebsch, G. R. Martin and P. J. Butler
(2006). "Moving towards acceleration for estimates of activity‐specific metabolic rate in free‐
living animals: the case of the cormorant." Journal of Animal Ecology 75(5): 1081-1090.

24

CHAPTER 2
A PROCESS FOR FILTERING HYPERBOLICALLY POSITIONED UNDERWATER
TELMETRY DATA WITH HORIZONATAL POSITIONG ERROR (HPE)

Meckley T.D., Holbrook C.,Wagner C.M., Binder T. 2014. An approach for filtering
hyperbolically positioned underwater acoustic telemetry data with position precision
estimates. Animal Biotelmetry 2:7. (doi:10.1186/2050-3385-2-7).
ABSTRACT
Background: Telemetry systems that estimate animal positions with hyperbolic
positioning algorithms also provide a technology-specific estimate of position precision (e.g.,
horizontal position error (HPE) for the VEMCO positioning system). Position precision
estimates (e.g., dilution of precision for a global positioning system (GPS)) have been used
extensively to identify and remove positions with unacceptable measurement error in studies of
terrestrial and surfacing aquatic animals such as turtles and seals. Few underwater acoustic
telemetry studies report using position precision estimates to filter data in accordance with
explicit data quality objectives because the relationship between the precision estimate and
measurement error is not understood or not evaluated. A four-step filtering approach which
incorporates data-filtering principles developed for GPS tracking of terrestrial animals is
demonstrated. HPE was evaluated for its effectiveness to remove uncertain fish positions
acquired from a new underwater fine-scale passive acoustic monitoring system. Results: Four
filtering objectives were identified based on the need for three sequential future analyses and
four data quality criteria were developed for evaluating the performance of individual filters
(step 1). The unfiltered, baseline position confidence from known-position test tags was
considered to determine if filtering was necessary (step 2). An HPE filter cutoff of 8 was selected
to meet the four criteria (step 3), and it was determined that one analysis may need to be adjusted
25

for use with this dataset. The data quality objectives, criteria, and filter selection rationale were
reported (step 4). Conclusions: The use of position precision estimates that reflect the confidence
in the positioning process should be considered prior to the use of biological filters that rely on a
priori expectations of the subject’s movement capacities and tendencies. Position confidence
goals should be determined based upon the needs of the research questions and analysis
requirements versus arbitrary selection, in which filters of previous studies are adopted. Data
filtering with this approach ensures that data quality is sufficient for the selected analyses and
presents the opportunity to adjust or identify a different analysis in the event that the requisite
precision was not attained. Ignoring these steps puts a practitioner at risk of reporting errant
findings.
INTRODUCTION
Background
Advances in animal-tracking systems have made high-resolution, high-density position
data increasingly available, affording the scientific community a substantial opportunity to finely
examine the habits of difficult-to-detect individuals and species [1-4]. Any technology that
assigns spatiotemporal positions to an animal produces values that contain varying degrees of
measurement error. Determining an appropriate method to detect and remove positions with
excessive error from datasets presents a real challenge to researchers as over-filtering (i.e., the
omission of ‘good positions’) is a significant source of error in animal movement studies [5-7].
Each time location data are filtered, researchers must also be cognizant of the tradeoff between
increasing confidence in position accuracy and the possibility of introducing systematic biases
[8]. A useful technique must therefore discern between measurement error and true variation in
animal movement. Two common approaches involve the use of biological filters, which equate
26

to detecting violations of an animal’s expected maximum movement capacity (e.g., maximum
observed velocity) or known habits (e.g., observing a fish on land), or a position precision
estimate (PPE) in which each measured position has an associated position-specific estimate of
measurement error or confidence. We outline a widely applicable approach to use a PPE to
remove positions with unacceptable measurement error, and provide an example of a specific
application to a telemetry dataset collected from an underwater acoustic positioning system, the
VEMCO Positioning System (hereafter VPS; VEMCO Division of AMIRIX Systems; Halifax,
Nova Scotia).

The VPS system
VEMCO provides two-dimensional positions derived from raw detection data using
proprietary hyperbolic positioning algorithms that estimate positions from the time-differenceof-arrival (TDOA) of a coded signal at three or more receivers in the VPS array [9]. Each
detection record in a receiver consists of a transmission code that is unique to a tag, the time of
signal reception, and, depending on the tag configuration, an environmental variable measured
by the tag (e.g., pressure, temperature, etc.). As with all TDOA-based processes, time
synchronization of the autonomous receivers is critical. The VPS system makes use of
synchronization transmitters (synch-tags) placed at stationary locations throughout the array to
account for time drift in individual receiver clocks [9,10] (Figure 2.1). Like other multi-receiver
positioning systems [11], the VPS provides a weighted average location from every combination
of three receivers that detected each tag transmission (hereafter referred to as a position ‘fix’),
and an associated PPE termed horizontal position error (HPE) [9,12].

27

Filtering spatial data with HPE
HPE is assigned by VEMCO to positions for both synch-tags and animal-implanted tags,
but practitioners must perform auxiliary testing to be able to discern when and how HPE can be
used to filter data. HPE is a unitless estimate of position precision based on the relationship
between theoretical position error sensitivities and observed measurement errors (measured
horizontal position error in meters, HPEM) for synchronization tags [9]. If the relation between
HPE and measurement error developed for synch-tags is representative of the relation for animal
implanted tags, then the dimensionless HPE value can be used as a confidence estimate for each
fish position.
Choosing an appropriate HPE cutoff to remove inaccurate fixes is the primary challenge
in filtering VPS data, but few studies have provided an exposition of the filtering approach [13].
Studies either used a conservatively high HPE cutoff to remove the major problematic positions
but retain most data (e.g., for an HPE number of 20, 83% retention [14]), or employed a lower
HPE cutoff for the study and sacrificed large amounts of data (e.g., for an HPE number of 10,
58% retention [15]). In other cases, HPE was used but the level of data reduction was not
reported (for an HPE number of 15 [12]), no filtering occurred and the HPE estimated for synchtags was assumed to characterize the precision of all animal positions in the array [10], or the
authors referenced an example of the accuracy attained from a similarly configured array used in
an unrelated study [16]. Despite an extensive literature on PPE estimates [5-8], a consistent
framework does not exist for developing and applying an HPE-based filter to positional
telemetry data, and many studies using HPE do not clearly report the selection criteria for the
cutoff other than reference to prior studies. Confusion exists among practitioners over what

28

filtering steps are broadly applicable to all studies and what components should be independently
derived due to unique research questions.
We propose a generalizable four-step process for filtering VPS spatial datasets via HPE
that accommodates the specific data quality objectives of each analysis. The steps include: 1)
establishing position confidence goals (maximum tolerable errors) contingent on the scientific
question and the chosen data analysis technique; 2) evaluating the unfiltered, baseline position
confidence from known-position test tags; 3) selecting a filter cutoff and assessing if data
reductions have introduced biases (a-c); and, 4) reporting the data filtering criteria and process in
the final manuscript or report. The details that comprise the four steps are not interchangeable for
all studies but completing each step will improve the rationale and transparency of data filtering
for many telemetry studies. Importantly, the process does not set rigid standards for data
inclusion; rather, it requires the assignment of standards that are specific and defensible for each
project.
To demonstrate our approach, an extensive data set was derived from a 41-receiver array
placed in the waters of Lake Huron (Michigan, USA) near the outlet of the Ocqueoc River (45°
29ʹ 24.53ʺ N, 84° 4ʹ 23.45ʺ W). The array had an internal area of 0.92 km2 and a maximum
coverage of 8.22 km2 in waters that ranged from 0.2 to 20.8 m deep. The dataset included
recorded positions taken from stationary tags (9 synch-tags and two additional stationary position
tags), three mobile tag tests (submerged tags dragged behind a slow-moving boat with
continuous GPS recording), and animal location data recorded for 76 free-swimming adult
female sea lampreys (Petromyzon marinus) with surgically-implanted tags (V9P-2H, 15 to 45
second transmission rate). With this dataset we demonstrate a framework for filtering VPS data

29

with HPE, provide a comparison to current filtering approaches used by other VPS studies, and
contrast the advantages of PPE based filters and biological filters.

METHODS
A telemetry system (VPS, VEMCO) composed of 41 receivers (VR2W) in diamond
formations with between-receiver spacing ranging from 125 m to 250 m covering 2 km2 was
used to position a stationary tag at two locations, a mobile tag, and 76 tagged sea lamprey. Sea
lamprey were implanted with acoustic transmitters (VEMCO model V9P-2H, 69 kHz, 150 db, 47
mm (length), 3.5 g (dry weight), random interval: 15 to 45 s, burst length: 3.54 s or 3.26 s,
between 09 April 2010 and 09 July 2010 (Figure 2.1). The depth of the site ranged from 0 to 5.8
m within the array but sea lamprey were positioned at greater depths on the fringes of the array.
Nine synchronization transmitters (VEMCO model V16-2H, 69 kHz) were deployed in
stationary positions through the array. Compared to fish tags, synch-tags emitted a more
powerful signal (160 db) less frequently (500 to 700 s). The performance of the VPS array was
tested by comparing the position estimates of one stationary V9P-2H transmitter (stationary
testing) at two locations (Location 1: 17 June 2010 19:43:20 to 01 July 2010 14:34:36, Location
2: 01 July 2010 14:52:47 to 08 July 2010 17:03:20) and a slow moving V9P-2H transmitter
(mobile testing) inside the body cavity of a sea lamprey that was recently deceased, located 1 m
off the bottom and below a floating boat powered by an electric motor with a GPS (Trimble Geo
XH) mounted directly above the tag monitoring the tags’ true position. Due to a GPS
malfunction we did not have GPS measured positions for the stationary tests and were forced to
use position averages for comparison. This is not an ideal solution. The GPS and tag were
maintained in alignment with a down rigger and there was no visual layback of the tag. The
mobile test was comprised of three transects through the array starting near the coast and ending
30

outside of the internal array area on 06 July 2010 at an average speed of 0.60 ms−1 (range: 0.25
to 1.15 ms−1) to mimic sea lamprey ground speed: (i) Boat Path 1, 13:12:59 to 13:57:24 (distance
934 m); (ii) Boat Path 2, from 14:22:19 to 15:10:08 (distance 754 m); and (iii) Boat Path 3, from
15:35:08 to 15:59:42 (distance 309 m) (Figure 2.1; left to right). The GPS positions were postprocessed with three local reference stations that have one second reference intervals (CHB5,
MIAL, MIMC) and the positions had an estimated accuracy of 0.36 m, 0.29 m, and 1.03 m
(mean, min., max.). In our analysis, we assume that these GPS positions represent the true
mobile tag position. We were not able to synchronize the acoustic receivers to the GPS clock at
the time of tests, so we applied a constant offset to the timestamps of GPS positions until error
between the GPS track and mobile tag track (derived from VPS) was minimized. Error was
minimized using the Optim function in R (Additional file 1), which used the Broyden–Fletcher–
Goldfarb–Shanno method [28]. The test tag was identical to the tag inserted into female sea
lamprey (V9P-2H, 69 kHz, 150 db, 47 mm (length), 3.5 g (dry weight)), which were monitored
as they approached, entered or bypassed the Ocqueoc River mouth (45.490278°, −84.072931°) in
Hammond Bay (Lake Huron), where lake water temperatures ranged from 6.0 to 19.7°C
throughout the study (April 9, 2010 to June 7, 2010). The use of sea lampreys was approved by
the Michigan State University Institutional Animal Use and Care Committee via animal use
permit 02/10-020-00.

Error evaluation based on test data
To evaluate the baseline array accuracy the position yield (number of positions
observed/maximum possible number of positions) based on transmission rate, the proportions of
data that were below 6 m error and the number of positions with greater than 15 m error were
enumerated for stationary and mobile test tags. The position yield was calculated by determining
31

the total number of locations positioned by the VPS versus the number of acoustic tag signals
emitted in the time period (maximum transmission number). The true number of fixes can be
readily counted. To estimate position yield, the average time for one complete transmission cycle
must be determined by adding the average programmed transmission interval between the start
and end of a transmission plus the length of the transmission (burst length). The V9P-2H tag had
a pre-set transmission interval that ranged from 15 to 45 seconds (average: 30 seconds) and a
burst length that alternated between 3.26 and 3.54 seconds. The burst length was always
composed of a series of eight pings, but varied because two different alternating signals were
transmitted from these tags as one provided a depth code and a short tag identifier (two digit) and
the other provided a longer tag identifier (five digit) that is more robust to individual
misidentification. The resulting average length for one transmission cycle was 33.4 seconds. The
time between transmissions is readily available; however, the burst length should be obtained
prior to performing this calculation. Second, the total transmission period must be divided by the
length of one transmission cycle to determine the maximum transmission number. Average error
was calculated and the presence of positions with unacceptably large error values (>15 m error)
were enumerated for all three tests (two stationary and one mobile).

HPE evaluation with test data
The average error (criteria 1) and number of positions with large error were estimated
(>15 m, criteria 2) for all HPE cutoff possibilities (3 to 15) to evaluate the first two criteria
(Table 2.1). To evaluate if the HPE value provided for fish tags and calculated by VEMCO using
synch-tag detections was a representative estimate of locational error for fish tags, we calculated
2DRMS for each stationary test tag by first calculating the Euclidian distance between each
individual test position and the median or best estimate of the ‘true’ test tag location. The HPE
32

calculated by VEMCO is scaled to an approximately 1:1 relationship between HPE and
measured error in m for synch-tags [9]. 2DRMS can be used to compare the relationship to fish
tags and is reported in the base line evaluation of the dataset. The 2DRMS for the median value
is actually a measure of precision [8], but it was also the best available estimate of accuracy. The
2DRMS linear model was calculated by first binning all data by one unit HPE increments and an
average of the HPE within the bin was calculated to represent the bin. The error in the x direction
(Xe) and y direction (Ye) was estimated for each location within a bin. The 2DRMS was then
estimated for the 95% confidence interval using Xe and Ye within each bin and a line was fit to



the 2DRMS data 2DRMS Error = 2*

SD(Xe)

2



+ SD(Ye) 2  . A linear model fit to those

2DRMS values for each HPE (2DRMS line) of interest (3 to 15) was used to predict the proper
HPE filter for 95% confidence about a given target error value. The 2DRMS regression
developed for the mobile test was not considered due to the small number of recorded
transmissions (138). The 2DRMS model was used to estimate the HPE at which there was 95%
confidence that positions had less than 6 m accuracy (criteria 3). A range of HPE cutoffs (3 to
15) for removing unacceptably erroneous positions (>6 m error) while avoiding the loss of
acceptable positions (<6 m error) were considered based on the percentage of incorrectly
retained (of all retained) and incorrectly rejected positions (percentage of acceptable positions)
(Table 2.2). The goal for the incorrectly retained percentage of all positions was >99% and
incorrectly rejected of all positions was >95%, as the loss of data was considered more
acceptable than allowing the erroneous positions to remain. This was an arbitrary proportion,
though the loss of positions when evaluating trajectory data can quickly become problematic and
retaining unacceptably erroneous positions could result in incorrect habitat assignment, so we

33

selected high target proportions. The variation in temporal gaps was not considered but could be
very important to a study.

Filtering fish data
We examined all fish data classified by HPE in a schematic to assess any spatial pattern
in HPE and to identify areas of poor array coverage. For each candidate HPE cutoff, we
calculated the amount of data rejected inside and outside of the array area as defined as a
polygon drawn around the outer most receivers. The post-filtering spatial coverage of the array
was also tabulated for all candidate HPE values with the expectation that locations outside of the
array are likely to have an increased HPE and lower accuracy [12] with increasing distance from
the array periphery. Coverage was calculated by drawing a polygon around the outer most fish
positions retained after filtering, and calculating the area (km2). Lastly, the evenness of filtered
positions, referring to how grouped filtered positions were in the array, was considered for each
individual fish to determine if specific subjects were more prone to being rejected. We visually
assessed whether these positions were associated with any particular four-receiver diamond or if
they were missed throughout the array, occurring in multiple four-receiver diamonds.

RESULTS

We demonstrate our approach by filtering the example dataset according to the data
quality objectives of the study.

Step 1: Establishing data quality objectives
We adopted four data quality objectives. Data quality objectives are explicit assignments
of acceptable errors in precision that derive from the requirements of the data analysis technique.

34

Any defensible objective could have been adopted to fulfill this step because the details are
study-specific and could be simplified to a single criterion.
Our four objectives included the ability to detect changes in trajectories (e.g., ground
speed, turn angle, etc.), perform basic behavioral assignments (e.g., is the animal moving or
stationary?), assign fish positions to habitat types (point data), and avoid the loss of acceptably
accurate data while ensuring the removal of unacceptable positions if filtering is necessary to
perform habitat assignment (Table 2.1). These objectives include both position specific criteria
(i.e., <15 m error) and global criteria (i.e., mean error <1.77 m), though in each case a position
specific filter (i.e., precision target) must be identified. The filter value should represent the
minimum value that must have been met by a filtered dataset in order to achieve each objective.
Among candidate filters, we identified the preferred filter as the one with the largest HPE that
best met all criteria. The criteria provided a clear framework with which to select the best filter.

Step 2: Estimating the baseline position confidence in the array
Baseline position confidence refers to the quality of the unfiltered position data and is
partly a function of array coverage (i.e., low quality positions are likely more abundant in regions
of the array with poor coverage). Baseline confidence must be identified to determine if filtering
is required. The baseline position confidence was described by mean and median array accuracy;
position yield, the proportion of data that was below 6 m error, and the number of positions with
greater than 15 m error were enumerated for stationary and mobile test tags that collectively
determined if the array met the first three criteria. The fourth criteria dealt with data reduction.
The initial evaluation was based on position yield (percentage of positions estimated vs. expected
during the time period), average data accuracy, presence of large errors (>15 m), the twice the

35

distance root mean square (2DRMS) model equations, the proportion of data that was below 6 m
error, and array coverage based on actual fish telemetry data.
There were three tests of accuracy and each test varied in length, including 29,355,
16,400, and 138 positions (stationary site one, two, and mobile test) (Figure 2.1). The position
yield for the stationary tags were 82.3% (location #1) and 89.4% (location #2), and 94.8% for the
mobile test; this indicates that 5.2 to 17.7% of the tag transmissions did not result in a position.
The estimated positions of the stationary tags were generally clustered around a central position;
each exhibited a ‘tail’ of increasingly erroneous positions offset to the west. Without filtering,
the mean error was 11.7 m for stationary test one (median: 2.9 m, range: 2.7 to 29,289.3 m), 4.2
m for stationary test two (median: 2.6 m, range: 2.5 to 1,425.8 m), and 5.81 m for the mobile test
(median: 2.95 m, range: 2.50 to 186.40 m). There were 625, 134, and 0 positions with greater
than 15 m of error (stationary test one, stationary test two, mobile test). The linear models for the
2DRMS regression line relating HPE and measured error obtained from the stationary-location
tags were y = 0.18x – 0.89 (r2 = 0.92) for location #1 and y = 1.1x – 0.57 (r2 = 0.78) for location
#2 (Figure 2.2). The slopes were both near 1 (0.81 and 1.1), indicating a similar relationship of
HPE and synchronization tags to HPE and fish tags, with HPE potentially representing a
conservative estimate of position error for fish tags (the slope of the 2DRMS line was less than
1). The proportion of test positions with 6 meter accuracy exceeded 90% for all three tests
(Figure 2.3). Maximum array coverage was 8.22 km2, estimated by forming a polygon around
the outer most estimated fish positions.

Step 3a: Evaluating HPE filters consistent with the data quality objectives and guides
The necessary filter cutoff for each criterion to be met must be determined prior to
selection of a filter cutoff. Although the highest acceptable HPE value will retain the most data,
36

the range of HPE values that fulfill each criterion to reach each objective should be evaluated
and reported. If no value meets all criteria, either the analysis associated with an objective must
be reevaluated, a different analysis should be selected, or the research question will prove
difficult to address with the original planned approach.

Criterion 1
Any HPE cutoff from 3 (minimum observed in the test data) to 15 (maximum
considerable due to objective 2) was sufficient to meet criterion 1 for stationary tests but not for
the mobile test, although the mobile test methodology was susceptible to inherently high average
error (Table 2.2).

Criterion 2
The number of positions that violated criterion 2 increased greatly between HPE filters of
6 and 15 (max error: location 1: 10.0 to 26.7 m, location 2: 9.2 to 25.0 m; violating positions:
location 1: 0 to 21, location 2: 0 to 10) (Figure 2.4). Only HPE cutoffs less than 7 (stationary
location 1) or 8 (stationary location 2) met criteria 2, although no position had greater than 15 m
error for the mobile test (Table 2.2).

Criterion 3
The 2DRMS equation for desired 95% confidence in 6 m accuracy (Criterion 3) returned
an HPE cutoff of 8.5 for stationary test one and 6.0 for stationary test two (Figure 2.2). There
was not enough mobile test data to calculate this metric.

37

Criterion 4
The incorrectly retained proportion of all positions was less than 1% for HPE cutoffs of 3
to 10 for stationary test one and was 1% for all HPE cutoffs 3 to 15 for stationary test two
(Figure 2.5). For the mobile test, only an HPE of 3 reduced the incorrectly retained proportion
below though only four positions were incorrectly retained (3% of total; max. error = 7.7 m)
when HPE cutoffs ranged from 6 to 10 (Figure 2.6; Table 2.2). Incorrectly rejected positions
occurred at less than 5% of all positions for any HPE cutoff greater than 8 for stationary test one,
greater than 3 for stationary test two, and greater than 5 for the mobile test, representing a
minimal loss of acceptable data. When comparing the relationship between incorrectly retained
and incorrectly rejected, acceptable HPE cutoffs of 8 to 10 (stationary test one), 4 to 15
(stationary test two), and no HPE was effective for the mobile test, although 5 to 10 was closest.

Step 3b: Selecting an HPE cutoff
The highest HPE cutoff that met or was the closest to meeting all criteria was an HPE of
8, although from this dataset HPE cutoffs between 6 to 8 all fit the criteria similarly. There were
only two positions remaining with error greater than 15 m and an HPE less than 8 (criterion 2)
during the combined 45,744 transmissions during the two stationary tests, which may be good
enough for first passage time analysis if visual inspection of the remaining data points allows the
remaining problematic values to be easily identified or the analysis for this step may need to be
adapted. If we selected a lower HPE to meet this criterion perfectly, an HPE cutoff of 6 would
have been required. An HPE cutoff of 6 would have incorrectly rejected 7.7% of data from
stationary test one. Although a practitioner could make a case for this level of filtering, we
preferred the increased coverage as data loss is a major issue with trajectory based analyses
(Figure 2.7); 4,124 more positions would have been lost with an HPE of 6 vs. 8 (28.35% lost
38

outside array and 11.22% lost inside array for HPE 6) and total areal coverage would have
reduced from 2.11 km2 (HPE 8) to 1.89 km2 (HPE 6) (Figure 2.8).

Step 3c: Evaluating the potential for introduced bias through application of HPE cutoff
There was evidence for spatial bias in filtering inside versus outside of the array and
potentially behavioral or habitat filter bias. Only three mobile test positions with acceptable
accuracy (<6 m) were rejected and all occurred at the array edge, suggesting that HPE may overfilter at the periphery of coverage (Figure 2.6a). In the sea lamprey data set, the HPE cutoff (8)
rejected 8.8% of 58,025 total positions inside the array and 23.5% of 53,435 total positions
outside of the array, though this comparison does not discriminate between removal of inaccurate
and accurate positions (Figure 2.8). There was no evidence for spatial filter bias inside of the
array as no region appeared more prone to poor positioning across individuals (Figure 2.8). A
behavioral or habitat filter bias was evident, as the majority of positions rejected from inside of
the array (3,070 positions, 74.6% of all rejections) were associated with seven lampreys during
daylight hours that were likely stationary, as sea lamprey are nocturnal. This observation is
consistent with previous observations that sea lamprey may settle in locations that interfere with
acoustic tag signal transmission, blocking the line of sight between receivers and tags [19].

DISCUSSION
We developed a straightforward conceptual approach for using a PPE to filter
hyperbolically positioned data and demonstrated this approach with a technology (underwater
acoustic telemetry) in which users evaluate position accuracy with HPE. The framework
included selection of defensible data quality objectives, evaluation of the array’s positioning

39

accuracy with an independent dataset, and determination of the relationship of the selected PPE
to measured accuracy and data retention.
Data-filtering with a PPE estimate of position accuracy has certain conceptual advantages
over a biological filter if the PPE is properly evaluated. PPE’s are calculated for each position
obtained from the telemetry apparatus, whereas biological filters ignore the positioning process
and only evaluate resultant positions based on an expectation of what is biologically reasonable
for the study species. Biological filters can be useful and are frequently used because they are
conceptually straightforward, and at times, the only available option. Some habitat filters are
quite reasonable (e.g., fish do not swim 500 m onto land); although the rule could introduce filter
bias as positions closer to the physical habitat edge are more likely to be rejected. Similarly, as
HPE represents a 95% confidence value, HPE becomes large when solutions become less
precise, as typically occurs outside the array periphery where the overlapping of parabolas allows
for multiple potential solutions [9]. Failure to select a proper filter is more problematic for
calculating movement trajectories (e.g., maximum ground speed), and identifying a useful
biological filter is especially challenging for aquatic species for which maximum movement
capacities are often unknown [22], or poorly estimated. Selection of a biological filter cutoff that
is high allows incorrectly retained positions to remain in the dataset, whereas cutoffs that are low
near the average speed remove valuable data. Either case will serve to infuse the data with a
perceived improvement in accuracy that is not supportable (i.e., the rejected positions may be no
less accurate than many retained positions). Unlike biological filters, PPE filters are position
specific and rely principally on the assumption that animal-integrated tags match the
performance of stationary or towed tags, and that the array is well-constructed to ensure
sufficient areal coverage, avoidance of obstructions, etc. [9,23]. If these assumptions are
supported, carefully selected PPEs that are based on the position quality should be used prior to
40

biological filters that are based upon the biological plausibility of resultant positions and not the
quality of the positioning process.
PPE filter selection ranges from choosing an arbitrary HPE based on subjective operator
preferences to developing a complete algorithm that would output a filter threshold based on a
set of inputs (i.e., the model determines both performance and value). The approach we suggest
clearly falls in the middle (i.e., performance measurement is objective but value judgment can be
subjective) but still represents a significant improvement over the use of arbitrarily selected filter
values. The selection of an HPE of 8 fell within the most effective range for the criteria and
ensures high confidence in 6 m accuracy (criteria 3). Although we identified three different
criteria to complete a complex analysis, a single criterion could be chosen for a single analysis,
or, if multiple analyses are contemplated, a different cutoff could be chosen for each, which
would make selection and reporting straightforward. In our example, the criteria for objective 2
was not met, as it required no positions to have greater than 15 m of error, clearly representing an
ineffective criteria for an objective. At very low numbers of violating positions, only extreme
positions remain and improving the filter to remove these positions came at a high cost (increase
in positions incorrectly rejected). These extreme outliers may not even represent predictable
performance of the system. If the filter is to be useful, the proper response is to adjust the first
passage time analysis, which is very easy and would have only required shifting the moving
designation by a few meters (17.7 m was the largest remaining error). However, perhaps a better
criterion would be to choose a high percentile (e.g., 99.5%) that represents an acceptable level of
risk or error in your future analysis. With the VPS system, little effort has been made to defend
filtering cutoffs beyond reference to prior use [13], and ambiguous filter criteria are at risk of
inadvertently becoming acceptable practice through the accumulation of use. In our case,
adoption of an ambiguous filter based on a previous study (HPE 10 to 20 [14-16]), would have
41

been less useful than the carefully evaluated filter cutoff of 8 and indefensible (Table 2.2).
Telemetry technology represents a very different tool than typical scientific instruments as its
design is rarely consistent, is difficult to standardize, and does not generate data points with fixed
accuracy and precision. For this reason, we suggest that data filtering should be a flexible
process that may progress towards more concrete rules if some level of standardization of use
occurs for telemetry equipment as has occurred for the use of positional dilution of precision in
certain GPS applications (e.g., [24]).
A VPS system is capable of attaining accurate positions (2DRMS <6 m, <2 m average)
with a high position yield (>82%) via autonomous receivers that are capable of covering large
areas (>2 km2), although these results were specific to this system and environment. As with all
systems, the VPS was susceptible to spatial, temporal, and behavioral or habitat bias in position
yield and position precision which could cloak important biological phenomena [6,24]. For
example, when an animal occupies a habitat that blocks the line of sight from tags to receivers,
as was suspected for the seven stationary sea lamprey that composed 74.6% of our filtered data
inside of the array [19], there is a potential habitat and behavioral bias [9]. Observed HPE values
also increased with distance outside of the test array, consistent with observations from other
studies [9,13]. Confirming a temporal or spatial bias is challenging because, depending on
placement, stationary tags may not reveal systematic spatial biases in the array [9,25], and may
be sensitive to regular variation in environmental conditions (e.g., louder waves in shallow vs.
deep waters) that differentially impact receiver performance. Mobile range testing is
recommended for spatial evaluations of filter performance, though results may not be
representative of the full range of environmental conditions encountered by tagged animals due
to the typically short duration of mobile tests, selection of favorable boating conditions, etc. The

42

mobile test appeared to be biased towards higher error estimates, though the spatial patterns were
consistent between fish tracking and mobile test data.
Both mobile and stationary tests presented unforeseen challenges and unexpected
findings. Mobile testing suggested an HPE of 8 was overly conservative outside of the internal
array area, though we lacked stationary tests in this region that would have confirmed this
conclusion (Figure 2.6). The mobile test presented some challenges to evaluation because the
GPS clock was not synchronized with VPS time and only recorded a position every second. To
minimize the effects of clock differences on position error estimates, we applied a constant time
offset to all mobile test positions that minimized error between the mobile test tag tracks and
corresponding GPS tracks (Additional file 1). In addition, we likely over-estimated error by
assuming that the GPS track represented the true path of the test tag because we only collected a
single post-processed position every second along the track. Collecting positions with sub-meter
accuracy usually requires averaging several GPS points at a given location, but point averaging
over time is not an option while moving. The tail of positions that were observed in the
stationary tests could not be explained with the example dataset but was likely the result of a
specific set of receivers with poor geometry that consistently estimated inaccurate positions in
one direction. The tails were not troubling to us as these positions were easily filtered with HPE.
Many challenges can be avoided in advance by careful project planning. We recommend
multiple mobile and stationary tag tests with the animal tag during the same time period that the
animals are being monitored, and ideally, at least some fraction of the synch-tags would be the
same model as the fish tags. Receivers should be positioned outside the maximum spatial extent
of interest, or at a minimum, stationary animal tags should be monitored in any area of interest to
ensure the chosen HPE cutoff has met the filter goals.

43

Although we present a process for selecting a single fixed HPE cutoff and provide the R
code necessary for performing this approach (Additional file 1), a dynamic filter in which the
HPE cutoff could be either temporally (e.g., [26]), or both temporally and spatially flexible,
might prove useful, although it also would require synchronization tags that match fish tags.
Regardless of how the HPE selection process is fine-tuned, the key is following a standard
process like the one we have described, including reporting the process for others to properly
assess the research findings. Even if the necessary accuracy required to perform the mostpreferred analysis cannot be attained with the data collected, PPEs can be used to tailor the
selection of an analytical tool, or the spatial scale at which the behavior is considered. Either
adaptation to the study is preferred over the reporting of errant observations [17,27].

Conclusions

PPE error estimates are frequently available from animal telemetry systems that rely on
hyperbolic positioning and can be used to evaluate data quality prior to analysis. When using
PPE to filter data, practitioners should undertake (a) a priori determination of data accuracy
requirements; (b) independent assessment of the telemetry system performance; (c) a
determination of how well the PPE represents measured accuracy; (d) selection of a filter cutoff
based on the balance between accuracy improvement and data retention; and (e) explicit
consideration of spatial, behavioral, and habitat bias associated with the telemetry system and the
animal under observation. A carefully constructed PPE filter is more defensible than biological
filters that can improve data accuracy but require (1) an interpretation of the data vs. an
assessment of its precision, and (2) are only applied to a subset of the data collected (extreme
movements). HPE offers the intriguing possibility for direct use in the analysis as an error
estimate (vs. a criterion for data retention); akin to bench apparatus precision estimates, though
44

there is no evidence that this approach has been used in such a manner with other hyperbolic
positioning systems. Because data analysis requirements are likely to be as varied as the
movement data to which they are applied, complete exposition of the selection process and
criteria should be included in the methods section of any subsequent reports or publications. The
minimum level of a reporting should include a description of the data quality objectives, criteria,
rational for cutoff selection, and evidence of reaching the criteria, which could come in the form
of a paragraph, table, or appendix and does not need to be a substantial component of the report
or paper.
ACKNOWLEDGEMENTS
This manuscript is contribution number 1843 of the Great Lakes Science Center. This
work was funded by the Great Lakes Fishery Commission by way of Great Lakes Restoration
Initiative appropriations (GL-00E23010-3). This paper is contribution 7 of the Great Lakes
Acoustic Telemetry Observation System (GLATOS).

45

APPENDIX

46

Table 2.1: Filtering Objectives. The four specific criteria adopted to establish the data quality objectives for the project.
Objective

Criteria

Rationale

1

Detect changes in
trajectory

Mean error ≤1.77 m

Many trajectory based analyses are only preserved when the average position
error is <10% of the mean step length [17]. Transmission delay was (33.4 s),
sea lamprey ground speed is 0.53 ms−1 [18,19], equating to a step length of
17.7 m.

2

Assign behavioral state Max error <15 m
(moving or stationary) to
fish positions

3

Assign habitat to fish
positions

2DRMS ≤6 m

To ensure accurate assignment of each estimated fish position to habitat types
that are defined at the resolution of 18 × 18 m grid cells we selected a twice
the distance root mean square (2DRMS) error of <6 m for our selected HPE
values.

4

Balance loss of
acceptable data and
retention of
unacceptable data

Retain 95% of
acceptable data, Reject
≥99% of unacceptable
data

To ensure accurate assignment of each estimated fish position to habitat types
while avoiding loss of accurate data. We aimed to keep 95% of acceptable
data (≤6 m) while retaining <1% of unacceptable data (>6 m) of all retained
positions (unacceptable data remaining/all retained positions × 100).

Simple behavioral assignment of moving or stationary with first passage time
analysis, which is a measure of the time it takes an individual to leave a circle
of fixed radius r drawn around each measured location to determine if
movement is occurring [20,21]. Based on prior analyses, a displacement of
greater than 15 m per transmission at the average tag transmission rate (33.4
s) could result in a false designation of moving when the fish was stationary.

47

Table 2.2: Criteria for selection of an HPE filter cutoff

Criteria

Stationary
Mobile
test
test
1
2

3

This criterion was met for stationary tests (Figure 2.4), but the mobile test would have
required a very low HPE to attain below 1.77 m of average error as the unfiltered
accuracy was lower (Figure 2.3). See review of the mobile test methodology in the
Discussion.

≤15

An HPE of 8 did not meet the criteria for stationary tests, but only 1 of 625 positions (test
one) and 1 of 134 positions (test two) remained. The mobile test criteria were met. Only 2
of 45,744 positions were problematic (<0.001%) for the combined stationary tests (Figure
2.4). The maximum point remaining was 17.7 m error.

NA

The criterion was met for stationary test one, which was the longer test that covered more
variable weather conditions. The HPE 8 bin only was 95% confident in an HPE of 8.25
meters for test two but the estimate was based on data calculated within each bin and few
points had an HPE of >6 for test two (3% of the data), which results in a less reliable
2DRMS prediction.

Criterion 1
3–15 3–15
Mean ≤1.77 m
Criterion 2
<7

<8

Max error <15 m
Criterion 3
2DRMS ≤6 m

8.5

6.0

Rational for selection of a HPE filter cutoff of 8

Criterion 4

None

Percentage incorrectly 8–10 3–15
retained vs. percentage
incorrectly rejected

5–10

An HPE cutoff of 8 met the criteria for both stationary tests, and although the mobile test
did not have a suitable range, the range from 5–10 was equally effective.

48

Figure 2.1: Acoustic telemetry activities at the Hammond Bay field site. A schematic of the VPS array that was located in Lake Huron
around the mouth of the Ocqueoc River (blue line). Triangles represent receiver (VR2W) positions. VPS array testing in 2010 included
two stationary tag tests (Gray dots, with median point as a black dot) and three mobile test transects (black dots forming lines). The
schematic is oriented with north up and the black line running from left to right (east to west) represents the coast.

49

Figure 2.2: The stationary test schematics and 2DRMS plots of all stationary test positions. Two
schematics depict all VPS positions during two stationary tag tests, including (a) 29,355
50

Figure 2.2: (cont’d)
positions between the dates (6/17/2010 to 7/01/2010), and (b) 16,400 positions between the
dates (7/01/2010 to 7/08/2010), allowing us to evaluate array performance at two locations
through an extended period of time. The white dot in the center of the clusters is the median
location. The HPE versus measured error to the median point is shown for each estimated
position during test one (c) and test two (d). The white circles with black outline and red x
represent twice the distance root mean square error of x and y components of error within an
HPE bin of one; 95% of tag detections have an error less than this point within each bin. Note
there is a minimum HPE of 2.7 and 2.5 within the data. The line running between these points
represents the 2DRMS and the equation and fit for this line are shown in the top left corner of (c)
and (d), respectively. Data points above the 15 m bin, which can be seen in (a) and (b) are not
shown in (c) or (d), because they are outside of the zone of interest.

51

Figure 2.3: Proportion of test positions with measured error from 1 to 10 m. Percent of positions
with accuracy equal to or less than each measured value for the mobile tag test (circle),
stationary tag at location one (triangle), and the stationary tag at location two (square) depicted
an array with most positions having accuracy better than 6 m. The average error of the unfiltered
data for the stationary tag at location one (1.98 m), location two (1.11 m), and the mobile test
(6.83 m) are marked by representative symbols along the x-axis.

52

Figure 2.4: Resultant data quality for HPE cutoffs of 3 to 15. The mean and maximum measured
error is below 1.77 m for all HPE thresholds, sufficient to meet criteria 1 for both stationary test
one (black) and test two (red). The maximum error exceeds 15 m at an HPE of 7 for test one and
8 for test two (violating criteria 2). The number of positions violating each test is located at the
top of the figure for HPE cutoffs of 3 to 15.

53

Figure 2.5: Data loss versus error retention for HPE cutoffs of 3 to 15. The relationship of the
percent of incorrectly rejected positions of all acceptable positions and percent of incorrectly
retained positions of all retained positions suggested that HPE cutoffs of 8 to 10 for stationary
test one and 3 to 15 for stationary test two, met criteria 4.

54

Figure 2.6: An evaluation of the performance of the VPS array with a mobile tag. (a) A
schematic depicting receiver positions (+) and the coast (black line) during the 2010 research
55

Figure 2.6: (cont’d)
season. A mobile tag test was completed on 7/6/2010. The small dots represent the VPS
estimated positions during the mobile test. There were a total of 126 correctly retained positions
(black , <6 m error, <HPE 8), 5 correctly rejected positions (blue), 3 incorrectly rejected
positions (orange), and 5 incorrectly retained positions (red). The incorrectly rejected (n = 3)
positions occurred consecutively and were located at the furthest distance from the array in the
left transect. (b) A graph scaled to cover positions with an HPE between 0 and 20 depicts
correctly rejected, correctly retained, incorrectly rejected, and incorrectly retained positions. All
values with an HPE greater than 20 were correctly rejected (n = 3, not shown). Euclidean
distance is our best estimate of measured error.

56

Figure 2.7: HPE values for all fish positions. The schematic depicts the HPE value for each fish
position during active monitoring of study subjects. The coast is represented by the black line,
receivers are represented by black triangles, and all sea lamprey positions are color coded by
their associated HPE estimate and it is clear that HPE greatly increases when the sea lamprey
position is outside of the array proximity, other than for a few positions inside the array.

57

a) All positions

b) Positions retained (HPE≤8)

c) Positions removed (HPE>8)

Figure 2.8: Fish positions filter by an HPE cutoff of 8. A depiction of how many of all fish
locations in 2010 would be rejected by an HPE filter of 8.0. Fish locations are represented by
small black dots, receiver positions are shown by red triangles and the coast is depicted by a gray
line. These are cumulative graphs with all sea lamprey tags shown in (a), fish positions with an
HPE below or equal to 8.0 are displayed in (b), and fish positions with an HPE greater than 8.0
are shown in (c).
58

Figure 2.8: (cont’d)
Supplementary Material: R code for filtering VEMCO positioning system data with Horizontal
Positioning Error (HPE).
Available at: http://www.animalbiotelemetry.com/content/2/1/7/additional

59

REFERENCES

60

REFERENCES

1. Cagnacci F, Boitani L, Powell RA, Boyce MS: Animal ecology meets GPS-based
radiotelemetry: a perfect storm of opportunities and challenges. Philos Trans R Soc London
[Biol] 2010, 365(1550):2157–2162.
2. Cooke SJ, Hinch SG, Wikelski M, Andrews RD, Kuchel LJ, Wolcott TG, Butler PJ:
Biotelemetry: a mechanistic approach to ecology. Trends Ecol Evolut 2004, 19(6):334–343.
3. Cooke SJ, Midwood JD, Thiem JD, Klimley P, Lucas MC, Thorstad EB, Eiler J, Holbrook C,
Ebner BC: Tracking animals in freshwater with electronic tags: past, present and future.
Anim Biotelem 2013, 1:5.
4. Giuggioli L, Bartumeus F: Animal movement, search strategies and behavioural ecology:
a cross‐disciplinary way forward. J Anim Ecol 2010, 79(4):906–909.
5. D’Eon RG: Effects of a stationary GPS fix rate bias on habitat selection analyses. J
Wildlife Manage 2003, 67:858–863.
6. D'Eon RG, Delparte D: Effects of radio‐collar position and orientation on GPS radio‐
collar performance, and the implications of PDOP in data screening. J Appl Ecol 2005,
42(2):383–388.
7. Frair JL, Nielsen SE, Merill EH, Lele SR, Boyce MS, Munro RHM, Stenhouse GB, Beyer HL:
Removing GPS collar bias in habitat selection studies. J Appl Ecol 2004, 41:201–212.
8. Whithey JC, Bloxton TD, Marzluff JM: Effects of tagging and location error in wildlife
telemetry studies. In Radio Tracking and Animal Populations. Millspaugh JJ, Marzluff JM
(Eds.). Waltham, MA: Academic Press Inc.; 2001, pp. 43–75.
9. Smith F: Understanding HPE in the VPS Telemetry System. VEMCO Tutorials; 2013.
[http://vemco.com/wp-content/uploads/2013/09/understanding-hpe-vps.pdf]
10. Andrews KS, Tolimieri N, Williams GD, Samhouri JF, Harvey CJ, Levin PS: Comparison
of fine-scale acoustic monitoring systems using home range size of a demersal fish. Mar Biol
2011, 158(10):2377–2387.
11. White GC, Garrott RA: Analysis of Wildlife Radio-Tracking Data. Waltham, MA: Academic
Press Inc.; 1990.
12. Espinoza M, Farrugia TJ, Webber DM, Smith F, Lowe CG: Testing a new acoustic
telemetry technique to quantify long-term, fine-scale movements of aquatic animals. Fish
Res 2011, 108(2):364–371.

61

13. Roy R, Beguin J, Argillier C, Tissot L, Smith F, Smedbol S, De-Oliveira E: Testing the
VEMCO Positioning System: spatial distribution of the probability of location and the
positioning error in a reservoir. Anim Biotelm 2014, 2:1.
14. Scheel D, Bisson L: Movement patterns of giant Pacific octopuses, Enteroctopus dofleini
(Wülker, 1910). J Exp Mar Biol Ecol 2012, 416–417:21–31.
15. Furey NB, Dance MA, Rooker JR: Fine‐scale movements and habitat use of juvenile
southern flounder Paralichthys lethostigma in an estuarine seascape. J Fish Biol 2013,
82:1469–1483.
16. McMahan MD, Brady DC, Cowan DF, Grabowski JH, Sherwood GD: Using acoustic
telemetry to observe the effects of a groundfish predator (Atlantic Cod, Gadus morhua) on
movement of the American lobster (Homarus americanus). Can J Fish Aquat Sci 2013,
70(11):1625–1634.
17. Bradshaw CJA, Sims DW, Hays GC: Measurement error causes scale-dependent
threshold erosion of biological signals in animal movement data. Ecol Appl 2007, 17(2):628–
638.
18. Vrieze LA, Bergstedt RA, Sorensen PW: Olfactory-mediated stream-finding behavior of
migratory adult sea lamprey (Petromyzon marinues). Can J Fish Aquat Sci 2011, 68:523–533.
19. Meckley TD, Wagner CM, Gurarie E: Coastal movements of migrating sea lamprey
(Petromyzon marinus) in response to a partial pheromone added to river water:
implications for management of invasive populations. Can J Fish Aquat Sci 2014, 71(4):533–
544.
20. Fauchald P, Tveraa T: Using first-passage time in the analysis of area-restricted search
and habitat selection. Ecology 2003, 84(2):282–288.
21. Barraquand F, Benhamou S: Animal movements in heterogeneous landscapes: identifying
profitable places and homogenous movement bouts. Ecology 2008, 89:3336–3348.
22. Katopodis C, Gervais R: Ecohydraulic analysis of fish fatigue data. River Res Applic 2012,
28:444–456.
23. Espinoza M, Farrugia TJ, Lowe CG: Habitat use, movements and site fidelity of the gray
smooth-hound shark (Mustelus californicus, Gill 1863) in a newly restored southern
California estuary. J Exp Mar Biol Ecol 2011, 401(1):63–74.
24. Lewis JS, Rachlow JL, Garton EO, Vierling LA: Effects of habitat on GPS collar
performance: using data screening to reduce location error. J Appl Ecol 2007, 44(3):663–
671.

62

25. Biesinger Z, Bolker BM, Marcinek D, Grothues TM, Dobarro JA, Lindberg WJ: Testing an
autonomous acoustic telemetry positioning system for fine-scale space use in marine
animals. J Exp Mar Biol Ecol 2013, 448:46–56.
26. Coates JH, Hovel KA, Butler JL, Klimley AP, Morgan SG: Movement and home range of
pink abalone Haliotis corrugate: implications for restoration and population recovery. Mar
Ecol Prog Ser 2013, 486:189–201.
27. Hayes GC, Kesson SA, Godley BJ, Luschi P, Santidrian P: The implications of location
accuracy for the interpretation of satellite-tracking data. Anim Behav 2001, 61:1035–1040.
28. Nocedal J, Wright SJ: Numerical Optimization. New York, NY: Springer; 1999.

63

CHAPTER 3

HOW DO NON-HOMING FISHES FIND THE SHORE? EVIDENCE FOR BATHYMETRIC
ORIENTATION IN MIGRATING SEA LAMPREY

ABSTRACT

How non-homing fishes like the sea lamprey (Petromyzon marinus) find distant targets is
poorly understood. Following translocation by host fishes during parasitic feeding in the Great
lakes, sea lamprey must find the coast and enter a river to spawn. Three-dimensional paths of 22
captured and released spawning-phase sea lamprey were quantified from the commencement of
movement from a settled state until exiting an acoustic array with 3 km2 of coverage, centered
3.3 km from the western coast of Lake Huron in greater than 33 m of water. 81 % of individuals
arrived at the nearest coast within 72 hours, suggesting efficient orientation to the coast termed
y-axis orientation. Two phases of movement were documented for most individuals: i) in phase
1, sea lamprey moved slowly on the bottom and turned in a consistent direction and made
vertical ascents to the surface and returned to the bottom; and, ii) in phase 2, sea lamprey moved
more quickly and steered a straight course while making occasional heading corrections and
pirouettes and moved primarily at the surface, with occasional descents to the bottom. During
phase 2, sea lamprey moved in the direction of the bathymetric slope towards shallower water
rather than towards the nearest coast, and no apparent effect was observed due to where the sea
lamprey had previously entered a river in Michigan. We postulate that these sea lamprey used the
absolute hydrostatic pressure sensed at successive sites along the lake bottom to pick a heading
towards shallower water during Phase 1 and used declining hydrostatic pressure at the lake
bottom to steer towards shallower water during Phase 2. In contrast to a natal homing migration,

64

the non-homing sea lamprey migration is consistent with a growing number of observations of
organisms that orient to a region of resource availability with a simple set of movement rules
based on topographic features of the environment.

INTRODUCTION

Efficient navigation to a distant target requires that a searching animal must first orient
itself to environmental features (referents) that reveal the direction to the target, undertake
directed movements in accordance with the referent(s) (Able 1991, Åkesson and Hedenström
2007, Luschi 2013), and recognize the target upon arrival (Dittman and Quinn 1996, Walker et
al. 2002). Animals can reach resources via: i) non-oriented search (random), ii) oriented search
with information that is independent of the resource but related to a region of high target
availability, iii) a cue directly associated to the resource, or iv) based on prior knowledge of the
target location (Mueller and Fagan 2008). To understand how animals orient during migration,
large scale detailed data on the environment, the movement path and an understanding of the
internal state of the animal is necessary (Nathan et al. 2008). Studies of this nature in aquatic
environments are rare (Lohmann et al. 2008, Cooke et al. 2013), as underwater positioning
technologies and methodologies capable of large-scale coverage with high resolution are just
becoming available (Jonsen et al. 2003, Espinoza et al. 2011, Biesinger et al. 2013).
The emerging theory for homing migrations is that migrants switch between navigation
methods at different scales but rely predominantly on geomagnetic information to arrive near
their natal home before relying on other senses, notably olfaction, to move more specifically to a
reproduction site (Lohmann et al. 2008, Putman et al. 2014). A major difference in non-homing
oriented search is that individuals can rely on topography to locate a distinct geophysical region
65

with high resource availability (e.g., Pe'er et al. 2004). Orientation to topographic features is
common to many species including hilltopping butterflies, Melitaea trivia, that move upslope to
locate mates, sand hoppers, Orchestoidea corniculata, that rely on slope and other features to
maintain perpendicular orientation to the shoreline in a preferred tidal zone (Craig 1973, Scapini
et al. 1996), or Monarch butterflies, Danaus plexippus, that rely on simple geographic barriers to
funnel their migration towards Mexican overwintering sites (Mouritsen et al. 2013). Geophysical
features that inform an organism of its position are commonplace, though they may be masked
by direct cues or not useful at all in a homing migration due to the specificity of their target.
Specifically, many aquatic organisms orient themselves perpendicularly to the coast in a special
case of orientation termed y-axis orientation (Ferguson 1971). Y-axis orientation has been
observed in many species including multiple amphibians that rely on celestial features to orient
to shore (Ferguson et al. 1965, Landreth and Ferguson 1967) or green turtles, Chelonia mydas,
that use windblown odors to orient to a specific beach (Luschi et al. 2001). Over large distances
in large bodies of water, moving opposite a gradient such as slope along the lake bottom could
theoretically provide directional cues leading to a coast, which could be considered analogous to
navigating up a hill.
Orientation in a non-homing species is more difficult to study than in homing species
because the animal has greater flexibility to use indirect information to orient that may not be
directly associated to a specific target (Benhamou 2006). For this reason, orientation mechanisms
have principally been studied in species with prior knowledge of the target, wherein referents can
be manipulated to reveal if the change in the referent alters how the animal searches for a
specific site (Lohmann et al. 2008, Holland et al. 2010, Luschi 2013). The time required for an
animal to orient depends on how long it takes to gather directional information and calibrate an

66

internal compass. Such a situation often results in an initial period of sinuous exploratory
movement (Wendler and Scharstein 1986, Ueda et al. 1998, Pe'er et al. 2004). When searching
for distant targets, oriented animals usually move in a straighter fashion. The heading of the
straight movement or direction of the final point observed in a study (vanishing point) can be
compared with environmental features to determine if there is a consistent pattern and what
information the animal is using to orient (Girard et al. 2004, Holland et al. 2010).
The obligation of non-homing adult sea lamprey to efficiently find a coastline and
subsequently enter a suitable river to spawn after feeding has terminated makes them an
excellent candidate for studying y-axis orientation in fishes. The Great Lakes adult sea lamprey
completes a non-homing reproductive migration from lacustrine feeding grounds, where they
parasitize large bodied fishes, to riverine spawning habitat (Bergstedt and Seelye 1995, Waldman
et al. 2008). The semelparous migration initiates when they release from their final host, which
may occur at a location never previously visited, and may represent an uncontrolled
displacement (Clemens et al. 2010, Silva et al. 2014). Any coast is a viable target for fishes in a
lake (Meckley et al. 2014b), whose target is a river labeled locally by olfactory cues (Waldman
et al. 2008, Vrieze et al. 2010). Therefore, sea lamprey could feasibly rely on indirect cues that
allow for straight movement in any direction or cues directly associated to the nearest coast.
Once arriving at the coast, sea lamprey move parallel to the coastal edge while searching for
river water at the lake edge (Vrieze et al. 2011, Meckley et al. 2014b). If sufficient larvae occupy
a river, water at the mouth will contain larval odor indicating adequate proven spawning and
rearing habitat within the watershed (Teeter 1980, Sorensen et al. 2005, Wagner et al. 2006). Sea
lamprey rely heavily on non-visual information during their nocturnal migration limiting
potential orientation mechanisms (Binder and McDonald 2007, Keefer et al. 2013). Based on the

67

life history and sensory capacity of sea lamprey, we postulate that sea lamprey rely on an aspect
of the local bathymetric gradient to determine the direction to shore (y-axis orientation).
We investigated the three dimensional path of sea lamprey in Lake Huron by following
their release 3.3 km from the nearest coast in 30 m of water to test whether they complete y-axis
orientation and return to the coast by following the bathymetric gradient. We specifically
evaluated a bundled set of predictions to address our complex hypothesis, (1) sea lamprey will
appear at the nearest coastline within 72 hours, indicating that they performed y-axis orientation
and approached the nearest coast first; (2) sea lamprey move in a two-phase pattern after
displacement consistent with local exploration (high sinuosity) followed by search for a distance
target (straight movement), indicating that we observed an orientation process; (3) the cardinal
direction of movement during the second phase across individuals is not in a consistent direction,
signifying a variable search strategy or use of a locally varying feature; (4) the cardinal direction
will be consistent with following the local bathymetric gradient ; and (5) the direction will be
based on local information in the array and not on which river the sea lamprey was trapped in
Michigan.

METHODS
Ethics Statement
The treatment and acoustic tagging of sea lamprey was approved by the Michigan State
University Institutional Animal Use and Care Committee via animal use permit 02/10-020-00.

Study Design
The tracks of 22 sea lamprey were analyzed after their release in the center of an acoustic

68

positioning array located 3.3 km from the nearest coast (Figure 3.1). The array positioned fish
over a region of the coast that had shallower bathymetric contours to the south and west
depending on where movement commenced (Figure 3.2). Sea lamprey did not always stop
quickly after release so movement could commence at different regions within the array despite
release occurring at the center of the array. We bundled predictions to test our hypothesis that sea
lamprey rely on an aspect of the bathymetric gradient to determine the direction to shore
regardless of the compass direction of the bathymetric slope in front of the river in which they
were trapped. To determine if sea lamprey performed y-axis orientation we relied on the first
detection of each sea lamprey at acoustic receivers located along the coast (Figure 3.1). To
determine if sea lamprey employed a two-phase strategy indicating that we captured the
orientation process, we relied on visual inspection and evaluation of track sinuosity. The heading
of the sea lamprey from the point it began moving in a straight path until exiting the array was
extracted and used to determine if sea lamprey move directly towards the closest coast or follow
the local bathymetric contours, and whether there was an effect of past experience, in terms of
the river they were caught. Lastly we characterized the vertical movement of sea lamprey to
build hypotheses why sea lamprey make excursions through the entire water column and what
role it may have in search behavior.
Typically sea lampreys were transported to the field site in aerated coolers and released in
the center of the acoustic array at 15:00 when weather permitted (09:00 - 17:30 EST) by being
lowered in a release cage at a rate of 6 m·min-1 that opened at a fixed depth of 30 m when door
clamps were triggered by pressure. Three sea lamprey were released at the surface to determine
the minimum rate at which lamprey could descend without apparent harm and all reached the

69

bottom in less than 2 min, equating to a descent rate of faster than 17.5 m·min-1. The surfacereleased sea lamprey were not considered in any other test.
Sea lamprey came from six different sources, located in three different scenarios where
the contours from deeper to shallower water (bathymetric slope) occurred in three different
compass directions (Figure 3.1). If sea lamprey orient to any aspect of the coast and relied on
geomagnetic information from previous experience we would expect these conditions to
influence their selected headings. Adult female sea lamprey were obtained from five rivers in
Michigan via barrier traps (Manistee River: 44.249981, -86.344531 ( N=8, trapped 15-May2012), Cheboygan River: 45.656202, -84.464478 (N=21, trapped 02-May-2012), Manistique
River: 45.945189, -86.247733 (N=21, trapped 05-May-2012), Betsie River: 44.630058, 86.252273 (N=3, trapped 21-May-2012), Ocquoec River: 45.490246, -84.072981 (N=4)), and
parasites were caught while attached to fish in the lake (Lake Huron, Hammond Bay (N=10,
caught January 03,2012 February 28, 2012)) (Figure 3.1). Individuals from the Manistee and
Betsie rivers did not move for several days after release and were dropped from the study. These
animals were likely more mature and no longer migratory, as they came from rivers that were
warmer. As a result, we compared only the Manistique and Cheboygan River sea lamprey to
directly evaluate experience, in terms of the river in which they were trapped.

Study Site
The center of the study site (45.527799, -84.044466) was in Lake Huron and located 3.3
km from the nearest shoreline in an area of the lake where bathymetry varied only in two
directions (Figure 3.1). A telemetry system (VPS, VEMCO, Halifax, NS) composed of 43
receivers (VR2W) in diamond formations with between receiver spacing of 275 m and an

70

internal area of 3 km2 recorded sea lamprey position (Figure 3.2; see Meckley et al. 2014a).
Approval for temporary placement of equipment was provided by the United States Coast Guard
(Permit: 16518, Ser. 09-12). Receivers were placed 3 m from the lake bottom in a region of the
lake ranging from 30.6 m – 37.2 m deep, although some fish positions were obtained for fish
located outside of the array in an area ranging from 21-39 m deep. Nine synchronization
transmitters (VEMCO model V16-2H, 69 kHz, 160 db, 500 to 700 s transmission interval) were
collocated with receivers to maintain time synchronization of the receiver clocks. In addition to
the VPS array, individual receivers provided detection data along multiple locations on the coast
and in the Black Mallard, Ocquoec and, St. Mary’s Rivers as part of the Great Lakes Acoustic
Telemetry Observation System (GLATOS, Figure 3.1). Detections on these receivers informed
us of whether sea lamprey reached the nearest coast prior to other coasts. Solitary receivers were
positioned at the Black Mallard River Mouth (BM, 45.532888, -84.120801) and Ocqueoc River
Mouth (OCQ, 45.491893, -84.071879). Receiver strings were treated as individual detection
points and were located near the Cheboygan River mouth (CHB, 45.67239, -84.429368, N=5),
40 mile point (FMP, Outer receiver: 45.507563, -83.901379, N=3), Presque Isle (PRS, Outer
receiver: 45.333842, -83.458343, N=3), and Detour Pass, which leads to the St. Mary’s River
(SMR, West side of Drummond Island, centered at: 45.984929, -83.891787, N=6).
Detection data revealed sea lamprey presence and was used to determine the fate of sea
lamprey on the coast. Data were obtained from shared receivers as part of GLATOS. The
detection range of a tag with 150 db of power placed in 5 m of water will vary and is based on
line of sight to a receiver, and the environment. A 2-D range test in Hammond Bay under calm
conditions revealed a 95% detection efficiency at 155 m and a 23% detection at 1.1 km.
Detections exceeding 1.5 km and passing the nearest neighbor filter were unlikely. The detection

71

range of a receiver placed in < 3 m of water directly in front of the Ocqueoc and Black Mallard
rivers, as was done in this study, was likely lower than observed in the range test, especially
because the receiver was located in the wave zone. A nearest neighbor filter of 30 minutes was
used to remove spurious detections and rare detections near the maximum extent from the
GLATOS detection data.
Bathymetry in the detection region of the array was measured along transects spaced 50
m apart and oriented northwest to southeast, and then repeated crisscrossing the first grid at an
orientation of southwest to northeast. A Lowrance depth sounder (HDS-8) streamed depth and
GPS location to a laptop to record bottom depth during calm conditions and was matched to
more accurate post processed GPS positions collected by a Trimble GeoXH and Tornado
antennae. Total water column depths at fish positions were interpolated in program R from the
georeferenced depth data using inverse distance weighting (function “interp”, package akima)
(Akima 1978, Gebhardt et al. 2013).

Tagging Procedure
All sea lamprey were held in 150 L flow-through tanks that cycled ambient Lake Huron
water (100% exchange every 2 h) under a natural light cycle. Two types of acoustic transmitters
transmitted signals every 15-45 s, though one contained a pressure sensor (n=49, model V9P2H,Vemco, Halifax, Nova Scotia, Canada 9mm D x 47 mm L, mass: 6.4 g in air, 3.5 g in water,
power output 150 dB (re 1 µPa at 1 m)) and the other only provided horizontal position (n= 18,
model V9-2H,Vemco, Halifax, Nova Scotia, Canada, 9mm diameter, 29 mm length: mass: 3.6 g
in air, 2.2 g in water, power output 151 dB (re 1 µPa at 1 m)). Sea lamprey were 273-577 mm
long (mean 499 mm) and weighed 139-398 g (mean 274 g). Acoustic tagging procedures

72

followed Meckley et al. (2014a, b). Prior to surgery, sea lamprey were anesthetized by
immersion in 0.2 mL·L−1 clove oil solution. The anesthetic solution was produced through
dilution of 2 mL of clove oil (minimum 84%-88% eugenol, Lot No. HB9387, Hilltech Canada
Inc. Vankleak Hill. Ontario, Canada) into 18 mL of 70% ethanol and vigorously mixed into 10 L
of Lake Huron water. Sea lamprey were removed from the bath upon reaching stage four of
anesthesia, determined by individuals that did not respond to handling but retained gill
movement (mean time to stage ± 1 SE, 559 ± 14.15 s, maximum 846 s). The surgery was
performed in a PVC pipe with continuous water flow that allowed gill irrigation to be maintained
by completely submerging the head and gills. We inserted the transmitter into the peritoneal
cavity through a 20 mm incision approximately 10 mm off the ventral midline that ended in line
with the anterior insertion of the first dorsal fin. The incision was closed with three independent
interrupted surgeon knots (3-0 Ethicon sterile monocryl monofilament) and each knot was sealed
with veterinary adhesive (Vetbond, n-butyl cyanoacrylate adhesive). The surgical procedure took
an average of 283 ± 4 s to complete. Each subject was monitored in a postoperative holding tank
until we observed the animal regain equilibrium and begin natural swimming movements
(recovery time, mean ± 1 SE, 360 ± 28 s, maximum 1240 s). Transmitter-implanted lamprey
were held for 72 h prior to release to ensure metabolism of stress compounds (Close et al. 2003).

Data quality assessment and position filtering
The position accuracy of the VPS array was tested by comparing the VPS position
estimates to GPS measured positions (Trimble Geo XH, post processed) of two transmitters at
fixed locations (Fixed test 1: June 17, 2014 – June 26, 2014; Fixed test 2: June 17, 2014 – July
01, 2014) and two transmitters pulled through the array (Drag 1:June 13, 2014; Drag 2: June

73

17,2014) (V9P-2H transmitter). To avoid the effects of positioning error on path sinuosity, a data
quality objective was set to attain 95 % confidence that reversals did not erroneously occur in the
data. VEMCO positioning systems provide a position precision estimate for each position,
horizontal positioning error (HPE), if evaluated the HPE can be used to remove erroneous
positions (Smith 2013, Meckley et al. 2014a, Roy et al. 2014). The step length was 22 m in calm
wave conditions, which equated to an objective of 95% confidence in an error less than 11 m.
We estimated the twice the distance root mean square error (2DRMS) equations for each fixed
tag and calculated maximum HPE values that would allow 95 % confidence in 11 m of error. We
selected an HPE filter of 15.17 (Appendix 1).
To classify sea lamprey positions as active or stopped, a first passage-time classification
method was performed with the program R. The first passage-time tool classified a position as
moving if it left a radius of 10 meters in 250 s, had a minimum displacement of 15 m in a 3
position moving average, and at least 3 consecutive moving observations (function “fpt”,
package adehabitatLT) (Calenge 2006, Gurarie et al. 2015). The assignment accuracy was
verified for each sea lamprey through visual inspection and was robust to imprecision in acoustic
positioning largely because we did not observe position error of greater than 10 m for three
consecutive positions.
Sea lamprey were assigned to one of 5 categories based on post-release behavior: (A)
those that stopped in the array and exited the array on the first night (n=22); (B) those that
stopped at the edge of the array and began on the first night but few active points were captured
because they were near the edge of the detection range (n=5); (C) those that stopped in the array
and exited on the second night or later (n=10); (D) those that immediately abandoned the array
on the bottom after release (n=8); and (E) those that immediately abandoned the array after

74

release but moved throughout the water column (n=22). Only those in the predefined group A,
that stopped before leaving the array on the first night were evaluated for orientation behavior
and vertical movement patterns although we compared groups A and E in terms of their use of
the water column when moving vertically at day and night as there was an apparent daytime
avoidance of movement in the upper area of the water column. Sea lamprey are nocturnal and the
internal state of the subjects that exited the array during the day immediately after release cannot
be reliably interpreted as searching for the coast or searching for cover (Almeida et al. 2002,
Vrieze et al. 2011, Meckley et al. 2014b). Only considering fish that followed a progression of
settlement and inactivity during the day and activity at night protects our findings from adverse
tagging effects commonly seen immediately after release (Frank et al. 2009).

Statistical analyses
(1) Do Sea lamprey perform y-axis orientation?
To determine if sea lamprey approached the nearest southern coastline, we analyzed
detection of tags at receivers on the coast within 72 hours of exiting the array. This timeframe
ensured that sea lamprey could not have reached a northern coast prior to reaching the nearest
southern coast. Lamprey observed on the southern coast, were classified as successfully
orienting; the null expectation was that absent an orientation mechanism the probability of
success would be 50%.

(2) Are there two phases of movement?
To determine if a transition occurred between two phases of movement and whether
trajectory parameters differed between the postulated phases, we evaluated the paths of 22 sea

75

lamprey from the commencement of movement from a settled state until exiting the region of
VPS array coverage. To classify the two phases of: i) apparent initial undirected (sinuous
movement), followed by ii) directed movement (extensive straight movement), we analyzed the
sinuosity of the path in terms of the progression of the backward beeline distance to total
backward path length at each position along the track to the vanishing point for each sea lamprey
(Bovet and Benhamou 1988, Girard et al. 2004). A broken stick model was fit to each path and
optimized across four parameters including initial slope (𝛽1), break point (𝜏), sigma (σ), and
final slope (𝛽2), using maximum likelihood estimation with function “optim” in program R (R
Development Core Team 2015). In the model backward path length (𝑥) was the explanatory
variable for estimating the response variable backward beeline distance (𝑦). The y intercept (𝛽0 )
was fixed at 0 due to the nature of the parameter always returning to (0, 0). Confidence intervals
were estimated from the Hessian of the log-likelihood. The model used:
𝑦 = 𝑦̂ + ε; ε ~ Normal(0, σ2 )
𝑦̂ = 𝛽1 𝑥1 + 𝛽2 𝑥2 + ε
𝑥1 = {

𝛽(𝑥),
𝜏,

0,
𝑥2 = {
(𝑥 − 𝜏),

𝑥≤𝜏
𝑥>𝜏
𝑥≤𝜏
𝑥>𝜏

When 𝑥 ≤ 𝜏 the model is linear with slope 𝛽1 and when 𝑥 > 𝜏 the model linear with
slope 𝛽2 and intercept constrained by continuity to the first part of the stick. In the special case of
𝛽1 = 𝛽2 the model is a simple linear relationship between x and y (White et al. 2008). The initial
break point parameter provided was estimated through visual inspection of each of the 22 tracks.
Individual pairwise t-tests were used to determine if trajectory parameters that are not directly
associated with track sinuosity (e.g. ground speed, turning bias, average depth, and variation in

76

depth) varied between phase one (before break point) and phase two (after break point). We
predict that if a distinct exploratory phase occurs, the initial behavior will be accompanied by
slower more sinuous movement followed by straighter faster movement.

(3) Is there evidence for orientation towards a consistent cardinal direction?
We tested if there was a significant sample mean direction with a Raleigh test based on
an unspecified mean (function “raleigh.test”, R package circular) or a significant non-uniform
distribution via a Watson’s test (function “watson.test”, R package circular) for the portion of
tracks between the transition point and departure point from the array (Agostinelli and Lund
2013). These tests are similar in nature and answer the same question by either using a z-test
statistic (Raleigh test) or goodness of fit to a circular uniform distribution (Watson’s test). If sea
lamprey orient to a consistent broadly available feature such as visual recognition of the closest
coast, we predicted sea lamprey would have a significantly clustered mean direction and a nonuniform distribution. If sea lamprey were not orienting, oriented to multiple cues at different
times, or oriented to a locally varying feature such as bathymetry in our array, we predicted sea
lamprey would not have a significant mean heading or a uniform distribution in any compass
direction.

(4) Do sea lamprey orient directly to a coastal feature or the bathymetric gradient?
To evaluate if sea lamprey orient to a feature of the closest coast or bathymetry, we
evaluated the heading of each individual during phase 2, defined as the portion of tracks between
the transition point and departure point from the array. We tested if sea lamprey moved towards
the closest coast (coast test) or towards shallower water (bathymetry test). For the bathymetry

77

test, the bearing of the local bathymetric slope was first calculated from the bearing of the
deepest to shallowest depth point on a 1000 point ring of positions in a 100 m radius around the
final point to capture the general slope for each sea lamprey leading to the exit. After all
bathymetric headings were verified visually on a map, the heading was subtracted from the phase
two headings for each individual so that a bathymetry test heading of near 0 o (or 360 o) indicated
that the sea lamprey was moving in the ideal bathymetric direction and 180 o represented moving
the opposite direction. If sea lamprey oriented in the direction of the local bathymetric slope
towards shallower water we predicted the Rayleigh test of the turn angle between fish heading
and shallower slope heading with a specified mean turn angle of 0 would be significant and a
Raleigh test of the fish heading with a specified mean of 172o for the direction to shallower water
would be insignificant.

(5) Did the river the sea lamprey was trapped in influence orientation?
To determine if the lamprey’s capture river influenced orientation (experience), the
distribution of Phase 2 headings between river sources was compared using a Watson-Wheeler
test (“Watson.wheeler.test”, R package circular) and the differences in the distance traveled
during Phase 1 were considered (two sample t-test) (Agostinelli and Lund 2013). We compared
the sea lamprey trapped in the Manistique River and Cheboygan River. If the river they were
trapped in had an effect, we would expect sea lamprey migrants from the Cheboygan River to
reach the nearest coast after a shorter orientation process, and have a different distribution of
directed headings. Although the Watson Wheeler test only tests for a difference between groups,
our expectation was that the phase 2 heading would be more southerly towards the local coast,

78

than the Manistique source subjects. If a sea lamprey relied on local information to orient, we
predicted no effect of river source on orientation.

Characterizing the process: Do lamprey show consistent vertical excursions?
To characterize the vertical movements of sea lamprey through the water column beyond
average depth and standard deviation of depth, vertical occupancy was characterized for each
step (two consecutive points) into one of three classes of vertical movement (ascent, descent, or
horizontal phases) based on pitch (descent < -5o, horizontal ≥ -5o and ≤ 5o, ascent > 5o). Because
sea lamprey move from surface to bottom during each phase, we compared vertical movement
with the time spent moving horizontally within each sinuous phase, the depth of sea lamprey
when moving horizontally in each phase, and the number of casts, defined as the number of
times sea lamprey ascended and descended at least 10 m in succession. These variables were
used to assess whether the vertical movements represented constant and consistent oscillations or
primarily horizontal movements on the surface or bottom with intermittent movements to either
the surface or bottom.
Individual pairwise t-tests for each sea lamprey were checked for differences in depth
between Phase 1 and 2 during horizontal movements (horizontal ≥ -5o and ≤ 5o). To test for
differences in elapsed time spent moving horizontally in the water column during phases 1 and 2
and if a difference existed in the rate of vertical displacement or pitch during ascent and descent,
we used a logistic regression analysis. A single response variable including time moving
horizontally(s), pitch angle (o), or vertical displacement rate (m/s)), respectively. Each was fit
using a generalized linear mixed effects modeling framework. For testing for differences in time
spent moving horizontally, the movement phase was a fixed effect, individual sea lamprey

79

(identification number) was a random effect and we weighted the model by the number of
observations. For testing displacement rates and pitch during ascent and descent, ascent and
descent were fixed effects, while individual sea lamprey was a random effect. The explanatory
variables, pitch and vertical movement rate, were considered positive values, regardless of
whether the sea lamprey was ascending or descending in the water column. Finally the
shallowest depth reached for sea lamprey that showed regular vertical movements during the day
versus at night were tested with a simple two way t-test. This was the only test where data for sea
lamprey that did not stop immediately after release were used. All analyses were performed in R
(R Development Core Team 2015), including the lme4 package (Bates et al. 2014) for fitting the
mixed effects model.

RESULTS
(1) Do Sea lamprey perform y-axis orientation?
Evidence suggested that sea lamprey perform y-axis orientation, although only 22 of 67
released sea lamprey stopped in the array after release and left on the first night (Group A). An
additional five individuals stopped at the edge of the array where only intermittent positions were
attained and nocturnal activity was not recorded (Group B). Prior to stopping at the edge of the
array, two fish made vertical excursions of greater than 10 m prior to stopping, and one fish
moved laterally on the bottom; the remaining individual did not have depth data. Of 10 fish that
did not exit the array on the first night, six were from the eight total Manistee River fish (Group
C). Sea lamprey that did not stop included eight (Group D) individuals that exited the array on
the bottom and 22 that showed vertical movements of at least 10 m off of the bottom (Group E).
Of the 27 sea lamprey (Group A and Group B), 21 (78%) arrived at the nearest coastline within
72 hours and most that were detected at the nearest two receivers were detected on the first night
80

(12 of 15) (Table 3.1). If considering all tagged lamprey, 43 of the 67 (61%) sea lamprey were
detected on the nearest coast (Table 3.1).

(2) Are there two phases of movement?
We observed a post-settlement orientation pattern consistent with initial sinuous
movements (Phase 1) followed by straight movements (Phase 2) for 21 of the 22 sea lamprey
(Figure 3.3, Figure 3.2). The remaining individual immediately departed in an apparent straight
line without an exploratory phase (T26, Figure 3.2). Phase 1 persisted for an average track
length of 808 m ± 178 m (mean ± 1SE) or an interquartile range of 409-837 m (1st IQR-3rd
IQR). Phase 2 lasted 975 m ± 110 m with an Interquartile range of 625-1311 m. Several aspects
of the sea lamprey movement varied between the classified phases, including persistence in
turning in one direction, amount of turning, mean depth, variation in depth, and ground speed
(Table 3.2, Table 3.3). During Phase 1, 10 of the 21total individuals turned with a left-bias while
four individuals were right-biased and the magnitude of persistence in turning of biased
individuals was greater before than after the transition (Table 3.2, Figure 3.4). Sea lamprey
turned less during Phase 2 though this observation could be confounded by the fact that the break
point was defined in terms of the path sinuosity (Table 3.2). Mean depth was shallower after the
break point for 14 of 16 individuals and the standard deviation in depth was significantly greater
for five of 16 individuals vs. significantly reduced for two of 16 individuals, who appeared to
swim near the surface without surface to bottom casting in Phase 2. Lastly, ground speed was
significantly faster for 18 of 21 individuals during phase two when compared to phase one
(Table 3.4).

81

(3) Is there evidence for orientation towards a consistent cardinal direction?
During Phase 2, sea lamprey assumed an average absolute heading of 239o ± 96o (mean ±
2STD, 95 % CI: 201-276 deg.; Figure 3.5). Sea lamprey did not show an absolute orientation in
any cardinal direction (r = 0.24, p = 0.26; Rayleigh Test, unspecified mean) and their circular
distribution was not significantly different from random (test statistic = 0.091, critical value at a
significance of 0.05 = 0.187; Watson’s Test) (Figure 3.6).

(4) Do sea lamprey orient directly to a coastal feature or the bathymetric gradient?
No evidence indicated that sea lamprey oriented towards the nearest coast (r = 0.11, p =
0.22; Rayleigh Test, specified mean: 173 deg.). By contrast, Phase 2 headings (mean: 353 deg.,
95% CI: 307-49 deg., were not randomly distributed with respect to bathymetry (Watson’s Test:
test statistic= 2.244, critical value= 0.187, reject Null) and were significantly oriented to local
bathymetry in the region where they exited the array (r = 0.32, p = 0.017; Rayleigh Test,
specified mean: 0o).

(5) Did the river from which the sea lamprey was trapped influence orientation?
The Watson-Wheeler test of homogeneity of angles found no significant difference
between sea lamprey trapped in different places, although the strength of this conclusion is
limited because of the small sample size (N<10) for each group (W = 0.76, p = 0.69). No
significant difference occurred between the distances traveled prior to the break point for
Cheboygan River fish (mean ± 1SE, 972 m ± 207 m) or Manistique River fish (mean ± 1SE, 976
m ± 244 m) (p=0.99) .

82

Characterizing the process: Do lamprey show consistent vertical excursions?
Sea lamprey moved vertically throughout the entire water column (e.g., Figure 3.7). On
average, they made seven ascents or descents of at least 10 meters (6.7 ± 3.2 dives per h Phase 1,
7.3 ± 4.5 dives per h Phase 2). Sea lamprey maintained a straight course (no turns >15o) for >
500 m during the day, at night, and while moving at the surface, moving on the bottom, and
while ascending and descending through the water column. Sea lamprey ascended at a pitch of
15.0 ± 7.8 and vertical speed of 5.1 ± 2.8 m·min-1 during Phase 1 and a pitch of 13.0 ± 6.8 o and
vertical speed of 5.6 ± 2.5 m·min-1 during Phase 2. Sea lamprey descended at a pitch of 14.9.0 ±
8.0 o and vertical speed of 7.1 ± 7.9 m·min-1 during phase one and a pitch of 14.9 ± 7.7 o and
vertical speed of 8.3 ± 5.2 m·min-1 during phase two. During Phase 1, sea lamprey primarily
moved on the bottom and made occasional vertical excursions to the surface (Figure 3.8). During
Phase 2, sea lamprey made more vertical excursions and spent more time at the surface than the
bottom (Table 3.4). Mixed effects logistic regression modeling revealed that sea lamprey spent
significantly less time moving horizontally during Phase 2 than Phase 1 (estimate: - 0.42, p <
0.01), and more time moving vertically through the water column. Mixed effects logistic
regression revealed that no difference occurred in the pitch of sea lamprey during ascent or
descent (p = 0.08), but a difference did exist in the vertical rate of ascent and descent (p < 0.001).
A two way t-test revealed that vertical excursions during the day did not extend as close to the
surface as at night, rarely entering the upper 10 meters of the water column during the day (p:
<0.01; mean minimum depth ± SD; Day: 10.4 m ± 5.4m, Night: 2.6 m ± 3.6 m).

DISCUSSION
The offshore movement of sea lamprey was consistent with y-axis orientation to the local
bathymetric gradient towards shallower water; we will refer to this class of search as
83

bathokinesis. We addressed this complex hypothesis through a bundle of predictions. (1) The
majority of sea lamprey (78%) arrived at the nearest coast following release. (2) Sea lamprey
(21/22) displayed a two-phase pattern that varied in sinuosity consistent with orientation (Phase
1) followed by directed search (Phase 2). During Phase 1, sea lamprey moved slowly on the
bottom and turned, often in a consistent direction. These tracks included vertical ascents to the
surface and returns to the bottom, building from partial to full excursions through the water
column. During Phase 2, sea lamprey moved more quickly and steered a straight course while
making occasional heading corrections and small looping turns (pirouettes), and moved primarily
at the surface, with occasional descents to the bottom that returned to the surface. (3) Sea
lamprey did not orient in a consistent cardinal direction, signifying a variable search strategy or
use of a varying local feature. (4) Sea lamprey did move perpendicular to the local bathymetric
contours to shallower water. (5) The orientation process and resulting heading was not different
for sea lamprey trapped in different rivers. In contrast to a natal homing migration, the nonhoming sea lamprey migration joins a growing number of organisms that orient to a region of
resource availability with a simple set movement rules.
These findings revealed that sea lamprey oriented consistently with the bathymetric
slope, though the findings did not indicate an orientation mechanism or how sea lamprey
maintained a straight course. We postulated that these sea lamprey: i) used the absolute
hydrostatic pressure sensed at successive sites along the lake bottom to pick a heading towards
shallower water during Phase 1, and ii) used declining hydrostatic pressure at the lake bottom to
steer towards shallower water during Phase 2. Because sea lamprey moved straight during phase
two regardless of their position in the water column, it would be reasonable to infer that sea
lamprey rely on additional directional features to maintain course, rather than relying on a

84

gradient based feature of the environment. We have identified four possible mechanisms for how
sea lamprey follow the bathymetric gradient based on the characteristics of the physical water
column and the nature of sea lamprey movement.
Sea lamprey could orient to a feature of bathymetry when moving at the bottom such as
absolute hydrostatic pressure, a physical aspect of the bottom substrate or bottom pitch, or by
moving vertically through the water column and using a vertically collected feature like dynamic
hydrostatic pressure or the physical extent (e.g., time or distance) traversed through the water
column (Taylor et al. 2010). An internal clock method keeping track of the time or distance
between the surface and bottom was not supported as sea lamprey did not maintain pitch and
vertical movement speed. Vertical speed was significantly faster when descending and pitch was
nearly significant (p=0.08) shallower during ascent. Bottom substrate did not contain features
like sand waves found in shallower water that would inherently indicate the direction to the coast
(Figure 3.9). Fishes have been shown to orient to hydrostatic pressure (Cain 1995, Holbrook and
Burt de Perera 2009, Holbrook and Burt de Perera 2013), and can discern differences in absolute
or differential hydrostatic pressure using the lateral line, inner ear (labyrinth), and swim bladder
(Fraser 2002, Fraser et al. 2008, Bleckmann and Zelick 2009). Sea lamprey lack a swim bladder
and have a primitive lateral line and inner ear that lacks a lagena (otolith endorgan); however,
sea lamprey possess semicircular canals that are structurally similar to the otolith of other
pressure sensitive fishes (Hammond and Whitfield 2006, Fraser et al. 2008, Khorevin 2008),
which can detect changes on the order of 0.5-2.0 kPa (5-20 cm) (Blaxter 1980). We postulate
that sea lamprey use absolute hydrostatic pressure at the maximum depth to orient towards
shallower water through turning and correcting towards shallow water (bathokinesis), and short
looping pirouettes during the directed phase to reaffirm their selected heading. The movement

85

pattern was consistent with kinesis rather than taxis, as initial headings were undirected
(appearing random) and turns were made until movement occurred towards shallower water
(Miller et al. 2009).
The purpose of excursions to the surface is unclear and introduces two challenges to
using hydrostatic pressure to orient; first, sea lamprey must be capable of recalling previous
dives and second they must be able to maintain a course when moving off of the bottom. Some
fishes are capable of remembering information in three-dimensional space to locate targets
(Holbrook and Burt de Perera 2013) and in open-water environments lacking horizontal edges
(walls), the vertical component of the water column may be more informative to navigation
(Holbrook and Burt de Perera 2009, Holbrook and Burt de Perera 2013). Recalling the maximum
hydrostatic pressure would allow for orientation but maintenance of a straight path requires a
second directional feature. Two common navigational mechanisms include reliance on
geomagnetic fields or current direction to maintain course. Whether sea lamprey are capable of
geomagnetic navigation is unknown, though this form of navigation is commonly used by other
fishes that make large migrations and should be tested in this basal vertebrate. Water currents in
this environment are highly variable vertically and horizontally and usually weak (< 2 cm/s)
(Beletsky et al. 1999), making for a potentially poor directional cue in the lake. On the other
hand, current is strongest at the surface (Beletsky and Schwab 2001), a potential reason for
moving to the surface. Combining this with knowledge that sea lamprey do not vertically cast in
deep rivers (16 m depth, Holbrook et al. 2015), but do vertically cast in the presence of river
water (i.e., Meckley et al. 2014b), we propose that vertical movement is designed to encounter
stronger currents, or navigate absent strong currents.

86

Animals could make vertical excursions to improve migration performance, to use
navigation cues located at the surface and bottom, to search for a missing cue stratified in the
water column, or to avoid sensory habituation to a cue (Westerberg 1982, Klimley et al. 2002).
We had no evidence for an improvement in migratory performance as the movement increases
the total path length and sea lamprey do not spend more time falling than rising as would likely
be required to receive an energetic advantage (Carey et al. 1990, Klimley et al. 2002). The
occurrence of vertical excursions in a shallow fully mixed water column (Vrieze et al. 2011,
Meckley et al. 2014b), suggests against the presence of a thermoregulatory advantage (Katz
2002). Sensory habituation can occur if a receptor is constantly exposed to a stimulus and
becomes fatigued resulting in a loss of sensitivity (Ferrari et al. 2010). In theory, vertical
excursions in and out of a cue could prevent sensory habituation. Sea lamprey could encounter
information that is stratified at the surface or when present would reside at the surface (missing
cues). River water can be thermally pinned at the lake surface or bottom of the water column
(Masse and Murthy 1990, Churchill et al. 2003), or mixed throughout the water column, which
led Vrieze et al. 2011) to conclude that sea lamprey vertically cast to encounter olfactory
evidence of a river. This is a plausible explanation, though it is likely that additional navigational
benefits are gained by moving vertically (Meckley et al. 2014b), but from our observations we
cannot report a dominant purpose.
The sea lamprey migration in the Great Lakes appears to rely on a simple set of rules. Sea
lamprey follow the local bathymetry to return to the coast, potentially relying on absolute
hydrostatic pressure. Once arriving at the coast, they move parallel to the coast until
encountering a river plume, and then transition to deflecting along the local coast around the
river mouth to ensure encounter with the river outflow, as they assess or enter the river (Vrieze et

87

al. 2011, Meckley et al. 2014b). The leading hypothesis is that once encountering a river plume
individuals either enter, stage, or search for a different river based on the presence of larval odor
(quality) and river temperature (timing)(Binder et al. 2010, Clemens et al. 2010, Meckley et al.
2012). The odor of river water and larvae from previous generations is attractive to sea lamprey
and is surmised to inform sea lamprey of the presence of a river mouth and the quality of the
spawning and rearing habitat upstream (Moore and Schleen 1980, Bjerselius et al. 2000, Vrieze
and Sorensen 2001). Simple orientation rules based on features of the geophysical environment
that can reliably lead to a target rich region appear to be a common attribute of non-homing
animal search, rather than reliance on direct information associated to a target as is found in
homing migrations.

Future Directions
As commonly occurs in novel studies of animal migration, we identified two hypotheses
that should be tested further and additional features of the sea lamprey migration worth
consideration. First, we postulate that sea lamprey can use gradients of hydrostatic pressure at the
bottom of the water column to determine the bathymetric gradient which provides them the
direction towards shallower water. Second, we postulate that sea lamprey move to the surface to
gather other navigational information and/or to encounter additional cues (e.g. chemical, current)
that could be structurally stratified at some point throughout the water column, while also
preventing habitation to sensory cues in the environment.
Unexpectedly a larger proportion of sea lamprey in this study did not stop immediately
after release and still moved vertically through the water column (33 %), a finding conflicting
with two years of study that found in shallow water (<10 m) in both years only 17 % of sea

88

lamprey didn’t stop and the movement was nearly exclusively on the bottom (unpublished data).
Though in this study, when sea lamprey did move after release they did not move into the upper
10 meters of the water column during the day but moved all the way to the surface at night. It
remains unclear if sea lamprey are more active during daylight hours in deeper water (>10 m).
During the open water migration if sea lamprey avoid the upper water column, the strike range of
the sea lamprey’s most documented predator (avian predators, Close et al. 2002) could be
avoided. It is also possible that this observation was just an artifact of being released during the
day coupled with an inherent aversion to strong light and less inclination to stop immediately in
deeper darker water.
Three other observations of uncertain relevance occurred including the persistent turning
that created loops in a consistent direction during phase one, the observation that few fish went
east after reaching the coast, and the failure of fish from more southerly rivers to exit the array.
The handedness of turning appeared to be dominated by counterclockwise turners, which could
be an artifact of our small data set or exist due to an ecological advantage to handedness.
Handedness in turning during search has been suggested to improve energetic efficiency (e.g.,
Kells and Goulson 2001), making the behavior worth additional consideration. Secondly, fish
appeared less likely to go east after reaching the coast (Table 3.1). This directionality could be
the result of moving opposite the dominant lake current in this region of the lake (Beletsky et al.
1999) or some other bias. The final interesting feature of the study came from the failure of sea
lamprey from more southerly rivers to exit the array on the first night. These individuals did
leave the array after a period of several days. These individuals were caught farther upstream
prior to tagging, and later in the year than the other individuals, and had visibly more developed

89

gonad. It appeared that these individuals required some type of reset period before they began to
migrate out of the array.
Bathokinesis should be examined in ocean-run sea lamprey. Both Pacific lamprey and sea
lamprey have been are neither panmictic nor philopatric (Waldman et al. 2008, Spice et al. 2012,
Hess et al. 2013), suggesting that some aspect of their life cycle constrains them to a region but
also disperses them. Additional evidence of restricted movement exists, as North American and
European populations do not mix (Rodriguez-Munoz et al. 2004, Genner et al. 2012), and move
broadly throughout the ocean when parasitizing other fishes, in some cases moving over 815 km
from the nearest coast (Silva et al. 2014). Additional mechanisms may occur to prevent
navigational mistakes that result in the mixing of populations within the Atlantic basin sea
lamprey or some control over when a parasite will release from a host to prevent large
displacements. If sea lamprey select hosts that migrate back to rivers, it is plausible that sea
lamprey could hitch a ride back to rivers in some instances; a behavior termed phoresy, although
we suspect this is not the dominant strategy. The only evidence available is consistent with
bathokinesis by ocean-run sea lamprey, as parasitic sea lamprey feeding within a general basin
were constrained by basin topography during their return migration (Lanca et al. 2014).

ACKNOWLEDGEMENTS
This research was funded through grants from the Great Lakes Fishery Commission
including by way of the Great Lakes Restoration Initiative appropriations (GL-00E23010-3) and
was accomplished through considerable assistance from the US Fish and Wildlife Service and
the USGS Hammond Bay Biological Station.

90

APPENDIX

91

Table 3.1: Sea lamprey were primarily detected on the nearest coast within 72 hours of release.
This included 78 % of individuals that were observed stopping in the array and moving on the
first night (n=22) and 61% of all individuals released (n=67). Note that Time to reach the
receiver mean and min/max are only listed for the 22 individuals. Only 4 of 67 sea lamprey’s
first detections occurred outside of the nocturnal movement period, with 3 of 4 occurring on the
first day of release with 2 at Forty Mile point (East), 1 at the Ocquoec river mouth (South) and 1
near Cheboygan river mouth (West).
No. of
Receiver/
Distance
Time to reach
Time to reach
unique
detection from array
receiver hours
receiver hours
individuals
Site
(km)
(mean ±SE)
(min, max)
(N=22,
N=67)
OCQ
4.3
(9, 17)
(6.9 ± 5.0)
(0.83 - 46.7)
(12, 26)
BM
6.2
(6, 11)
(17.7 ± 10.7)
(1.7 – 69.3)
(9, 17)
CHB
33.8
(2, 9)
(51.8 ± 0.4)
(51.4 - 52.2)
(4, 15)
SMR
51.6
(2, 5)
(141.0 ± 86.3)
(54.7, 227.3)
(2 , 5)
FMP
11.1
(0, 4)
(1, 5)
PRS
48.9
(0, 0)
(1, 1)
The number of sea lamprey detected on any receiver was 19 of 22 and 46 of 65). Ten of 23 and
19 of 65 sea lamprey entered the Ocquoec River. 1 of 22 and 2 of 65 entered the Black Mallard
River. Four of 23 and seven of 67 sea lamprey were trapped in the Ocqueoc River and one of 67
was trapped in the Cheboygan River.
No. of 1st
Detections in 72
hours
(N=22, N=67)

92

Table 3.2: Individual t-tests for each sea lamprey of whether sea lamprey
exhibited a persistence in turning (theta; -pi to pi), different from 0, revealed
that more sea lamprey had a bias of left (negative) or right (positive) turns
during phase 1 (P1), but not during phase 2 (P2). Most persistent turns were
counterclockwise. The magnitude is also greater for those that are significant
in phase 1 than phase 2. A pairwise t-test of the concentration in turning (Rho)
for each individual between phase 1 and phase 2 revealed that individuals
tended to go straighter during phase 2. Bold values are significant.
Theta

ID
P1
T02
-0.03
T04
-0.06
T05
-0.17
T07
-0.81
T08
0.07
T12
-0.58
T17
0.14
T22
-0.43
T25
0.50
T27
-0.26
T31
-0.47
T32
0.07
T36
-0.58
T37
0.02
T42
0.14
T44
-0.28
T47
0.68
T54
-0.12
T56
-0.13
T61
-0.01
T63
-0.20
Total ( ρ < 0.05):
Total Possible:

P1 ρ

P2

P2 ρ

0.488
0.494
0.214
0.000
0.689
0.000
0.625
0.001
0.001
0.004
0.001
0.792
0.000
0.878
0.082
0.836
0.028
0.107
0.006
0.903
0.005
10
21

-0.08
0.00
0.01
-0.21
0.04
0.06
-0.10
-0.04
-0.07
-0.06
0.14
0.05
0.01
0.33
-0.14
0.03
-0.08
0.01
-0.18
0.06
-0.07

0.04
1.00
0.79
0.29
0.46
0.48
0.48
0.51
0.27
0.67
0.24
0.52
0.88
0.00
0.25
0.61
0.30
0.76
0.01
0.56
0.02
4
21

93

Concentration of turns
(Rho)
P1
P2
ρ
0.81
0.70
0.74
0.74
0.48
0.61
0.54
0.59
0.64
0.61
0.63
0.65
0.67
0.69
-0.28
0.62
0.72
0.69
0.75
0.65

0.83
0.86
0.86
0.65
0.84
0.77
0.75
0.88
0.81
0.75
0.73
0.69
0.75
0.85
0.78
0.79
0.84
0.81
0.72
0.73
0.86

0.61
0.01
0.18
0.47
<0.01
0.18
0.66
<0.01
0.01
0.18
0.36
0.77
0.19
0.12
0.34
0.21
0.14
0.10
0.37
0.82
<0.01
5
21

Table 3.3: Pairwise t-tests to determine if there was a difference in the mean depth, standard deviation in depth, mean depth when

Mean Depth (m)
ID
T02
T04
T05
T07
T08
T12
T17
T22
T25
T27
T31
T32
T36
T37
T42
T44
T47
T54
T56
T61
T63

P1
P2
32.85
17.98
32.69
15.85
29.80
21.59
33.86
23.76
35.07
28.09
21.38
8.04
23.52
14.67
22.72
4.97
32.91
5.99
23.16
7.51
27.31
13.95
30.48
11.31
31.98
31.06
28.93
26.25
15.02
3.84
26.67
5.94
Total Significant (p <
0.05):
Total Possible:

ρ
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.09
<0.01
<0.01
<0.01
<0.01
<0.01
0.13
0.04
<0.01
<0.01

Standard Deviation Depth
(m)
P1
P2
ρ
4.86
9.38
0.00
6.60
9.89
0.08
8.79
11.45
0.38
0.15
8.30
0.00
0.92
6.84
0.00
10.54
6.69
0.14
8.42
10.96
0.62
12.08
3.62
0.00
3.90
3.85
0.95
11.88
6.93
0.03
6.51
11.37
0.13
3.44
9.39
0.00
1.79
5.61
0.00
2.90
3.11
0.87
9.06
2.14
0.06
9.94
7.45
0.28

Hz Mean Depth (m)
P1
35.1
36.8
34.1
33.9
35.3
26.5
32.0
24.2
34.4
25.0
25.8
21.5
32.1
30.2
8.6
-

P2
14.6
7.1
20.3
24.8
27.5
4.4
6.2
3.8
4.8
3.9
4.7
3.4
32.5
26.1
2.3
4.5

Ground Speed (m/s)

ρ
0.05
<0.01
0.16
0.21
0.23
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.12
0.35
0.28
-

P1
0.38
0.38
0.38
0.33
0.29
0.33
0.59
0.30
0.17
0.35
0.34
0.19
0.36
0.40
0.42
0.19
0.56
0.37
0.37
0.46
0.27

P2
0.49
0.48
0.49
0.50
0.51
0.51
0.69
0.52
0.30
0.47
0.57
0.40
0.43
0.44
0.73
0.44
0.69
0.60
0.52
0.51
0.53

ρ
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.08
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.15
<0.01
<0.01
0.08
<0.01
<0.01
<0.01
<0.01

12

7

7

18

16

16

14

21

94

Table 3.3: (cont’d)
maintaining vertical depth (Hz Mean Depth), or ground speed during phase 1(P1) and before the transition to phase 2 (P2). The tests
support our observations that these two phases are different. Sea lamprey had greater mean depths and less variation during phase 1,
and when moving at a particular depth in the water column, sea lamprey spent time on the bottom during phase 1 and closer to the
surface during phase 2. The ground speed (GS) was also significantly lower during phase 1 than phase 2. The only individuals with
less variation in depth during phase two were those individuals swimming mostly at the surface with occasional vertical excursions.
Individuals T42-T56 did not have pressure sensitive tags.

95

Figure 3.1: Adult female sea lamprey were obtained from 5 rivers via barrier traps (Manistee
River (2a), Cheboygan River (1), Manistique River (3), Betsie River (2b), Ocquoec River
(“OCQ”), and parasites were caught attached to fish in the lake. A telemetry system (VPS,
VEMCO) composed of 41 receivers was used to position sea lamprey (“array”). There were
individual receivers that provided detection data along multiple locations on the coast and in
front of as well as in the Black Mallard (BM), Ocquoec (OCQ) and, St. Mary’s Rivers (SM).
Receiver strings were treated as individual detection points and were located near the Cheboygan
River mouth (CHB), 40 mile point (FMP), and Presque Isle (PRS).
96

Figure 3.2: Each sea lamprey was released near the bottom in the middle (yellow star) of a 43
receiver array (orange dots) with 3 km2 coverage. A receiver in the center of the array at the
release point and on the east side of the array were not recovered, resulting in an irregular design.
The black arrow demonstrates the direction to the nearest coastline 172o (3 radians).

97

Figure 3.3: The track of sea lamprey T04 is depicted as it exits the array (a), and below is the
relationship between backward path length and backward beeline distance of the track (b). An
arrow is drawn from the breakpoint to the end of the phase two to demonstrate the direction
headed during Phase 2 (a). In the lower graph the yellow line is a one to one relationship from 0.
A broken stick model fit is shown with the red lines representing the standard error and the black
line is the line fit (b). The break point separates phase one (orange) from phase two (blue). The
22 tracks of Group A are shown in Figure 3.10.
98

Figure 3.4: Dotted lines connect phase 1 (orange) and phase 2 (blue), data points for each sea
lamprey subject, separated by a transition characterized by backward path length versus
backward beeline distance sinuosity. Both the mean value is shown (dot) as well as the standard
deviation in each axis (+). There is an increase in ground speed and a transition from a turning
bias in one direction to no bias in turning direction (a), as turning persistence overlaps 0 during
phase 2. Also apparent was a change in the mean depth from mostly movement on the bottom in
phase one with brief vertical excursions to the surface followed by movement on the surface
during phase two and brief vertical excursions to the bottom (b).

99

Figure 3.5: The overall heading of 22 sea lamprey during phase two is depicted as a black dot
within one of 60 bins separating the circle and shows an average heading of 239o (gray arrow)
with a 95% confidence interval of 201-276o (shaded area) (a). The heading of sea lamprey with
respect to the ideal heading if following local bathymetry at the exit of the array is depicted by
the gray arrow with most sea lamprey clustered near 0 (mean= -7, 95% CI: 307-49o), repsenting
a movement towards shallower water, while a few individuals were observed moving opposite of
the ideal bathyemtric controur. The phase two headings were significantly oriented to local
bathymetry in the region where they exited the array (r = 0.32, p = 0.017; Rayleigh Test,
specified mean: 0o).

100

Figure 3.6: Sea lamprey (n=22) settled in the array during the day and began moving at
night (yellow), transitioning (green circle) between phase one (red arrow) and phase two
(black arrow) before they exited the array. Sea lamprey that moved north (b), east (c),
south (d), and west (e) are shown in individual frames with symbols indicating where
each individual was first detected on the coast. The array is color coded by depth 2039(m).

101

Figure 3.7: The track of sea lamprey T04 is depicted as it exited the array (a), and below is the
water column depth (b) and ground speed during the track (c). The graphs are color coded by 15
minute intervals. In inset “a” small circles represent receiver positions and the small colored dots
represent fish positions with a line showing the path. In inset “b” the red line represents the total
water column depth and the squares show the depth of the fish with respect to the total water
column depth through time.. All 22 tracks are shown in Figure 3.10.
102

Figure 3.8: Overall most observations occurred near the bottom (All observations; left panel).
During phase 1 most sea lamprey moved on the bottom and made occasional excursions to the
surface that returned to the bottom (center panel). Phase 2 consisted of more surface movements
with occasional excursions to the bottom, although a few individuals did move vertically and
only moved on the bottom (right panel).

103

Figure 3.9: Near shore sand waves in shallow water of Lake Huron (a) versus offshore detritus
and algae covered bottom without clear features associated to a common direction (b).

104

Figure 3.10: There are 2 pages dedicated to each subject and only the data is shown for the
points when fish began to move following settlement, with the exception of transmitter 21 that
was shown moving immediately after release. The two pages correspond to the pathways of 22
subjects. Page one, “Movement Summary”, depicts the path of each sea lamprey through the
array when active (a) and is color coded by 15 minutes intervals corresponding to the lower
graph (b,c). Figure 3.10: (cont’d) Time is in Eastern Standard Time (UTC-4) and shown in a 24
105

Figure 3.10: (cont’d)
hour time scale. The lower graph depicts the vertical movements of sea lamprey for those
individuals that had transmitters with pressure sensors corresponding to the left y axis (b) and
ground speed corresponding to the right y axis for all individuals (c). The red line is the total
water column depth at each fish position. Note only T21 is shown moving during the day as it
was notable because it appears to orient during the day and then repeated the process

106

Figure 3.10: (cont’d)
at night but the night time track was too short to consider in the movement classification. This
suggests that this is a daily orientation process. Page 2, “Movement Classification”, the lower
graph depicts the backward path length versus the backward beeline distance for each track. The
point at (0, 0) represents the point that the subject left the array. The yellow line is a one to one

107

Figure 3.10: (cont’d)
relationship from 0. A perfectly straight path from the exit would move along this yellow line.
The broken stick model fit is shown with the black line and the red lines represent the standard
error. The break point is the point we use to separate phase one from phase two. Phase one is
highlighted in the in orange and phase two is depicted in blue. In the upper graph, the path of
each individual is shown and an arrow is drawn from the breakpoint to the end of phase two

108

Figure 3.10: (cont’d)
when the fish exits the array coverage to demonstrate the heading of phase two that was used to
test against the direction to shallower water and the direction to the nearest coast using Raleigh
tests.

109

Figure 3.10: (cont’d)

110

Figure 3.10: (cont’d)

111

Figure 3.10: (cont’d)

112

Figure 3.10: (cont’d)

113

Figure 3.10: (cont’d)

114

Figure 3.10: (cont’d)

115

Figure 3.10: (cont’d)

116

Figure 3.10: (cont’d)

117

Figure 3.10: (cont’d)

118

Figure 3.10: (cont’d)

119

Figure 3.10: (cont’d)

120

Figure 3.10: (cont’d)

121

Figure 3.10: (cont’d)

122

Figure 3.10: (cont’d)

123

Figure 3.10: (cont’d)

124

Figure 3.10: (cont’d)

125

Figure 3.10: (cont’d)

126

Figure 3.10: (cont’d)

127

Figure 3.10: (cont’d)

128

Figure 3.10: (cont’d)

129

Figure 3.10: (cont’d)

130

Figure 3.10: (cont’d)

131

Figure 3.10: (cont’d)

132

Figure 3.10: (cont’d)

133

Figure 3.10: (cont’d)

134

Figure 3.10: (cont’d)

135

Figure 3.10: (cont’d)

136

Figure 3.10: (cont’d)

137

Figure 3.10: (cont’d)

138

Figure 3.10: (cont’d)

139

Figure 3.10: (cont’d)

140

Figure 3.10: (cont’d)

141

Figure 3.10: (cont’d)

142

Figure 3.10: (cont’d)

143

Figure 3.10: (cont’d)

144

Figure 3.10: (cont’d)

145

Figure 3.10: (cont’d)

146

Figure 3.10: (cont’d)

147

Figure 3.10: (cont’d)

148

REFERENCES

149

REFERENCES

Able, K. P. (1991). "Common Themes and Variations in Animal Orientation Systems."
American Zoologist 31(1): 157-167.
Agostinelli, C. and U. Lund (2013). R package 'circular':circular statistics. R package version
0.4-7.
Åkesson, S. and A. Hedenström (2007). "How Migrants Get There: Migratory Performance and
Orientation." BioScience 57(2): 123-133.
Akima, H. (1978). "A Method of Bivariate Interpolation and Smooth Surface Fitting for
Irregularly Distributed Data Points." ACM Trans. Math. Softw. 4(2): 148-159.
Almeida, P. R., B. R. Quintella and N. M. Dias (2002). "Movement of radio-tagged anadromous
sea lamprey during the spawning migration in the River Mondego (Portugal)." Hydrobiologia
483(1-3): 1-8.
Bates, D., M. Maechler, B. M. Bolker and S. Walker (2014). lme4: Linear mixed-effects models
using Eigen and S4. R package version 1.1-7.
Beletsky, D., J. H. Saylor and D. J. Schwab (1999). "Mean Circulation in the Great Lakes."
Journal of Great Lakes Research 25(1): 78-93.
Beletsky, D. and D. J. Schwab (2001). "Modeling circulation and thermal structure in Lake
Michigan: Annual cycle and interannual variability." Journal of Geophysical Research: Oceans
106(C9): 19745-19771.
Benhamou, S. (2006). "Detecting an orientation component in animal paths when preferred
direction is individual-dependent." Ecology 87(2): 518-528.
Bergstedt, R. A. and J. G. Seelye (1995). "Evidence for Lack of Homing by Sea Lampreys."
Transactions of the American Fisheries Society 124(2): 235-239.
Biesinger, Z., B. Bolker, D. Marcinek, T. Grothues, J. Dobarro and W. Lindberg (2013).
"Testing an autonomous acoustic telemetry positioning system for fine-scale space use in marine
animals." J Exp Mar Biol Ecol 448: 46 - 56.
150

Binder, T. R. and D. G. McDonald (2007). "Is there a role for vision in the behaviour of sea
lampreys (Petromyzon marinus) during their upstream spawning migration?" Canadian Journal
of Fisheries and Aquatic Sciences 64(10): 1403-1412.
Binder, T. R., R. L. McLaughlin and D. G. McDonald (2010). "Relative Importance of Water
Temperature, Water Level, and Lunar Cycle to Migratory Activity in Spawning-Phase Sea
Lampreys in Lake Ontario." Transactions of the American Fisheries Society 139(3): 700-712.
Bjerselius, R., W. Li, J. H. Teeter, J. G. Seelye, P. B. Johnson, P. J. Maniak, G. C. Grant, C. N.
Polkinghorne and P. W. Sorenson (2000). "Direct behavioral evidence that unique bile acids
released by larval sea lamprey (Petromyzon marinus) function as a migratory pheromone."
Canadian Journal of Fisheries and Aquatic Sciences 57: 557-569.
Blaxter, J. H. S. (1980). The Effect of Hydrostatic Pressure on Fishes. Environmental Physiology
of Fishes. M. A. Ali, Springer US. 35: 369-386.
Bleckmann, H. and R. Zelick (2009). "Lateral line system of fish." Integrative Zoology 4(1): 1325.
Bovet, P. and S. Benhamou (1988). "Spatial-analysis of animal movements using a correlated
random-walk model " Journal of Theoretical Biology 131(4): 419-433.
Cain, P. (1995). "Navigation in Familiar Environments by the Weakly Electric Elephantnose
Fish, Gnathonemus petersii L. (Mormyriformes, Teleostei)." Ethology 99(4): 332-349.
Calenge, C. (2006). "The package "adehabitat" for the R software: A tool for the analysis of
space and habitat use by animals." Ecological Modelling 197(3-4): 516-519.
Carey, F. G., J. V. Scharold and A. J. Kalmijn (1990). "Movements of blue sharks (Prionace
glauca) in depth and course." Marine Biology 106(3): 329-342.
Churchill, J. H., E. A. Ralph, A. M. Cates, J. W. Budd and N. R. Urban (2003). "Observations of
a negatively buoyant river plume in a large lake." Limnology and Oceanography 48(2): 884-894.
Clemens, B. J., T. R. Binder, M. F. Docker, M. L. Moser and S. A. Sower (2010). "Similarities,
Differences, and Unknowns in Biology and Management of Three Parasitic Lampreys of North
America." Fisheries 35(12): 580-594.

151

Close, D. A., M. S. Fitzpatrick and H. W. Li (2002). "The ecological and cultural importance of
a species at risk of extinction, Pacific lamprey." Fisheries 27(7): 19-25.
Cooke, S., J. Midwood, J. Thiem, P. Klimley, M. Lucas, E. Thorstad, J. Eiler, C. Holbrook and
B. Ebner (2013). "Tracking animals in freshwater with electronic tags: past, present and future."
Anim Biotelem 1: 5.
Craig, P. C. (1973). "Orientation of the sand-beach amphipod, ." Animal Behaviour 21(4): 699706.
Dittman, A. and T. Quinn (1996). "Homing in Pacific salmon: mechanisms and ecological
basis." The Journal of Experimental Biology 199(1): 83-91.
Espinoza, M., T. Farrugia, D. Webber, F. Smith and C. Lowe (2011). "Testing a new acoustic
telemetry technique to quantify long-term, fine-scale movements of aquatic animals." Fish Res
108(2): 364 - 371.
Ferguson, D. E. (1971). "Sensory basis of orientation in amphibians." Annals of the New York
Academy of Sciences 188(NDEC): 30-&.
Ferguson, D. E., H. F. Landreth and M. R. Turnipseed (1965). "Astronomical Orientation of the
Southern Cricket Frog, Acris gryllus." Copeia 1: 58-66.
Ferrari, M. C. O., C. K. Elvidge, C. D. Jackson, D. P. Chivers and G. E. Brown (2010). "The
responses of prey fish to temporal variation in predation risk: sensory habituation or risk
assessment?" Behavioral Ecology.
Frank, H. J., M. E. Mather, J. M. Smith, R. M. Muth, J. T. Finn and S. D. McCormick (2009).
"What is "fallback"?: metrics needed to assess telemetry tag effects on anadromous fish
behavior." Hydrobiologia 635(1): 237-249.
Fraser, E. J., and Stacey, N. E. (2002). "Isolation increases milt production in goldifish." Journal
of Experimental Biology 293: 511-524.
Fraser, P. J., S. F. Cruickshank, R. L. Shelmerdine and L. E. Smith (2008). "Hydrostatic Pressure
Receptors and Depth Usage in Crustacea and Fish." Navigation 55(2): 159-165.

152

Gebhardt, A., T. Petzoldt and M. Maechler (2013). akima: Interpolation of irregularly spaced
data. R Package Version 0.5-11.
Genner, M. J., R. Hillman, M. McHugh, S. J. Hawkins and M. C. Lucas (2012). "Contrasting
demographic histories of European and North American sea lamprey (Petromyzon marinus)
populations inferred from mitochondrial DNA sequence variation." Marine and Freshwater
Research 63(9): 827-833.
Girard, C., S. Benhamou and L. Dagorn (2004). "FAD: Fish Aggregating Device or Fish
Attracting Device? A new analysis of yellowfin tuna movements around floating objects."
Animal Behaviour 67: 319-326.
Gurarie, E., C. Bracis, M. Delgado, T. D. Meckley, I. Kojola and C. M. Wagner (2015). "A
comparison of methods and practical guide to the behavioral analysis of animal movements."
Journal of Animal Ecology Special Issue.
Hammond, K. L. and T. T. Whitfield (2006). "The developing lamprey ear closely resembles the
zebrafish otic vesicle: otx1 expression can account for all major patterning differences."
Development 133(7): 1347-1357.
Hess, J. E., N. R. Campbell, D. A. Close, M. F. Docker and S. R. Narum (2013). "Population
genomics of Pacific lamprey: adaptive variation in a highly dispersive species." Molecular
Ecology 22(11): 2898-2916.
Holbrook, C. M., R. Bergstedt, N. S. Adams, T. W. Hatton and R. L. McLaughlin (2015). "FineScale Pathways Used By Adult Sea Lampreys during Riverine Spawning Migrations."
Transactions of the American Fisheries Society 144(3): 549-562.
Holbrook, R. I. and T. Burt de Perera (2009). "Separate encoding of vertical and horizontal
components of space during orientation in fish." Animal Behaviour 78(2): 241-245.
Holbrook, R. I. and T. Burt de Perera (2013). "Three-dimensional spatial cognition: freely
swimming fish accurately learn and remember metric information in a volume." Animal
Behaviour 86(5): 1077-1083.
Holland, R. A., I. Borissov and B. M. Siemers (2010). "A nocturnal mammal, the greater mouseeared bat, calibrates a magnetic compass by the sun." Proceedings of the National Academy of
Sciences 107(15): 6941-6945.

153

Jonsen, I. D., R. A. Myers and J. M. Flemming (2003). "Meta-analysis of animal movement
using state-space models." Ecology 84(11): 3055-3063.
Katz, S. L. (2002). "Design of heterothermic muscle in fish." Journal of Experimental Biology
205(15): 2251-2266.
Keefer, M. L., C. C. Caudill, C. A. Peery and M. L. Moser (2013). "Context-dependent diel
behavior of upstream-migrating anadromous fishes." Environmental Biology of Fishes 96(6):
691-700.
Kells, A. and D. Goulson (2001). "Evidence for Handedness in Bumblebees." Journal of Insect
Behavior 14(1): 47-55.
Khorevin, V. I. (2008). "The lagena (the third otolith endorgan in vertebrates)." Neurophysiology
40(2): 142-159.
Klimley, P. A., S. C. Beavers, T. H. Curtis and S. J. Jorgensen (2002). "Movements and
Swimming Behavior of Three Species of Sharks in La Jolla Canyon, California." Environmental
Biology of Fishes 63(2): 117-135.
Lanca, M. J., M. Machado, C. S. Mateus, M. Lourenco, A. F. Ferreira, B. R. Quintella and P. R.
Almeida (2014). "Investigating Population Structure of Sea Lamprey (Petromyzon marinus, L.)
in Western Iberian Peninsula Using Morphological Characters and Heart Fatty Acid Signature
Analyses." Plos One 9(9): 14.
Landreth, H. F. and D. E. Ferguson (1967). "Movements and orientation of the tailed frog,
Ascaphus Truei." Herpetologica 23(2): 81-93.
Lohmann, K. J., C. M. F. Lohmann and C. S. Endres (2008). "The sensory ecology of ocean
navigation." Journal of Experimental Biology 211(11): 1719-1728.
Luschi, P. (2013). "Long-Distance Animal Migrations in the Oceanic Environment: Orientation
and Navigation Correlates." ISRN Zoology 2013: 23.
Luschi, P., S. Åkesson, A. Broderick, F. Glen, B. Godley, F. Papi and G. Hays (2001). "Testing
the navigational abilities of ocean migrants: displacement experiments on green sea turtles
(Chelonia mydas)." Behavioral Ecology and Sociobiology 50(6): 528-534.

154

Masse, A. K. and C. R. Murthy (1990). "Observations of the Niagara River thermal plume (Lake
Ontario, North America)." Journal of Geophysical Research: Oceans 95(C9): 16097-16109.
Meckley, T., C. Holbrook, C. Wagner and T. Binder (2014a). "An approach for filtering
hyperbolically positioned underwater acoustic telemetry data with position precision estimates."
Animal Biotelemetry 2(1): 7.
Meckley, T., C. Wagner and E. Gurarie (2014b). "Coastal movements of migrating sea lamprey
(Petromyzon marinus) in response to a partial pheromone added to river water: implications for
management of invasive populations." Can J Fish Aquat Sci 71(4): 533 - 544.
Meckley, T. D., C. M. Wagner and M. A. Luehring (2012). "Field evaluation of larval odor and
mixtures of synthetic pheromone components for attracting migrating sea lampreys in rivers."
Journal of Chemical Ecology 38(8): 1062-1069.
Miller, J. R., P. Y. Siegert, F. A. Amimo and E. D. Walker (2009). "Designation of Chemicals in
Terms of the Locomotor Responses They Elicit From Insects: An Update of Dethier et al.
(1960)." Journal of Economic Entomology 102(6): 2056-2060.
Moore, H. H. and I. P. Schleen (1980). "Changes in spawning runs of Sea Lamprey
(Petromyzon-Marinus) in selected streams of Lake Superior after chemical control." Canadian
Journal of Fisheries and Aquatic Sciences 37(11): 1851-1860.
Mouritsen, H., R. Derbyshire, J. Stalleicken, O. Ø. Mouritsen, B. J. Frost and D. R. Norris
(2013). "An experimental displacement and over 50 years of tag-recoveries show that monarch
butterflies are not true navigators." Proceedings of the National Academy of Sciences 110(18):
7348-7353.
Mueller, T. and W. F. Fagan (2008). "Search and navigation in dynamic environments - from
individual behaviors to population distributions." Oikos(117): 654-664.
Nathan, R., W. M. Getz, E. Revilla, M. Holyoak, R. Kadmon, D. Saltz and P. E. Smouse (2008).
"A movement ecology paradigm for unifying organismal movement research." Proceedings of
the National Academy of Sciences 105(49): 19052-19059.
Pe'er, G., D. Saltz, H.-H. Thulke and U. Motro (2004). "Response to topography in a hilltopping
butterfly and implications for modelling nonrandom dispersal." Animal Behaviour 68(4): 825839.

155

Putman, N. F., E. S. Jenkins, C. G. J. Michielsens and D. L. G. Noakes (2014). Geomagnetic
imprinting predicts spatio-temporal variation in homing migration of pink and sockeye salmon.
R Development Core Team (2015). R: A language and environment for statistical computing.
Vienna, Austria, R Foundation for Statistical Computing.
Rodriguez-Munoz, R., J. R. Waldman, C. Grunwald, N. K. Roy and I. Wirgin (2004). "Absence
of shared mitochondrial DNA haplotypes between sea lamprey from North American and
Spanish rivers." Journal of Fish Biology 64(3): 783-787.
Roy, R., J. Beguin, C. Argillier, L. Tissot, F. Smith, S. Smedbol and E. De-Oliveira (2014).
"Testing the VEMCO Positioning System: spatial distribution of the probability of location and
the positioning error in a reservoir." Anim Biotelm 2: 1.
Scapini, F., M. Fallaci and M. C. Mezzetti (1996). "Orientation and migration in Talitrus
saltator." Revista Chilena de Historia Natural 69: 553-563.
Silva, S., M. J. Araújo, M. Bao, G. Mucientes and F. Cobo (2014). "The haematophagous
feeding stage of anadromous populations of sea lamprey Petromyzon marinus: low host
selectivity and wide range of habitats." Hydrobiologia 734(1): 187-199.
Smith, F. (2013). "Understanding HPE in the VEMCO positioning system (VPS)." Halifax, NS:
VEMCO.
Sorensen, P. W., J. M. Fine, V. Dvornikovs, C. S. Jeffrey, F. Shao, J. Z. Wang, L. A. Vrieze, K.
R. Anderson and T. R. Hoye (2005). "Mixture of new sulfated steroids functions as a migratory
pheromone in the sea lamprey." Nature Chemical Biology 1(6): 324-328.
Spice, E. K., D. H. Goodman, S. B. Reid and M. F. Docker (2012). "Neither philopatric nor
panmictic: microsatellite and mtDNA evidence suggests lack of natal homing but limits to
dispersal in Pacific lamprey." Molecular Ecology 21(12): 2916-2930.
Taylor, G. K., R. I. Holbrook and T. B. de Perera (2010). Fractional rate of change of swimbladder volume is reliably related to absolute depth during vertical displacements in teleost fish.
Teeter, J. H. (1980). "Pheromone communication in sea lampreys (Petromyzon marinus):
implications for population management." Canadian Journal of Fisheries and Aquatic Sciences
37: 2123-2132.

156

Ueda, H., M. Kaeriyama, K. Mukasa, A. Urano, H. Kudo, T. Shoji, Y. Tokumitsu, K. Yamauchi
and K. Kurihara (1998). "Lacustrine sockeye salmon return straight to their natal area from open
water using both visual and olfactory cues." Chemical Senses 23(2): 207-212.
Vrieze, L., R. Bergstedt and P. Sorensen (2011). "Olfactory-mediated stream-finding behavior of
migratory adult sea lamprey (Petromyzon marinues)." Can J Fish Aquat Sci 68: 523 - 533.
Vrieze, L. A., R. Bjerselius and P. W. Sorensen (2010). "Importance of the olfactory sense to
migratory sea lampreys Petromyzon marinus seeking riverine spawning habitat." Journal of Fish
Biology 76(4): 949-964.
Vrieze, L. A. and P. W. Sorensen (2001). "Laboratory assessment of the role of a larval
pheromone and natural stream odor in spawning stream localization by migratory sea lamprey
(Petromyzon marinus)." Canadian Journal of Fisheries and Aquatic Sciences 58(12): 2374-2385.
Wagner, C. M., M. L. Jones, M. B. Twohey and P. W. Sorensen (2006). "A field test verifies that
pheromones can be useful for sea lamprey (Petromyzon marinus) control in the Great Lakes."
Canadian Journal of Fisheries and Aquatic Sciences 63(3): 475-479.
Waldman, J., C. Grunwald and I. Wirgin (2008). Sea lamprey Petromyzon marinus: an exception
to the rule of homing in anadromous fishes.
Walker, M. M., T. E. Dennis and J. L. Kirschvink (2002). "The magnetic sense and its use in
long-distance navigation by animals." Current Opinion in Neurobiology 12(6): 735-744.
Wendler, G. and H. Scharstein (1986). "The orientation of grain weevils Sitophilus Granarius influence of spontaneous turning tendencies and of gravitational stimuli " Journal of
Comparative Physiology a-Sensory Neural and Behavioral Physiology 159(3): 377-389.
Westerberg, H. (1982). "Ultrasonic tracking of Atlantic salmon (Salmo salar L.), 2: Swimming
depth and temperature stratification." Report-Institute of Freshwater Research, Drottningholm
(Sweden).

157

CHAPTER 4

DOES LARVAL SEA LAMPREY ODOR AID NAVIGATION OR GUIDE HABITAT
SELECTION DECISIONS BY ADULTS

ABSTRACT

The ways in which non-homing fishes, such as sea lamprey (Petromyzon marinus), locate
and select reproductive habitat is poorly understood. Prior studies indicated the presence of
larval odor in river water increased the likelihood that a sea lamprey entered a river. However, it
was not known whether larval odor played a role in navigation (guiding the migrant to the river
mouth) or mediated habitat selection by labeling the suitability of a river for spawning. Fixed
acoustic telemetry captured sea lamprey movement and a hydrodynamic model coupled with a
dye concentration model provided estimates of the Ocqueoc River plume position through time.
This allowed for the reconstruction of the hydrodynamic experience of sea lamprey as the
approached the coast, encountered river water, and entered a river in the Great Lakes. Once
reaching a coast and encountering river water, sea lamprey deflected along the physical coastal
edge within the river water labeled region of the lake and approached the river, consistent with a
thigmotactic search mechanism rather than a chemotactic search mechanism, where gradients of
odorants are important. These observations compliment other evidence to suggest that encounter
with river water absent larval odor is sufficient for migrants to transition to local search, while
larval odor guides the river entry decision. The river entry decision may have occurred before
entry, though it was never reversed following entry. Three synthesized components found in the
full larval odor had no apparent effect on the likelihood of river entry when compared to the
entry rate previously observed in the Ocquoec River between 1 and 2 years after pesticide based

158

removal of larvae. These known components of larval odor appeared insufficient to bias river
entrance; however, the results suggested that the use of the entire larval odor could be a powerful
management tool, capable of manipulating the river entry decision.

INTRODUCTION
Animals can use physical and chemical information to find and select reproductive
habitat. Olfactory signals associated with a resource can label directional information that leads
to a target and remotely inform the searcher of the profitability of the target (Vickers 2000).
When studying animal search, determining whether a particular odor is important to navigation
to a target or assessment of target quality is difficult. Sea lamprey adults do not home and rely on
the odor of larval sea lamprey from previous generations residing in a stream to select a river
(Morman et al. 1980, Teeter 1980, Bergstedt and Seelye 1995, Waldman et al. 2008). Once in a
river, sea lamprey are hypothesized to search using odor mediated rheotaxis to locate habitat and
mates (Bjerselius et al. 2000, Vrieze et al. 2011, Johnson et al. 2012, Meckley et al. 2014b).
However, it is unclear how sea lamprey search in a river plume, defined as the dynamic region of
a lake where river water mixes with lake water, or how they enter rivers. Lab tests designed to
mimic a river plume demonstrated that sea lamprey are attracted to river water lacking larval
odor (Vrieze and Sorensen 2001). This opens the question of whether larval odor or river water
plays a role in the decision to enter a river or whether larval odor is important to both river
selection and navigation in a river plume.
The sea lamprey migration in the lake is just beginning to be described. On the coast in
the Great Lakes, sea lamprey display extensive movements parallel to the coast, typified by
straight movements while completing vertical excursions from the surface to the bottom of the
water column until encountering a river plume (Vrieze et al. 2011, Meckley et al. 2014b), during
159

their non-homing migration (Bergstedt and Seelye 1995, Waldman et al. 2008). The vertical
movement is hypothesized to improve encounter with migration cues found in stratified layers of
the water column and may be involved in orientation (Vrieze et al. 2011, Meckley et al. 2014b,
Chapter 3). Upon encountering a river plume sea lamprey transitioned to intensive movements
typified by horizontal turning and continued vertical excursions, as well as regular deflections
off the coast in the area of the river plume, as individuals moved through the river plume and
localized the river mouth (Meckley et al. 2014b).
A general characterization exists of how sea lamprey navigate in rivers and move along
the physical coastal edge to encounter river plumes (Vrieze et al. 2011, Meckley et al. 2014b),
though how sea lamprey move in river plumes and the functional role of larval odor remains
unclear due to the limitations of the tracking equipment of earlier studies. Some have suggested
sea lamprey search in river plumes through kinesis (e.g., Vrieze et al. 2011), an undirected search
absent directional information, while others have suggested a directed search mechanism reliant
on physical edges (e.g., Thigmotaxis, Meckley et al. 2014b). Larval odor could increase selection
of a river by labeling good habitat that guides decision making or by playing a role in navigation,
allowing sea lamprey to localize the river mouth more easily. Larval odor influences the number
of sea lamprey to enter a river, demonstrated by the large reduction in sea lamprey trapped in
rivers the year after mass removal of lamprey larvae (Moore and Schleen 1980).
Olfaction is thought to be integral to sea lamprey finding and entering rivers in the Great
Lakes (Vrieze et al. 2010). Reliance on odor for habitat selection sets up the potential to switch
from traditional management (e.g., dams, pesticide) to odor based manipulation of migration
routes, guiding sea lamprey into habitat that is easy to manage or is unfit for larval survival (Li et
al. 2007). Manipulation of the invasive sea lamprey’s habitat selection with the use of larval odor

160

has been greatly anticipated as an integrated pest management solution, although it requires
identification and synthesis of the components in larval odor to be able to activate rivers
(Sorensen et al. 2003, Johnson et al. 2006, Hoye et al. 2007). To our knowledge, only three
components of the larval odor have been identified and synthesized (Hoye et al. 2007). Field
experiments indicate that sea lampreys can detect and respond to incomplete odors to facilitate
the search for reproductive opportunities (Fine and Sorensen 2008, Meckley et al. 2014b), but
ultimately require encounter with a complete composite of the components that represent a signal
prior to selecting spawning habitat (Wagner et al. 2009, Meckley et al. 2012, Meckley et al.
2014a) or a mate (Johnson et al. 2006, Johnson et al. 2009, Luehring et al. 2011). The river
mouth would be the most likely location for a partial odor to inform a decision, as the complete
odor could be the most dilute rendering the most concentrated components the most reliable
indicators of larval presence (Meckley et al. 2012). If a collection of partial odors fails to elicit
entry at the river mouth, it would indicate the need for identification and synthesis of additional
compounds found in larval odor.
Here, we report a field experiment that evaluated whether larval odor influences the riverentry decision or whether the cue is important to navigation. Secondarily we provided a test of
whether the addition of synthetic larval odor increases entry beyond what is expected two years
after TFM treatment. During a two-year study in 2010 and 2011, we documented the three
dimensional movements of acoustically tagged sea lamprey as they approached, entered, or
bypassed a river under one of two conditions: (1) low larval odor, following larval removal with
a pesticide (TFM), below (2008) and above a barrier (2009) (2010); and, (2) higher larval odor,
created by larval recruitment, plus synthetic larval odor components Petromyzonamine disulfate
(PADS), petromyzosterol disulfate (PSDS), and petromyzonol sulfate (PZS) to a 1 x 10 -12 M

161

concentration (2011). A coupled hydrodynamic model and dye concentration model provided
environmental information including water current direction and speed, water temperature, and
the presence of river water at each fish position for the entire study period allowing us to
characterize the sea lamprey migration experience with unprecedented precision. We specifically
tested three hypotheses. First, sea lamprey will encounter the river plume at the same rate
between years, although more will enter in 2011 due to a higher larval odor concentration a year
after any TFM treatment in the Ocqueoc River and following natural recruitment of larvae (H1).
Second, higher larval odor will not increase how close sea lamprey come to the river mouth but
will influence how sea lamprey move in the river plume for those that encounter river water
(H2). Third, the addition of the known synthetic odor components of the full larval odor
composition will be insufficient to increase entry rate beyond the expected river entry rate (H3).
The expected river entry rate was based on the increase historically observed between one and
two years after pesticide treatment in the Ocqueoc River (H3).
METHODS
General Methods
Acoustically tagged sea lamprey were released at three staggered locations 700 m north
of the Ocquoec River (45.490278°, −84.072931°) in Hammond Bay (Lake Huron) at the center
of a VEMCO positioning array capable of providing 3-D positions every 30 seconds ( Figure
4.1). Sea lampreys were released over two different spawning seasons (N=72, 2010; N=79,
2011) under two conditions including an assumed lower (2010) and higher (2011) larval
population condition. Prior to the first year of the study (April 2010) both the lower (6 October
2008) and upper (22 August 2009) sections of the Ocqueoc river, separated by a dam, were
treated with the pesticide 3-trifluoromethyl-4-nitrophenol (TFM). TFM treatment vastly reduced
162

the population in the Ocquoec River, although the Ocquoec River still receives a reduced but
reliable sea lamprey run the year after treatment, as treatment is required every four years prior
to larvae transforming to parasites and out migrating from the river to the lake (Aaron Jubar, US
Fish and Wildlife Service, personal communication).
Study Site
The Ocqueoc river had an annual spring median discharge during trials of 2.64 m3·s-1 in
2010 and 3.9 m3·s-1 in 2011 (range: 0.6-9.7 m3·s-1 2010, 1.8-8.6m3·s-1 2011). A telemetry system
(VPS, VEMCO) with a similar configuration of acoustic receivers (VR2W) and coverage in
2010 and 2011 was composed of 41 receivers located in a 0.90 km2 in 2010 and 37 receivers
covering 1.57 km2 in 2011 and centered on the mouth of the Ocqueoc River (45.490278°,
−84.072931°). The receiver configuration was in diamond formations with between receiver
spacing ranging from 75 m to 250 m (Figure 4.1). The depth of the site ranged from 0 to 5.8 m
within the array but sea lamprey were positioned at greater depths on the fringes of the array.
Nine independently moored synchronization transmitters (VEMCO model V16-2H, 69 kHz)
were deployed in stationary positions through the array. In addition to the receivers in the
acoustic array, receivers were positioned in the Ocqueoc River at six locations monitoring
upstream progress in 2010 and 2011.
Experimental Subjects
The ethical treatment and acoustic tagging of sea lamprey was approved by the Michigan
State University Institutional Animal Use and Care Committee via animal use permit 02/10-02000. Adult female sea lamprey were obtained from the Cheboygan River at a barrier integrated
trap (2010: N=72, trapped from 15, April to 10, May; 2011: N=79, trapped from 30, April to 15,
May). In 2010, Sea lampreys were 417-567 mm in length (mean 495 mm) and weighed 148-409
163

g (mean 247.7 g). In 2011, Sea lampreys were 458-588 mm in length (mean 506 mm) and
weighed 191-389 g (mean 274 g). All Sea lamprey were held in 150 L flow through tanks that
cycled ambient Lake Huron water (100% exchange every 2 h) and experienced a natural light
cycle. Acoustic transmitters transmitted signals every 15-45 s, though one contained a pressure
sensor that transmitted pressure every transmission (2011) and the other transmitted pressure
every other transmission (2010) (model V9P-2H,Vemco, Halifax, Nova Scotia, Canada 9mm D
x 47 mm L, mass: 6.4 g in air, 3.5 g in water, power output 150 dB (re 1 µPa at 1 m)) (model V92H,Vemco, Halifax, Nova Scotia, Canada, 9mm diameter, 29 mm length: mass: 3.6 g in air, 2.2
g in water, power output 151 dB (re 1 µPa at 1 m)). The model with alternating pressure
transmissions was selected in 2010 due to its shorter duration transmission length to ensure
transmission collisions did not prevent position acquisition. Acoustic tagging procedures
followed methods by Meckley et al. 2014b. Prior to surgery sea lamprey were anesthetized by
immersion in 0.2 mL·L−1. The anesthetic solution used in this experiment was composed of 2
mL of clove oil (minimum 84%-88% eugenol, Lot No. HB9387, Hilltech Canada Inc. Vankleak
Hill. Ontario, Canada) diluted in 18 mL of 70% ethanol and vigorously mixed into 10 L of Lake
Huron water. Sea lamprey took 634 ± 11 s to reach stage four of anesthesia, as denoted by
unresponsiveness to touch but retained gill movement, at a temperature of 8.0 ± 0.1 O C (mean ±
1 SE). The surgical procedure took an average of 319 ± 7 s (mean ± 1 SE) and equilibrium was
regained in 527 ± 32 s (mean ± 1 SE). Transmitter-implanted subjects were held for 72 h prior to
release to ensure metabolism of stress compounds (Close et al. 2003).
Pheromone application
Synthesized PADS (MW: 722 g), PSDS (MW: 625 g), and PZS (MW: 473.4 g) were
applied continuously at a ratio of 1:1:1 based on molecular weight, from 4 May to 6 June 2011,
164

900 m from the river mouth to the middle of the water column in the center of the stream.
PADS, PSDS and PZS were synthesized by Bridge Organics, Inc. (Kalamazoo, Michigan, USA).
The purity of each product was confirmed at >95% via high-pressure liquid chromatography
(HPLC) with a purity evaporative light scattering detector and mass spectrometry by Xiodan Xi
at Michigan State University. Pheromone applications were permitted by the US Environmental
Protection Agency through experimental use permit 75437-EUP-1 and the Michigan Department
of Environmental Quality pursuant to Rule 97. Prior to application, we created 1.0 mg·mL−1
stock solutions for each compound via dissolution into 1:1 v/v methanol:water. A single stock
solution for each compound was used during the course of the study. The final odor was
prepared by mixing the required amount of each stock solution into a necessary amount of river
water before being pumped into the stream at a fixed rate of 600 mL·h−1 via a battery-operated
programmable peristaltic pump (Admiral Reef Dosing Pump, Norwich, Connecticut, USA).
VPS Array Performance and data treatment
The performance of the VPS array in terms of position accuracy was tested by comparing
the VPS position estimates of V9P-2H transmitter to GPS measured positions (Trimble Geo XH,
post processed). Array performance was tested by either attaching a transmitter to an anchored
line as described in Meckley et al. 2014a (stationary testing) that provided longer term
information (2010: Location 1: 17 June 2010 19:43 01 July 2010 14:34 (N= 29,355 detections),
Location 2: 01 July 2010 14:52 to 08 July 2010 17:03 (N= 16,400 detections); 2011: 09 June
2011 016:30 to 24 June 2011 04:07 (N= 26,628 detections)) or by dragging transmitters while
inside the body cavity of a sea lamprey that was recently deceased, located 1 m off the bottom
and below a floating boat powered by an electric motor with the GPS mounted directly above the
tag monitoring the tags’ true position (2010: N=138; 2011: N=309). The mean unfiltered
165

accuracy estimates from stationary tests were 2.0 ± 6.6 m (2010 Test 1), 1.1 ± 5.3 m (2010 Test
2), 1.5 ± 1.4 m (2011 Test 1) (mean ± SD). The mean unfiltered accuracy from the mobile tests
was 3.0 ± 6.2 m (2010) and 5.4 ± 12.5 m (2011) (mean ± SD).
Three data quality objectives were selected, including the need to ensure that trajectory
data (e.g. turn angles) could be considered, designations of when the sea lamprey were stopped
or moving were reliable based on first passage time; large losses in data did not occur while
achieving the first two objectives that could result in misrepresenting the tracks (Table 4.1).
Trajectory based analyses often break down when average position error is > 10 % of the mean
step length (Bradshaw et al. 2007). In this case the mean step length was 17.7 m, resulting in a
filter objective of less than 1.77 m average error (Meckley et al. 2014 a; Table 4.2). To avoid
incorrect behavioral assignment of swimming or stopped we wanted to remove large erroneous
positions (e.g. >15 m (citation)). Finally we wanted to maximize the removal of inaccurate data
and retention of accurate data to ensure proper assignment of when the fish encountered river
water in our dye concentration model, a step that should be possible in our array conditions, this
meant retention of 95 % of acceptably accurate (<6 m) data and removal of 99 % of inaccurate
data (> 6 m). If a large amount of data was lost, the analysis that required the data quality
objective would have been replaced. An HPE filter of 8 was selected in 2010 and an HPE filter
of 7 was selected in 2011 based on the three criteria regarding the position precision estimates
and data retention (Table 4.1, Table 4.2). The filtered coverage was 8.22 km2 in 2010 and 10.95
km2 in 2011. The filter removed 9% of fish positions inside the array in 2010 and 7% of
positions inside the array in 2011.
Hydrological Data
Bathymetry in the region of the array was recorded via 50 meter transects oriented
166

towards and away from shore and then repeated parallel to shore. Bathymetric data were
collected with an acoustic Doppler current profiler (ADCP; Workhorse Rio Grande 1200 kHz,
Teledyne RD Instruments, Poway, CA). Data from four ADCP beams were combined in real
time with geographic locations from a WAAS-enabled GPS (model Geo-XT, Trimble
Navigation Ltd., Sunnyvale, CA) using the software program BathMapper (courtesy of A. Blake,
U.S. Geological Survey, California Water Science Center). Total water column depths at fish
positions were interpolated from the georeferenced depth data using inverse distance weighting.
A 2 MHz upward facing Aquadopp Current Profiler (ADCP, Nortek-USA) was deployed on the
bottom in 7.9 m of water at two nearby locations asynchronously from the 8 May to 30 May
(45.50207737, -84.07552895) and from the 30 May to 17 June, 2011 (45.50207846, 84.07552753). The ADCP recorded water movement in 3 dimensions at twenty slices through
the water column and was used to check the hydrodynamic model (see below for more
information). Five HOBO temperature loggers were deployed each at three locations in front of
the river mouth in the lake (from mouth; left: 300 m, center: 150 m, right: 350 m) farther
offshore (700 m) and in the river for the duration that receivers were deployed. We estimated the
stream discharge daily at the application site using the midsection transect method with a
Doppler flow meter (Flo-Mate Model 2000, Marsh-McBirney)(Gore 2007).
Hydrodynamic Models: Two models describe the circulation of lake water in Hammond
Bay of Lake Huron (Numerical Model for 2010 and 2011) and how river water mixes with lake
water forming a zone of river influenced lake water, which we refer to as the river plume (dye
concentration model). We estimated the water conditions at each fish position and through the
region including current speed and direction, water temperature, and dye concentration as an
estimate of the presence of river water.

167

Numerical Model: The three-dimensional unstructured grid numerical model (FVCOM;
minimum grid size: 9.15 m2) is used to describe circulation and thermal structure in Hammond
Bay, Lake Huron (Nguyen et al. 2014) (Figure 4.2). The model solves the hydrodynamic
primitive equations using the hydrostatic assumption in the vertical direction with the Boussinesq
simplification for convective flows. The continuity, momentum, and temperature equations are
shown in Equation 1 to Equation 5.
𝜕𝑢 𝜕𝑣 𝜕𝑤
+
+
=0
𝜕𝑥 𝜕𝑦 𝜕𝑧

(1)

𝜕𝑢
𝜕𝑢
𝜕𝑢
𝜕𝑢
1 𝜕𝑃 𝜕
𝜕𝑢
+𝑢
+𝑣
+𝑤
− 𝑓𝑣 = −
+ (𝐾𝑀 ) + 𝐹𝑢
𝜕𝑡
𝜕𝑥
𝜕𝑦
𝜕𝑧
𝜌0 𝜕𝑥 𝜕𝑧
𝜕𝑧

(2)

𝜕𝑣
𝜕𝑣
𝜕𝑣
𝜕𝑣
1 𝜕𝑃 𝜕
𝜕𝑣
+𝑢
+𝑣
+𝑤
+ 𝑓𝑢 = −
+ (𝐾𝑀 ) + 𝐹𝑣
𝜕𝑡
𝜕𝑥
𝜕𝑦
𝜕𝑧
𝜌0 𝜕𝑦 𝜕𝑧
𝜕𝑧

(3)

𝜕𝑃
= −𝜌𝑔
𝜕𝑧
𝜕𝑇
𝜕𝑇
𝜕𝑇
𝜕𝑇
𝜕
𝜕𝑇
+𝑢
+𝑣
+𝑤
=
(𝐾𝐻 ) + 𝐹𝑇
𝜕𝑡
𝜕𝑥
𝜕𝑦
𝜕𝑧 𝜕𝑧
𝜕𝑧

(4)
(5)

Here (𝑢, 𝑣, 𝑤) are velocity components in horizontal and vertical directions, respectively; 𝜌 is
the density; 𝜌0 is the reference density; 𝑇 is the temperature; 𝑃 is the pressure; 𝑓 is the Coriollis
parameter;𝑔 is the acceleration due to gravity; (𝐹𝑢 , 𝐹𝑣 , 𝐹𝑇 ) are the horizontal momentum and
thermal diffusion terms. Vertical eddy viscosity and diffusivity(𝐾𝑀 , 𝐾𝐻 )are modeled using the
Mellor-Yamada 2.5 level turbulence closure scheme (Mellor and Yamada 1982, Galperin et al.
1988). The horizontal diffusion coefficients are calculated using the Smagorinsky turbulence
closure model (Smagorinsky 1963). Comparisons between vertically averaged current and
temperature from the ADCP observations and simulations provided a good fit suggesting the

168

models effectively predicted water conditions in Hammon Bay (Nguyen et al. 2014) (Figure 4.3,
Figure 4.4).
Dye Concentration Model; the dye concentration model is coupled with the
hydrodynamic model and was solved using the following equations:
𝝏𝑫𝑪
𝝏𝒕

+

𝝏𝑫𝒖𝑪
𝝏𝒙

+

𝝏𝑫𝒗𝑪
𝝏𝒚

+

𝝏𝒘𝑪
𝝏𝝈

𝟏 𝝏

𝝏𝑪

− 𝑫 𝝏𝝈 (𝑲𝒉 𝝏𝝈) − 𝑫𝑭𝒄 = 𝑫𝑪𝟎 (𝒙, 𝒚, 𝝈, 𝒕)

where C is the concentration of the dye, D is the total depth, u, v, and w are the x,y, and 
components of the water velocity, Kh is the vertical diffusion coefficient, Fc is the horizontal
diffusion term, and C0 is the initial dye concentration (Chen et al., [2012]).The dye concentration
is calculated at every grid cell throughout all layers in the water column in which the initial
condition of dye concentration is zero at every grid cell of the FVCOM grid. The dye
concentration from river was kept at 100ppm throughout the simulation time though the amount
of water released is based on river discharge and influences C0. The river plume is then
determined based on the values of the vertically-integrated concentration at every grid cell. The
comparisons for velocity and temperature from hydrodynamic model are acceptable, providing
confidence for the dye concentration model as it is coupled with hydrodynamic model. The fact
that this model is coupled with the hydrodynamic models and those models have a good fit with
observed data, we have confidence in the predictions of the dye concentration model (Figure 4.3,
Figure 4.4).
Identifying active sea lamprey
To characterize the periods of sea lamprey activity in the lake, we first had to robustly
classify sea lamprey positions as active or stopped. We achieved this with a first passage time
classification method performed in R (R Development Core Team 2015). The first passage time

169

tool classified a position as moving if it left a radius of 10 meters in 250 seconds, had a minimum
displacement of 15 m in a 3 position moving average, and at least 4 consecutive moving
observations. The characterization method was checked through visual inspection and was robust
to imprecision in acoustic positioning after filtering (Gurarie et al. 2015).
Each night of activity that an individual sea lamprey was observed in the array was
characterized into a unique bout of activity for that individual. For each step comprised of two
consecutive positions that was separated by less than ten minutes was used to calculate trajectory
data. The midpoint of each step was used to estimate environmental information experienced
during the step and the distance to key features (e.g. nearest coast or river). The trajectory data
included heading (Phi; radians 0-2Π), turn angle (Theta, phase angle; radians Π (right turn) – 0
(straight) - - Π (left turn), and ground speed (m/s). The lake water temperature, lake water current
magnitude and direction, whether during a bout a sea lamprey had yet encountered river water,
the distance to the nearest coast and the distance to the river were all determined in R for each
mid-point.
Characterizing river plume encounter and entry
We separated individuals into two categories including individuals that stopped in the
array after release (A) and all individuals released (B), as sea lamprey that stopped in the array
after release provided a more complete depiction of the start of the lake migration. Both
categories were further separated into sea lamprey that encountered river water on the first night
and those that did not encounter river water. To determine if differences occurred in the rate of
entry between years on the first night or ever, the rates at which sea lamprey encountered river
water and entered the Ocqueoc River for both groups (A and B) were evaluated with a chi square
tests of independence (χ2)(R Development Core Team 2015). To consider if the rate of river
170

entry changed when individuals encountered the river plume on more than one night, entry of all
sea lamprey (B) were tabulated in light of the number of nights they encountered the river plume
before entry (rate of all available). Similarly the specific entry rates of individuals that returned
and encountered river water on 1, 2, or more than 3 nights were quantified.
To evaluate the effect of a host of variables on river entry, a mixed effects logistic
regression with a binary response variable (river entry: 1 or 0) was implemented in R using
glmer (lme4 package, Bates et al. 2014). The binary response variable river entry (E) was
considered for each night of activity that a sea lamprey encountered river water based on the dye
model (dye concentration > 5); where individual (ID) was a random effect since multiple nights
of activity could occur for the same sea lamprey. The fixed effects included year (Y), iterative
number of nights encountering river water (N), a year by night number interaction(YxN),
average nightly discharge of the Ocqueoc River (DC: m3·S-1), average lake temperature
encountered (LT: C ̊), average nightly river temperature (RT; C ̊ ), the difference in experienced
lake temperature and river temperature(DFT; C ̊ ), minimum distanced reached from the river
mouth (D: km), and average water velocity experienced in the lake (V) (Table 4.3). We used
AIC (Akaike’s information criterion) to guide model selection (Burnham and Anderson 1998;
Burnham and Anderson 2002). We used AIC without criteria adjusted for sample sizes (AICC),
as AICC requires a known number of independent observations and this is difficult to determine
in mixed-effects models (Dodge et al. 2014; Burnham and Anderson 1998). All analyses were
performed in R (R Development Core Team 2015), including the lme4 package.
To determine what influenced the way sea lamprey move we quantified movement
behavior with ground speed (Vm, m·s-1) and path straightness. Activity was considered on the
first night of activity, for active sea lamprey that stopped in the array following release and

171

encountered river water on the first night. Each step was associated with a ground speed (V) and
all location associated variables were based on the mid-point of the step. Path straightness (S)
was estimated by looking at the total path length versus the beeline distance for sets of ten
consecutive points. We fit models with a nonlinear mixed-effects model in the formulation
described by Lindstrom and Bates (1990), using the function “nmle” in R (R Development Core
Team 2015). This method allowed for incorporation of the random effect of each individual
lamprey (ID) and first order auto-correlation in the residuals to be accounted for in the models.
The significance of effects within the best model was assessed by evaluating the t-values and pvalues; where large t values and p-values less than 0.05 were considered significant.
The closest distance each sea lamprey was observed to the river for all sea lamprey and
for those that encountered river water and did not enter was determined and tested for differences
between years with a Mann-Whitney U test with an alternative hypothesis that there was no
difference between the means. The time it took for entry to occur was also tested between years
with a Mann-Whitney U test that sea lamprey would take longer to enter in 2010. The upstream
progress of sea lamprey was monitored at 3 pairs (N=6) of acoustic receivers in the river that
could capture passage and allow for interpretation of upstream or downstream movement. We
used these receivers to determine river entry and the time it took for sea lamprey to reach the
lamprey barrier on the Ocqueoc River located 7916 m from the river mouth.
Movement Patterns
Ground speed was evaluated as a continuous response variable with respect to year (Y),
water temperature at fish location (T, C ̊ ), an interaction between current magnitude and the
angle between fish heading and current direction (as a turn angle in radians: 0-Π, where 0
indicated movement into the current and Π away from current) (CxF, radians 0- Π), whether the
172

fish position was inside or outside of the river plume (dye concentration of > 5) (P), and whether
encounter with river water occurred yet in each path (E). The average time difference between
points was 49 s with a max of 569 s (<10 min). We included the cumulative distance traveled (D)
for each sea lamprey path in each model as previous observations captured increasing ground
speed at the start of the night (Chapter 3). Animal identification number (ID) was included as a
random effect and first order autocorrelation in the residuals was taken into account. Ten
competing mixed effects models were fit (Table 4.4).
Path straightness (Sm) ranges from 0-1 and was evaluated as a continuous response
variable with respect to the fixed effects of whether sea lamprey had encountered river water in
the path yet (E), whether sea lamprey were in the river plume (P), year (Y), total cumulative path
length (D), and additional interactions within these variables. Changes in path straightness are
often used to identify local search behavior (Lohmann et al. 2008, Dodge et al. 2014). Animal
identification number (ID) was included as a random effect and first order auto correlation in the
residuals was taken into account. Ten competing mixed effects models were fit (Table 4.5).
To characterize how sea lamprey migrate outside of the river plume, we summarized path
length versus the approach ratio (distance moved towards the coast/total distance moved) to the
coast where -1 would indicate progress exactly away from the coast, 0 would be movement
parallel to the coast and 1 would represent movement towards the coast. The distance from
shore when sea lamprey exited the array while moving parallel to the coast was summarized and
the direction chosen with respect to the local current encountered was tabulated. To test if the
direction chosen was opposite the local water current encountered while moving in the array, we
tested the proportion of sea lamprey to move east on a night when the current was moving from
the east or from the west with a chi square test (χ2).

173

Effect of Synthetic Odor
The effect of synthetic larval odor is difficult to evaluate in light of the presence of larvae
in the system. The only metric for evaluating the influence of the three synthetic larval odor
components comes through comparison of the historical entry rate of wild sea lamprey one and
two years post TFM treatment and compared to the entry rate of the sea lamprey released in our
study. The percentage of sea lamprey to enter the Ocqueoc River of the total available migratory
sea lamprey in Lake Huron was achieved by comparing population estimates provided by Jessica
Barber of the Fisheries and Wildlife Service. The Ocquoec River population estimate was based
on mark recapture estimates from the Ocqueoc River alone and the Lake Huron estimate was
estimated based on extrapolating mark recapture data from a set of reference streams (Mullett et
al. 2003. Although there are uncertainties in comparing the magnitudes of estimates through
time, the general pattern post pesticide application is robust.

RESULTS
Sea Lamprey Activity
Sea lamprey primarily started moving during the half hour before nautical twilight and
stopped in the half hour before sunrise (Figure 4.5A). River entry was occasionally observed
prior to nautical twilight (set), although river entry was never observed after nautical twilight
(rise) in the morning (Figure 4.5D), despite activity in the lake extending until and occasionally
after sunrise (Figure 4.5B). There was some evidence of multiple pulses of river entry, including
a pulse of entry after nautical twilight (set), 2.5 hours after nautical twilight (set), and just prior
to nautical twilight (rise) (Figure 4.5C, Figure 4.5D). Stopping in the array during nautical
twilight was rare and prolonged periods of inactivity consistent with staging were not observed.
174

Characterizing river plume encounter and river entry
There was no difference in river plume encounter between 2010 and 2011, although all
measures suggested a higher rate of entry in 2011, in support of hypothesis 1. The majority of
sea lamprey stopped in the array after release in 2010 (53 of 72; 74%) and 2011 (59 of 79; 75%).
A low proportion of sea lamprey that exited the array coverage immediately after release were
not observed again (N=9, 2010; N=11, 2011). Our observations showed no difference in the
likelihood of encountering river water on the first night in 2010 (31/53; 58%) or 2011 (41/59;
69%) for trial animals that stopped in the array after release (χ2 = 1.03, df=1, p=0.3; R-core
Team 2015). Similarly, there was no significant difference in the likelihood of ever encountering
river water from the Ocqueoc River for all released in 2010 (60/72; 83%) and 2011 (61/79; 77%)
(χ2 = 0.54, df=1, p=0.46); which includes individuals that encountered river water while moving
after release, and those that encountered river water after the first night. Of those we observed
moving at night, 48 (66%; 2010) and 53 (67%: 2011) encountered river water during search, the
eventual river entry rate of these individuals was 0.5 in 2010 and 0.85 in 2011. Sea lamprey
entered the Ocqueoc river at a higher rate (1.72) in 2011 (45/79; 57%) than in 2010 (24/72; 33%)
(χ2 = 7.55, df=1, p=0.006). Sea lamprey that stopped in the array and encountered the Ocquoec
River water on the first night in 2011 were significantly more likely to enter the river on the first
night than in 2010 (2010: 7/31, 23%; 2011: 18/36, 50%; χ2 = 0.18, df=1, p=0.04). Sea lamprey
that ever encountered river water during the year eventually entered the Ocqueoc River at a
higher rate (1.85) in 2011 (45/61; 74%) than 2010 (24/60; 40%) (χ2 = 12.732, df=1, p<0.001).
The pattern of higher entry in 2011 than 2010 was consistent for those that did not enter on the
first night and were observed encountering river water on additional nights (Figure 4.6).

175

The best fitting models for river entry included year and minimum distance reached from
the river, and either the lake temperature experienced (Model 10) or the difference in lake
temperature experienced and the river temperature (Model 11) (Table 4.3). Only year (year 2011;
β=0.89), and the effect of the minimum river distance was significant (-2.1), where minimum
distance observed related to a reduced rate of entry (p<0.5) (Table 4.6, Model 11 output). River
entry did not appear to be influenced by discharge or absolute river temperature. The number of
nights observed in the array did not appear as an important factor as we expected, though the
pattern of sea lamprey entering if they returned in 2011, and not entering upon return in 2010
was present in the interaction effect between night and year (Figure 4.6).
The distance sea lamprey were observed from the Ocquoec River mouth did not vary
between 2010 and 2011 for all sea lamprey, in support of hypothesis 2 (mean ± 1 SD: 210 ± 266
m (2010); 129 ± 177 m (2011; Mann-Whitney U = 1065, p=0.54) or for those sea lamprey that
didn’t enter the river (mean ± 1 SD: 291 ± 281 m, 2010; 349 ± 236 m, 2011) following river
water encounter (Mann-Whitney U =229, p=0.25) (Function “wilcox.test”; R Development Core
Team 2015). Sea lamprey did take significantly longer to enter the Ocqueoc River in 2010 (116.9
± 125 hours; mean ± 1 SD) than in 2011 (86.8 ± 142 hours) based on a one sided Mann-Whitney
U with an alternative hypothesis that sea lamprey would enter more quickly in 2011 (W=1006,
p=0.046) (Function “wilcox.test”; R Development Core Team 2015).
In 2010, 24 of 72 sea lamprey (33%) entered the Ocqueoc River and were detected
passing a receiver located 342 m from the river mouth and never reversed course past that point
until late in the year once spawning had occurred. Sea lamprey that moved downstream were
likely dead or dying, as determined by tags that were stationary in the array and based on
recovery of some dead individuals. In addition, two sea lamprey moved close enough to the river

176

mouth or briefly entered the main channel and were detected by a receiver located 103 m from
the river mouth. In 2011, 45 of 79 (57%) sea lamprey entered the Ocqueoc River and were
detected passing a receiver located 342 m from the river mouth, and never reversed course and
exited the river. One sea lamprey was detected on the lowest receiver and did not continue
upstream. Two non-trial subjects (2 of 12) released prior to the start of odor pumping in 2011 in
an attempt to capture staging behavior, did enter the river and go within the detection range of
the receiver located 342 meters upstream before exiting the river, with one of the two returning
to the river and progressing upstream in the same night. These two individuals were released in
the visible river plume. In 2010, 19 of 24 sea lamprey reached the upper receiver, including 4
that reached the upper receiver on the first night, equaling an upstream ascent of 0.44 ± 0.11 m·s1

(mean ± 1 SE) over the 7.9 km stretch of river. In 2011, 41 of 45 sea lamprey reached the

upper receiver including 5 on the first night that had a ground speed of 0.38 ± 0.08 m·s-1 (mean ±
1 SE). The average time to reach the upper receivers (5 or 6) after reaching the second receiver
was 44.8 ± 14.6 hours (2010) and 35.5 ± 6.8 hours (2011) (mean ± 1 SE). The time to reach the
upper receivers varied from 3.9 – 290.5 hours in 2010 and 4.1 – 187 hours in 2011.
Movement Patterns
Ground speed did not vary between 2010 and 2011, in support of hypothesis 2 (body
length/Second; BL·S-1). The model comparison suggested that Model 9 had the best fit. The best
model included temperature (T), if the sea lamprey was in the river plume (P), whether river
plume encounter occurred (E), the interaction between lake current magnitude and whether the
fish was moving towards it or away from it (CxF), and the interaction effect of cumulative path
length and before and after encounter. Year had no apparent effect, as the two worst models
included year (Table 4.4). The Model 9 fixed effects parameter estimates revealed that ground
177

speed increased with warmer water temperature (β=0.02, p < 0.05), increased following river
plume encounter (β=0.36, p < 0.05), was influenced by the current direction fish movement
interaction (β=0.29, p < 0.05), and prior to river encounter speed increased with cumulative path
length (β=0.16, p < 0.05) but not after encounter (β=-0.001, p=0.66). Despite being part of the
best model the effect of being in the river plume was small (β=0.002) and insignificant (p = 0.88)
(Table 4.7, model 9 output).
Similarly, path straightness did not vary between 2010 and 2011, in support of hypothesis
2. The model comparison suggested that Model 9 and Model 6 were the best fit. The best two
models included whether the sea lamprey was in the plume (P) and Model 9 included the
interaction effect between cumulative distance traveled (D) and whether the river plume had
been encountered (E) (Table 4.5). Model 9 fixed effects parameter estimates indicated that paths
were more sinuous in the plume (β= -0.0045, p < 0.05), and that the path became straighter after
settlement and before encounter with the plume as the path progressed (β= 0.048, p < 0.05), and
again became straighter after encounter as the path progressed but the effect was weaker (β=
0.018, p < 0.05) (Table 4.8; model output).
Sea lamprey that did not encounter river water (N=33) in the array on the first night
moved in a circuitous path before approaching the coast and then turning and moving parallel to
the coast at some distance (Figure 4.7, Figure 4.8). The average distance from the coast at the
final point exiting the array when moving parallel to the coast was 303 ± 221 m. More sea
lamprey exited the array to the west (N=17) than to the east (N=7), while 9 were last observed
moving to the north or south (χ2 = 4.17, df=1, p< 0.05). Individuals were not more likely to move
east when current came from the east rather than the west (0.38 moved east with east current,

178

N=8; 0.15 moved east with west current, N=13) (χ2 = 0.39, df=1, p=0.735, R function
“prop.test”).
Effect of synthetic odor
An annual increase in the number of sea lamprey estimated to enter the Ocquoec River of
the entire Lake Huron population is observed. Rates increased on the order of 1.8 (1995-1996),
1.3 (1999-2000), and 3.1 (2003-2004), an average historical rate increase of 2.1 times from one
to two years after treatment. The observed rate increase of untagged sea lamprey was 2.1 for
2010 to 2011 for the Ocqueoc River population estimate for our two-year study, matching the
historical average. To test if our observed total entry rate of sea lamprey that stop in the array on
the first night and encounter river water was higher than the expected 2.1 historical average
increase in entry, we ran a chi square test with an alternative hypothesis that the rate observed in
2010 of 0.35 would rise to greater than the expected 0.71 rate in 2011. 41 sea lamprey stopped in
the array after release and encountered river water with 28 eventually entering the river. We
would expect significantly more than a 29.11 individuals to enter if the entry increased with the
average increase between year one and year two. There was no evidence of a greater rate of entry
than would be expected two years after TFM treatment, which suggested no effect of synthetic
larval odor (hypothesis 3)(χ2 = 0.0, df=1, p=0.50, alternative H0: “greater than”; function
prop.test).
DISCUSSION
These results support our hypothesis that sea lamprey enter rivers with higher larval sea
lamprey odor at a higher rate (H1). The effect of larval odor is consistent with altering the
decision of sea lamprey to enter a river but does not have a measurable influence on how sea
lamprey navigated in the river plume, as ground speed, path straightness, and the minimum
179

distance sea lamprey were observed from the river mouth was not influenced by the effect of
year (H2). No evidence was obtained supporting an effect of synthetic larval odor on river entry
(H3). We characterized the sea lamprey migration and report which aspects of the prior
descriptions of sea lamprey migration in the lake are consistent with our observations.
We observed four general patterns of movement. First sea lamprey performed maneuvers
consistent with the offshore pattern described as an orientation mechanism (Chapter 3), in which
sea lamprey turn in a persistent direction and move slowly before approaching the coast at a
faster ground speed and with a relatively straight course (Figure 4.9). Second, if sea lamprey did
not encounter river water at some point when approaching the coast (mean 303 m from shore),
they turned and moved parallel to the coast (Figure 4.8) more often to the west (Figure 4.7), with
straight movements, presumably to encounter a river plume (Vrieze et al. 2011, Meckley et al.
2014b). The dominant current of Lake Huron would come from the west (Beletsky et al. 1999),
though the local current experienced by individuals did not relate to the direction individuals
exited the array (Figure 4.7). Third, when sea lamprey did encounter river water most individuals
made repeated passes up and down the coast in the river plume, which switching between
movements towards and away from the coast in an apparent rebounding pattern and counter
turning at presumably the river plume edge in a behavior we term, coastal rebounding that was
associated with faster ground speed and a more sinuous path. A series of tight rebounds were
often observed before sea lamprey entered the river and we postulate with others (e.g., Vrieze et
al. 2011, Johnson et al. 2012, Meckley et al. 2014b) that this is the point where sea lamprey
transition from edge guided search to odor mediated rheotaxis.

180

Movement outside of the river plume
After stopping on the bottom in the array for the day most sea lamprey began moving
shortly after nautical twilight, those that did not encounter river water approached the coast and
then turned and moved parallel to the coast while either making regular vertical movements
between the surface and bottom, alternating between periods at the surface and periods at the
bottom, or by moving primarily on the bottom. The majority of sea lamprey exited the array
towards the west (55% west, 25% north, 21% east), consistent with observations that most sea
lamprey released 3.3 km from shore in Hammond Bay went west if they didn’t enter the closest
river and were observed at least 10 km away from the release point (60% (west), 20% (east),
20% (north) (Chapter 3). No advantage to moving in a consistent direction was identified unless
a heading in a particular direction allows for the movement parallel to the coast to be maintained
more easily or a direction is energetically more efficient. There was no apparent influence of the
local current encountered on the direction sea lamprey moved along the coast (Figure 4.7),
although the dominant lake current in this area of the lake would be from the west to the east
(Beletsky et al. 1999). It is unclear if more sea lamprey moved west as a result of an inherent
bias associated with handedness (Kells and Goulson 2001) or if it is due to a geographic or
hydrological feature that is valuable to navigation. Observations of the sea lamprey migration on
a coast oriented in a different direction would be valuable, as all observations of the lake
migration to our knowledge have occurred in or near Hammond Bay, Michigan. It is important
to note that eight of the 33 sea lamprey that did not encounter river water and exited the array on
the first night returned and entered the river on a subsequent night, clouding the picture of
whether sea lamprey move in a consistent compass direction or what features guide the direction
of search along the coast.

181

Sea lamprey are likely capable of relying on multiple features of the coastal
environmental as they traverse the coast in search of river water, as redundancy in search
mechanisms is common (Able 1991). The same environmental features including the
bathymetric slope that sea lamprey were postulated to use to return to a coast in chapter 3 could
allow sea lamprey to move parallel to the coast; including, water column depth, absolute
hydrostatic pressure, or water current. In addition to these feature, physical features of the lake
bottom such as the coastal wave zone comprised of a comparatively steep series of banks and
troughs that run horizontal to the coast and are easily observed on satellite images could be
valuable (Figure 4.10). Use of the bottom structure could explain the periods of movement
observed on the bottom before or after periods of vertical casting. A thermal bar along the coast
also forms in spring when water above and below 4 degrees Celsius mixes, creating a physically
distinct water mass along the coast (2-5 km wide, Rao and Schwab 2007), though this zone
should extend beyond the distance sea lamprey were observed moving parallel to shore it still
could label the coastal zone. Although blinded sea lamprey were capable of completing a river
migration (Binder and McDonald 2007), vision may be valuable to the nocturnal migration as
sea lamprey’s eyes are sensitive to light, although it is unclear if this equates to strong night
vision (Morshedian and Fain 2015). Our observations confirm the pattern of extensive search
parallel to the coast but do not confirm a mechanism.
River plume encounter
River entry hinges on sea lamprey first encountering river water in the lake, localizing the
river mouth, and finally entering the river. The number of sea lamprey to encounter a river plume
will depend on both the rate of encounter and the number of sea lamprey to move past the river
mouth. Olfaction is postulated as the primary mechanism for detecting river plume encounter;
182

however, other factors such as temperature or anomalies in river current at the mouth could be
useful and could explain how nasally occluded sea lamprey still entered rivers in the Great Lakes
(Vrieze et al. 2010). The rate of encounter will vary with the size of the river plume, principally
how far river water extends away from the coast and how far from shore sea lamprey traverse the
coast in that region. The size of river plumes vary with river discharge, coastal bathymetry,
coastline configuration at the river mouth (e.g., embayment trapping water), and wind driven
lake current (Churchill et al. 2003, Rao and Schwab 2007). Wind-driven current and the coastal
boundary are the dominant features regulating the extent of the river plume in the Great Lakes
(Churchill et al. 2003). River plume size is likely a key factor but is not explicitly linked with
river discharge a potential explanation to why Moore and Schleen (1980) failed to see a
relationship between sea lamprey entry and river discharge for rivers that were known to contain
sea lamprey larvae. Sea lamprey traversing the coast at 300 m from shore in search of river water
may miss many smaller river plumes, such as the Black Mallard River plume that was observed
extending less than 250 m from shore on 80 % of nights in spring (Meckley et al. 2014b). The
number of sea lamprey available to encounter the plume will depend on the number of sea
lamprey to return to the particular region after the parasitic stage and whether they reach a given
river or stop at a previous river. It remains unclear if the start of the migration could bias
encounter with a specific region of a lake or if release from a host occurs broadly throughout the
lake without major patterns, as virtually nothing is known about the behavior of parasitic sea
lamprey beyond wounding rates (Jorgensen and Kitchell 2005, Silva et al. 2014). Additional
research evaluating the conditions associated with the number of sea lamprey to enter a river
would be valuable.

183

Movement in the river plume
Once in the river plume sea lamprey regularly moved towards and away from the coast as
they made persistent progress along the coast; they reversed direction in shallow water and at
consistent distances from the coast, a pattern termed coastal rebounding (Meckley et al. 2014)
(Figure 4.11, Figure 4.12). We hypothesize that coastal rebounding relies on thigmotaxis within
an odor defined region of the coast, where sea lamprey turn once reaching the edges of the river
plume and the physical coastal edge to progress up and down the coastline. The behavior is
consistent with localization and assessment, as sea lamprey often approached very close to the
river mouth on multiple occasions without entering the river. A river mouth will almost always
occur near one end of the river plume although some lake conditions can consistently create a
more centrally located river mouth with respect to the river plume (García Berdeal et al. 2002,
Churchill et al. 2003), and the river mouth may not be located at the strongest concentration of
river water in cases where water is trapped within a basin. Coastal rebounding would allow for
localization of the river mouth regardless of the dynamic river plume conditions.
Sea lamprey do not follow the coast in shallow water but instead repeatedly move toward
the coast and offshore as progress is made along the coast. There are at least three explanations.
First the rebounding nature may enable sea lamprey to avoid entrapment in complex shorelines
but second, it could result in avoidance of habituation to olfactory information in the river plume.
Sensory habituation can occur if a receptor is constantly exposed to a stimulus and becomes
fatigued resulting in a loss of sensitivity (Ferrari et al. 2010). Olfactory habituation can be
induced by multiple paradigms, although the type of habituation this type of movement would
best avoid is habituation by a change in the function of the olfactory bulb, a process which takes
longer than 30 minutes (Chaudhury et al. 2010). For example sea lamprey habituate to the odor

184

of dead adult sea lamprey after exposure for between two and four hours (Jason Bals,
unpublished data). To address whether sea lamprey are reaching the edge of the river plume, a
more accurate representation of the river plume than a dye concentration model would be
required, which could be provided by a particle diffusion model as a large number of particles
provide a more reliable edge structure (Mantha S. Phanikumar, personal communication). Third,
the search could be a sampling period and sea lamprey could be collecting other information
such as the presence of mates. Given that sea lampreys conceptually need to locate two things,
spawning habitat and mates; it seems that recognition of mates could be important.
The role of larval odor
Our leading hypothesis is that once encountering a river plume individuals either enter,
stage, or search for a different river based on the quality of the river and river temperature
(Binder et al. 2010, Clemens et al. 2010, Vrieze et al. 2011, Meckley et al. 2014b). Odor and in
particular larval odor has support from lab and field tests as an important cue that informs a
migrant of the quality of a river (Moore and Schleen 1980, Vrieze and Sorensen 2001, Vrieze et
al. 2010, Meckley et al. 2014b). In this study sea lamprey entered the Ocqueoc River at a higher
rate in 2011 than 2010, two years removed from any TFM treatment. We saw no evidence that
an environmental variable was associated with the way sea lamprey moved in the river plume or
entry rate, although features like river discharge were higher in 2011 (discharge mean (m3·S-1)
(minimum, maximum) 2010: 2.64 (0.6-9.7); 2011: 3.9 (1.8-8.6)) and river temperature varied
between years (Figure 4.13). Our observations of reduced entry after treatment is consistent with
historical observation in the Ocqueoc River (Figure 4.14), and is a wide spread pattern
throughout rivers in the Great Lakes (Moore and Schleen 1980). These observations provide

185

additional support that river water alone is sufficient to induce local search, while larval odor
influences the river entry decision.
Search behavior did not differ between years, consistent with the hypothesis that river
water alone induced the transition from extensive search to local search in the river plume and
larval odor influenced the river entry decision. This could only have been confirmed if no sea
lamprey larvae were in the river in 2010, which was not the case in this study (Personal
communication Aaron Jubar of the U.S Fish and Wildlife Service). Previous suggestions that sea
lamprey moved faster outside of the river plume, searched the river plume through klinokinesis
(Vrieze et al. 2011), and stopped more frequently in the river plume (Meckley et al. 2014b), were
unsupported and likely the result of inaccuracies in manual acoustic telemetry and short-term
observations. The interpretation that sea lamprey moved slower in the river plume was likely a
result of the order of observations as activity in the river plume was often captured at the start of
the night and movement outside of the river plume was observed later in the night. This resulted
in the slow ramp up in speed we observed after settlement being attributed to movement in the
river plume. The in plume search does not appear to be a kinesis as there is clear persistence and
reliance on the coastline during search. The persistent exploration up and down the coastline in
the river plume is consistent with an absolute olfactory search mechanism where the presence of
river water defines the search boundary, rather than a differential mechanism where the cue
intensity is important and gradients are used to provide directional information (Benhamou and
Bovet 1989). The pattern of movement is inconsistent with reliance on odor gradients for
navigation but we are unable to fully rule out chemotaxis as increases in a chemical cue gradient
while rebounding could inform the migrant if it is moving towards a river mouth and could be
undetected in our study as sea lamprey do make repeated passes up and down the coast. For

186

example, Blue crabs (Callinectes sapidus) use a combined mechanism in which they locate odor
sources with rheotaxis but require chemotaxis to adjust to gradients across the current stream
(Zimmer-Faust et al. 1995).
River entry
The decision of a sea lamprey to enter a river could occur prior to reaching the river
mouth, though in the Ocqueoc River when other migrants were present and spawning and rearing
habitat are available, individuals were never observed reversing course after moving past the
immediate river mouth. Sea lamprey that reached the immediate river mouth did not always
enter. In both years sea lamprey came to the immediate river mouth without entering and
encountered river water on multiple nights, though in 2011 sea lamprey that did either task
entered at a high rate (Figure 4.6). River entry most often occurred following a series of very
tight rebounds along the coast before moving into the river mouth. Sea lamprey transition from
rebounding movement to straight upstream movements on the bottom, suggesting that sea
lamprey transition from edge guided search to odor mediated rheotaxis as they enter the river
(Vrieze et al. 2011, Johnson et al. 2012, Meckley et al. 2014b). Vertical excursions also cease at
this point in the migration (Chris Holbrook, personal communication).
Two previously reported aspects of entry were not regularly observed in this study,
including staging in front of the Ocqueoc River or repeated entries and exits at the river mouth
(Applegate 1950). Sea lamprey rarely remained stationary during nautical twilight and only in
three instances of 151, they entered the river mouth and then exited. One of the three individuals
reentered the river. This does not support or discredit the occurrence of staging or repeated entry
as the sea lamprey in this study had already entered a river and the river temperatures were
between 10-20 C ̊ for the majority of the study period, well above the 4.4 C ̊ temperature in
187

which these behaviors were reported by Applegate (1950) and above the 10 C ̊ thermal threshold
for active migration in the river (Binder and McDonald 2008). Given that many individuals were
reported staging together, and sea lamprey often spent multiple days moving in the plume in
front of the river, there is a possibility that social information is important to the decision to enter
a river. Sea lamprey would not have to actively follow a sea lamprey, by maintaining proximity
to an individual, to use social information to enter a river. Social information could come in the
form of visual observations of other individuals or odor recognition of mates in the area. The
odor or visual recognition of mates could be an important piece of information, as it is known
that the odor of maturing migrants synchronizes the migration (Chung-Davidson et al. 2013).
Although channel bias was not observed in immature female lamprey in response to the male
released sex pheromone (Siefkes et al. 2005), migrants are believed to increase upstream
migration in the odor’s presence (Johnson et al. 2013), suggesting that odor-based mate
recognition is possible.
Odor mediated manipulation of sea lamprey habitat selection
In addition to the goal of manipulation of sea lamprey river selection by encouraging sea
lamprey to enter streams with poor spawning or rearing habitat as suggested by Li et al. (2007),
we suggest that a second potential method would be encouraging sea lamprey to enter a river the
year after TFM application, in rivers treated every three years. Ideally recruitment could be
raised in the year after treatment to a rate similar to the second or third year entry rates following
treatment. It is unclear if the high rate of recruitment in the third year is due to a further increase
in entry between the second and third year post TFM treatment or if additional encounter occurs
due to TFM treatment of other neighboring rivers on a different pesticide treatment schedule.
The only weak comparison for evaluating the effect of the three synthetic compounds found in
188

larval odor provided no evidence of an effect on river entry in this study. Combined with other
field tests it appears that the full larval odor or at minimum additional or different components
found in larval odor will be required to manipulate tributary selection or river entry (Meckley et
al. 2012, Meckley et al. 2014b).
Summary
The sea lamprey lake migration has now begun to be characterized over a narrow region
of Lake Huron in Hammond Bay (Vrieze et al. 2010, Vrieze et al. 2011, Meckley et al. 2014b),
suggesting the evaluation of the migration on other coastlines could be a valuable future
contribution. All paths of sea lamprey movement observed on the first night that did not
encounter river water and did encounter river water can be observed in Figure 4.15 and Figure
4.16, respectively. This study represents the finest scale exploration of in plume search to date,
however an improvement would be the development of a hydrodynamic model that depicts the
plume edges more finely and captures the exact role of the river plume edge in the search
process. Notably, do sea lampreys exit the plume before counterturning back toward the coast or
do sea lamprey reverse back towards the coast once reaching a physical feature of the water
column? Observing the behavior of sea lamprey near the end of the river plume not near the
river mouth could also be informative as our array only offered a window into the river mouth
side of plume search. It will be difficult to reveal a chemotactic search mechanism in the river
plume even if it exists because sea lamprey often make many passes up and down the coast,
providing the impression that the movement is not explicitly in the river plume or performed to
localize the river mouth but may be a direct assessment of the river quality (e.g., searching for
larval odor or presence of mates). Sea lamprey appear to be a physical edge-guided species, that
relies heavily on the surface, bottom, and coastal edges to guide the migration first to the coast
189

(Chapter 3), and then along the coast in search of rivers. Simple movement rules leading sea
lamprey to the coast, along the coast, and within a river plume, appear sufficient to guide sea
lamprey to a river mouth in a similar way that simple movement rules guide other species to
geographically distinct spawning habitat and to find mates (e.g. Pe'er et al. 2004).
These findings suggest that the currently identified components of larval odor are
insufficient to influence river entry, although a more thorough test would be completed in a river
lacking any larval odor of any lamprey species. Given the likely lack of coevolution that would
bias how sea lamprey larvae release the larval signal, it is possible that components of larval
odor that are used by migrants is composed of a more complex number of components than
would be common to a sex pheromone and may be harder to replicate. However our observations
do suggest that the use of the entire synthetic larval odor could be a powerful management tool, a
designation that appears less likely for the male released sex pheromone.
ACKNOWLEDGEMENTS
This research was funded through grants from the Great Lakes Fishery Commission
including by way of the Great Lakes Restoration Initiative appropriations (GL-00E23010-3) and
was accomplished through considerable assistance from the US Fish and Wildlife Service and
the USGS Hammond Bay Biological Station.

190

APPENDIX

191

Table 4.1: Three specific criteria were adopted to establish the data quality objectives based on
an extensive review of data quality (Meckley et al. 2014a).

1

2

3

Objective
Criteria
Detect changes in Mean error ≤1.77
trajectory
m

Rationale
Many trajectory based analyses are only preserved
when the average position error is <10% of the mean
step length (Bradshaw et al. 2007). Transmission
delay was (33.4 s), sea lamprey ground speed is 0.53
ms−1 (Vrieze et al. 2011, Meckley et al. 2014b),
equating to a step length of 17.7 m.
Assign behavioral Max error <15 m Simple behavioral assignment of moving or
state (moving or
stationary with first passage time analysis, which is a
stationary)
measure of the time it takes an individual to leave a
circle of fixed radius r drawn around each measured
location to determine if movement is occurring
(Fauchald and Tveraa 2003, Barraquand and
Benhamou 2008).
Balance loss of
Retain 95% of
To ensure representation of fish tracks while
acceptable data
acceptable data,
avoiding loss of accurate data. We aimed to keep
and retention of
Reject ≥99% of
95% of acceptable data (≤6 m) while retaining <1%
unacceptable data unacceptable data of unacceptable data (>6 m) of all retained positions.

192

Table 4.2: Criteria for selection of an HPE filter cutoff. In the data values listed for each
criterion the HPE cutoffs are listed acceptable HPE cutoffs are listed for 2010 above and 2011
below.
Criteria

Stationary Mobile
Rational for selection of a HPE filter cutoff of 8
test
test
1
2
All All 3
This criterion was met for stationary tests. The mobile test
Criterion 1
Mean ≤1.77 m
All NA None would have required a very low HPE to attain below 1.77 m.
Although for a number of reasons stationary tests are more
reliable estimators for this metric (Meckley et al. 2014a).
<7 <8 ≤19
In 2010: An HPE of 8 did not meet the criteria for mobile
Criterion 2
tests, but only 1 of 134 positions remained. The stationary
Max error <15 m <7 NA ≤1
test criteria were met. Only 2 of 45,744 positions were
problematic (<0.001%) for the combined stationary tests
(Figure 4.4). In 2011 an HPE of 7 removed all except one
position for the stationary test, and one position for the
mobile test exceeding 15 m error.
8–10 3–15 None In 2010: An HPE cutoff of 8 met the criteria for both
Criterion 3
% incorrectly
3-15 NA None stationary tests, and although the mobile test did not have a
retained vs. %
suitable range, the range from 5–10 was equally effective. In
incorrectly
2011: An HPE of 7 met criteria for the stationary test as it
rejected
fell within at least 3-15.

193

Table 4.3: Akaike’s information criterion (AIC) values for linear mixed effects models using
river entry (Em) as a binary response variable to various combinations of fixed effects including
year (Y), iterative number of nights the sea lamprey encountered river water (N), a year by night
number interaction (YxN), average nightly discharge of the Ocqueoc River (DC: m3·S-1),
average lake temperature encountered (LT: C ̊), average nightly river temperature (RT; C ̊ ), the
difference in experienced lake temperature and river temperature (DFT; C ̊ ), minimum distanced
reached from the river mouth (D: km), and average water velocity experienced in the lake (V).
Model

Fixed Effects

df

Vm AIC (ΔAIC)

0

1

2

209.8 (13.0)

1

Y

3

202.6 (5.8)

2

N

3

211.7 (14.9)

3

Y + YxN

5

205.7 (8.9)

4

DC

3

211.8 (15.0)

5

Y + DC

4

204.5 (7.7)

6

LT + RT

4

207.5 (10.7)

7

Y + LT

4

199.7 (2.9)

8

Y + RT

4

204.3 (7.5)

9

Y + DFT

4

199.6 (2.8)

10

Y + LT + D

5

197.1 (0.3)

11

Y + DFT +D

5

196.8 (0)

12

Y + DFT +LT

5

201.5 (4.7)

13
14

Y + DC + LT + RT
Y+V

6
6

203.5 (6.7)
199.2 (2.4)

15

Y + LT + RT + V +DC + D + YxN

10

205.8 (9)

194

Table 4.4: Akaike’s information criterion (AIC) values for a nonlinear mixed-effects model in
the formulation described by Lindstrom and Bates (1990), using the function “nmle” in R (R
Development Core Team 2015). Ground speed (Vm) as a response variable to various
combinations of fixed effects including year (Y), the log of the cumulative distance traveled (D),
whether encounter with river water occurred yet in the path (E), or if the point was currently in
the plume (P), temperature at fish position (T), and an interaction effect between water current
magnitude and direction by fish heading interaction (CxF); and finally the random effect of
individuals (animal ID).
Model

Fixed Effects

Df

Vm AIC (ΔAIC)

0

E

5

178.3 (99.2)

1

E + E:D

6

89.2 (10.1)

2

Y + E:D

5

276.2 (197.1)

3

T

5

279.5 (200.4)

4

P +Y

6

290.9 (211.7)

5

T+P+E

7

168.9 (89.4)

6

T+P+Y

7

279.7 (200.6)

7

T + CxF

6

278.1 (199.0)

8

P + CxF + E:D

8

98 (19.7)

9

T + P + CxF + E + E:D

9

79.1 (0)

195

Table 4.5: Akaike’s information criterion (AIC) values for a nonlinear mixed-effects model in
the formulation described by Lindstrom and Bates (1990), using the function “nmle” in R (R
Development Core Team 2015). Path straightness (Sm) as a response variable to various
combinations of fixed effects including year (Y), the log of the cumulative distance traveled (D),
if whether encounter with river water occurred yet in the path (E), or the point was currently in
the plume (P); and finally the random effect of individuals (animal ID). Model 6 and 9 had the
lowest AIC scores.
Model

Fixed Effects

Df

Vm AIC (ΔAIC)

0

D+Y

6

-354.6 (7.5)

1

D+E

6

-358.0 (4.1)

2

E+Y

7

-340.4 (21.7)

3

P+Y

7

-348.9 (13.2)

4

P+E

6

-355.2 (6.9)

5

E

5

-353.2 (8.8)

6

P

5

-360.4 (1.7)

7

D+E +P

7

-356.8 (5.3)

8

D+E+P+Y

8

-351.9 (10.2)

9

DxE + P

-362.0 (0)

196

Table 4.6: To evaluate the effect of a host of variables on river entry (Em), a mixed effects
logistic regression with a binary response variable (river entry: 1 or 0) was implemented in R
using glmer (lme4 package, Bates et al. 2014). The best fitting models, as determined by AIC,

R code: m11<-glmer(Em ~ Y+ DFT+ D+ (1|ID), data=entry, binomial)
Family: binomial ( logit )
Formula: Em ~ Y + DFT + D + (1 | ID)
Data: entry
AIC
196.8

BIC logLik deviance df.resid
211.9 -93.4 186.8
144

Scaled residuals:
Min
1Q Median
3Q Max
-1.1786 -0.7532 -0.5324 0.9475 2.8220
Random effects:
Groups Name
ID (Intercept)
Number of obs:

Variance
Std.Dev.
1e-14
1e-07
149, groups: ID, 98

Fixed effects:
(Intercept)
year2011
DFT
D
--Signif. codes:

Estimate
-0.64956
0.88611
0.02464
-2.08616

Std. Error
0.41248
0.35433
0.04996
1.05819

z value
-1.575
2.501
0.493
-1.971

0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation of Fixed Effects:
(Intr)
yr2011
year2011
-0.526
DFT
-0.615
0.017
M
-0.436
0.050

DFT

-0.015

197

Pr(>|z|)
0.1153
0.0124 *
0.6219
0.0487 *

Table 4.6: (cont’d)
for river entry included year and minimum distance reached from the river and either the lake
temperature experienced (Model 10) or the difference in lake temperature experienced and the
river temperature (Model 11). The fixed effects included year (Y), iterative number of nights
encountering river water (N), a year by night number interaction(YxN), average nightly
discharge of the Ocqueoc River (DC: m3·S-1), average lake temperature encountered (LT: C ̊),
average nightly river temperature (RT; C ̊ ), the difference in experienced lake temperature and
river temperature(DFT; C ̊ ), minimum distanced reached from the river mouth (D: km), and
average water velocity experienced in the lake (V). Here is Model 11 output.

198

Table 4.7: Output from the summary of the best fit nonlinear mixed-effects model for the
continuous response variable ground speed (body lengths per second), Model 9. The model

Linear mixed-effects model fit by REML
Data: GirardEnc
AIC
BIC
logLik
79.13117
146.8898
-29.56558
Random effects:
Formula: ~1 | ID
(Intercept) Residual
StdDev: 0.2078676 0.2983834
Correlation Structure: AR(1)
Formula: ~1 | ID
Parameter estimate(s):
Phi: 0.5945631
Fixed effects: Vm ~ T+ P + C:F + E + I(D/1000): E
Value
Std.Error
DF
t-value
(Intercept)
0.6803459
0.04771404 6437 14.258818
t
0.0198588
0.00355488 6437 5.586343
as.factor(P)Y 0.0018479
0.01246258 6437 0.148279
as.factor(E)b 0.3591986
0.02492367 6437 14.411945
C:F
-0.2506198 0.11486012 6437 -2.181956
Ea:I(D)
0.1645548
0.01549279 6437 10.621383
Eb:I(D)
-0.0016136 0.00364662 6437 -0.442490
Correlation:
t
PY
E
C:F
Ea:D
Eb:D

(Intr)
-0.606
0.167
-0.297
-0.006
-0.238
0.021

t
-0.250
0.040
-0.163
-0.021
-0.046

P

-0.273
-0.027
0.007
0.107

Standardized Within-Group Residuals:
Min
Q1
Med
Q3
Max
-4.9337612 -0.4086086 0.1300112 0.6036556 8.6886942
Number of Observations: 6483
Number of Groups: 40

199

p-value
0.0000
0.0000
0.8821
0.0000
0.0291
0.0000
0.6581

E

C:F

Ea:D

0.042
0.458
-0.482

0.042
-0.017

0.073

Table 4.7: (cont’d)
included the temperature experienced (T), whether encounter with the river plume had occurred
in the path (E), whether the sea lamprey was in the river plume (P), and an interaction between
water current and if the fish was moving towards or away from the current direction. The
correlation statement accounts for first order autocorrelation in the residuals that occurred from
analyzing multiple simultaneous steps relating to the same individual. The random effect of
animal ID is included.

200

Table 4.8: Output from the summary of the best fit nonlinear mixed-effects model for path
straightness (0-1), Model 9. The model included the effect of the interaction between cumulative

R code: m9 <- lme(fixed = S ~ I(D/1000):as.factor(E) + as.factor(P),
random = ~ 1 | ID,
correlation = corAR1(form = ~ 1|ID),
data = sin)
Linear mixed-effects model fit by REML
Data: sin
AIC
BIC
logLik
-362.0493
-329.653
188.0247
Random effects:
Formula: ~1 | ID
(Intercept)
StdDev: 0.06001644

Residual
0.1812923

Correlation Structure: AR(1)
Formula: ~1 | ID
Parameter estimate(s): Phi 0.1026954
Fixed effects: sin ~ I(D/1000):as.factor(E) + as.factor(P)
Value Std.Error
DF
(Intercept)
0.7951321 0.01479099 684
as.factor(P)1
-0.0450035 0.01790223 684
I(D/1000):as.factor(E) 0 0.0482397 0.01201762 684
I(D/1000):as.factor(E)1 0.0179331 0.00434604 684
Correlation:
as.factor(plume)1
I(stepcum/1000):as.factor(enc)
I(stepcum/1000):as.factor(enc)1

(Intr)
-0.318
0
-0.415

t-value
53.75785
-2.51385
4.01408
4.12631

p-value
0.0000
0.0122
0.0001
0.0000

as.()1

I(/1000):.()0

-0.552
-0.262

Standardized Within-Group Residuals:
Min
Q1
Med
Q3
Max
-4.1849828 -0.3403049 0.3338883 0.6488679 1.5624139
Number of Observations: 760
Number of Groups: 73

201

0.206
0.324

Table 4.8: (cont’d)
distance traveled (D) and Encounter with the river plume and the fixed effect of being in the
river plume. The correlation statement accounts for first order autocorrelation in the residuals
that occurred from analyzing multiple simultaneous steps relating to the same individual. The
random effect of animal ID is included. The distance is divided by 1000 to put the variable in km
rather than m for easier interpretation.

202

Figure 4.1: The study site was located in Hammond Bay of Lake Huron (A, B). Sea lamprey
were released at three locations (open circles). The study site was covered by a two kilometer
wide array that reached a kilometer from shore (C). The receiver positions in 2010 are
represented by triangles, and in 2011 they are represented by circles. In each year nine sync tags
were distributed in two rows across the array. During peak performance a transmitter could be
positioned every 15-45 s. Two of the six river receivers used for monitoring passage are
represented by black squares. The receivers in the river could not detect tags in the lake.
203

Figure 4.2: The hydrodynamic model makes predictions of the current speed and direction and
lake temperature at each node (N=4346) of an unstructured grid fit to Hammond Bay (FVCOM
grid).The size of the triangular grid cells decreases near the river mouth to provide a higher
precision near our fish tracking array (minimum grid size: 9.15 m2). The hydrodynamic model
provides predictions of current speed and direction and water temperature. The model is three
dimensional and provides predictions in 2010 and 2011 at 20 slices through the water column
water column, independent of depth (m), which is color coded.

204

Figure 4.3: Comparisons between vertical averaged currents form ADCP observations and
simulations taken in 2011.

205

Figure 4.4: Comparisons between observed temperature recorded by the ADCP and
hydrodynamic model simulations in 2011.

206

Figure 4.5: Sea lamprey mostly began moving between sunset (orange vertical lines) and
nautical twilight NT (gray box) in most cases (A). Activity continued until between NT and
sunrise, although activity continued in lighter conditions than under which it commenced (B).
The double vertical lines represents sunrise and sunset as the time from NT changes with night
length. Sea lamprey entered the river throughout the night but the earliest they were observed
entering was just before NT (C) and despite activity being observed at sunrise in the lake, river
entry was never observed outside of NT in the lake (D). Two graphs show river entry as night
length changes with day length, this explains the dark gray areas (NT, all year), and light gray
areas (NT during some parts of the year). Focus on the 0 side of insets C and D.

207

≥

Figure 4.6: The entry rate based on the total number of different nights that sea lamprey were
observed encountering river water in 2010 and 2011. The nights do not have to be consecutive
and don’t have to include the first night after release. The entry rate on the first night “1st night”
refers to the entry rate of sea lamprey that encountered river water on the first night of release
after settlement. The number of sea lamprey that fall into each category for each year are found
at the top of the graph.

208

Northing

Easting
Figure 4.7: The 33 tracks of sea lamprey that did not encounter river water on the first night are
color coded and the arrow of the heading from the 10th to last point to final point are shown (A).
209

Figure 4.7: (cont’d)
The circular plot in A (top) shows that most fish moved west out of the array. The exit arrow
(thick arrow) is shown for each sea lamprey and the average current experienced is shown by the
thin arrow, where the arrow points the current direction and the length represents current strength
(max: 27 cm·s-1,average: 4.9 cm·s-1) (B). The circle plots show the current directions observed
on the 33 nights (“Current”), the direction sea lamprey moved when the current was from the
East (“Fish-East Current”), the direction the sea lamprey moved when the current was from the
west (“Fish-West Current”).

210

Figure 4.8: Each transect (set of two points) for sea lamprey that never encountered river water
were binned by 100 m increments from the coast based on their approach ratio (distance moved
towards the coast/total distance traversed). When sea lamprey were 500-1000 m from the coast
movement was directed towards the coast (1) before turning and moving parallel to the coast (0500 m). An increase in movement away from the coast occurs between 100-300 m from the coast
resulting in an average parallel movement along the coast of 303 m ± 221 m when exiting the
array (centered at an approach ratio of 0).

211

Figure 4.9: Ground speed (Body lengths per second) is shown for each step over the cumulative
path length for all sea lamprey (N=33) that did not encounter river water on the first night. Each
line represents a different individual. The pattern for before river plume encounter (black circles)
and after river plume encounter (red squares) is shown. Observations are restricted to a
cumulative path length of 2500 meters. A ramp up in speed to one body length per second is
evident in the first 500 meters of activity.

212

Figure 4.10: The google earth image depicts the distinct nearshore wave zone which includes a
comparatively steep series of banks and troughs that run horizontal to the coast compared the
shallow bathymetric gradient in Hammond Bay. The image captures the coastline in Hammond
Bay in front of the Ocquoec River Mouth, located in the bottom right of the image. The edge of
the banked zone extends from 204-308 m from shore in this area of the array.

213

Figure 4.11: A smoothed utilization distrution in 50 m cells of all active sea lamrpey points following river plume encounter as
defined by encounter with a dye concentraiton of greater than 5 ppm in 2010 and 2011. The white line represents the Ocquoec River.
Movement extended farther east and west along the coast in both cases but was not captured by the array.

214

Figure 4.12: The river plume occupied three general forms in Hammond Bay as represented by the output of the dye concentration
model. The image to the left represents a less common river plume extending to the west. The central schematic depicts the river
plume that occurs when the plume switches from either east to west or occasionally when currents were moving in northerly or
southerly directions and pushed the plume out from shore. The most common river plume extended to the east and extended out of
Hammond bay. The river water was often trapped to the east and west due to the shape of the coast.

215

Figure 4.13: The temperature by Julian date for 2010 (Blue Line) and 2011 (Orange Line) is
shown. A symbol represents each fish if it was in the array on a given night in 2010 (Blue) and
2011 (Orange) at the maximum temperature encountered by the fish. Triangles represent fish that
entered the river on that night and circles representing fish that do not enter that night. An
individual fish is represented multiple times if it returns on multiple nights.

216

Figure 4.14: The percentage of the entire population of adult sea lamprey in Lake Huron that
return to the Ocqueoc River to spawn from 1995 to 2015 based on mark recapture (black circles)
and the change in entry between years (blue circles). A change of entry rate of 2 represents a
doubling in the rate of entry. Vertical lines represent the times of TFM treatment to the Ocquoec
River. The change in entry rate from one year after TFM treatment to the second year shows that
the increase in entry rate between 2010 and 2011 (D) fits within the historical increase in rate (A:
1.8, B: 1.3, C: 3.1, D: 2.1).

217

Time

Figure 4.15: The 40 sea lamprey that passed through the river plume based on the estimates of
the dye concentration at each fish position are shown in a schematic of Hammond bay, where the
black line represents the coast and the red dot represents the river mouth. Each page is labeled by
the year (10 or 11) and animal ID. The upper graph shows the sea lamprey movement on the first
night from where the sea lamprey settled in the array after release (orange dot). In the schematic
of movement, the gray arrow points the heading during the last 5 points of the track and the track
218

Figure 4.15: (cont’d)
is color coded by 15 minute intervals of activity. The label of “Entered” indicates that sea
lamprey eventually entered the Ocqueoc River and the label of “Did Not Enter” indicates that the
sea lamprey never entered the Ocqueoc River. The number of bouts follows the label indicating
the number of different nights the sea lamprey was observed in the array. If the sea lamprey

Time

219

Figure 4.15: (cont’d)
entered and the bout is one, the individual entered at the end of the track shown. In some cases
entry is not captured well by the array (e.g. 10-17). The lower graph depicts the depth of the sea
lamprey during movement through time and the color of the points corresponds to the schematic
in the upper graph. The red line represents the estimated total water column depth. When the fish
is below the line it can be assumed the fish is on the bottom.

Time
220

Figure 4.15: (cont’d)

Time

221

Figure 4.15: (cont’d)

Time

222

Figure 4.15: (cont’d)

Time

223

Figure 4.15: (cont’d)

Time

224

Figure 4.15: (cont’d)

Time

225

Figure 4.15: (cont’d)

Time

226

Figure 4.15: (cont’d)

Time

227

Figure 4.15: (cont’d)

Time

228

Figure 4.15: (cont’d)

Time

229

Figure 4.15: (cont’d)

Time

230

Figure 4.15: (cont’d)

Time

231

Figure 4.15: (cont’d)

Time

232

Figure 4.15: (cont’d)

Time

233

Figure 4.15: (cont’d)

Time

234

Figure 4.15: (cont’d)

Time

235

Figure 4.15: (cont’d)

Time

236

Figure 4.15: (cont’d)

Time

237

Figure 4.15: (cont’d)

Time

238

Figure 4.15: (cont’d)

Time

239

Figure 4.15: (cont’d)

Time

240

Figure 4.15: (cont’d)

Time

241

Figure 4.15: (cont’d)

Time

242

Figure 4.15: (cont’d)

Time

243

Figure 4.15: (cont’d)

Time

244

Figure 4.15: (cont’d)

Time

245

Figure 4.15: (cont’d)

Time

246

Figure 4.15: (cont’d)

Time

247

Figure 4.15: (cont’d)

Time

248

Figure 4.15: (cont’d)

Time

249

Figure 4.15: (cont’d)

Time

250

Figure 4.15: (cont’d)

Time

251

Figure 4.15: (cont’d)

Time

252

Figure 4.15: (cont’d)

Time

253

Figure 4.15: (cont’d)

Time

254

Figure 4.15: (cont’d)

Time

255

Figure 4.15: (cont’d)

256

Figure 4.15: (cont’d)

Time

257

Time

Figure 4.16: The 33 sea lamprey that did not encounter the river plume on the first night based
on the estimates of the dye concentration at each fish position are shown in a schematic of
Hammond bay, where the black line represents the coast and the red dot represents the river
mouth. Each page is labeled by the year (10 or 11) and animal ID. The upper graph shows the
sea lamprey movement on the first night from where the sea lamprey settled in the array after
release (orange dot). In the schematic of movement, the gray arrow points the heading during the
258

Figure 4.16: (cont’d)
last 5 points of the track and the track is color coded by 15 minute intervals of activity. The label
of “Entered” indicates that sea lamprey eventually entered the Ocqueoc River and the label of
“Did Not Enter” indicates that the sea lamprey never entered the Ocqueoc River. The number of
bouts follows the label indicating the number of different nights the sea lamprey was observed in

Time

259

Figure 4.16: (cont’d)
the array. If the sea lamprey entered and the bout is one, the individual entered at the end of the
track shown. In some cases entry is not captured well by the array (e.g. 10-17). The lower graph
depicts the depth of the sea lamprey during movement through time and the color of the points
corresponds to the schematic in the upper graph. The red line represents the estimated total water
column depth. When the fish is below the line it can be assumed the fish is on the bottom.

Time
260

Figure 4.16: (cont’d)

Time

261

Figure 4.16: (cont’d)

Time

262

Figure 4.16: (cont’d)

Time

263

Figure 4.16: (cont’d)

264

Figure 4.16: (cont’d)

265

Figure 4.16: (cont’d)

Time

266

Figure 4.16: (cont’d)

Time

267

Figure 4.16: (cont’d)

Time

268

Figure 4.16: (cont’d)

Time

269

Figure 4.16: (cont’d)

Time

270

Figure 4.16: (cont’d)

Time

271

Figure 4.16: (cont’d)

Time

272

Figure 4.16: (cont’d)

Time

273

Figure 4.16: (cont’d)

Time

274

Figure 4.16: (cont’d)

Time

275

Figure 4.16: (cont’d)

Time

276

Figure 4.16: (cont’d)

Time

277

Figure 4.16: (cont’d)

Time

278

Figure 4.16: (cont’d)

Time

279

Figure 4.16: (cont’d)

Time

280

Figure 4.16: (cont’d)

Time

281

Figure 4.16: (cont’d)

Time

282

Figure 4.16: (cont’d)

Time

283

Figure 4.16: (cont’d)

Time

284

Figure 4.16: (cont’d)

Time

285

Figure 4.16: (cont’d)

Time

286

Figure 4.16: (cont’d)

Time

287

Figure 4.16: (cont’d)

Time

288

Figure 4.16: (cont’d)

Time

289

Figure 4.16: (cont’d)

Time

290

REFERENCES

291

REFERENCES

Able, K. P. (1991). "Common Themes and Variations in Animal Orientation Systems."
American Zoologist 31(1): 157-167.
Applegate, V. C. (1950). The natural history of the sea lamprey in Michigan. Washington, D. C,
U. S. Department of Interior Fish & Wildlife Service, Special Scientific Report Fisheries.
Barraquand, F. and S. Benhamou (2008). "Animal movements in heterogeneous landscapes:
identifying profitable places and homogenous movement bouts." Ecology 89: 3336 - 3348.
Bates, D., M. Maechler, B. M. Bolker and S. Walker (2014). lme4: Linear mixed-effects models
using Eigen and S4. R package version 1.1-7.
Beletsky, D., J. H. Saylor and D. J. Schwab (1999). "Mean Circulation in the Great Lakes."
Journal of Great Lakes Research 25(1): 78-93.
Benhamou, S. and P. Bovet (1989). "How animals use their environment: a new look at kinesis."
Animal Behaviour 38(3): 375-383.
Bergstedt, R. A. and J. G. Seelye (1995). "Evidence for Lack of Homing by Sea Lampreys."
Transactions of the American Fisheries Society 124(2): 235-239.
Binder, T. R. and D. G. McDonald (2007). "Is there a role for vision in the behaviour of sea
lampreys (Petromyzon marinus) during their upstream spawning migration?" Canadian Journal
of Fisheries and Aquatic Sciences 64(10): 1403-1412.
Binder, T. R. and D. G. McDonald (2008). "The role of temperature in controlling diel activity in
upstream migrant sea lampreys (Petromyzon marinus)." Canadian Journal of Fisheries and
Aquatic Sciences 65(6): 1113-1121.
Binder, T. R., R. L. McLaughlin and D. G. McDonald (2010). "Relative Importance of Water
Temperature, Water Level, and Lunar Cycle to Migratory Activity in Spawning-Phase Sea
Lampreys in Lake Ontario." Transactions of the American Fisheries Society 139(3): 700-712.
Bjerselius, R., W. Li, J. H. Teeter, J. G. Seelye, P. B. Johnson, P. J. Maniak, G. C. Grant, C. N.
Polkinghorne and P. W. Sorenson (2000). "Direct behavioral evidence that unique bile acids
292

released by larval sea lamprey (Petromyzon marinus) function as a migratory pheromone."
Canadian Journal of Fisheries and Aquatic Sciences 57: 557-569.
Bradshaw, C., D. Sims and G. Hays (2007). "Measurement error causes scale-dependent
threshold erosion of biological signals in animal movement data." Ecol Appl 17(2): 628 - 638.
Chaudhury, D., L. Manella, A. Arellanos, O. Escanilla, T. A. Cleland and C. Linster (2010).
"Olfactory bulb habituation to odor stimuli." Behavioral neuroscience 124(4): 490-499.
Chung-Davidson, Y. W., H. Y. Wang, M. J. Siefkes, M. B. Bryan, H. Wu, N. S. Johnson and W.
M. Li (2013). "Pheromonal bile acid 3-ketopetromyzonol sulfate primes the neuroendocrine
system in sea lamprey." Bmc Neuroscience 14: 13.
Churchill, J. H., E. A. Ralph, A. M. Cates, J. W. Budd and N. R. Urban (2003). "Observations of
a negatively buoyant river plume in a large lake." Limnology and Oceanography 48(2): 884-894.
Clemens, B. J., T. R. Binder, M. F. Docker, M. L. Moser and S. A. Sower (2010). "Similarities,
Differences, and Unknowns in Biology and Management of Three Parasitic Lampreys of North
America." Fisheries 35(12): 580-594.
Dodge, K. L., B. Galuardi, T. J. Miller and M. E. Lutcavage (2014). "Leatherback Turtle
Movements, Dive Behavior, and Habitat Characteristics in Ecoregions of the Northwest Atlantic
Ocean." PLoS ONE 9(3): e91726.
Fauchald, P. and T. Tveraa (2003). "Using first-passage time in the analysis of area-restricted
search and habitat selection." Ecology 84(2): 282 - 288.
Fine, J. M. and P. W. Sorensen (2008). "Isolation and biological activity of the multi-component
sea lamprey migratory pheromone." Journal of chemical ecology 34(10): 1259-1267.
Galperin, B., L. H. Kantha, S. Hassid and A. Rosati (1988). "A Quasi-equilibrium Turbulent
Energy Model for Geophysical Flows." Journal of the Atmospheric Sciences 45(1): 55-62.
García Berdeal, I., B. M. Hickey and M. Kawase (2002). "Influence of wind stress and ambient
flow on a high discharge river plume." Journal of Geophysical Research: Oceans 107(C9): 3130.

293

Gore, J. A. (2007). Chapter 3 - Discharge Measurements and Streamflow Analysis. Methods in
Stream Ecology (Second Edition). F. R. Hauer and G. A. Lamberti. San Diego, Academic Press:
51-77.
Gurarie, E., C. Bracis, M. Delgado, T. D. Meckley, I. Kojola and C. M. Wagner (2015). "A
comparison of methods and practical guide to the behavioral analysis of animal movements."
Journal of Animal Ecology Special Issue.
Hoye, T. R., V. Dvornikovs, J. M. Fine, K. R. Anderson, C. S. Jeffrey, D. C. Muddiman, F.
Shao, P. W. Sorensen and J. Wang (2007). "Details of the structure determination of the sulfated
steroids PSDS and PADS: new components of the sea lamprey (petromyzon marinus) migratory
pheromone." Journal of Organic Chemistry 72(20): 7544-7550.
Johnson, N., A. Muhammad, H. Thompson, J. Choi and W. Li (2012). "Sea lamprey orient
toward a source of a synthesized pheromone using odor-conditioned rheotaxis." Behavioral
Ecology and Sociobiology 66(12): 1557-1567.
Johnson, N. S., M. A. Luehring, M. J. Siefkes and W. M. Li (2006). "Mating pheromone
reception and induced behavior in ovulating female sea lampreys." North American Journal of
Fisheries Management 26(1): 88-96.
Johnson, N. S., M. J. Siefkes, C. M. Wagner, H. Dawson, H. Y. Wang, T. Steeves, M. Twohey
and W. M. Li (2013). "A synthesized mating pheromone component increases adult sea lamprey
(Petromyzon marinus) trap capture in management scenarios." Canadian Journal of Fisheries and
Aquatic Sciences 70(7): 1101-1108.
Johnson, N. S., S. S. Yun, H. T. Thompson, C. O. Brant and W. M. Li (2009). "A synthesized
pheromone induces upstream movement in female sea lamprey and summons them into traps."
Proceedings of the National Academy of Sciences of the United States of America 106(4): 10211026.
Jorgensen, J. C. and J. F. Kitchell (2005). "Growth potential and host mortality of the parasitic
phase of the sea lamprey (Petromyzon marinus) in Lake Superior." Canadian Journal of Fisheries
and Aquatic Sciences 62(10): 2343-2353.
Kells, A. and D. Goulson (2001). "Evidence for Handedness in Bumblebees." Journal of Insect
Behavior 14(1): 47-55.
Li, W., M. Twohey, M. Jones and M. Wagner (2007). "Research to Guide Use of Pheromones to
Control Sea Lamprey." Journal of Great Lakes Research 33, Supplement 2(0): 70-86.
294

Lohmann, K. J., C. M. F. Lohmann and C. S. Endres (2008). "The sensory ecology of ocean
navigation." Journal of Experimental Biology 211(11): 1719-1728.
Luehring, M. A., C. M. Wagner and W. M. Li (2011). "The efficacy of two synthesized sea
lamprey sex pheromone components as a trap lure when placed in direct competition with natural
male odors." Biological Invasions 13(7): 1589-1597.
Meckley, T., C. Holbrook, C. Wagner and T. Binder (2014a). "An approach for filtering
hyperbolically positioned underwater acoustic telemetry data with position precision estimates."
Animal Biotelemetry 2(1): 7.
Meckley, T., C. Wagner and E. Gurarie (2014b). "Coastal movements of migrating sea lamprey
(Petromyzon marinus) in response to a partial pheromone added to river water: implications for
management of invasive populations." Can J Fish Aquat Sci 71(4): 533 - 544.
Meckley, T. D., C. M. Wagner and M. A. Luehring (2012). "Field evaluation of larval odor and
mixtures of synthetic pheromone components for attracting migrating sea lampreys in rivers."
Journal of Chemical Ecology 38(8): 1062-1069.
Mellor, G. L. and T. Yamada (1982). "Development of a turbulence closure model for
geophysical fluid problems." Reviews of Geophysics 20(4): 851-875.
Moore, H. H. and I. P. Schleen (1980). "Changes in spawning runs of Sea Lamprey
(Petromyzon-Marinus) in selected streams of Lake Superior after chemical control." Canadian
Journal of Fisheries and Aquatic Sciences 37(11): 1851-1860.
Morman, R. H., D. W. Cuddy and P. C. Rugen (1980). "Factors Influencing the Distribution of
Sea Lamprey (Petromyzon marinus) in the Great Lakes." Canadian Journal of Fisheries and
Aquatic Sciences 37(11): 1811-1826.
Morshedian, A. and Gordon L. Fain (2015). "Single-Photon Sensitivity of Lamprey Rods with
Cone-like Outer Segments." Current Biology 25(4): 484-487.
Mullett, K. M., J. W. Heinrich, J. V. Adams, R. J. Young, M. P. Henson, R. B. McDonald and
M. F. Fodale (2003). "Estimating Lake-wide Abundance of Spawning-phase Sea Lampreys
(Petromyzon marinus) in the Great Lakes: Extrapolating from Sampled Streams Using
Regression Models." Journal of Great Lakes Research 29, Supplement 1(0): 240-252.

295

Nguyen, T. D., P. Thupaki, E. J. Anderson and M. S. Phanikumar (2014). "Summer circulation
and exchange in the Saginaw Bay-Lake Huron system." Journal of Geophysical Research:
Oceans 119(4): 2713-2734.
Pe'er, G., D. Saltz, H.-H. Thulke and U. Motro (2004). "Response to topography in a hilltopping
butterfly and implications for modelling nonrandom dispersal." Animal Behaviour 68(4): 825839.
R Development Core Team (2015). R: A language and environment for statistical computing.
Vienna, Austria, R Foundation for Statistical Computing.
Rao, Y. R. and D. J. Schwab (2007). "Transport and mixing between the coastal and offshore
waters in the Great Lakes: a review." Journal of Great Lakes Research 33(1): 202-218.
Siefkes, M. J., S. R. Winterstein and W. Li (2005). "Evidence that 3-keto petromyzonol sulphate
specifically attracts ovulating female sea lamprey, Petromyzon marinus." Animal Behaviour 70:
1037-1045.
Silva, S., M. J. Araújo, M. Bao, G. Mucientes and F. Cobo (2014). "The haematophagous
feeding stage of anadromous populations of sea lamprey Petromyzon marinus: low host
selectivity and wide range of habitats." Hydrobiologia 734(1): 187-199.
Smagorinsky, J. (1963). "General circulation experiments with the primiate equations." Monthly
Weather Review 91(3): 99-164.
Sorensen, P. W., L. A. Vrieze and J. M. Fine (2003). "A multi-component migratory pheromone
in the sea lamprey." Fish Physiology and Biochemistry 28(1-4): 253-257.
Teeter, J. H. (1980). "Pheromone communication in sea lampreys (Petromyzon marinus):
implications for population management." Canadian Journal of Fisheries and Aquatic Sciences
37: 2123-2132.
Vickers, N. J. (2000). "Mechanisms of animal navigation in odor plumes." Biological Bulletin
198(2): 203-212.
Vrieze, L., R. Bergstedt and P. Sorensen (2011). "Olfactory-mediated stream-finding behavior of
migratory adult sea lamprey (Petromyzon marinues)." Can J Fish Aquat Sci 68: 523 - 533.

296

Vrieze, L. A., R. Bjerselius and P. W. Sorensen (2010). "Importance of the olfactory sense to
migratory sea lampreys Petromyzon marinus seeking riverine spawning habitat." Journal of Fish
Biology 76(4): 949-964.
Vrieze, L. A. and P. W. Sorensen (2001). "Laboratory assessment of the role of a larval
pheromone and natural stream odor in spawning stream localization by migratory sea lamprey
(Petromyzon marinus)." Canadian Journal of Fisheries and Aquatic Sciences 58(12): 2374-2385.
Wagner, C. M., M. B. Twohey and J. M. Fine (2009). "Conspecific cueing in the sea lamprey: do
reproductive migrations consistently follow the most intense larval odour?" Animal Behaviour
78(3): 593-599.
Waldman, J., C. Grunwald and I. Wirgin (2008). Sea lamprey Petromyzon marinus: an exception
to the rule of homing in anadromous fishes.
Zimmer-Faust, R. K., C. M. Finelli, N. D. Pentcheff and D. S. Wethey (1995). "Odor plumes and
animal navigation in turbulent water flow: a field study." The Biological Bulletin 188(2): 111116.

297

CHAPTER 5

ISOLATING OPPORTUNITIES FOR MANIPULATING SPAWNING HABITAT
SELECTION OF INVASIVE SEA LAMPREY (PETROMYZON MARINUS) IN THE
LAURENTIAN GREAT LAKES

ABSTRACT
Recent advances in the understanding of the sea lamprey lake migration have shed light
on how sea lamprey encounter, enter, and ascend rivers in the Great Lakes. Herein, a simple
framework is provided for discussion of the factors that influence the transition between each
step of the migration and how these factors could be manipulated for improved management of
the invasive sea lamprey. A profitable alternative management effect either reduces the amount
of pesticide applied, increases the effect of the pesticide (reduces cost per kill), or allows for the
removal of barriers increasing stream continuity, while maintaining or decreasing the sea
lamprey population. Six management scenarios are discussed including what behaviors would be
manipulated in each scenario, what aspects of the migration need to be better understood to
determine the probability of success, and what scientific advancements are necessary for each
manipulation to be performed in nature.
INTRODUCTION
The invasive sea lamprey has been historically controlled by preventing access to habitat
with dams and killing larvae with pesticides in the areas of watersheds sea lamprey could reach,
driving the population towards, but not to, management targets designed to protect the ecosystem
(Christie and Goddard 2003). Alternatives to traditional management options or maximization
of current control methods could be valuable if they achieve sea lamprey management objectives
or maintain the current efficacy with fewer negative effects. The current methods include
298

lampricides which kill non-target species (Boogaard et al. 2003) and are expensive (Christie and
Goddard 2003), and physical barriers that prevent access to other migratory species. Particularly
promising is the potential to use natural odorants that are part of the habitat and mate selection
decisions of sea lamprey to manipulate their behavior for control (Li et al. 2007).
Three elements make behavioral manipulation of a pest for protection of a resource
possible: a known behavior, a known stimulus that initiates the behavior, and a scenario in which
the stimulus can be presented to induce the behavior subsequently resulting in protection of a
resource (Foster and and Harris 1997). Behavioral manipulation of the potadromous Great Lakes
sea lamprey is promising, given their known reliance on odorants to select habitat (Teeter 1980).
However, a lack of understanding of much of their life cycle and the challenge of identifying and
replicating the odorants has made this alternative control method difficult to achieve.
Sea lamprey are known to alter behavior in response to the odor of larval lamprey, adult
sea lamprey, and dead adult or larval sea lamprey as they complete a non-homing return
migration from offshore feeding grounds to riverine spawning grounds (Bjerselius et al. 2000,
Siefkes 2000, Fine et al. 2004, Waldman et al. 2008, Wagner et al. 2011, Bals 2012). The
odorants represent a potential tool for controlling habitat selection, trapping sea lamprey, or
disrupting the migration system by masking odorants in the system through pumping synthetic
replicates of the odors at high concentrations (Teeter 1980), classic concepts of insect
pheromone-based control strategies (Steiner 1952, Carde and Minks 1995). Odorants can act as a
releaser (e.g. a male sex pheromone component, 3kPZS, induces upstream search for mates), and
(or) as a primer, (e.g., 3kPZS modulates release of gonadotropic releasing hormone, resulting in
maturation of immature sea lamprey), helping to synchronize the migration (Brennan and Zufall
2006, Chung-Davidson et al. 2013). We are focused on releaser effects for manipulating

299

behavior. Though progress has been made in understanding the migration system and
identification of some odorants, an understanding of how to best use odor in a control scenario
has not materialized. Through a simple migration model that is analogous in some respects to
trap catch steps previously developed for sea lamprey (Bravener and McLaughlin 2013), we
identify the steps of the migration, although in this case the sea lamprey reaches spawning
habitat instead of entering a trap and we identify the steps of the migration, the factors that
influence transitions, and how the process can be manipulated.
RIVER ENTRY FRAMEWORK
The framework covers river plume encounter, river entry, and river ascension to the
spawning grounds. It allows us to outline what factors influence each step and to discuss the
potential to manipulate these steps to better manage sea lamprey in the Great Lakes (Figure 5.1).
River Plume Encounter
The number of sea lamprey to encounter a river plume will depend on the number of
individuals that pass the river mouth and how far they are from shore compared to the distance
the river plume extends from shore (Figure 5.2). The probability of encounter for each river will
vary with the size of the river plume, principally how far river water extends from the coast and
how far from shore sea lamprey traverse parallel to the coast in that region (Chapter 4). River
plume size varies with river discharge, coastal bathymetry, coastline configuration at the river
mouth (e.g., embayment trapping water), and wind driven lake current (Churchill et al. 2003,
Rao and Schwab 2007). Wind driven currents and the coastal boundary are the dominant features
regulating the extent of the river plume in the Great Lakes (Churchill et al. 2003). The number of
sea lamprey available to encounter the plume will depend on the number of sea lamprey to return

300

to the particular region following the parasitic stage and whether they reach a given river or stop
at a previous river.
River water alone is attractive to sea lamprey in two-choice tests and appears to be the
primary feature that initiates a transition to local search behavior around the river mouth in the
lake (Vrieze and Sorensen 2001, Vrieze et al. 2011, Meckley et al. 2014, Chapter 4). However
nose plugged sea lamprey were still able to locate a nearby river, albeit much less efficiently
(12.5 %) (Vrieze et al. 2010), indicating that other features such as conspecific attraction or
physical features such as substrate, water color (tannins), temperature gradients, or outflow may
be sufficient for a migrant to recognize their proximity to a river mouth and enter the river.
River Entrance, Upstream Movement, and Tributary Selection
The timing of river entry and upstream movement is related to river temperature, changes
in river temperature, stream flow, and sexual maturation (Almeida et al. 2002, Binder et al.
2010). River and tributary selection appears most reliant on the presence of larval sea lamprey
odor, a form of public information that provides evidence of spawning and rearing habitat quality
(Bjerselius et al. 2000, Sorensen et al. 2005, Waldman et al. 2008). However, additional odorants
influence sea lamprey behavior and may influence the migration. For example, it is unclear how
public information such as larval odor of other lamprey species or decaying lamprey larvae
modulates river entry.
Unlike specialized sex pheromones that are often a compilation of a few compounds at a
specific ratio (Ando et al. 2004), tuned as a result of receiver bias, when an odor releaser is
rewarded with more mates because of releasing a more attractive odor (Buchinger et al. 2013);
the unspecialized release of larval odor is unlikely to be discernable as a species specific signal
by adults and may be a larger number of components (Li et al. 1995). The composition of the
301

gut microbiome of sea lamprey is just beginning to be explored (Tetlock et al. 2012), but if the
larval diet, gut microbiome, and subsequent bile salts released by larvae are similar between
different types of larvae, larval odor may be a general lamprey habitat quality cue. This is
supported by evidence that different species of lamprey showed attraction to different lamprey
species larval odors and larvae of different species release similar components (Fine et al. 2004).
Although a blend of active compounds must be present (Fine and Sorensen 2008, Meckley et al.
2012), odors are released to a constrained environment and ratios may not be important (Derby
and Sorensen 2008). Little is known about the influence of the odor of dead larvae on the sea
lamprey migration, beyond that the odor of dead lamprey larvae is repulsive to migratory sea
lamprey in lab tests (Imre et al. 2010, Wagner et al. 2011, Bals and Wagner 2012), and has been
hypothesized to cause habitat avoidance where larval sea lamprey have died during the winter.
Migratory sea lamprey avoid the odor of dead adults and larvae in the lab, though it is unclear if
different compounds are responsible for the avoidance (Bals and Wagner 2012). These odor cues
are unlikely to be species specific.
The second type of information widely relied upon by animals but poorly described in sea
lamprey is reliance on interactions with mates, known as conspecific attraction (Doligez et al.
2003). The conspecific attraction strategy always coexists with other strategies because it allows
individuals to utilize the decisions of other individuals and ensures the availability of mates
(Doligez et al. 2003). The presence of mates could be determined visually or through olfaction as
sea lamprey migrants can smell the odor of other migrants at different stages of the migration
and the odor of dead mates upstream. To date the natural progression of the sea lamprey
migration has been poorly observed as studies utilize subjects trapped in a river (Vrieze et al.
2010, Vrieze et al. 2011, Meckley et al. 2014 ), and activities like staging at a river mouth could

302

be a major component of the migration (Applegate 1950, Clemens et al. 2010), as it could signal
the presence of mates and synchronize the start of the migration. Unlike larval odors these cues
are more likely to be species specific migration cues.
The odor of adult conspecifics synchronizes adult maturation in the river (ChungDavidson et al. 2013) and the presence of a male released sex pheromone may encourage female
river entry (Meckley et al. 2014) and tributary selection. However, sex pheromone was not found
to be attractive to immature migratory sea lamprey caught in early spring (Siefkes et al. 2005),
and may never be available at a river mouth to migrants experiencing a natural migration
progression. More progress has been made towards synthetically replicating the sex pheromone
(Johnson et al. 2005), though an effective large scale management scenario has not been
identified to date (Luehring et al. 2011, Johnson et al. 2013) and as a trap bait where an effect
has been shown, the method may only be effective at low sea lamprey densities (El-Sayed et al.
2006, Johnson et al. 2013).
The odor of dead migratory sea lamprey is repulsive to non-mature male and female sea
lamprey in lab tests (Imre et al. 2010, Wagner et al. 2011, Bals and Wagner 2012) and may
represent predator avoidance or the end of spawning, indicating sea lamprey have already
reproduced and senesced. The function of dead odor is further confounded by the finding that
migrants become habituated to the odor of dead adults after approximately four hours of
exposure (Bals 2012), and initial field tests suggest that the odor of dead sea lamprey is
insufficient to prevent river entry and instead increased river entry (Tom Luhring, Personal
Communication). The odor functions to bias the side of the stream channel occupied by sea
lamprey during upstream movement when the odor can be avoided (Bals 2012), but may increase

303

upstream movement past the source of the odor when the odor is mixed throughout the entire
channel (Tom Luhring, Personal Communication).
River Retention
If a river contains a large larval population, spawning habitat, and conspecific mates, sea
lamprey are unlikely to reverse course after river entrance. This is supported by a two-year study
that never observed sea lamprey exiting a river after they entered and moved more than 300 m
upstream, and some individuals enter the immediate river mouth without ascending the river
(Applegate 1950, Meckley et al. 2014, Chapter 4), together leading to the suggestion that the
decision to enter occurs at the river mouth (Chapter 4). Alternatively, some sea lamprey return to
a river plume over multiple nights without ever entering the river, regularly reversing between
movement inside and outside of the river plume, and moving very close to the river mouth,
although this occurred less with higher larval odor (Chapter 4). Observations that sea lamprey
moved to the river mouth without entering supports that the concept that the decision to enter a
river usually occurs at the river mouth.
MANIPULATING THE MIGRATION
Management of an invasive species is difficult because a strong understanding of the
species biology and behavior in the new environment is often unavailable. Because eradication
of established invasive species is rare (Simberloff 2003), an explicit definition of what
constitutes a profitable outcome is required. We define a profitable alternative management
effect as one that reduces the amount of pesticide applied, increases the effect of the pesticide
(reduces cost per kill), or allows for the removal of barriers, while maintaining or decreasing the
sea lamprey population. The results must be measureable to determine whether a strategy has a
profitable effect, a difficult proposition given the dynamic pesticide treatment schedule. An
304

effective odor-based control method must identify and synthesize the active compounds cheaply,
a very elusive process (Li et al. 2007), that we will not discuss in detail. We discuss six
management scenarios for manipulating the sea lamprey migration for improved management
(Table 5.1).
River Plume Encounter
Encounter with a river plume must occur for a sea lamprey to enter the river. River plume
encounter could be manipulated by changing the number of individuals to pass the river,
physically influencing the distance that sea lamprey move from the coast while passing the river
plume, through providing chemical information indicative of a river plume farther from shore, or
by physically increasing the distance the river plume extends into the lake. Physically altering
the distance sea lamprey travel from shore or the size of the river plume would likely require
large construction investments and extensive research. However, there are navigational features
that may provide suggestions for how manipulation of the river plume size could be achieved
(Figure 5.3). An alternative would be the release of river odor out in front of a river mouth that
indicates the presence of a nearby river, inducing local search (Table 5.1, Option 1). The
encounter rate may be challenging to alter; thereby changing the number of sea lamprey to pass a
river is the most plausible approach if entry rates of nearby rivers can be influenced. River plume
encounter is an event not a decision. For this reason it may not have management value unless
encounter can be prevented, however if it is combined with some decision manipulation
following encounter, the encounter rate becomes important.
Cohort-Size Maintenance
If synthetic larval odor was available, the most immediate value could come in the form
of maintaining the number of sea lamprey to enter streams that require pesticide treatment every
305

four years (Table 5.1, Option 2; Figure 5.2, iii). If the larval odor signal is maintained, migrant
dispersion could be abated and could reduce the cohort size of sea lamprey to enter other rivers,
potentially including small river plumes that may get irregular sea lamprey recruitment. It is
unconfirmed whether the increased river entry rates observed each year after treatment is due
only to the increase in larval odor, or if for example an increase in river recruitment of 2.4-6.8%
of the total population of Lake Huron (i.e., as observed between 1 and 3 years post treatment in
the Ocqueoc River; Figure 5.4), is sufficient for a lake wide management value. A potential
downside to the strategy is the need for continued pesticide treatments. There are many smaller
streams that have sea lamprey populations that are too small, or are too dispersed to be treated
effectively and never or rarely meet the standard to qualify for treatment (Jones et al. 2003).
This strategy works under the premise that the constant presence of larval odor in rivers with
large river plumes that are regularly treated would reduce the spreading of migrants to other
rivers by reducing the coastal search time and associated likelihood of encountering smaller
rivers. This strategy is only important if sea lamprey are spread to non-expert treatment streams
that do not get treated every four years. If sea lamprey are only being pushed between regularly
treated streams, then the cohort-size maintenance approach is not valuable.
Recruitment to habitat deficient for larval survival
Synthetic larval odor could encourage entry into rivers with poor spawning or rearing
habitat (Li et al. 2007; Table 5.1, Option 3). We suspect a river with limited habitat, such as a
river with a physical barrier to movement low in the river before spawning habitat is reached, is
unlikely to retain migrants, as sea lamprey have been observed exiting large rivers with
waterfalls low in the river (Teeter 1980). However if a river has ample spawning habitat and
poor larval habitat, sea lamprey may stay and spawn. It would need to be confirmed that larvae
306

do survive poorly in any river site considered for this strategy. Encouraging sea lamprey into a
river lacking larvae or with few larvae could be counterproductive if enough larvae survive to
warrant pesticide treatment of the river or if it results in an unacceptable number of parasites. It
is unclear if there are rivers that fit this strategy as most rivers that do not contain larvae, are
intermittent or have very low flow (Teeter 1980).
Chemical barrier and Push-Pull
The potential to create a chemical barrier at a river mouth or at tributaries has begun to be
tested (Bals 2012, Bals and Wagner 2012, C. Michael Wagner personal communication; Figure
5.2, iv). Currently only physical barriers are plausible for blocking sea lamprey access to rivers.
Dead adult odor does not prevent river entry, though the dead odor of larvae has not been tested
in this scenario (Table 5.1, Option 6). Parts of a watershed may be blocked through application
of dead migratory sea lamprey odor alone, although the application method will need to avoid
extended exposure resulting in odor habituation which was observed after sea lamprey were
exposed to the odor for approximately four hours (Bals 2012, Bals and Wagner 2012, Tom
Luehring personal communication). The effect may be strengthened with a “pull” comprised by
larval odor, as sea lampreys prefer to move in the stronger larval signal (Wagner et al. 2009,
Meckley et al. 2012; Table 5.1, Option 4; Figure 5.2, v, vi).
This method of providing a chemical barrier to part of the watershed has promise to limit
sea lamprey to a narrow region of the watershed that could reduce the pesticide necessary for
treatment. The level of success necessary for a management value would need to be evaluated
and the method will fail to manipulate individuals that spawn before encountering the selected
stream bifurcation. If sea lamprey are sufficiently repelled to delay the need for pesticide

307

treatment, or allow for a physical barrier to be removed or placed higher in the watershed, the
method would have value.
Larval Odor Removal (TFM)
We only highlight the use of TFM as an alternative control measure because the way it is
applied removes the larval signal from a river and has a measurable effect on entry into
neighboring rivers (e.g., Teeter 1980; Table 5.1, Option 5). The more effective the removal of
lamprey larvae in the river the more effectively the treatment will likely reduce entry the
following year, creating a repellent from the absence of a cue. In the circumstance pesticide
treatment occurs in many rivers, the reduction in entry by migrants the year after treatment is a
negative management effect. If TFM treatment is required every four years before larvae can
transform to the parasitic stage, no management advantage is gained from the reduction in entry
and instead sea lamprey are spread to other rivers. Although where migrants are dispersed the
year after treatment, is poorly understood. In many cases, only a few rivers receive a large
proportion of the sea lamprey population estimated in the lake, removing the larval signal from
one of these rivers can cause a substantial shift in where migrants go to spawn and the more time
spent searching may increase the likelihood of encountering smaller river plumes that are rarely
encountered (Meckley et al. 2014).
Based on similar logic, we would recommend against the treatment of multiple large
rivers in a region in the same year. Ideally alternating pesticide application between large rivers
may result in shifting recruitment between the major rivers rather than encouraging encounter
and entry with many smaller rivers, a concept that needs further testing. Similarly we
recommend against concepts like treatment of all rivers in back to back years, as this is likely to
result in spreading migrants broadly. Encouraging dispersal of a pest is counterproductive to
308

chemical based treatment. Finally shifting treatments to after the spawning season would be the
most valuable as the larval signal that acts to congregate migrants would congregate sea lamprey
in rivers that were about to be treated. It is likely infeasible to treat all rivers after spawning
occurs although the more the treatments can be shifted to after the spawning season the more
effectively the treatment program would be using the natural sea lamprey habitat selection
decisions to the control program’s advantage.
CONCLUSION
The importance of pesticide application and physical barriers to sea lamprey movement
will not be quickly replaced. As knowledge of the sea lamprey migration increases more options
become available for how the sea lamprey migration can be manipulated. There are variants of
these six potential alternative management strategies, although they cover the primary aspects of
the migration that can be manipulated (Figure 5.1; Figure 5.2).

309

APPENDIX

310

Table 5.1: Six general strategies for manipulating sea lamprey migration behavior are presented
including the stimulus affected, how the stimulus is affected (action), the goal of the
manipulation and the type of manipulation. Only option one does not stand alone as a
management strategy as it manipulates behavior but not a decision to select habitat.

Stimulus
Altered

Manipulation Action

Intended
Result

Type of
Manipulation

1

River water

Release river water out in
front of a river mouth

River Plume
Assessment

Pull/transition

2

Larval Odor

Add larval odor to “expert
judgement” stream

River Entry

Pull

3

Larval Odor

Add larval odor to stream
with poor larval survival

River Entry

Pull

4

Larval Odor
And
Dead Odor

Add larval odor and dead
lamprey odor

Tributary
selection

Push-Pull

5

Larval Odor

Remove larval odor signal
via lampricide application
(TFM)

River rejection

Push

6

Dead Larval
Odor

Add dead lamprey odor to
entire river

River avoidance

Push

311

Figure 5.1: The steps of the spawning migration are analogous to the steps involved in trapping
sea lamprey (Bravener and McLaughlin 2013), although in this case the river takes the place of
the trap and the model depicts the transitions that occur when a sea lamprey enters a river in the
great lakes. The arrows refer to the probabilities of progressing from the lake to spawning habitat
or reversing through any prior step. The P refers to the probability of the event.

312

Figure 5.2: The distance sea lamprey travel from shore (a, dotted arrow), the distance the river plume extends from the coast (b), the
presence of other rivers on the coast and the number of rivers encountered prior to reaching a river (c), are all important to a sea
lamprey encountering a particular river plume (Region of Coast). Following encounter with a river plume (d) sea lamprey become

313

Figure 5.2: (cont’d)
available for manipulation of their habitat selection decisions including river entry (e), or tributary entry (f). Five general
manipulation scenarios are shown including the manipulation of river plume encounter (event) (ii) or habitat selection (decision) (iiivi). The operator (+, - , =), indicates how the manipulation of adding an attractant (“attract”), a repellant, or removing an attractant
(“repel”), influences encounter rates at nearby rivers and entry of those rivers. The manipulation of decisions through the addition of
odorants (iii-vi), uses general terms based on the effect desired and not specific odorants that will cause the desired effect. Encounter
is shown to change due to the order of sequential river plumes. Encounter at a river can change due to a manipulation at a prior river
but not the current river manipulated (decrease, iii; increase, iv). It is unclear how past experience might influence entry of future
rivers (iii, “?”).

314

Figure 5.3: Many river mouths have preexisting structures that may extend the distance the river
plume would otherwise reach from shore (i.e., as seen west of river mouths in Ludington, MI
(left) and Sagatuck, MI (right)). Structures of this type designed for boat traffic could present
unique solutions for influencing the sea lamprey migration and may be the only scenario that
would create a river plume extending a greater distance from shore than would otherwise occur
for that river. In each image the river plume can be faintly seen due to darker turbidity.

315

Figure 5.4: The percentage of the entire population of adult sea lamprey in Lake Huron that
return to the Ocqueoc River to spawn from 1995 to 2015 based on mark recapture (black circles)
and the change in entry between years (blue circles). A change of entry rate of 2 represents a
doubling in the rate of entry. Vertical lines represent the times of TFM treatment to the Ocquoec
River. The change in entry rate from one year after TFM treatment to the second year shows that
the increase in entry rate between 2010 and 2011 (D) fits within the historical increase in rate (A:
1.8, B: 1.3, C: 3.1, D: 2.1).

316

REFERENCES

317

REFERENCES

Almeida, P. R., B. R. Quintella and N. M. Dias (2002). "Movement of radio-tagged anadromous
sea lamprey during the spawning migration in the River Mondego (Portugal)." Hydrobiologia
483(1-3): 1-8.
Ando, T., S.-i. Inomata and M. Yamamoto (2004). Lepidopteran Sex Pheromones. The
Chemistry of Pheromones and Other Semiochemicals I. S. Schulz, Springer Berlin Heidelberg.
239: 51-96.
Applegate, V. C. (1950). The natural history of the sea lamprey in Michigan. Washington, D. C,
U. S. Department of Interior Fish & Wildlife Service, Special Scientific Report Fisheries.
Bals, J. D. (2012). The sea lamprey alarm response: field and laboratory investigations. M.S.,
Michigan State Unviersity.
Bals, J. D. and C. M. Wagner (2012). "Behavioral responses of sea lamprey (Petromyzon
marinus) to a putative alarm cue derived from conspecific and heterospecific sources."
Behaviour 149(9): 901-923.
Binder, T. R., R. L. McLaughlin and D. G. McDonald (2010). "Relative Importance of Water
Temperature, Water Level, and Lunar Cycle to Migratory Activity in Spawning-Phase Sea
Lampreys in Lake Ontario." Transactions of the American Fisheries Society 139(3): 700-712.
Bjerselius, R., W. Li, J. H. Teeter, J. G. Seelye, P. B. Johnson, P. J. Maniak, G. C. Grant, C. N.
Polkinghorne and P. W. Sorenson (2000). "Direct behavioral evidence that unique bile acids
released by larval sea lamprey (Petromyzon marinus) function as a migratory pheromone."
Canadian Journal of Fisheries and Aquatic Sciences 57: 557-569.
Boogaard, M. A., T. D. Bills and D. A. Johnson (2003). "Acute toxicity of TFM and a
TFM/niclosamide mixture to selected species of fish, including lake sturgeon (Acipenser
fulvescens) and mudpuppies (Necturus maculosus), in laboratory and field exposures." Journal of
Great Lakes Research 29: 529-541.
Bravener, G. A. and R. L. McLaughlin (2013). "A behavioural framework for trapping success
and its application to invasive sea lamprey." Canadian Journal of Fisheries and Aquatic Sciences
70(10): 1438-1446.

318

Brennan, P. A. and F. Zufall (2006). "Pheromonal communication in vertebrates." Nature
444(7117): 308-315.
Buchinger, T. J., H. Wang, W. Li and N. S. Johnson (2013). Evidence for a receiver bias
underlying female preference for a male mating pheromone in sea lamprey.
Carde, R. T. and A. K. Minks (1995). "Control of moth pests by mating disruption: successes
and constraints." Annual review of entomology 40(1): 559-585.
Christie, G. C. and C. I. Goddard (2003). "Sea Lamprey International Symposium (SLIS II):
Advances in the integrated management of sea lamprey in the Great Lakes." Journal of Great
Lakes Research 29: 1-14.
Chung-Davidson, Y. W., H. Y. Wang, M. J. Siefkes, M. B. Bryan, H. Wu, N. S. Johnson and W.
M. Li (2013). "Pheromonal bile acid 3-ketopetromyzonol sulfate primes the neuroendocrine
system in sea lamprey." Bmc Neuroscience 14: 13.
Churchill, J. H., E. A. Ralph, A. M. Cates, J. W. Budd and N. R. Urban (2003). "Observations of
a negatively buoyant river plume in a large lake." Limnology and Oceanography 48(2): 884-894.
Clemens, B. J., T. R. Binder, M. F. Docker, M. L. Moser and S. A. Sower (2010). "Similarities,
Differences, and Unknowns in Biology and Management of Three Parasitic Lampreys of North
America." Fisheries 35(12): 580-594.
Derby, C. and P. Sorensen (2008). "Neural Processing, Perception, and Behavioral Responses to
Natural Chemical Stimuli by Fish and Crustaceans." Journal of Chemical Ecology 34(7): 898914.
Doligez, B., C. Cadet, E. Danchin and T. Boulinier (2003). "When to use public information for
breeding habitat selection? The role of environmental predictability and density dependence."
Animal Behaviour 66(5): 973-988.
El-Sayed, A. M., D. M. Suckling, C. H. Wearing and J. A. Byers (2006). Potential of Mass
Trapping for Long-Term Pest Management and Eradication of Invasive Species.
Fine, J., L. Vrieze and P. Sorensen (2004). "Evidence That Petromyzontid Lampreys Employ a
Common Migratory Pheromone That Is Partially Comprised of Bile Acids." Journal of Chemical
Ecology 30(11): 2091-2110.

319

Fine, J. M. and P. W. Sorensen (2008). "Isolation and biological activity of the multi-component
sea lamprey migratory pheromone." Journal of chemical ecology 34(10): 1259-1267.
Foster and, S. P. and M. O. Harris (1997). "Behavioral manipulation methods for insect pestmanagement." Annual Review of Entomology 42(1): 123-146.
Imre, I., G. E. Brown, R. A. Bergstedt and R. McDonald (2010). "Use of Chemosensory Cues as
Repellents for Sea Lamprey: Potential Directions for Population Management." Journal of Great
Lakes Research 36(4): 790-793.
Johnson, N. S., M. J. Siefkes and W. M. Li (2005). "Capture of ovulating female sea lampreys in
traps baited with spermiating male sea lampreys." North American Journal of Fisheries
Management 25(1): 67-72.
Johnson, N. S., M. J. Siefkes, C. M. Wagner, H. Dawson, H. Y. Wang, T. Steeves, M. Twohey
and W. M. Li (2013). "A synthesized mating pheromone component increases adult sea lamprey
(Petromyzon marinus) trap capture in management scenarios." Canadian Journal of Fisheries and
Aquatic Sciences 70(7): 1101-1108.
Jones, M. L., R. A. Bergstedt, M. B. Tvohey, M. F. Fodale, D. W. Cuddy and J. W. Slades
(2003). "Compensatory mechanisms in great lakes sea lamprey populations: Implications for
alternative control strategies." Journal of Great Lakes Research 29: 113-129.
Li, W., M. Twohey, M. Jones and M. Wagner (2007). "Research to Guide Use of Pheromones to
Control Sea Lamprey." Journal of Great Lakes Research 33, Supplement 2(0): 70-86.
Li, W. M., P. W. Sorensen and D. D. Gallaher (1995). "The olfactory system of migratory adult
sea lamprey (Petromyzon marinus) is specifically and acutely sensitive to unique bile-acids
released by conspecific larvae." Journal of General Physiology 105(5): 569-587.
Luehring, M. A., C. M. Wagner and W. M. Li (2011). "The efficacy of two synthesized sea
lamprey sex pheromone components as a trap lure when placed in direct competition with natural
male odors." Biological Invasions 13(7): 1589-1597.
Meckley, T., C. Wagner and E. Gurarie (2014). "Coastal movements of migrating sea lamprey
(Petromyzon marinus) in response to a partial pheromone added to river water: implications for
management of invasive populations." Can J Fish Aquat Sci 71(4): 533 - 544.

320

Meckley, T. D., C. M. Wagner and M. A. Luehring (2012). "Field evaluation of larval odor and
mixtures of synthetic pheromone components for attracting migrating sea lampreys in rivers."
Journal of Chemical Ecology 38(8): 1062-1069.
Rao, Y. R. and D. J. Schwab (2007). "Transport and mixing between the coastal and offshore
waters in the Great Lakes: a review." Journal of Great Lakes Research 33(1): 202-218.
Siefkes, M. J. (2000). Spermiating male sea lampreys release a sex pheromone that attracts postovulatory female sea lampreys. Masters, Michigan State University.
Siefkes, M. J., S. R. Winterstein and W. Li (2005). "Evidence that 3-keto petromyzonol sulphate
specifically attracts ovulating female sea lamprey, Petromyzon marinus." Animal Behaviour 70:
1037-1045.
Simberloff, D. (2003). "How Much Information on Population Biology Is Needed to Manage
Introduced Species?" Conservation Biology 17(1): 83-92.
Sorensen, P. W., J. M. Fine, V. Dvornikovs, C. S. Jeffrey, F. Shao, J. Wang, L. A. Vrieze, K. R.
Anderson and T. R. Hoye (2005). "Mixture of new sulfated steroids functions as a migratory
pheromone in the sea lamprey." Nature Chemical Biology 1(6): 324-328.
Steiner, L. F. (1952). "Methyl eugenol as an attractant for oriental fruit fly." Journal of Economic
Entomology 45(2): 241-248.
Teeter, J. H. (1980). "Pheromone communication in sea lampreys (Petromyzon marinus):
implications for population management." Canadian Journal of Fisheries and Aquatic Sciences
37: 2123-2132.
Tetlock, A., C. K. Yost, J. Stavrinides and R. G. Manzon (2012). "Changes in the Gut
Microbiome of the Sea Lamprey during Metamorphosis." Applied and Environmental
Microbiology 78(21): 7638-7644.
Vrieze, L., R. Bergstedt and P. Sorensen (2011). "Olfactory-mediated stream-finding behavior of
migratory adult sea lamprey (Petromyzon marinues)." Can J Fish Aquat Sci 68: 523 - 533.
Vrieze, L. A., R. Bjerselius and P. W. Sorensen (2010). "Importance of the olfactory sense to
migratory sea lampreys Petromyzon marinus seeking riverine spawning habitat." Journal of Fish
Biology 76(4): 949-964.

321

Vrieze, L. A. and P. W. Sorensen (2001). "Laboratory assessment of the role of a larval
pheromone and natural stream odor in spawning stream localization by migratory sea lamprey
(Petromyzon marinus)." Canadian Journal of Fisheries and Aquatic Sciences 58(12): 2374-2385.
Wagner, C. M., E. M. Stroud and T. D. Meckley (2011). "A deathly odor suggests a new
sustainable tool for controlling a costly invasive species." Canadian Journal of Fisheries and
Aquatic Sciences 68(7): 1157-1160.
Wagner, C. M., M. B. Twohey and J. M. Fine (2009). "Conspecific cueing in the sea lamprey: do
reproductive migrations consistently follow the most intense larval odour?" Animal Behaviour
78(3): 593-599.
Waldman, J., C. Grunwald and I. Wirgin (2008). Sea lamprey Petromyzon marinus: an exception
to the rule of homing in anadromous fishes.

322