Assessing distribution and movement of blue catfish in Kansas reservoirs (DRAFT)

              U.S. Fish and Wildlife Service
Assessing Distribution and
Movement of Blue Catfish in
Kansas Reservoirs
Martha Mather1
Kayla Gerber2
and Zachary Peterson2
1 U.S. Geological Survey, Kansas Cooperative Fish and Wildlife
Research Unit, Kansas State University, Manhattan, KS
2 Kansas Cooperative Fish and Wildlife Research Unit, Kansas State
University, Manhattan, KS
Cooperator Science Series # 117-2015
COOPERATOR SCIENCE SERIES
ii
About the Cooperator Science Series:
The Cooperator Science Series was initiated in 2013. Its purpose is to facilitate the archiving and
retrieval of research project reports resulting primarily from investigations supported by the U.S. Fish
and Wildlife Service (FWS), particularly the Wildlife and Sport Fish Restoration Program. The online
format was selected to provide immediate access to science reports for FWS, state and tribal
management agencies, the conservation community, and the public at large.
All reports in this series have been subjected to a peer review process consistent with the agencies
and entities conducting the research. Authors and/or agencies/institutions providing these reports are
solely responsible for their content. The FWS does not provide editorial or technical review of these
reports. Comments and other correspondence on reports in this series should be directed to the
report authors or agencies/institutions. In most cases, reports published in this series are preliminary
to publication, in the current or revised format, in peer reviewed scientific literature. Results and
interpretation of data contained within reports may be revised following further peer review or
availability of additional data and/or analyses prior to publication in the scientific literature.
The Cooperator Science Series is supported and maintained by the FWS, National Conservation
Training Center at Shepherdstown, WV. The series is sequentially numbered with the publication year
appended for reference and started with Report No. 101-2013. Various other numbering systems
have been used by the FWS for similar, but now discontinued report series. Starting with No. 101 for
the current series is intended to avoid any confusion with earlier report numbers.
The use of contracted research agencies and institutions, trade, product, industry or firm names or
products or software or models, whether commercially available or not, is for informative purposes
only and does not constitute an endorsement by the U.S. Government.
Contractual References:
This document fulfills reporting requirements for Federal Aid in Sport Fish Restoration Grant F-58-R-
19. Previously published documents that partially fulfilled any portion of this contract are referenced
within, when applicable. (USGS IPDS #: IP-065549)
Recommended citation:
Mather, M., K. Gerber and Z. Peterson. 2015. Assessing Distribution and Movement of Blue Catfish in
Kansas Reservoirs. U.S. Department of Interior, Fish and Wildlife Service, Cooperator Science Series
FWS/CSS-117-2015, Washington, D.C.
For additional copies or information, contact:
Martha Mather
U.S. Geological Survey
Kansas Cooperative Fish and Wildlife Research Unit
205 Leasure Hall, Division of Biology
Kansas State University
Manhattan, KS 66506-3501
Phone: (785) 532-6070
Email: mmather@ksu.edu
1
ASSESSING DISTRIBUTION AND MOVEMENT OF BLUE CATFISH IN KANSAS
RESERVOIRS
Martha Mather1, Kayla Gerber2, Zachary Peterson2
1U. S. Geological Survey, Kansas Cooperative Fish and Wildlife Research Unit, Division of
Biology, Kansas State University, Manhattan, KS
2Kansas Cooperative Fish and Wildlife Research Unit, Division of Biology, Kansas State
University, Manhattan, KS
This draft manuscript is distributed solely for purposes of scientific peer review. Its content is
deliberative and predecisional, so it must not be disclosed or released by reviewers. Because the
manuscript has not yet been approved for publication by the U.S. Geological Survey (USGS), it
does not represent any official USGS finding or policy
OVERVIEW
This report is organized into three chapters that address six objectives. The first chapter
addresses objectives 1-3. The second chapter addresses objectives 4-5. The third chapter
addresses objective 6. The objectives for the project are listed below for reference.
OBJECTIVES
1. Develop and test tagging protocols for blue catfish.
2. Develop and test protocols for setting up and calibrating stationary receivers.
3. Summarize tagging and tracking protocols for use in other systems with other species.
4. Determine where tagged blue catfish spend their time within Milford reservoir.
5. Determine when, size distribution, and how many blue catfish exit Milford reservoir.
6. Quantify potential drivers of distribution
Chapter 1 - Methodology – Objectives 1-3
2
1 OBJECTIVES 1-3
2
3 DEVELOPMENT / EVALUATION OF METHODOLOGIES FOR EFFECTIVE
4 ACOUSTIC TAGGING AND STATIONARY RECEIVER ARRAY SET-UP
5
6 INTRODUCTION
7 Benefits of Tagging Fish for Research and Management. Knowing fish location is useful
8 for many questions related to research and management (Hubert 1999; Millspaugh and Marzluff
9 2001). The variable distribution patterns that result from movement are the foundation for
10 effective fisheries, ecology, and conservation (Alldredge at al. 2011). In recent years, the number
11 of tagging studies has increased dramatically (Chapter 1 Figure 1). With the development of
12 smaller and lighter transmitters and other technological advances (Knaepkens et al. 2005;
13 Metcalfe 2006; Hitt and Angermeier 2008; Albanese et al. 2009), biotelemetry has become one
14 of the most popular methods to study fish in their natural environment (Bridger and Booth 2003).
15 Lack of Detections. Changes in timing and location of detections are the essential pieces
16 of information that radio or acoustically tagged fish provide. Thus, lack of detections is a
17 problem for telemetry studies. Lack of detections can occur when a tagged fish: (a) naturally
18 leaves the detection system temporarily or permanently; (b) dies from natural causes; (c) dies
19 from tagging or handling associated with tagging; or (d) loses its tag via egestion (mouth, anus)
20 or ejection (incision site). Lack of detections from each of these sources has different
21 implications for data interpretation. Identifying why tagged fish are undetected in the field is
22 difficult. However, a good tagging methodology and sound research design for detection of
Chapter 1 - Methodology – Objectives 1-3
3
tagged fish can reduce some of the uncertainty related to tagging 23 mortality and tag loss (c-d
24 above).
25 Methodological Challenges for Tagging. Surgically implanting acoustic tags within the
26 coelomic cavity of a fish is generally regarded as the most appropriate method for long-term
27 biotelemetry applications (Jepsen et al. 2002; Bridger and Booth 2003; Brown et al. 2011; Cooke
28 et al. 2011; Thiem et al. 2011). However, the surgical implantation of acoustic tags has the
29 potential to cause infection, alter behavior, and ultimately lead to mortality (Bridger and Booth
30 2003). To ensure that the data generated from tagged fish are relevant to untagged conspecifics,
31 fish tracking research can benefit from methodological synthesis and refinement (Cooke et al.
32 201). Thus, sound tagging methodology is important for all tracking studies. Here we evaluate a
33 tagging methodology for Blue Catfish (Ictalurus furcatus) and Channel Catfish (Ictalurus
34 punctatus).
35 Tag Loss. Tag loss (c-d above) is a problem for all fish and especially for catfish. Several
36 studies have tracked Blue Catfish (e.g., Fischer et al. 1999; Grist 2002; Lee 2009; Garrett 2010;
37 Garrett and Rabeni 2011) in the field. However, only a limited number of studies have developed
38 or evaluated tagging methodologies for Blue Catfish (e.g., Holbrook et al. 2012; Bodine et al.
39 2014) and Channel Catfish (e.g., Summerfelt and Mosier 1984, Marty and Summerfelt 1986,
40 1990).
41 In this literature, tag retention (% tags retained) in evaluations of recreationally-important
42 catfish species (Blue Catfish and Channel Catfish) is variable but usually low [Blue Catfish: 33,
43 60% (Holbrook et al 2012); 100, 42% (Bodine et al. 2014); Channel Catfish 29% (Summerfelt
44 and Mosier 1984); 44, 2% (Marty and Summerfelt 1986, 1990)]. Through controlled hatchery
45 and laboratory studies in which tags were found outside of previously-tagged catfish, we know
Chapter 1 - Methodology – Objectives 1-3
4
some catfish tag loss occurs via ejection (i.e. loss through incision s 46 ite; Summerfelt and Mosier
47 1984; Marty and Summerfelt 1986). Even though new methods are being developed and
48 evaluated (Bodine et al. 2013), a high-survival, high-retention methodology for tagging catfish
49 has still not been identified.
50 Goals. Here, we (a) refine a methodology that minimizes stress and maximizes retention
51 of acoustic tags for catfish, (b) evaluate this methodology four times for two catfish species over
52 three years in two settings (hatchery and field), and (c) and describe the receiver array and range
53 test we used for field evaluation of Blue Catfish tags.
54
55 METHODS
56 Study System. Milford Reservoir (39°08'42"N, 96°56'54"W) is an impoundment of the
57 Republican River (Dickinson, Clay, and Geary counties, KS) and is part of the Lower
58 Republican watershed, KS. Milford reservoir has a surface area of 6,555 ha, 262 km of shoreline
59 dominated by limestone cobble and boulders, an average depth of 6.7 m, and a maximum depth
60 of 19.8 m (Reinke 2001).
61 Tagging Overview and Summary. We tagged Blue Catfish (BC) and Channel Catfish
62 (CC) four times over three years (2012-2014) in two settings (Milford Hatchery and Milford
63 Reservoir) (Chapter 1 Table 1). These trials served three purposes: to practice tagging techniques
64 (2012, BC, Milford Hatchery); to evaluate field distribution (2012, 2013, BC, Milford
65 Reservoir); and to test three variables in the hatchery that might affect tag retention (2014, CC,
66 Milford Hatchery). We used the same tagging methodology for all evaluations.
67 2012 – Blue Catfish, Milford Hatchery, Technique Practice and Evaluation. After
68 reviewing the literature, developing a surgical protocol, and practicing incision and suturing
Chapter 1 - Methodology – Objectives 1-3
5
techniques in the laboratory, we tested our tagging protocol on live catfish 69 [estimated range: 150-
70 250 mm Total length (TL)] at Milford hatchery (Chapter 1 Table 1). Each individual tagger
71 sequentially tagged five fish, following the procedures in our written protocol. Tagged fish were
72 held in a hatchery tank for seven days. Then tag placement was evaluated through euthanasia and
73 dissection. This qualitative evaluation was an opportunity to standardize and improve our
74 tagging technique.
75 2012, 2013 - Blue Catfish, Milford Reservoir, Field Evaluation of Distribution. In both
76 2012 and 2013, for our test of distributional patterns of Blue Catfish in Milford Reservoir, we
77 targeted the size range of fish that was common in the reservoir (400-600 mm TL; additional
78 details are provided in Chapter 2). In 2013, we added a limited number of smaller and larger fish
79 to the study (Chapter 1 Table 2). In 2012, the average fish size tagged was 487 mm TL [range
80 383-1020, Standard Error (SE) 14.5, n=48]. In 2013, the average size of Blue Catfish tagged was
81 517 mm TL (range 343-1090, SE 17.8, n=75). In 2012, for field tagging, we used V9 tags
82 (length: 29-47 mm, weight in air: 4.7-6.4 g, weight in water: 2.9-3.5 g). In 2013, we also tagged
83 fish with V13 tags (length: 36-48 mm, weight in air: 11-13 g, weight in water: 6-6.5 g). We
84 evaluated survival of tagged Blue Catfish and retention of tags in two ways (Chapter 1 Table 1).
85 First, we plotted detections for the first 10 days when post-tagging mortality and loss to acute
86 stress was most likely to occur. For this plot, we first checked that fish moved across multiple
87 receivers to make sure they were not dead. Second, we plotted the number of fish detected per
88 month (%) across the first five months of the study for both years. We predicted that fish that
89 were repeatedly detected at different locations survived the tagging process and retained their
90 tags. No statistics were used for this evaluation.
Chapter 1 - Methodology – Objectives 1-3
6
2014 - Channel Catfish, Milford Hatchery, Evaluation. 91 In 2014, we tested how three
92 factors (incision location, antibiotics, and surgery time) affected tag loss for 70, age-0, hatchery-93
reared channel catfish (Chapter 1 Table 1). The tagging protocol was the same as for other
94 tagging evaluations except that we used smaller dummy tags to keep tag weight < 2% fish body
95 weight (Bridger and Booth 2003).
96 In a review of tagging methodologies, Cooke (2011) noted that the importance of incision
97 location and antibiotics are rarely tested. First, we chose to test the incision location because we
98 used a lateral incision whereas most other tagging studies have used a ventral incision. We also
99 chose to test if antibiotics have an effect on tag loss and survival because many catfish tagging
100 studies do not use antibiotics. We chose to test surgery time because we suspect surgery time
101 varies across surgeons and studies, and longer surgery time may increase post tagging stress. Our
102 five treatments contained 14 fish each that were given different combinations of incision,
103 antibiotics, and surgery time. Treatment 1 was the treatment we describe below for our field
104 tagging [lateral incision, antibiotics, quick surgery time (2-3 min)]. Treatment 2 was similar to
105 treatment 1 but used a ventral incision (ventral incision, antibiotics, quick surgery time).
106 Treatment 3 used a lateral incision, no antibiotics, and a quick surgery time. Treatment 4 used
107 alternative options to treatment 1 [ventral incision, no antibiotics, longer surgery time (about 8
108 min)]. Treatment 5 was a control in which tagging was simulated but no fish were tagged.
109 Before tagging, all dummy VEMCO tags were engraved with the tag number. Post-110
tagging, all fish were Floy tagged. We recorded treatment, VEMCO dummy tag number, and
111 Floy tag number so we could link tag loss to a treatment. We held all 70 fish in a single (4 m X 4
112 m) compartment of a hatchery raceway for 12 weeks. We recorded general individual fish
113 condition weekly, in addition to incision condition (suture present, redness at incision, redness at
Chapter 1 - Methodology – Objectives 1-3
7
suture insertions, and general condition and healing of the incision), F 114 loy tag number, and Floy
115 tag insertion condition. We also took pictures of all fish. Each week we searched the bottom of
116 the hatchery compartment visually and manually four times (two times each by two people) to
117 recapture ejected tags. At the end of 12 weeks, we euthanized all fish, measured and weighed
118 fish, recovered tags, and photographed tag position within the body cavity. To summarize data,
119 we plotted tag loss data by treatment. We used a Chi square test with 2,000 Monte Carlo
120 simulations to evaluate if tag loss was distributed equally across all treatments. Two thousand
121 simulations is a default value for a simulated P-value (chisq.test function; R Core Team 2013).
122 Tagging Methodology. We used an 8-step tagging procedure that included: 1-preparation
123 before field work; 2-preparation in the field to allow quick and minimal stress tagging; 3-
124 minimal stress fish collection and holding; 4-pre-surgery considerations; 5-quick, minimal stress
125 surgery; 6-prophylaxis after surgery; 7-recovery and release; and 8-evaluation (Chapter 1 Figure
126 2). The same procedures were used for field and hatchery tagging.
127 1. Pre-field preparations. To minimize stress, preparation before field work was
128 essential. Existing literature on tagging studies, tagging techniques in general, fish morphology
129 and fish physiology were reviewed and summarized. We also contacted authors who had
130 published on catfish tagging via email for additional insights. As with most research facilities,
131 we were required to submit an Institutional Animal Care and Use Committee (IACUC) protocol
132 (#3151 and #3151.1). Insights from a university veterinarian were very useful relative to
133 anesthetic and surgical techniques.
134 In addition to the literature and technical expert consultations, practicing incisions and
135 suturing was essential. Many useful print and online tutorials exist on surgical techniques.
136 However, practice was perhaps the most important component of our protocol. Incision and
Chapter 1 - Methodology – Objectives 1-3
8
suturing can be practiced on inanimate objects (oranges and 137 bananas) any time. Dead fish added
138 a new dimension to incision and suturing practice. A very important component of our technique,
139 however, was tagging live fish prior to field tagging. This tagging of hatchery fish was followed
140 by an evaluation of survival, healing, and tag placement in the hatchery for seven days. In
141 summary, a good literature review, thoughtful protocols, and extensive practice before field
142 tagging were important parts of our protocol.
143 2. Preparation in the Field. For field preparation of the surgical area, pre-sampling
144 organization was critical (Chapter 1 Figure 2). For our field sampling, we used jon boats as
145 mobile surgical stations that were beached adjacent to the collection area. This allowed us to
146 minimize the time fish were confined during transport before surgery. This setup also allowed us
147 to release fish near the location where they were captured. For tagging in the field, workspace
148 will be limited, so we pre-planned all steps for fish processing to make sure that a two-person
149 surgical team could easily transfer fish from the capture boat to anesthesia tank to the operating
150 arena to recovery tanks then to the lake for release. Often, this required thought about placement
151 of tanks and work stations. We chose to use two operating teams in two separate jon boats with a
152 shared salt bath recovery tank to process our fish quota more rapidly. We also ensured that all
153 holding and recovery tanks were large enough to accommodate the length of the fish body
154 (typically 60 cm diameter circular bucket; 64 liter capacity). We monitored temperature in each
155 bucket and compared it to ambient lake temperatures. When bucket temperature exceeded
156 reservoir temperature we changed the water. When sun was intense, patio umbrellas over the
157 holding and recovery tanks provided shade for the fish. This preparation and organization
158 allowed us to process fish quickly with minimal stress.
Chapter 1 - Methodology – Objectives 1-3
9
3. Minimal Stress Fish Capture and Pre-surgery Holding. 159 We collaborated with State
160 colleagues on tagging. State biologists captured fish using boat electrofishing (1 stationary boat,
161 2 capture boats) with low pulse DC current (15 pulses/s, 3-5 amps) (Bodine and Shoup 2010).
162 All fish were collected in pre-identified areas. Fish were held on State electrofishing boats post-163
sampling in large aerated live wells. We only tagged 5-10 fish at a time so that fish were held on
164 board our boat < 60 minutes post-capture. This step in our protocol allowed us to tag fish of
165 predetermined size from known locations that were captured with minimal stress and held in low
166 stress conditions for a relatively short time per surgery.
167 4. Pre-surgery, 5. Surgery, 6. Prophylaxis, 7. Recovery and Release. Individual fish were
168 anesthetized one at a time with Aqui-S 30 mg-L in a single fish tank until they lost orientation
169 (2012: Average: 2 min. 16 sec. SE = 12 sec; 2013: Average = 2 min. 30 sec. SE = 7 sec). Doses
170 of anesthetic were tested in hatchery trials before field tagging. Two people processed each fish.
171 One acted as the surgeon and never moved from the operating station. The other acted as the
172 anesthesiologist and moved the fish from pre-tagging tank to the anesthesia tank to operating
173 station to the recovery tank. The anesthesiologist also constantly applied ambient water (with
174 Aqui-S if needed) to the fish skin and gills during surgery and made sure the fish remained in the
175 optimal position for a quick and stress-free surgery.
176 After anesthesia, fish were weighed (hanging scale with a cradle of soft mesh) and
177 measured on a wet measuring board. A 15-30 mm lateral incision was made below the pectoral
178 fin about ¾ of the way to the tip of the fin (15-20 mm – 300-700 mm TL Blue Catfish; 20-30
179 mm– >700 mm TL Blue Catfish). We used surgical scalpels of size 12 for fish < 700 mm TL and
180 22 for fish > 700 mm TL). As catfish intestines are very close to a thin body wall, we were
181 careful to make the incision into fish body wall in increments so that only skin and muscle, not
Chapter 1 - Methodology – Objectives 1-3
10
intestines, were cut. A sterile tag was carefully inserted into the body 182 cavity. The incision was
183 closed with 2-4 sutures (Ethicon braided, coated Vicryl, 3-0, FS-1, 24 mm 3/8 c reverse cutting –
184 fish > 700 mm TL; Ethicon, braided, coated Vicryl, 3-0, FS-2, 19 mm 3/8 c, reverse cutting –
185 fish < 700 mm TL). Surgery time was relatively short (2012 Average = 2 min. 38 sec, SE = 7
186 sec; 2013 Average = 2 min. 54 sec, SE = 5 sec).
187 As a prophylaxis, after surgery we gave all fish an intramuscular injection of antibiotic
188 (Liquamycin - 0.1 mg/kg fish; Pautzke et al. 2010), then allowed the tagged fish to recover in an
189 individual tank with oxygenated, ambient water until the fish was upright and swimming
190 (Recovery times 2012: Average = 5 min. 7 sec, SE = 24 sec; 2013 Average = 7 min. 14 sec, SE =
191 13 sec). Next, tagged fish were transferred to a larger community recovery tank with a 0.05%
192 salt solution to aid in slime coat recovery. After at least 15 minutes in a salt bath (Long et al.
193 1977), fish were individually captured with a soft mesh trout net, placed in the lake close to
194 where they were captured, and allowed to swim away (Chapter 1 Figure 2). All times were
195 recorded.
196 Receiver Placement. In 2012 and 2013, we tracked tagged Blue Catfish with a benthic
197 20-stationary receiver array (discussed in Chapter 2) and a 57-site monthly manual receiver
198 survey (discussed in Chapter 3). For the stationary array, data were collected using VEMCO
199 (VR2W-69kHz) receivers which received coded pings from tags each time a tagged fish came
200 within range of the receiver. In 2012, we deployed receivers in June (Chapter 1 Table 3);
201 receivers were placed at 18 locations within the reservoir and two locations adjacent to the
202 reservoir exits (Chapter 1 Figure 3). The upper river receiver (receiver 1) and the upper within-203
reservoir receiver (receiver 2) formed a two-tier gate to detect upriver egress from the reservoir.
204 The southernmost receivers in the reservoir (receiver 19) and the river receiver below the dam
Chapter 1 - Methodology – Objectives 1-3
11
(receiver 20) formed another two tier gate to detect downriver 205 egress (Chapter 1 Figure 3). We
206 also had two 3-stationary receiver gate arrays (receivers 6-8, 11-13) across the mid-reservoir
207 constriction (i.e., the limited width allowed complete coverage of the entire reservoir as
208 confirmed by range tests) to detect any fish that moved through the middle region of the
209 reservoir. In 2012, for data analysis, we removed data from 2 of the 3 receivers in these gates (7,
210 8, 11, 13) to obtain a more even distribution of detections (Chapter 1 Figure 3A- dashed squares
211 indicate receivers that were removed). Thus, in 2012, of the 18 within reservoir receivers, 14
212 were used for data analysis. In 2013, we deployed receivers similarly (May-November 2013;
213 Chapter 1 Table 3). However, receiver 1 was vandalized in August, 2013. Receivers 16-17 were
214 lost due to vandalism or boating collisions. Gate receiver 13 replaced gate receiver 12 because
215 receiver 12 was lost. As in 2012, in 2013, we also removed data from 2 of the 3 gate receivers
216 (7,8, 11, 12) (Chapter 1 Figure 3B- dashed squares indicate receivers that were removed). Thus,
217 in 2013, of the 18 within reservoir receivers, 12 were used for data analysis. Receivers were
218 grouped into five regions based on general size and location (upper, upper middle, Madison,
219 lower middle, and lower; Chapter 1 Figure 4).
220 We also collected data on acoustically tagged Blue Catfish at 57 (0.8 km2) manual
221 tracking sites (Chapter 1 Figure 5). Tracking sites were positioned to cover the maximum
222 amount of surface area while preventing overlap among adjacent sites (i.e., < maximum range)
223 (e.g., limited spatial arrangements were possible to cover the entire reservoir with sampling units
224 of this size). We chose this design to quantify spatial heterogeneity. The choice of 57 spatially-225
explicit sampling locations that covered the entire reservoir provided good resolution for
226 quantifying Blue Catfish distribution, allowed us to construct detailed spatial maps of Blue
Chapter 1 - Methodology – Objectives 1-3
12
Catfish, and resulted in substantial statistical power. The manual 227 tracking survey was conducted
228 in June through November in 2013 (described in detail in Chapter 3).
229 Stationary Receiver Range Test. We conducted range tests using two methods. Both tests
230 provided information on the distance at which a tag can be detected under field conditions. First,
231 we conducted a range test using the methods provided by the receiver manufacturer, VEMCO.
232 For this, we deployed an array of receivers in an 800-m straight line, separated by 100-m
233 intervals. A test tag, vertically oriented, was located near the first receiver. Receivers at 100-800
234 m were constantly exposed to the repetitive pinging of this tag. Over a week, adequate data were
235 collected at each receiver to get a probability of detection at 100 m intervals. These range test
236 data were processed using VEMCO software.
237 We also conducted a second set of range tests at three receiver locations within Milford
238 Reservoir. We chose these three receivers because they were at sites with similar bathymetry
239 (e.g., water depth), so we could get an estimate of range variation associated with individual
240 sites. For this range test, we drove a boat in four cardinal directions (N,S,E,W) from a centrally-241
deployed receiver for up to 1,000 m (or until we encountered the shore). At 100-m intervals, we
242 submerged test tags in the water for a count of five detection pings, determined using the manual
243 tracker. From this design, we could determine distances that a tag was detected in four different
244 directions. Data for the second range test were processed using Excel.
245
246 RESULTS
247 2012 – Blue Catfish, Milford Hatchery, Technique Practice and Evaluation. In our initial
248 tagging during which we tested our protocols and evaluated our tagging techniques, all tagged
249 fish survived seven days, all tags remained within the body cavity, incisions healed well, and we
Chapter 1 - Methodology – Objectives 1-3
13
observed no differences among taggers. Based on this result, few 250 changes were made to our field
251 protocol.
252 2012, 2013 – Blue Catfish, Milford Hatchery, Technique Practice and Evaluation. For
253 our field tagging of Blue Catfish, tagged fish suffered little short-term tag loss. In 2012, all 48
254 tagged fish were detected at least once in the first ten days (black squares per row=detection per
255 fish; Chapter 1 Figure 6). A fish was not scored as detected for this tag evaluation unless it
256 moved between at least two receivers. This ensured that we did not score dead fish as live fish
257 that had retained their tags. Seventy three percent of tagged fish were detected for five or more
258 days during the first ten days (Chapter 1 Figure 6). Apart from methodological considerations,
259 tagged fish had different patterns of distribution as some fish were detected more often than
260 others (variation in black squares per row = variation in detections per fish; Chapter 1 Figure 6).
261 For example, fish 12 was detected across five days (days 1, 5, 6, 9, 10) whereas fish 47-48 were
262 detected daily (Chapter 1 Figure 6). In 2013, all 75 tagged fish were detected at least once in the
263 first ten days (Chapter 1 Figure 7). Ninety six percent of all fish tagged in 2013 were detected
264 for five or more days within the first ten days post-tagging (Chapter 1 Figure 7).
265 In 2012, 95% of the fish were detected in early July and August (Chapter 1 Figure 8).
266 About 90% were detected in September and October. In November, 85% of the tagged Blue
267 Catfish continued to be detected (Chapter 1 Figure 8). In 2013, about 90% of the fish we tagged
268 were detected in July (Chapter 1 Figure 9). We continued to detect over 85% of the tagged fish
269 from August through October, 2013 (Chapter 1 Figure 8).
270 2014 - Channel Catfish, Milford Hatchery Tagging Experiment. Age-0 channel catfish
271 from Milford Hatchery suffered little tag loss or mortality in any treatment during our 12-week
272 study. No mortality occurred in treatment 1 (our methodology), treatment 3 (no antibiotics), and
Chapter 1 - Methodology – Objectives 1-3
14
the control (Treatment 5) (data not shown). Fish in treatment 273 2 (ventral incision) had an overall
274 mortality of 21% while those in treatment 4 [[ventral incision, no antibiotics, longer surgery time
275 (about 8 min)]. had an overall mortality of 7%. Differences in mortality were not statistically
276 significant, possibly because mortality was low for all fish in all treatments.
277 All tag loss occurred within the first week (Chapter 1 Figure 9) with the exception of one
278 fish in treatment 3. Treatment 1, the treatment we used for field tagging, had no tag loss (Chapter
279 1 Figure 10). Treatments 2 and 3 had an overall tag loss of 21% (3 individuals in each treatment
280 lost tags). Treatment 4 had an overall tag loss of 29% (4 individuals lost their tags; Chapter 1
281 Figure 10). Our tagging methodology (treatment 1) had a significantly lower tag loss than
282 treatment 4, based on a chi square test (Chapter 1 Figure 10). Other differences described above
283 were not statistically significant, (P> 0.05), possibly because tag loss was low for all fish in all
284 treatments.
285 Range Test Results. Both V9 and V13 tags were detected over 80% of the time at
286 distances from 0-300 m (Chapter 1 Figure 11). Percent detections decreased to about 75%
287 between 300-500 m. Detections declined to 70% at 600 m from the tag (Chapter 1 Figure 11).
288 VEMCO recommends selecting a receiver range that corresponds to at least 70% of the
289 detections. In our range test, the 70% detection range corresponded to a radius of 600 m
290 (Chapter 1 Figure 11).
291 For our second range test, individual detection radii varied from 300-650 m (average 462
292 m) for receiver 4. Individual detection radii varied from 500-1,000 m (average 775 m) for
293 receiver 7 (Chapter 1 Figure 12A). Individual detection radii varied from 700-900 m (average
294 825 m) for receiver 12 (Chapter 1 Figure 12B). Overall, the average range radius in the second
295 range test (average 687 m) was similar to the range found in the VEMCO recommended range
Chapter 1 - Methodology – Objectives 1-3
15
test (average 600 m) (Chapter 1 Figure 12C(Chapter 296 1 Figure 12A).). Based on these combined
297 tests, we used a receiver range of 600 m.
298
299
300 DISCUSSION
301 High Tag Retention. A primary goal of this research was to develop a high-survival, high-302
retention tagging methodology for catfish. High retention of tags increases the quality and cost
303 effectiveness of a tagging dataset. Conversely, a large proportion of undetected fish raises
304 questions about fish stress during tagging and whether tagged fish behave like untagged fish (an
305 assumption of tagging). For these reasons, we made high tag retention and a high detection rate
306 priorities. In our hatchery trial of Channel Catfish tagging, our methodology (Treatment 1)
307 resulted in no mortality and no tag loss. In one of the early studies that internally implanted tags
308 into Channel Catfish, Marty and Summerfelt (1986) found that 22 of 39 (44% retention) and 45
309 of 46 (2% retention) fish expelled their tags in 19 and 20 days respectively after being tagged
310 with traditional (non-anchored) implantation methods. In response to this tag ejection, complex
311 internal anchoring procedures were developed (e.g., Siegwarth and Pitlo 1999) that had better,
312 but still low, tag retention rates. However, this anchored implantation technique can be
313 physiologically stressful to tagged fish. For example, in preparation for using ultrasonic
314 telemetry on Blue Catfish in Lake Texoma, OK, Lee (2009) used both traditional and anchored
315 attachment methods (n= 5 fish per attachment method). After 120 days in the hatchery pond, all
316 fish retained their tags but 90% died from both methods. Seven of 10 fish died within 48 h of
317 surgeries (Lee 2009). Recently, transmitter retention for adult Blue Catfish (> 600 mm TL) was
318 again evaluated for traditional and anchored implantation methods (n=15 per attachment
Chapter 1 - Methodology – Objectives 1-3
16
methods). Ten and six fish respectively expelled their tags 23-319 243 days post-surgery, resulting in
320 retention rates of 33 and 60%, respectively, for traditional and anchored tag attachment methods
321 (Holbrook et al. 2012). In a recent test of a new technique that externally attaches tags to skeletal
322 structure, Bodine et al. (2014) had mixed retention rates. In two hatchery trials, tagged Blue
323 Catfish had 100% (n=20; TL range = 435-638mm) then 41.7% retention (n=24, TL range = 600-
324 995) after two months. Thus, our tag retention rate exceeds that of most existing Blue Catfish tag
325 evaluations.
326 High Detection. Our tagging methodology was also very successful in detecting fish in
327 the reservoir, in that we repeatedly detected 85% of our tagged Blue Catfish in Milford Reservoir
328 through five months across two years (n= 48, 75). Other Blue Catfish tagging studies have not
329 detected such a high proportion of tagged fish. In Lake Norman, NC, only 15 of 29 (52%) Blue
330 Catfish (500-900 mm TL) with externally attached radio tags were alive and retained their tags
331 throughout the study (Grist 2002). In Lake Texoma, only 22 of 50 (44%) tagged Blue Catfish
332 (639-1305 mm TL) were successfully tracked. Eight tagged fish were confirmed dead and 20
333 were not detected (Lee 2009). In the lower Missouri River, Garrett (2010) implanted radio tags
334 into 40 Blue Catfish in each of two years (mean=872, range =569-1260 mm TL). Annual
335 movement cycle data were based on only 12 fish in each year (30% detection of tagged fish
336 throughout the study) because of the large number of tagged fish that were missing. Finally, for a
337 field evaluation of 50 Blue Catfish (TL range = 600-995mm) in Lake Buchanan, Texas, Bodine
338 et al. (2014) redetected only 40% of all tagged fish at 6 months and 19% at 12 months.
339 Consequently, our methodology provides a more detailed dataset than has been previously
340 collected and suggests that our tagged fish were not stressed post tagging. Both of these results
Chapter 1 - Methodology – Objectives 1-3
17
increase confidence that our dataset will provide generalizable 341 insights about Blue Catfish
342 distribution.
343 Critical Attributes of Our Methodology. We attribute our success in tag retention to
344 several factors. Our protocol emphasized preparation, practice, and organization before the
345 tagging event, which allowed us to process fish quickly with minimal stress. A lateral incision
346 reduced our tag loss in the hatchery and was probably an important factor in successful field
347 tagging. Cooke et al. (2011) reviewed trends in intracoelomic tagging effects studies and found
348 that six of 108 studies compared elements of the incision, but only one study tested a ventral vs.
349 lateral incision. Although a ventral incision may be less likely to puncture the ovaries and may
350 be easier for the surgeon (Schramm and Black 1984), gravity may encourage tag loss in the
351 initial weeks before a ventral incision heals. Although the effect of antibiotics was unclear in our
352 hatchery evaluation, we suspect that antibiotics aided the survival and healing of our field caught
353 fish. In a review of tagging studies, only one study of 108 evaluated the effectiveness of
354 antibiotics. Specifically, Isely et al. (2002) found that the use of antibiotics was effective in
355 preventing initial post-surgery infection.
356 Receiver Array Effectiveness. Our receiver array detected fish throughout the lake.
357 Detection ranges of receiver arrays are important for understanding whether the data collected
358 represent an accurate estimate of a fish’s space use (Welsh et al. 2012; Klimley et al. 1998).
359 Detection ranges are often just assumed based on manufacturer specifications (Welsh et al. 2012;
360 Kessel et al. 2014); when tested by researchers they can deviate within different aquatic habitats
361 (Heupel et al. 2006) and across temporal, and spatial scales (Simpfendorfer et al. 2008; Payne et
362 al. 2010). Our two range evaluation methods provided similar range estimates which enhanced
363 our confidence in the range at which our tags could be detected. Data from the manual receiver
Chapter 1 - Methodology – Objectives 1-3
18
reinforced the results of the stationary receivers. Both regimes (364 stationary and manual) were
365 designed to detect lake-wide patterns. Our detection regimes covered the whole extent of Milford
366 Reservoir from the causeway in the upper reservoir to the dam. Neither of these regimes,
367 however, detected small-scale movements because of the large detection diameter of receivers
368 (1,200 m diameter) and the wide spacing between receivers.
369 The impetus for our field study was to understand broad-scale distributional patterns
370 throughout an entire reservoir. Receiver sites were designed to identify lake-wide aggregations,
371 not heterogeneity or frequent distribution changes within localized areas. When our field study
372 was initiated, little information existed about Blue Catfish distribution in Milford Reservoir.
373 Hence, an extensive sampling design with many samples across the reservoir was required.
374 Given the state of our knowledge when we initiated this study, we simply would not have known
375 where to place receivers to detect Blue Catfish. Conducting an extensive and intensive design
376 simultaneously is logistically unfeasible. Thus, the design we describe here (broad spatial scale,
377 low resolution) was well suited for our question and likely would be useful for initial studies in
378 other systems. Information goal, system morphometry, scientific question, and target species
379 behavior also need to be considered in tracking study designs.
380 Management Implications. We have provided information on how we tagged fish and set
381 up receiver arrays. Our intention was to provide guidance for future studies in other systems.
382 First, our tagging was quite successful because of the organization, preparation, and training we
383 invested. Because of the monetary and labor investment in a tagging program, we suggest this
384 level of preparation. The tagging protocol we describe should be directly applicable to other fish
385 species including but not limited to catfish. Second, because of across-fish variability, future
386 studies should seek to tag a large sample size with the high retention rate we have demonstrated
Chapter 1 - Methodology – Objectives 1-3
19
here. A large sample size is essential for generalizable 387 statistical analysis. Although the
388 anecdotal observations about the behavior of a few individuals are interesting, the scientific
389 generality of such isolated observations is low. Third, the choice of fish sizes should be made
390 carefully. Elsewhere (Chapter 2), we illustrated that distribution of same size fish varied widely.
391 Hence a lack of replication of similar--sized fish may result in the erroneous conclusion that
392 differences in distribution are related to size when in fact individual variation is responsible.
393 Fourth, to utilize the insights that we provide here in other systems, researchers and
394 managers should identify the question for which tagging is being used. As we note above, for a
395 reservoir-wide survey, the array setup we used (broad spatial coverage with relatively low
396 resolution at any specific location) worked well. We argue that this design is the best for the
397 initial study in any system when little knowledge exists about where fish are located. Likewise, if
398 egress is the goal, then gating all exists from the reservoir with multiple stationary receivers
399 would be advisable. Stationary receivers, especially in confined areas, are susceptible to human
400 (vandalism) and natural (high flow, high sedimentation) damage. Multiple receivers in sequence
401 can guard against study failure when receivers are lost and can also detect direction of
402 movement. If stationary receivers are used, downloading data regularly is essential. Receiver loss
403 is common in array studies. Once the receiver is gone, any unloaded data are also lost. Fifth, a
404 thoughtful evaluation of fish behavior relative to system bathymetry is suggested to apply the
405 insights provided here to other species and systems. Many fish travel along a channel (Pautzke et
406 al 2010; Kennedy et al 2014) so setting up receivers along this travel lane might be useful in
407 other initial tracking efforts. Confluences are also good locations for initial receiver placement. If
408 there is a central narrow constriction, setting up a series of gates that detect changes through the
409 entire system is useful. Our across-reservoir gates were essential for bounding patterns of
Chapter 1 - Methodology – Objectives 1-3
20
distribution for Blue Catfish in Milford Reservoir. Finally, the information 410 gained from tracking
411 studies will accelerate as more fish are tracked within a specific system. In any initial study, little
412 is known about where the fish are located or the study would not be needed. Recognizing that
413 every question cannot be answered in a single study will facilitate realistic expectations about the
414 steps needed for effective research or management planning relative to this issue.
Size (mm TL)
•Range
•Average
•SE
2012 CC 150-250* Hatchery V9 & V9TP 20 NA Euthanize / Dissect
400-600 Detections
487 •10 days
14.5 •5 months
300-1000+ Detections
517 •10 days
17.8 •5 months
Response
•Tag Loss
184-260 •Mortality
225 •Growth
2.3 Tested
•Incision
•Antibiotics
•Surgery Time
Reservoir V9 & V9TP 48 158
Year Species Location Tag Type No. Fish Average
Surgery
Time (s)
Chapter 1 Table 1. Summary of evaluation procedures used to develop and evaluate tagging protocols for catfish
including year, species, size (range, average, SE), location at Milford, KS, type of tag, number of fish used, surgery
time, and evaluation methods.
2014 CC Hatchery V6 70 114
2013 BC Reservoir V9, V13, &
V13TP 75 174
Evaluation
2012 BC
Fish Length (mm) Weight (kg) Release Location
1 430 0.66 School
2 480 0.88 School
3 430 0.56 School
4 480 0.82 School
5 430 0.72 School
6 500 1.05 School
7 489 0.97 School
8 434 0.64 School
9 512 1.26 School
10 384 0.41 School
11 411 0.73 School
12 452 0.77 School
13 490 1.12 School
14 510 1.09 School
15 420 0.66 Causeway
16 506 0.99 School
17 490 1.15 School
18 751 4.4 School
19 392 0.51 Causeway
20 383 0.43 Causeway
21 518 1.27 Causeway
22 484 1.1 Causeway
23 615 2.5 Madison
24 419 0.58 Causeway
25 516 1.08 Causeway
26 451 0.81 Causeway
27 471 1.01 Causeway
28 408 0.52 Causeway
29 419 0.63 Causeway
30 407 0.68 Madison
31 485 0.96 Madison
32 401 0.54 Madison
33 515 1.2 Madison
34 466 0.81 Madison
35 542 1.33 Madison
36 1020 9.52 Madison
37 487 0.88 Madison
38 489 2.01 Madison
39 439 0.67 Causeway
40 487 1 Causeway
41 531 1.41 Causeway
2012
Chapter 1 Table 2. Number, length (mm TL), weight (kg wet weight)
and release location for Blue Catfish tagged in 2012, 2013 in Milford
Reservoir, KS.
Fish Length (mm) Weight (kg) Tagging Location
42 436 0.68 Causeway
43 573 1.8 Causeway
44 504 1 Madison
45 480 1.21 Madison
46 421 0.6 Madison
47 532 1.33 Madison
48 469 1.01 Madison
1 370 0.44 Madison
2 377 0.64 Madison
3 372 0.36 School
4 392 0.57 Madison
5 396 0.47 Madison
6 361 0.35 Madison
7 369 0.35 Causeway
8 343 0.22 Causeway
9 393 0.41 School
10 375 0.43 School
11 369 0.33 Causeway
12 515 1.13 Madison
13 506 1.12 Madison
14 550 1.71 Madison
15 531 1.2 Madison
16 445 0.77 Madison
17 511 1.02 Madison
18 1030 17.9 School
19 451 0.74 School
20 591 1.91 School
21 403 0.53 School
22 505 1.04 Madison
23 470 0.98 Madison
24 425 0.94 Madison
25 820 6.59 Madison
26 413 0.6 Madison
27 440 0.74 Madison
28 405 0.54 Madison
29 472 0.85 Madison
30 446 0.66 Madison
31 443 0.68 Madison
32 438 0.68 School
33 449 0.77 School
34 519 1.44 Causeway
35 513 1.09 School
2013
Chapter 1 Table 2. Continued.
Fish Length (mm) Weight (kg) Tagging Location
36 455 0.71 School
37 430 0.56 School
38 490 1.2 School
39 415 0.51 School
40 530 1.35 School
41 450 0.87 School
42 735 4.77 School
43 765 5.9 Causeway
44 514 1.3 Causeway
45 845 8.6 Causeway
46 526 1.36 Causeway
47 705 4.54 Causeway
48 421 0.61 Causeway
49 421 0.63 Causeway
50 460 0.72 Causeway
51 440 0.82 Causeway
52 513 1.26 Causeway
53 423 0.67 Causeway
54 508 1.14 Causeway
55 521 1.22 Causeway
56 1090 20.4 Causeway
57 429 0.72 Causeway
58 900 9.54 Causeway
59 400 0.53 Causeway
60 513 1.27 Causeway
61 1000 15.4 Causeway
62 510 1.56 Madison
63 555 1.86 Madison
64 505 1.36 Madison
65 540 1.08 School
66 530 1.15 School
67 489 1.12 Madison
68 495 0.96 Madison
69 467 0.71 School
70 466 0.79 School
71 625 2.47 Causeway
72 730 5.68 Causeway
73 537 1.43 Causeway
74 510 1.13 School
75 528 1.26 Causeway
Chapter 1 Table 2. Continued.
Chapter 1 Table 3. Dates of stationary acoustic receiver deployment and
removal in Milford Reservoir, Kansas in 2012 and 2013 by receiver
number.
Receiver 2012
Deployment 2012 Removal 2013
Deployment 2
1 6-20-12 Dec. 2012 5-16-13
2 6-20-12 NA 5-16-13
3 6-20-12 Mar. 2013 5-16-13
4 6-20-12 July 2013 5-16-13
5 6-20-12 Mar. 2013 5-16-13
6 6-20-12 Mar. 2013 5-16-13
7 6-20-12 Mar. 2013 5-16-13
8 6-20-12 Mar. 2013 5-16-13
9 6-20-12 Mar. 2013 5-16-13
10 6-20-12 Mar. 2013 5-16-13
11 6-20-12 Jan. 2013 5-16-13
12 6-20-12 Mar. 2013 5-16-13
13 6-20-12 NA 5-16-13
14 6-20-12 Jan. 2013 5-16-13
15 6-20-12 Jan. 2013 5-16-13
16 6-20-12 Jan. 2013 5-16-13
17 6-20-12 Jan. 2013 5-16-13
18 6-20-12 Jan. 2013 5-16-13
19 6-20-12 Jan. 2013 5-16-13
20 6-20-12 Dec. 2012 5-16-13
Chapter 1 – Methodology - Figure Captions
21
DEVELOPMENT / EVALUATION OF 1 METHODOLOGIES FOR EFFECTIVE
2 ACOUSTIC TAGGING AND STATIONARY RECEIVER ARRAY SET-UP
3
4 FIGURE CAPTIONS
5 Chapter 1 Figure 1. Results of a Web of Science literature search on the key words “acoustic
6 tag” or “radio tag” and “fish” is shown. The results are sorted by calendar year.
7
8 Chapter 1 Figure 2. Shown is a flowchart that described the eight steps in our tagging protocol.
9 Each step is described in greater detail in the text.
10
11 Chapter 1 Figure 3. Distribution of 20 stationary acoustic receivers within Milford Reservoir is
12 shown for (A) 2012 and (A) 2013. Receiver 1 was deployed in the Republican River above the
13 inflow to the reservoir in order to detect egress out of the reservoir. Receiver 20 was deployed in
14 the Republican River below the dam in order to detect egress out of the reservoir. Receivers 2
15 and 19 were located within the reservoir and act as a second tier of egress gates. Receivers 6-8
16 and 11-13 formed two complete gates across the middle reservoir constriction to detect
17 distribution changes. (A) Receivers 7, 8, 11, 13 were removed for data analysis in 2012 to
18 provide a more even array distribution (red dashed boxes indicate the location of the receivers
19 that were removed). (B) Receivers 7, 8, 11, 12 were removed for data analysis in 2013 for the
20 same reason (red dashed boxes indicate the location of the receivers that were removed).
21 Vandalism and boater conflicts resulted in the loss of receivers 1, 16, and 17 in 2013. As a result,
22 in 2012 and 2013, we used 14 and 12 receivers for data analysis respectively.
23
Chapter 1 – Methodology - Figure Captions
22
Chapter 1 Figure 4. In order to more clearly explain reservoir wide 24 distribution patterns, Milford
25 Reservoir was divided into five regions. The main reservoir regions (upper, upper middle, lower
26 middle, lower) are approximately the same size. Madison Creek is a distinct region.
27
28 Chapter 1 Figure 5. Sample sites for manual tracking survey at 57 sites to quantify Blue Catfish
29 distribution in Milford Reservoir, KS. Sites were sampled once a month July through November,
30 2013. Details of the survey methodology are provided in the text.
31
32 Chapter 1 Figure 6. For 2012, shown are daily detections used to evaluate Blue Catfish response
33 to tagging. On the X axis are first ten days. On the Y axis are fish number. A filled square
34 indicates that a fish was detected by at least one stationary receiver in Milford Reservoir.
35
36 Chapter 1 Figure 7. For 2013, shown are daily detections used to evaluate Blue Catfish response
37 to tagging. On the X axis are first ten days. On the Y axis are fish number. A filled square
38 indicates that a fish was detected by at least one stationary receiver in Milford Reservoir. KS.
39
40 Chapter 1 Figure 8. For 2012 and 2013, shown are monthly detections of Blue Catfish in
41 Milford Reservoir, KS. The X axis is month and the Y axis is percent of tagged fish. Numbers of
42 fish tagged are also indicated.
43
44 Chapter 1 Figure 9. Tag retention by hatchery Channel Catfish through time is shown for five
45 treatments. (A) The X axis is week and the Y axis is number of fish that retained their tags (i.e.,
Chapter 1 – Methodology - Figure Captions
23
no tag loss). (B) The details of the treatments 1-5 are also s 46 hown related to incision location,
47 antibiotic use, and surgery time.
48
49 Chapter 1 Figure 10. Tag retention by hatchery Channel Catfish is shown. The X axis is
50 treatment and the Y axis is number of fish that retained their tags (i.e., no tag loss). Our five
51 treatments contained 14 fish each that were given different combinations of incision, antibiotics,
52 and surgery time. Treatment 1 was the treatment we describe below for our field tagging [lateral
53 incision, antibiotics, quick surgery time (2-3 min)]. Treatment 2 was similar to treatment 1 but
54 used a ventral incision (ventral incision, antibiotics, quick surgery time). Treatment 3 used a
55 lateral incision but no antibiotics (lateral incision, no antibiotics, and quick surgery time).
56 Treatment 4 used alternative options to treatment 1 [ventral incision, no antibiotics, longer
57 surgery time (about 8 min)]. Treatment 5 was a control in which tagging was simulated but no
58 fish were tagged.
59
60 Chapter 1 Figure 11. Distance at which VEMCO V9 and V13 tags were detected is shown.
61 Distance (m) is shown on the X axis and percent detections is shown on the Y axis. The arrow
62 indicates 70% detection, the range recommended by the tag manufacturer, VEMCO. The
63 VEMCO recommended range test is described in more detail in the text.
64
65 Chapter 1 Figure 12. Distances at which VEMCO tags were heard at three receivers (A) receiver
66 4, (B) receiver 7, and (C) receiver 2. The specific spatial pattern and mean, minimum, and
67 maximum distances are shown for each receiver. This second range test is described in more
68 detail in the text.
Fig. 3
0
10
20
30
40
50
60
70
80
90
100
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
Number of Papers Year
Peer Review Peer-Reviewed LLiitteerraatuturer e– –FisFhis Tha gTgaigngging
Chapter 1 – Figure 1
Fig. 5
Literature
Set-Up at Tagging Location to Minimize Time and Stress
Anesthesia
Incision Location
Surgery Time
Minimal Stress Handling
(Time, Immersion, Temperature)
Antibiotics
Surgeon Anesthesiologist
Capture and Holding (Methods, Timing) to Minimize Stress
IACUC Surgical Training
1. Preparation Pre Field
2. Preparation In Field
3. Fish Collection
4. Pre-Surgery
6. Prophylaxis
Salt Bath Temperature
7. Recovery and Release
Adequate Recovery Self Release
5. Surgery
Soft Net Release where Captured
8. Evaluation
Oxygen
Short Term Longer Term
Chapter 1 – Figure 2
1
2
3
4
5
6
9
20
16
17
18
19
14
15
10
12
2012 2013
N
4 km
1
2
3
4
5
6
9
20
18
19
14
15
10
13
7 8
11 13
7 8
1112
Chapter 1 – Figure 3
Upper
Upper
Middle
Madison
Lower
Middle
Lower
Region Receivers
Upper 2 & 3
Upper Middle 4, 5, & 6
Madison 9 & 10
Lower Middle 12, 14, & 15
Lower 16, 17, 18, & 19
N
4 km
Chapter 1 – Figure 4
n=57
Manual Tracking Locations
Chapter 1 – Figure 5
Daily Detections for First 10 Days Post Tagging - 2012
1 Day 10
Fish
1
48
Chapter 1 – Figure 6
1 Day 10
Fish
1
75
Daily DeDtaeilyc Dteiotenctsio nfos fro rF Fiirrsstt 1 01 D0a yDs aPoysst TPagogsintg T-a20g1g3ing – 2013
Chapter 1 – Figure 7
Chapter 1 – Figure 8
Detections of Blue Catfish Across 5 Months
Chapter 1 – Figure 9
A.
B.
Fish with Tags
Week
Retention of Tags by Hatchery Channel Catfish
Chapter 1 – Figure 10
Fish with Tags
Retention of Tags by Hatchery Channel Catfish
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700 800 Detection %
Distance (m)
V13
V9
Receiver Range Test
Chapter 1 – Figure 11
Chapter 1 – Figure 12
0%
20%
40%
60%
80%
100%
100 m
200 m
300 m
400 m
700 m 500 m
800 m
900 m
1000 m
Range
(m)
Average 825
Minimum 700
Maximum 900
North
West
East
South
A. Receiver 4 B. Receiver 7 C. Receiver 2
Range (m)
Average: 462
Minimum: 300
Maximum: 650
Range (m)
Average: 775
Minimum: 500
Maximum: 1000
Range (m)
Average: 825
Minimum: 700
Maximum: 900
Chapter 2 – Distribution and Egress - Objective 4-5
24
CHAPTER 2 – DISTRIBUTION OF BLUE CATFISH 1 WITHIN AND EGRESS OF BLUE
2 CATFISH FROM MILFORD RESERVOIR (OBJECTIVES 4-5)
3
4 INTRODUCTION
5 Overview. Flexibility in distribution is essential to the life history and ecological niche of
6 many taxa and is an adaptive response that allows animals to take advantage of spatial variation
7 in the fluctuation of resources (Baker 1978, Gross et al. 1988). However, mobility adds
8 complexity to quantifying distribution. Although many fish species change distributions for
9 spawning, foraging, and overwintering, little is known about geographically-localized
10 distribution patterns or the extent of individual or group variation within and across geographic
11 areas (Cadrin and Secor 2009). Until recently, researchers and managers had limited
12 methodological options for quantifying distributions of mobile organisms. This lack of
13 information on how mobile fish are distributed and if they move into and out of a study system
14 has been an obstacle for both research and management. Blue Catfish, Ictalurus furcatus, is a
15 model organism for addressing the tradeoffs between residency and mobility that influence
16 distribution patterns because of an array of life history features. Here, we use a newer technology
17 (acoustic telemetry and stationary receivers) to identify distributional patterns of Blue Catfish, if
18 tagged fish leave the reservoir in which they were tagged, and factors that may affect
19 distributional patterns (e.g., season, time of day, fish size, and individual variation).
20 Importance of Knowing Distribution. Knowing distribution is important for research and
21 management. Animals are not distributed evenly throughout their environments but instead
22 display spatially and temporally heterogeneous patterns (Albanese et al. 2004; Planque et al.
23 2011; Scheiner and Willig 2011). Understanding variation in distribution (Kennedy and Gray
Chapter 2 – Distribution and Egress - Objective 4-5
25
1993; Jackson et al. 2001; Metcalfe 2006; Roberts and Angermeier 24 2007) is foundational for
25 research and management. For example, knowing fish distribution is important for stock
26 assessment and for the collection of biological samples (e. g. diets, scales, otoliths). Without
27 knowing where fish are located, effective sampling for survival, recruitment, growth, and other
28 research and management objectives will be ineffective. Anything less than a complete census
29 (i.e., sampling) gives a very limited view of where the fish are located. Consequently, most
30 existing distributional data on fish give a limited view of where fish spend their time.
31 Mobility Adds a Special Challenge to Quantifying Distribution. Blue Catfish, native to
32 large rivers throughout the United States, can move tens of kilometers in reservoirs and several
33 hundreds of kilometers in rivers (Graham 1999). Blue Catfish may move upstream in the spring
34 and summer (Lagler 1961, Graham 1999) in reservoirs (Timmons 1999; Grist 2002) and rivers
35 (Garrett 2010). They also move downstream in the fall and winter (Lagler 1961; Pflieger 1997;
36 Graham 1999) in reservoirs (Grist 2002) and rivers (Garrett 2010), including downstream
37 emigration out of reservoirs (Graham and DeiSanti 1999). Seasonal patterns may vary (Lagler
38 1961, Pflieger 1997; Graham 1999; Timmons 1999; Fisher et al. 1999; Grist 2002, Garrett 2010).
39 In addition, diel conditions can alter catfish distribution (Graham 1999; Pugh and Schramm
40 1999; Baras and Laleye 2003; Nunn et al. 2010). Variation in distribution and movement across
41 systems reinforces the need to compare patterns across catfish populations (Kwak et al. 2011).
42 Blue Catfish distribution in reservoirs is not well known, whether Blue Catfish exit reservoirs is
43 not well known, and how season, diel period, size, and individual variation affect Blue Catfish
44 distribution are not well known. Although little quantitative data exist on these issues,
45 researchers and managers have assumed certain patterns of Blue Catfish distribution that have
Chapter 2 – Distribution and Egress - Objective 4-5
26
not been adequately tested, especially in KS reservoirs. As such, t 46 his research seeks to fill this
47 information gap on how Blue Catfish are distributed.
48 Smaller scale distribution patterns (e.g. daily, seasonal, non-breeding periods, ontogenetic
49 and habitat shifts; Werner and Gilliam 1984; Albanese et al. 2004; Roberts and Angermeier
50 2007; Albanese et al. 2009) and long distance migrations (Hobson 1999; Borcherding et al. 2002;
51 Roberts and Angermeier 2007) alter organismal distribution. New technology (e.g., electronic
52 tags) now allows for quantification of animal distributions (Hobson 1999; Metcalfe 2006). The
53 objectives of this chapter are to,: (1) document locations of tagged Blue Catfish within Milford
54 Reservoir, (2) assess if Blue Catfish migrate out of Milford Reservoir, (3) quantify changes in
55 distribution across months and diel periods, (4) test if Blue Catfish size affects distribution, and
56 (5) identify whether groups of same-sized individual Blue Catfish are distributed in the same
57 way.
58
59 METHODS
60 Study System. Milford Reservoir (39°08'42"N, 96°56'54"W) is an impoundment of the
61 Republican River (Dickinson, Clay, and Geary counties, KS) and is part of the Lower
62 Republican watershed, KS (Chapter 2 Figure 1). Milford Reservoir has a surface area of 6,555
63 ha, 262 km of shoreline dominated by limestone cobble and boulders, an average depth of 6.7 m,
64 and a maximum depth of 19.8 m (Reinke 2001).
65 Fish Tagging (Number, Size, Timing). In both 2012 and 2013, we targeted the most
66 common size of Blue Catfish in Milford Reservoir (about 400-600 mm) as determined from
67 previous field assessments (Chapter 1 Appendix Figure 1). In 2013, a limited number of smaller
68 and larger Blue Catfish were added (Chapter 1 Table 2). On 26-28 June, 2012, we internally
Chapter 2 – Distribution and Egress - Objective 4-5
27
implanted 48 Blue Catfish with VEMCO V9 acoustic tags (69 mean fish size = 487 mm TL, range
70 383-1020, SE 14.5, n=48). On 3-5 June, 2013, we internally implanted 75 Blue Catfish with
71 VEMCO 9 and V13 tags (mean fish size = 517 mm TL, range 343-1090, SE 17.8, n=75).
72 Tagging procedures are described in detail elsewhere (Chapter 1). Blue Catfish were collected at
73 three locations within Milford Reservoir: Causeway, Madison Creek, and School Creek. Fish
74 were released in the same location where they were caught and tagged. Equal numbers of fish
75 were tagged at each location on sequential days using identical protocols. We test whether
76 capture location affected distribution with a Kruskal-Wallis test and post-hoc multiple
77 comparison (kruskalmc, pgirmess package R).
78 Receiver Placement. In 2012 and 2013, we tracked tagged Blue Catfish with a 20-
79 stationary receiver array (deployed on the bottom) and a 57-site monthly manual receiver survey
80 (discussed in Chapter 3). For the stationary array, data were collected using VEMCO (VR2W-
81 69kHz) receivers which received coded pings from tags each time a tagged fish came within
82 range (i.e, 600 m of the receiver). In 2012, the receivers were placed at 18 locations within the
83 reservoir and two locations adjacent to the reservoir exits (Chapter 1 Figure 3). The upper river
84 receiver (receiver 1) and the upper within-reservoir receiver (receiver 2) formed a two-tiered gate
85 to detect upriver egress from the reservoir. The southernmost receivers in the reservoir (receiver
86 19) and the river receiver below the dam (receiver 20) formed a two-tiered gate to detect
87 downriver egress (Chapter 1 Figure 3). We also had two 3-stationary receiver gate arrays
88 (receivers 6-8, 11-13) across the mid-reservoir constriction (i.e., the limited width allowed
89 complete coverage of the entire reservoir as confirmed by range tests) to detect any fish that
90 moved through the middle region of the reservoir. In 2012, for data analysis, we removed data
91 from 2 of the 3 receivers in these gates (7, 8, 11, 13) to obtain a more even distribution of
Chapter 2 – Distribution and Egress - Objective 4-5
28
receivers. Thus, in 2012, of the 18 within reservoir receivers, 14 were 92 used for data analysis. In
93 2013, we deployed receivers similarly (May-November 2013; Chapter 1 Table 5). However,
94 receiver 1 was vandalized in August, 2013. Receivers 16-17 were lost due to vandalism or
95 boating collisions. Gate receiver 13 replaced gate receiver 12 because 12 was lost. As in 2012, in
96 2013, we also removed data from 2 of the 3 gate receivers (7, 8, 11, 12) for the same reasons.
97 Thus, in 2013, of the 18 within reservoir receivers, 12 were used for data analysis. Details of
98 array deployment and range testing are described in detail elsewhere (Chapter 1). Receivers were
99 grouped into five regions (upper, upper middle, Madison, lower middle, and lower; Chapter 1
100 Figure 4). The manual tracking survey, undertaken in June through November, 2013 (described
101 in detail in Chapter 3), was used to confirm stationary distribution data.
102 Data Format. When each receiver was downloaded, each individual tag detection was
103 recorded as a single data line including a date, time, and fish tag number. After field data
104 downloads were complete, data from all receivers were combined using VEMCO’s VUE
105 software, Microsoft ACCESS, and Microsoft EXCEL.
106 Egress. To test egress through the river up reservoir or past the dam down reservoir, the
107 four extreme receivers (1, 2, 19, 20) were downloaded regularly to check for detections. The
108 downloaded data for these receivers were examined for fish number. Discharge was examined
109 during the field season in both years (USGS 06857100 Republican River at Junction City, KS).
110 Overview of Experimental Design. Here, we first provide an overview of the research
111 design. Then we give more details for each component in subsequent sections. Because a
112 trajectory is too complex for quantitative analysis, to quantify distribution we focused on three
113 component metrics: unique individuals, residence time, and numbers of movements (Chapter 2
114 Figure 2). These responses are defined in more detail below. For distribution at each receiver, we
Chapter 2 – Distribution and Egress - Objective 4-5
29
examined two responses (numbers of unique individuals, mean 115 residence time) using maps and
116 Chi square analyses. Then we used one response (residence time at each receiver) to visually
117 depict and statistically test three treatments that might affect distribution: season, diel period, and
118 fish size. Numbers of movements were quantified for individual fish, receiver, and season.
119 Individual fish variation was examined with cluster analyses and box plots.
120 Responses. We used three specific components of trajectories (unique individuals,
121 residence time, and numbers of movements between receivers) to describe Blue Catfish
122 distribution within Milford Reservoir. Unique individuals, residence time, and movements were
123 summarized to provide a system-wide distribution pattern. Residence time was used to test all
124 treatments (season, diel, and size) and to calculate clusters.
125 Numbers of unique individuals, residence time, and numbers of movements are all
126 approaches to quantifying the distribution of tagged fish. To obtain this metric, the above
127 described data base was manipulated by fish number and date for each receiver and the presence
128 of individual fish at a specific location at a specific time was recorded. Residence time is a
129 relatively new metric for fish tracking and is only possible with an extensive array of stationary
130 receivers as we have deployed here. Residence time, likely our most useful response, quantifies
131 how much time each animal spends at each location. For fixed receivers that record data 24 h day
132 in the same location, residence time is the preferred metric and replaces home range, which
133 typically requires detections at random not fixed locations. To calculate residence time, raw
134 detection data from the receivers were transformed into residence times for each fish at each
135 receiver location using VTrack (R 2.15.2 software; R Core Team) (Campbell et al. 2012). This
136 program records a fish as present (or resident) at a specific location after two detections and until
137 it is not detected for a period of time specified by the researchers (here 1 h). Movements between
Chapter 2 – Distribution and Egress - Objective 4-5
30
receivers were also calculated by the VTrack program. For 138 this metric, detection between
139 receivers is tallied as a single movement.
140 Distribution. To quantify distribution, unique individuals and residence time were
141 calculated for the entire study period (June through November). These data were plotted on maps
142 of Milford Reservoir. Unique individuals and residence time were compared across receivers
143 using a Chi square analysis with 2000 Monte Carlo simulations in which the expected was an
144 even distribution. For unique individuals, an even distribution is calculated as the same number
145 of fish at each receiver. For residence time, an even distribution is calculated as an equal amount
146 of time spent at each receiver. For unique individuals, the Chi square analysis evaluated if fish
147 were evenly distributed. For residence time, Chi square analysis assessed if fish were spending
148 more time, less time, or the same amount of time at all receivers.
149 Tests of Season, Diel Period, and Fish Size Effects. We also tested if residence time
150 differed across season (months), diel period, and fish size. For season, residence time for June,
151 July, August, September, October, and November were calculated for each fish. Then differences
152 in residence time among months was tested with a Kruskal-Wallis test and post-hoc multiple
153 comparisons Individual fish were treated as replicates. For diel periods, residence times were
154 calculated for four daily time periods: (a) a 2 hour period centered around dawn, (b) day, (c) a 2
155 h period centered around dusk, and (d) night. Residence time was divided by hours in each diel
156 period before these four diel periods were compared with a Kruskal Wallis test. To test the effect
157 of fish size, we ran a univariate regression between fish total length (mm TL, treatment or X) and
158 residence time (response or Y).
159 Calculation of Clusters. To compare individual behavior, we used separate cluster
160 analyses on residence time for each month and all seasons combined. For cluster analysis,
Chapter 2 – Distribution and Egress - Objective 4-5
31
residence time data were log transformed and then a Euclidean 161 distance matrix was created. The
162 non-hierarchical method PAM (partitioning around medoids) was run on the data using the PAM
163 function in R (source) (‘cluster’ package) to determine if there were similar groups of fish
164 present throughout the reservoir. The optimal number of clusters was determined using silhouette
165 plots and Jaccard bootstrap mean values obtained from the bootstrap method (‘clusterboot’
166 function; ‘fpc’ package). Jaccard bootstrap mean values >0.60 confirmed cluster patterns
167 (Hennig 2010). The ecological meaning of the clusters was determined by receiver and season-168
specific boxplots for each cluster. For synthesis, we combined all monthly clusters into three
169 general movement patterns. This synthesis combined the voluminous original cluster data
170 (shown as monthly clusters in the appendix) into synthesis clusters.
171
172 RESULTS
173 Overall. In July - November, 2012, we recorded 1,139,515 detections. In June-October,
174 2013, we recorded 2,044,881 detections. These detections were made by 85% of the fish we
175 tagged. In 2012, five fish either died or lost their tags. In 2013, 11 fish died or lost their tags with
176 one fish a confirmed catch by an angler. These “missing” fish were not considered in the data
177 analysis.
178 Distribution: Unique Individuals and Residence Time. For both unique individuals and
179 mean residence time, tagged Blue Catfish did not spend equal amounts time in all areas of
180 Milford Reservoir. In 2012, for unique individuals, fish were concentrated in the upper middle
181 and lower middle regions of the reservoir with more fish than expected at receivers 4, 5, 6, 12,
182 14, 15 (Chapter 2 Figure 3A, B) and less fish than expected at receivers 2, 3, 9, 10, 17, 18, 19
183 (Chapter 2 Figure 3A, C). Chi square simulations statistically confirmed these patterns of
Chapter 2 – Distribution and Egress - Objective 4-5
32
aggregation (P<0.001; Chapter 2 Figure 3B, C). In 2013, for unique 184 individuals, fish were again
185 concentrated in the upper middle and lower middle regions of the reservoir as well as in the
186 upper reservoir region, with more fish than expected at receivers 2-6, 9, 13-14 (Chapter 2 Figure
187 4A, B) and less fish than expected at receivers 10, 15, 18-19 (Chapter 2 Figure 4A, C). Chi
188 square simulations again statistically confirmed patterns of aggregation (P<0.001; Chapter 2
189 Figure 4B, C).
190 In 2012, for mean residence time, fish were concentrated in the upper middle and lower
191 middle regions of the reservoir as well as Madison Creek with fish spending more time than
192 expected at receivers 6, 9, 10, 12 (Chapter 2 Figure 5A, B) and less time than expected at
193 receivers 2, 3, 4, 5, 14-19 (Chapter 2 Figure 5A, C). Chi square simulations statistically
194 confirmed these patterns of aggregation (P<0.001; Chapter 2 Figure 5B, C). In 2013, for mean
195 residence time, fish favored the upper middle region with fish spending more time than expected
196 at receivers 4, 6 (Chapter 2 Figure 6A, B) and less time than expected at receivers 2, 3, 5, 10, 14-
197 15, 18-19 (Chapter 2 Figure 6A, C). Chi square simulations statistically confirmed patterns of
198 aggregation (P<0.001; Chapter 2 Figure 6 B, C). For both responses in both years, this
199 clustering occurred in the funnel above the reservoir constriction (upper middle region) and
200 within the upper constriction (upper part of lower middle region).
201 Egress. In 2012 and 2013, no fish left Milford Reservoir through the downstream egress
202 via the dam (receiver 20; Chapter 2 Figure 7). In 2012, no fish left Milford Reservoir through
203 the upstream egress (receiver 1; Chapter 2 Figure 7; Chapter 2 Table 1). However, because of
204 the vandalized upstream receiver (receiver 1) in 2013, we had to rely on the inner gate (receiver
205 2) to detect potential upstream egress. In 2013, only five fish were last seen at the upstream
206 receiver 2 (receiver 20; Chapter 2 Figure 7). All five of these fish repeatedly traversed the upper
Chapter 2 – Distribution and Egress - Objective 4-5
33
and upper middle reservoir in spring as is shown by the repeated ve 207 rtical lines of detections
208 (Chapter 2 Figure 8). Two of these fish were not detected subsequently because receivers were
209 removed at the end of the study (Chapter 2 Figure 8A, B). The remaining three fish traversed
210 frequently between receiver 2 and other reservoir receivers. These repeated movements back and
211 forth through the upper reservoir (i.e. repeating vertical bands of detections) are unlike the quick
212 unidirectional movement (i.e., one single vertical line) that would be expected for long-distance,
213 unidirectional upstream migrants (Chapter 2 Figure 8C, E). In summary, no fish left through the
214 downstream egress in either year, no fish left through the upstream egress in 2012, and < 3 of 75
215 tagged fish could have left the reservoir through the upper egress in 2013. Because our 2012 and
216 2013 field seasons corresponded with a regional drought, discharge was relatively low in June
217 through November in either year (Chapter 2 Appendix Figure 2).
218 Seasonal Differences. Seasonal distribution varied across select receivers in 2012
219 (Chapter 2 Figure 9) and 2013 (Chapter 2 Figure 10). When comparing boxplots for residence
220 time across months, in 2012, fish spent more time at upper reservoir receiver 2 in October (2;
221 P<0.05; Chapter 2 Figure 9A), but less time at upper reservoir receiver 3 in November (3;
222 P<0.05; Chapter 2 Figure 9B). No statistically significant monthly differences existed across
223 other receivers in the upper middle region (4, 5, 6; P>0.05; Chapter 2 Figure 9C-E), Madison
224 Creek (9, 10; P>0.05; Chapter 2 Figure 9F, G) or in select lower middle reservoir receivers (12;
225 P>0.05; Chapter 2 Figure 9H). However, other lower middle reservoir receivers (14-15; P< 0.0;
226 Chapter 2 Figure 9I, J), and lower reservoir receivers (16-19; P< 0.05; Chapter 2 Figure 9K-N)
227 were significantly different across months. For these southern receivers, residence times were
228 higher in the fall. In general, these seasonal changes reflected decreases in residence time at
229 upper reservoir receivers and increases in residence time at lower reservoir receivers in fall as
Chapter 2 – Distribution and Egress - Objective 4-5
34
upper reservoir fish moved south to the middle reservoir and m 230 iddle reservoir fish moved south
231 to the lower reservoir.
232 Seasonal trends in 2013 were more variable. In 2013, upper reservoir receivers again had
233 variable visitation across months (2, 3; P< 0.05; Chapter 2 Figure 10A, B). In 2013, fish again
234 spent more time at lower reservoir receivers in the later fall (18, 19; P< 0.05; Chapter 2 Figure
235 10K, L) as fish moved from north to south. In 2013, upper middle receivers (4, 5; P< 0.05;
236 Chapter 2 Figure 10C, D) and Madison Creek receivers (9, 10; P< 0.05; Chapter 2 Figure 10F,
237 G) differed across months but a consistent overall trend was unclear. Other upper middle (6) and
238 lower middle reservoir receivers (13, 14) were not significantly different across months (P>
239 0.05; Chapter 2 Figure 10E, H, I). As in 2012, for 2013, this pattern generally reflected higher
240 use of the lower region of the reservoir in fall. In fact, more movements occurred at receivers in
241 the lower middle and lower reservoir (receivers 12-18) in the fall (Chapter 2 Figure 11) even
242 though movements were not greater for these lower reservoir receivers when all time periods
243 were combined (Chapter 2 Figure 12).
244 Diel and Size Differences. We found no significant differences among residence times
245 across diel periods at any of the receiver locations for 2012 (P> 0.05; Chapter 2 Figure 13A-N)
246 or 2013 (P>0.05; Chapter 2 Figure 14A-L). Neither residence time (P>0.05; Fig. 15A, C) nor
247 number of movements (P>0.05; Fig. 15B-D) differed by fish size. As a distribution of
248 movements across individuals in 2012 shows, even individual fish of similar sizes vary
249 substantially in the amount they move (Chapter 2 Figure 16).
250 Capture, Tag, and Release Location. In both 2012 and 2013, tagged Blue Catfish were
251 detected more often near the receivers where they were originally captured, tagged, and released
252 (Chapter 2 Figure 17-18). Tagged Blue Catfish that were captured, tagged and released at the
Chapter 2 – Distribution and Egress - Objective 4-5
35
Causeway site (near receiver 5; Chapter 2 Figure 17 D, 18D) were 253 detected more frequently at
254 receiver 5 (Chapter 2 Figure 17 D, 18D) and at the adjacent receivers 4 and 6 (Chapter 2 Figure
255 17C,17E, 18C. 18E). Tagged Blue Catfish that were captured, tagged and released at the
256 Madison site (near receiver 9; Chapter 2 Figure 17F, 18F) were detected more frequently at
257 receiver 9 (Chapter 2 Figure 17F, 18F) and at the adjacent receivers 6 and 10 (Chapter 2 Figure
258 17E, 17G, 18E, 18G). Tagged Blue Catfish that were captured, tagged and released at the School
259 Creek site (near receiver 15; Chapter 2 Figure 17J, 18J) were detected more frequently at
260 receiver 15 (Chapter 2 Figure 17J, 18J) and at the adjacent receiver 14 (Chapter 2 Figure 17I,
261 18I). These trends were not surprising since the fish were aggregated at Causeway, Madison, and
262 School Creek when there were captured and continued to stay in those aggregations after they
263 were tagged and released. These results do not alter any of the interpretations of our data because
264 we captured and released fish in the same location.
265 Cluster Synthesis. With cluster analysis, we identified that different groups of individual
266 fish existed. Within groups, individuals were distributed similarly, but across groups differences
267 in distribution existed. By combining clusters across seasons, we identified three types of
268 distribution. The first type of distribution included fish that changed their seasonal distribution
269 (Chapter 2 Figure 19). In July and August, these fish were most common at receiver 6 (Chapter
270 2 Figure 19A, B). In September, eight clusters emerged that were spread throughout the upper
271 middle, lower middle, and lower reservoir (Chapter 2 Figure 19C). In October and November,
272 these clusters merged into one mega cluster that frequented the lower middle and lower
273 reservoir, especially receivers 12-19 (Chapter 2 Figure 19D, E).
274 The second type of distribution included the non-migrating reservoir fish which were
275 regulars in the funnel just above and within the upper reservoir constriction (Chapter 2 Figure
Chapter 2 – Distribution and Egress - Objective 4-5
36
20A-E). This distribution group was composed of a single cluster 276 in July and August (Chapter 2
277 Figure 20A, B). This distributional group did not migrate south in fall, and across all seasons
278 remained in the upper middle and lower middle reservoir near receivers 6 and 12 (Chapter 2
279 Figure 20C-E).
280 A third type of distribution group included the Madison Creek fish (Chapter 2 Figure
281 21A-E) that stayed near Madison Creek receivers (9, 10) in July (Chapter 2 Figure 21A),
282 September (Chapter 2 Figure 21C), October (Chapter 2 Figure 21D), and November (Chapter 2
283 Figure 21E). These synthesis groups were derived from the original monthly clusters which are
284 presented here as an appendix but are not interpreted separately (Chapter 2, Appendix Figures 3-
285 32).
286 In summary, the uneven distribution, observed across the entire reservoir, is the result of
287 clusters of fish using upper, upper middle, lower middle, and lower regions of the reservoir
288 differently with southern movements by some fish in the fall.
289
290 DISCUSSION
291 Overview of Unique Contributions of Our Research. Our extensive Blue Catfish tracking
292 data set provided novel insights into a long-standing, but largely untested, question in fisheries
293 biology, fisheries management, and fish ecology (e.g., where are fish located?). Our unique data
294 set is unprecedented relative to the numbers of tagged fish, numbers of detections, temporal
295 extent of detections, and spatial distribution of detections. Specifically, our research design
296 included 123 fish tagged across 2 years, 85% tag retention over 5 months per year, continuous
297 24-h tag detections during summer and fall; 2 tiers of gates at each reservoir egress point; 2 3-
298 receiver, across-reservoir gates; and a 12-14-stationary receiver array distributed throughout the
Chapter 2 – Distribution and Egress - Objective 4-5
37
reservoir. With this data set of substantial spatial and temporal s 299 cope, we tested focused
300 questions about Blue Catfish distribution (e.g., nature of distributional patterns) and factors that
301 may change Blue Catfish distribution (e.g., existence of seasonal egress, role of seasonal and diel
302 time periods, influence of fish size, behavioral patterns of same-sized individuals). Although
303 many aspects of Blue Catfish distributional patterns are widely accepted, assumptions about the
304 distribution of this important sport fish have rarely been tested. This is because an effective and
305 affordable methodology to track large numbers of individuals over an entire system at a detailed
306 time scale was not available in the past.
307 Our quantification of Blue Catfish distribution was more detailed than any previous study
308 (e.g., Fisher et al. 1999; Edds et al. 2002; Grist 2002; Garrett 2010) because we used this newly
309 available fish tracking technology effectively (e.g., acoustic tags and a stationary receiver, a
310 substantial receiver array, a high sample size of tagged fish, strong research design). As a result,
311 our results on distributional patterns neither supports nor contradicts existing data on Blue
312 Catfish distribution simply because the novel level of detail we provide through our fish tracking
313 did not exist previously. However, our quantitative tests of treatments that might alter
314 distributional patterns (e.g. Blue Catfish egress, seasonal patterns, diel periodicity, fish size, and
315 variability in individual behavior) are comparable to questions asked previously (e.g., Fisher et
316 al. 1999; Grist 2002; Garrett 2010). Relative to these variables, our results suggest that many
317 assumptions about egress, season, diel periodicity, fish size, and individual variation may not be
318 widely applicable. We hope our research stimulates future tests of across system synthesis.
319 Together, these data (past descriptive research, this present study, and future studies) will
320 provide synthesis and generalization about distribution patterns of this important, popular, and
321 mobile sport fish predator.
Chapter 2 – Distribution and Egress - Objective 4-5
38
Distribution Patterns. Blue Catfish in Milford Reservoir were consistently 322 clustered in an
323 upper middle reservoir aggregation. This pattern was similar for two different fish responses
324 (e.g., numbers of unique tagged individuals, average residence time per individual). Specifically,
325 for all months and both years, more fish were present and individual fish spent more time in the
326 upper middle reservoir funnel that starts just above the upper reservoir constriction and ends just
327 below the Madison Creek confluence. Interestingly, this concentration of fish and elevated fish
328 residence is not in the geographic center of the reservoir and does not include the entire middle
329 reservoir constriction, but instead focuses on the geographic area leading into the constriction
330 funnel down through the upper constriction (through the first major tributary, Madison Creek).
331 Although fish were consistently concentrated in this funnel, they were not sedentary and
332 frequently moved to other locations before returning to the above described location.
333 The spatial resolution of our results far exceeds that provided by previous studies. Other
334 peer-reviewed Blue Catfish distributional studies do not provide detailed maps of system-wide
335 distributional patterns (e.g. Fisher et al. 1999; Edds 2002; Grist 2002; Garrett 2010). Although an
336 uneven distribution is probably common in fisheries and ecology, the detailed and consistent
337 view of an aggregated and clustered population, apparent from our data, is not frequently seen in
338 the existing fish ecology or fisheries management literature. Much scientific research discusses
339 and speculates about uncertainty in research results. Because of the design of our study and the
340 quality of our data, we know where Blue Catfish were located in Milford Reservoir. As seen in
341 the next chapter, manual tracking which covers more locations (n=57) for a shorter time
342 confirms this consistent aggregation in the mid-reservoir funnel and adds some additional details
343 on localized heterogeneity.
Chapter 2 – Distribution and Egress - Objective 4-5
39
Egress. We did not detect any tagged Blue Catfish migrating 344 out of Milford Reservoir
345 from June through November, 2012-2013, based on our continuous (24 h a day) tracking of 123
346 tagged fish at double egress gates at both upstream and downstream exits. We know that 85% of
347 the fish, tagged in both years, do not leave the reservoir because they were continually detected
348 at specific locations within the reservoir. We know for certain that no tagged Blue Catfish left
349 downstream past the dam in 2012 or 2013 because of our intact double gates at downstream
350 egress points (receivers 19 upstream of the dam; receiver 20 downstream of the dam) in both
351 years. We also know for certain that none of the 48 fish tagged in 2012 left the reservoir through
352 the upstream exit because of the presence of an intact double gate at the upstream egress point
353 (receiver 1; receiver 2). During the last part of the 2013 field season, receiver 1 was lost.
354 Unfortunately, receiver loss is common in tracking studies with fixed gear. However, the second
355 or inner tier of the upper gate (i.e., receiver 2) remained in place throughout the 2013 field
356 season and allowed us to evaluate if any tagged Blue Catfish might have exited the reservoir
357 using this route. Only five of 75 Blue Catfish, tagged in 2013, were last seen at receiver 2. Of
358 these, two were not redetected because the study ended and receivers were removed. Thus, the
359 ultimate fate of < 3 of 75 Blue Catfish tagged in 2013 is uncertain. Because these three fish
360 repeatedly moved back and forth between receiver 2 and other reservoir receivers, it is unlikely
361 that these three fish left the reservoir in 2013. Despite the unknown final disposition of these
362 three fish, our data clearly indicate that most Blue Catfish tagged in Milford Reservoir in 2012-
363 2013 did not make long distance migrations out of the study system in our summer-fall field
364 season.
365 In other studies, upriver or up-reservoir movements of Blue Catfish have been observed
366 in spring and downriver or down-reservoir movement have been observed in fall (Fisher et al.
Chapter 2 – Distribution and Egress - Objective 4-5
40
1999; Garrett 2010). In Milford, a few fish irregularly moved from 367 the lower receiver to the
368 upper receiver, but these rare movements for a few fish occurred over several weeks and were
369 not a common response. Spring movements are often associated with spawning, typically in
370 April-June at 21-24oC (Graham 1999). We did not track Blue Catfish in spring. If Blue
371 Catfish individuals left Milford Reservoir during June on a spawning migration, we would not
372 have captured them for tagging. In Milford Reservoir, during June 2014, water temperatures
373 exceeded 21o C, the optimal for spawning. If Blue Catfish spawned within Milford Reservoir,
374 likely our study missed that April-May period of spawning activity. Hence, if long distance
375 movement is associated with spring spawning, we would not detect these trends because of the
376 timing of our study. Discharge may be a variable influencing egress (Garrett 2010). In 2012 and
377 2013, stream flow and discharge from Milford Reservoir was low. If long distance migration out
378 of the reservoir is linked to changes in discharge, lack of hydrological variability during our
379 study may have prevented or reduced emigration.
380 When fish are tagged and not detected, stocked and never recovered, or just never
381 captured in standardized sampling, disentangling mortality and emigration is difficult.
382 Researchers and managers are often simply unable to answer whether fish die, leave, or evade
383 capture. Long distance movement may be erroneously suspected when simpler explanations
384 (e.g., mortality, sampling inefficiency) are in fact the underlying cause. If egress is variable
385 across fish within and across systems, system specific characteristics (system size, up and down
386 river configurations, availability of spawning and overwintering habitats within the reservoir,
387 population characteristics, and possible sampling design) may be responsible. Movement out of
388 reservoirs may be more common for stocked fish. Blue Catfish in Milford Reservoir are naturally
Chapter 2 – Distribution and Egress - Objective 4-5
41
reproducing (Goeckler et al. 2003), thus adequate spawning habitat 389 may be available within the
390 reservoir itself.
391 For most existing studies, extreme movements are described for a brief period for a few
392 fish. Unquestionably, Blue Catfish can move great distances (e.g., Lagler 1961; Garrett 2010).
393 Although an intriguing life history anecdote, a few observations of a few individuals provides
394 only a small piece of the distributional puzzle. Our depiction of how a large tagged population is
395 distributed over a long time period and a large spatial framework provides a different view of
396 Blue Catfish distribution that is perhaps more useful for research and management. Whether our
397 results of no egress are unusual for Blue Catfish in reservoirs or the more common pattern is
398 unclear. Tagging provides a way of testing these residency-migration patterns, but this
399 methodology requires resources (tags and receivers) and constant vigilance (i.e. labor intensive)
400 to maintain receivers.
401 Role of Season. Seasonal changes in distribution of Blue Catfish in Milford Reservoir
402 were more complex than previously assumed and varied across individuals. In Milford
403 Reservoir, some, but not all, tagged Blue Catfish moved south in fall. In addition, not all tagged
404 individuals moved down reservoir to the same extent. Others (Fisher et al. 1999; Garrett 2010)
405 have observed a southern shift in distribution in the fall and have speculated that this shift may
406 be related to overwintering. Most previous data on fall distributional shifts are based on a few
407 fish in a few locations (Fisher et al. 1999; Garrett 2010). Our data provide a much more detailed
408 view of seasonal changes in distribution. In our research, some tagged Blue Catfish in Milford
409 Reservoir moved south to the deepest part of the reservoir by the dam, as suggested by other
410 studies (Fisher et al. 1999). However, some of our tagged fish also moved to the middle and
411 lower middle region of the reservoir, south of their original location but not to the southernmost
Chapter 2 – Distribution and Egress - Objective 4-5
42
part of the reservoir. In addition, some tagged Blue Catfish 412 fish did not move down reservoir at
413 all but remained either in the middle reservoir or in Madison Creek. Without tagging and
414 tracking of individual fish of the same size, the complex and subtle details in this distributional
415 shift would not have been detected.
416 Individual Variation. Only a subset of individually-tagged Blue Catfish made a down-417
reservoir shift in distribution. Individuals of the same size have been assumed to behave in the
418 same general way. For the Blue Catfish that we tagged in Milford Reservoir, this was not true.
419 We observed clusters of similar-sized fish that were distributed differently both within and
420 across months. This pattern of clustering was complex. As a simplification of this individual
421 variation pattern revealed by the cluster analysis integrated across months, three types of spatial
422 distributions were observed. The first pattern was composed of Blue Catfish that used the upper
423 middle reservoir funnel in summer, then visited a range of southern locations in fall. The second
424 pattern was composed of Blue Catfish that used the upper middle reservoir funnel in summer and
425 fall and did not move south. The third pattern was composed of Blue Catfish that used the
426 Madison Creek region and also did not migrate seasonally. Our study is one of the first to
427 document these individual distributional groups for freshwater fish of the same size. This may be
428 a general pattern for predators as contingents of acoustically-tagged individuals have been
429 documented in coastal systems (e.g., striped bass, Pautzke et al. 2010). As the incidence of these
430 patterns increase, likely more sophisticated tools for analyzing and simplifying these data will
431 emerge (e.g., network analyses).
432 Behavioral syndromes occur when individuals or a group of individuals display
433 specialized traits or behaviors that vary from the population mean (Sih et al. 2004; Huntingford
434 et al. 2010). Behaviors exhibited by groups of individuals can have important ecological and
Chapter 2 – Distribution and Egress - Objective 4-5
43
evolutionary impacts, which can affect species distributions 435 and responses to environmental
436 change (Sih et al. 2004; Flaxman et al. 2011). Behavior of animals has been used in very few
437 studies to try to understand its influence on the spatial structure of populations (Knaepkens et al.
438 2005; Giuggioli and Bartumeus 2010; Fullerton et al. 2010). Within the behavioral syndrome
439 literature, few have used distribution patterns to distinguish groups of individuals. The patterns
440 we observed may be an example of behavioral syndromes based on distribution,
441 Effect of Diel Period. The distribution of the tagged Blue Catfish in Milford Reservoir
442 did not differ across diel period. Specifically, we observed no significant differences in residence
443 time at any receiver among the dawn, day, dusk, and night time periods for either year.
444 Differences in diel distribution of fish and other organisms has been a topic of interest in
445 fisheries and ecology for decades. However, diel patterns are rarely tested so much of this
446 speculation is based on limited quantitative data. In fisheries, many of our expectations are
447 influenced by angler experiences. In addition, traditional sampling across seasons, diel periods,
448 and locations, are unlikely to capture the full range of variability (i.e., diel differences or no diel
449 differences). For this reason, our data on residence time collected at 12-14 locations 24 hours a
450 day for 123 tagged fish over five months provide some of the most credible evidence available
451 that differential distribution did not occur among dawn, day, dusk and night time periods.
452 Physiological and diet generalists, like Blue Catfish, may take advantage of favorable conditions
453 for feeding, resting, and other activities without regard for time of day.
454 Effect of Fish Size. We also did not observe any difference in distribution and movement
455 related to Blue Catfish size. We included some smaller and some larger individuals, but most
456 fish we tracked were within the most common 400-600 mm TL size range. Substantial literature
457 exists to suggest that fish change their ecological role with size, but this ontogenetic niche shift is
Chapter 2 – Distribution and Egress - Objective 4-5
44
most pronounced when fish life stage or ecological habitats change w 458 ith size (e.g., Werner and
459 Gilliam 1984). Blue Catfish are reputed to spawn at 420-480 mm (Graham and DeiSanti 1999),
460 which suggests most fish we tagged were mature adults. For our data, although individual
461 distribution varied, fish size did not cause this this pattern. As suggested above, physiological
462 and diet generalists of a range of sizes may all take advantage of conditions for feeding, resting,
463 and spawning, as they occur. As such, other variables may affect distribution of Blue Catfish
464 more than size.
465 Management Implications. Our research on distribution has several management
466 implications. First, we have provided substantial information on where Blue Catfish are located.
467 Knowing distribution is critical for all management and research activities. Existing data on
468 distribution are very limited. Using a newer technology, we have compiled the best
469 understanding we have ever had of where Blue Catfish are located in Milford Reservoir. Our
470 spatially explicit approach suggests that fish are highly aggregated often in consistent locations.
471 Trends were surprisingly similar across years. If managers can identify the locations of these
472 Blue Catfish clusters in other reservoirs, they should be able to better assess the stock and more
473 effectively collect biological samples (e.g., diet, aging structures). To find these clusters,
474 managers might implement an extensive survey in which they systematically sample the entire
475 reservoir to identify patterns of aggregation. For example, in the future, managers might shock
476 50 locations once rather than 10 locations five times.
477 Second, we did not observe Blue Catfish leaving Milford Reservoir. Blue Catfish are
478 thought to be attracted by flow. Our study occurred during a regional drought so the absence of
479 movement out of the reservoir might be related to the lack of hydrological cues. If river
480 discharge or releases at the dam had been higher, our results might have been different. On the
Chapter 2 – Distribution and Egress - Objective 4-5
45
other hand, this lack of Blue Catfish egress may be typical 481 of Milford Reservoir and other
482 reservoirs. Many documented longer distance movements of Blue Catfish may be irregular
483 observations of relatively few individuals. Our results and those of others clearly document that
484 movement varies dramatically among individuals. Of course, tools exist to track long distance
485 movements. However, in Milford and other reservoirs, effort might be better used to map the
486 distribution of the Blue Catfish reservoir population that does not migrate which may be
487 comprised of as many or more individuals than the migrators.
488 Third, the number of empirical studies on Blue Catfish distribution, movement, and
489 habitat is increasing. However, at present, each one represents an isolated data point because of
490 system-specific differences in morphometry, bathymetry, habitat, and researcher-specific
491 methodological differences across studies. Researchers and managers would benefit from a
492 standardized synthesis of what is actually known about Blue Catfish distribution and movements
493 across a wide range of states and ecological systems. This synthetic working group effort could
494 formulate a range of broader questions of interest then use existing data to objectively test
495 hypotheses about distribution and movements.
496 Some management utility may arise from the awareness that discrete groups of same-497
sized fish can differ in their distribution. These results are novel in the field of freshwater fish
498 biology and management. As such, their present applications are unclear. However, knowledge
499 of this pattern could be useful in the future. For example, awareness that a subset of Blue Catfish
500 in Milford Reservoir remain within Madison Creek could influence habitat management,
501 restoration, and planning.
502 Finally, in its conception, this study was designed to look at the distribution of mobile
503 organisms in the most transparent way possible. Specifically, a decision was made to look at a
Chapter 2 – Distribution and Egress - Objective 4-5
46
system with a naturally reproducing population w 504 here there was no stocking to confound
505 patterns. Likely systems with other morphometric characters and fish that are stocked will show
506 different patterns. Our data provides a very strong baseline for across system comparison.
507 In summary, our data have addressed the research objectives of the original study. Of
508 course, as in any complex research and management area, a host of important questions about
509 distribution and movement remain. Nevertheless, our study has provided a wealth of information
510 on distribution and egress that was previously unknown.
Fish
Date
Receiver
Fish
Date
Receiver
1 Jan. 15 2013 19 1 July 21 2013 6
2 Jan. 15 2013 12 2 Dec. 21 2013 4
3 Jan. 6 2013 18 3 Dec. 4 2013 8
4 Jan. 17 2013 17 4 Nov.25 2013 8
5 Jan. 9 2013 18 5 June 21 2013 6
6 Jan. 9 2013 18 6 Nov. 17 2013 8
7 Jan. 15 2013 12 7 June 17 2013 4
8 Jan. 8 2013 18 8 Nov. 9 2013 18
9 Jan. 9 2013 18 9 Nov. 7 2013 13
10 Jan. 9 2013 18 10 Nov. 9 2013 15
11 Jan. 9 2013 18 11 Dec. 11 2013 4
12 Jan. 15 2013 12 12 June 9 2014 2
13 Jan. 15 2013 12 13 June 18 2014 8
14 Jan. 15 2013 12 14 June 18 2014 10
15 Dec. 28 2012 18 15 June 18 2014 8
16 Jan. 9 2013 18 16 June 18 2014 7
17 Jan. 9 2013 18 17 June 18 2014 8
18 Jan. 9 2013 18 18 June 1 2014 5
19 Jan. 9 2013 18 19 June 6 2014 5
20 Jan. 8 2013 18 20 May 20 2014 8
21 Jan. 10 2013 11 21 April 13 2014 8
22 Aug. 8 2012 5 22 June 16 2014 10
23 Jan. 16 2013 5 23 June 16 2014 10
24 Jan. 16 2013 12 24 June 18 2014 10
25 Jan. 16 2013 12 25 June 17 2014 10
26 Jan. 9 2013 18 26 June 18 2014 10
27 June 27 2012 5 27 April 28 2014 10
28 Jan. 9 2013 18 28 June 15 2014 7
29 Jan. 8 2013 16 29 June 11 2013 10
30 Oct. 5 2012 8 30 April 11 2014 8
31 Jan. 9 2013 17 31 June 18 2014 8
32 Aug. 6 2012 4 32 Feb. 26 2014 8
33 Aug. 20 2012 10 33 May 30 2014 5
34 Jan. 16 2013 12 34 June 19 2014 4
35 Jan. 9 2013 18 35 June 8 2014 5
36 Jan. 6 2013 7 36 May 8 2014 8
2012 Overall Last
Seen 2013 Overall Last Seen
Chapter 2 Table 1. Fish, date, and receiver at which tagged Blue
Catfish were last seen for 2012 and 2013 in Milford Reservoir,
Kansas. Fish last seen at receiver 2 in 2013 are boxed.
Fish
Date
Receiver
Fish
Date
Receiver
37 Jan. 10 2013 19 37 June 15 2014 5
38 Jan. 17 2013 8 38 June 15 2014 8
39 Dec. 5 2012 6 39 April 9 2014 5
40 Dec. 5 2012 16 40 June 22 2013 15
41 Dec. 5 2012 17 41 July 20 2013 14
42 Dec. 4 2012 18 42 June 7 2014 5
43 Dec. 5 2012 13 43 Aug. 30 2013 4
44 Dec. 5 2012 17 44 June 20 2014 4
45 Dec. 4 2012 16 45 June 19 2014 7
46 Dec. 6 2012 18 46 June 17 2014 8
47 Dec. 6 2012 8 47 June 21 2014 4
48 Dec. 23 2012 17 48 June 21 2014 4
49 June 10 2014 5
50 June 21 2014 4
51 April 27 2014 5
52 June 19 2014 8
53 June 20 2014 5
54 June 20 2014 4
55 June 21 2014 4
56 June 8 2014 2
57 April 20 2014 5
58 July 28 2013 6
59 June 20 2014 5
60 Jan. 1 2014 7
61 June 20 2014 8
62 Feb. 29 2014 2
63 Feb. 28 2014 5
64 Feb. 25 2014 4
65 Nov. 9 2013 14
66 Oct. 2 2013 13
67 Feb. 29 2014 2
68 June 16 2013 3
69 Nov. 9 2013 17
70 Nov. 9 2013 15
71 Feb. 27 2014 5
72 Feb. 30 2014 4
73 June 19 2013 3
74 Nov. 12 2013 7
75 Feb. 30 2014 2
2012 Overall Last Seen 2013 Overall Last Seen
Chapter 2 Table 1. Continued.
Chapter 2 – Distribution and Egress - Figure Captions
47
CHAPTER 2 – DISTRIBUTION BLUE CATFISH 1 WITHIN AND EGRESS OF BLUE
2 CATFISH FROM MILFORD RESERVOIR (OBJECTIVES 4-5)
3
4 CHAPTER 2 FIGURE CAPTIONS
5
6 Chapter 2 Figure 1. (A) Our study site, Milford Reservoir, is an impoundment of (B) the Lower
7 Republican River watershed in (C) northeastern Kansas.
8
9 Chapter 2 Figure 2. Examples of a trajectory made by a single tagged Blue Catfish that
10 illustrates select components of a complex trajectory pattern. Residence time quantifies how long
11 a tagged fish is at a single receiver location when detections for the entire time period of interest
12 are summed. Numbers of movements quantifies how many times a fish moves from receiver to
13 receiver for the entire period of interest. Numbers of unique individuals (i.e., the presence of a
14 single individual fish) and mean residence time are metrics that quantify the distribution of all
15 individuals together (i.e., the tagged population).
16
17 Chapter 2 Figure 3. (A) The spatial distribution of unique individuals (number) is shown for 48
18 tagged Blue Catfish at 14 receivers (18 receivers with four gate receivers removed) in 2012.
19 Each dot represents a receiver location. The size of the dot is proportional to numbers of unique
20 individuals. Also shown are the results of a Chi square analysis that identifies at which receivers
21 (B) more unique individuals occurred than were expected and (C) fewer unique individuals
22 occurred than were expected based on an even distribution (i.e., the same number of fish at all
23 receivers). In B-C, receiver numbers are shown. On the map in A, dark gray dots indicate more
Chapter 2 – Distribution and Egress - Figure Captions
48
unique individuals than expected and light gray dots 24 indicate fewer unique individual than
25 expected based on an even distribution.
26
27 Chapter 2 Figure 4. (A) The spatial distribution of unique individuals (number) is shown for 75
28 tagged Blue Catfish at 12 receivers (18 receivers with four gate and two missing receivers
29 removed) in 2013. Each dot represents a receiver location. The size of the dot is proportional to
30 numbers of unique individuals. Also shown are the results of a Chi square analysis that
31 identifies at which receivers (B) more unique individuals occurred than were expected and (C)
32 fewer unique individuals occurred than were expected, based on an even distribution (i.e., the
33 same number of fish at all receivers). In B-C, receiver numbers are indicated. On the map in A,
34 dark gray dots indicate more unique individuals than expected and light gray dots indicate fewer
35 unique individual than expected based on an even distribution.
36
37 Chapter 2 Figure 5. (A) The spatial distribution of mean residence time (h) is shown for 48
38 tagged Blue Catfish at 14 receivers (18 receivers with four gate receivers removed) in 2012.
39 Each dot represents a receiver location. The size of the dot is proportional to mean residence
40 time. Also shown are the results of a Chi square analysis that identifies at which receivers mean
41 residence time was (B) higher than that expected or (C) less than expected based on an even
42 distribution (i.e., fish spent the same amount of time at all receivers). In B-C, receiver numbers
43 are indicated. On the map in A, dark gray dots indicate a higher residence time than expected,
44 white dots indicate residence times equal to what was expected, and light gray dots indicate a
45 lower residence time than was expected based on an even distribution.
46
Chapter 2 – Distribution and Egress - Figure Captions
49
Chapter 2 Figure 6. (A) The spatial distribution of mean 47 residence time (h) is shown for 75
48 tagged Blue Catfish at 12 receivers (18 receivers with four gate and two missing receivers
49 removed) in 2013. Each dot represents a receiver location. The size of the dot is proportional to
50 mean residence time. Also shown are the results of a Chi square analysis that identifies at which
51 receivers mean residence time was (B) higher than that expected or (C) less than expected based
52 on an even distribution (i.e., fish spent the same amount of time at all receivers). In B-C,
53 receiver numbers are indicated. On the map in A, dark gray dots indicate a higher residence time
54 than expected, white dots indicate residence times equal to what was expected, and light gray
55 dots indicate a lower residence time than was expected based on an even distribution.
56
57 Chapter 2 Figure 7. For 2012 and 2013, numbers of tagged Blue Catfish detected at the upper
58 and lower reservoir egresses are shown. To assess egress, we examined the outer gates first
59 (receivers 1, 20). If data were missing from receivers 1, 20, we next examined the inner gates,
60 receivers 2 and 19. In 2012, no fish were detected at receiver 1. In 2013, receiver 1 was
61 vandalized and five fish were last seen at receiver 2. The numbers on the right side of the plot
62 indicate numbers of fish last detected at receivers 1, 2, 19, 20 in 2012 and 2013. A dashed line
63 indicates that the receiver was not examined because the outer gate was in place. More details on
64 these five fish are provided in Figure 8. In both 2012, 2013, no fish were detected at receiver 20,
65 which remained intact throughout the study for both years.
66
67 Chapter 2 Figure 8. The detections of the five fish last seen at receiver 2 in 2013 are shown. The
68 X axis depicts the time period and the Y axis shows receiver number. Diamonds are detections
69 of individual fish. Receiver 2, at the top of each plot, is indicated with an arrow. Shown in A-E
Chapter 2 – Distribution and Egress - Figure Captions
50
are five individuals. These plots should be interpreted as fish m 70 ovements through time (left to
71 right) and from the lower to the upper reservoir (bottom to top). For example, fish 12 (panel A)
72 in July repeatedly traversed the upper and upper middle reservoir. (A) Fish 12 and (B) fish 56
73 were not detected because the study ended and receivers were removed. (C) Fish 62, (D) 67, and
74 (E) 75 exhibited extensive movements between receiver 2 and other receivers which is more
75 typical of resident rather than migratory movements.
76
77 Chapter 2 Figure 9. For 2012, box plots depicting monthly changes in mean residence time (h)
78 are shown for (A) receiver 2, (B) receiver 3, (C) receiver 4, (D) receiver 5, (E) receiver 6, (F)
79 receiver 9, (G) receiver 10, (H) receiver 12, (I) receiver 14, (J) receiver 15, (K) receiver 16, (L)
80 receiver 17, (M) receiver 18, and (N) receiver 19. Gate receivers 7, 8, 11, 13 were removed for
81 analysis to ensure a more evenly distributed tracking array. The X axis is month. The Y axis is
82 average residence time at a receiver for all fish detected at that receiver. Y axes are standardized
83 in order to compare trends across receiver locations. Also shown are the results of a Kruskal
84 Wallis nonparametric ANOVA that tested the effect of season. P<0.05 was considered
85 significant.
86
87 Chapter 2 Figure 10. For 2013, box plots depicting monthly changes in mean residence time (h)
88 are shown for (A) receiver 2, (B) receiver 3, (C) receiver 4, (D) receiver 5, (E) receiver 6, (F)
89 receiver 9, (G) receiver 10, (H) receiver 13, (I) receiver 14, (J) receiver 15, (K) receiver 18, and
90 (L) receiver 19. Gate (7, 8, 11, and 12) and missing (16, 17) receivers were removed for analysis
91 to ensure a more evenly distributed tracking array. The X axis is month. The Y axis is average
92 residence time at a receiver for all fish detected at that receiver. Y axes are standardized in order
Chapter 2 – Distribution and Egress - Figure Captions
51
to compare trends across receiver locations. Also shown are the results 93 of a Kruskal Wallis
94 nonparametric ANOVA that tested the effect of season. P<0.05 was considered significant.
95
96 Chapter 2 Figure 11. Movements (number, Y axis) by receiver (X axis) averaged across
97 individual fish shown by month. Data are means.
98
99 Chapter 2 Figure 12. Movements (number, Y axis) by receiver (X axis) averaged across
100 individual fish. Data are mean and standard deviation.
101
102 Chapter 2 Figure 13. For 2012, box plots depicting diel changes in mean residence time (h) are
103 shown for (A) receiver 2, (B) receiver 3, (C) receiver 4, (D) receiver 5, (E) receiver 6, (F)
104 receiver 9, (G) receiver 10, (H) receiver 12, (I) receiver 14, (J) receiver 15, (K) receiver 16, (L)
105 receiver 17, (M) receiver 18, and (N) receiver 19. Gate receivers 7, 8, 11, 13 were removed for
106 analysis to ensure a more evenly distributed tracking array. The X axis is dawn, day, dusk, and
107 night diel periods. The Y axis is average residence time per hour per receiver. Y axes are
108 standardized in order to compare trends across receiver locations. Also shown are the results of a
109 Kruskal Wallis nonparametric ANOVA that tested the effect of diel period. P<0.05 was
110 considered significant.
111
112 Chapter 2 Figure 14. For 2013, box plots depicting diel changes in mean residence time (h) are
113 shown for (A) receiver 2, (B) receiver 3, (C) receiver 4, (D) receiver 5, (E) receiver 6, (F)
114 receiver 9, (G) receiver 10, (H) receiver 13, (I) receiver 14, (J) receiver 15, (K) receiver 18, and
115 (L) receiver 19. Gate (7, 8, 11, and 12) and missing (16, 17) receivers were removed for analysis
Chapter 2 – Distribution and Egress - Figure Captions
52
to ensure a more evenly distributed tracking array. The X 116 axis is dawn, day, dusk, and night diel
117 periods. The Y axis is average residence time per hour per receiver. Y axes are standardized in
118 order to compare trends across receiver locations. Also shown are the results of a Kruskal Wallis
119 nonparametric ANOVA that tested the effect of season. P<0.05 was considered significant.
120
121 Chapter 2 Figure 15. Residence time (h) (A, C) and movements (number) (B, D) are shown by
122 fish size (TL mm) for 2012 (A, B) and 2013 (C, D). Data points are individual fish. For each
123 plot panel also shown are the results of a univariate regression including the regression line
124 equation, R2, and P values. P<0.05 was considered significant.
125
126 Chapter 2 Figure 16. Movements (number, Y axis) made by individual fish (X axis) averaged
127 across receiver numbers. Data are mean and standard deviation.
128
129 Chapter 2 Figure 17. For 2012, shown are the relationships between capture-release location and
130 residence time (h) for (A) receiver 2, (B) receiver 3, (C) receiver 4, (D) receiver 5, (E) receiver 6,
131 (F) receiver 9, (G) receiver 10, (H) receiver 12, (I) receiver 14, (J) receiver 15, (K) receiver 16,
132 (L) receiver 17, (M) receiver 18, and (N) receiver 19. The X axis is location: C=Causeway, M=
133 Madison, S=School. The Y axis is average residence time at a receiver for all fish detected at that
134 receiver. Y axes are standardized in order to compare trends across receiver locations. Also
135 shown are the results of a Kruskal Wallis nonparametric ANOVA that tested the effect of
136 location. P<0.05 was considered significant. The Causeway release site was near receiver 5, the
137 Madison release site was near receiver 9, and the School release site was near receiver 15 Data
138 are means +/1 1 SE.
Chapter 2 – Distribution and Egress - Figure Captions
53
139
Chapter 2 Figure 18. For 2013, shown are the relationships between 140 capture-release location
141 and residence time (h) for (A) receiver 2, (B) receiver 3, (C) receiver 4, (D) receiver 5, (E)
142 receiver 6, (F) receiver 9, (G) receiver 10, (H) receiver 13, (I) receiver 14, (J) receiver 15, (K)
143 receiver 18, and (L) receiver 19. The X axis is location: C=Causeway, M= Madison, S=School.
144 The Y axis is average residence time at a receiver for all fish detected at that receiver. Y axes
145 are standardized in order to compare trends across receiver locations. Also shown are the results
146 of a Kruskal Wallis nonparametric ANOVA that tested the effect of location. P<0.05 was
147 considered significant. The Causeway release site was near receiver 5, the Madison release site
148 was near receiver 9, and the School release site was near receiver 15 Data are means +/1 1 SE.
149
150 Chapter 2 Figure 19. This is the first of three syntheses of individual by-month cluster analyses
151 created to show general distribution patterns. Individual panels show the months of (A) July, (B)
152 August, (C) September, (D) October, and (E) November. On the right side of each panel is a
153 map of the reservoir with individual clusters (circles) indicating where fish from each cluster
154 were detected. Bars on the left side of each plot are residence times (h, X axis) at each receiver
155 (Y axis) for each cluster. Cluster circles and bars are indicated by different colors. Cluster
156 numbers within the circles are listed at the bottom of each panel as C1-C8 and correspond to
157 individual cluster numbers in the monthly cluster analysis figures that follow. Also shown for
158 each cluster are Jaccard bootstrap values (JB), and numbers of fish (N). (We know this is
159 challenging to look at but it is the only way to integrate the numerous cluster figures. We present
160 this first because we know the individual clusters are difficult to process). This panel of clusters
161 depicts fish that are seasonal movers.
Chapter 2 – Distribution and Egress - Figure Captions
54
162
Chapter 2 Figure 20. This is the second of three syntheses 163 of individual by-month cluster
164 analyses that show general distribution patterns. Individual panels show the months of (A) July,
165 (B) August, (C) September, (D) October, and (E) November. On the right side of each panel is a
166 map of the reservoir with individual clusters (circles) indicating where fish from each cluster
167 were detected. Bars on the left side of each plot are residence times (h, X axis) at each receiver
168 (Y axis) for each cluster. Cluster circles and bars are indicated by different colors. Cluster
169 numbers within the circles are listed at the bottom of each panel as C1-C8 and correspond to
170 individual cluster numbers in the monthly cluster analysis figures that follow. Also shown for
171 each cluster are Jaccard bootstrap values (JB), and numbers of fish (N). This panel of clusters
172 depicts fish that are not seasonal movers but remain in the upper middle funnel constriction.
173
174 Chapter 2 Figure 21. This is the last of three syntheses of individual by-month cluster analyses
175 that show general distribution patterns. Individual panels show the months of (A) July, (B)
176 August, (C) September, (D) October, and (E) November. On the right side of each panel is a
177 map of the reservoir with individual clusters (circles) indicating where fish from each cluster
178 were detected. Bars on the left side of each plot are residence times (h, X axis) at each receiver
179 (Y axis) for each cluster. Cluster circles and bars are indicated by different colors. Cluster
180 numbers within the circles are listed at the bottom of each panel as C1-C8 and correspond to
181 individual cluster numbers in the monthly cluster analyses that follow. Also shown for each
182 cluster are Jaccard bootstrap values (JB), and numbers of fish (N). This panel of clusters depicts
183 fish that are not seasonal movers but remain in the Madison Creek Area.
184
Chapter 2 – Distribution and Egress - Figure Captions
55
185
186
187 CHAPTER 2 APPENDIX
188 Chapter 2 Appendix Figure 1. Frequency of Blue Catfish in Milford Reservoir in 2012 for the
189 size range 100-1000 mm TL. Survey sizes are compared to the sizes of Blue Catfish tagged in
190 this study in 2012 and 2013.
191
192 Chapter 2 Appendix Figure 2. Hydrograph from USGS gage 06857100 downstream of Milford
193 Reservoir for March-November (A) 2012 and (B) 2013. Discharge and median for 47 years are
194 shown. July-November corresponds to our field season in both
195 years. http://nwis.waterdata.usgs.gov/ks/nwis/uv?cb_00065=on&cb_00060=on&format=gif_stat
196 s&site_no=06857100&period=&begin_date=2012-03-01&end_date=2012-11-03
197
198 Chapter 2 Appendix Figure 3. Shown is a silhouette plot identifying clusters based on residence
199 time (h) for the combined July-November time period. Identity and Jaccard bootstrap values for
200 all clusters are indicated. Appendix Figures 2-6 depict a single cluster analysis.
201
202 Chapter 2 Appendix Figure 4. For the clusters in the combined July-November time period,
203 shown are boxplots of residence times for receivers 2-5. The Y axis is residence time (h); the X
204 axis is cluster number. These data are means for all individual fish in each cluster.
205
Chapter 2 – Distribution and Egress - Figure Captions
56
Chapter 2 Appendix Figure 5. For the clusters in the combined J 206 uly-November time period,
207 shown are boxplots of residence times for receivers 6, 9, 10, 12. The Y axis is residence time
208 (h); the X axis is cluster number. These data are means for all individual fish within a cluster.
209
210 Chapter 2 Appendix Figure 6. For the clusters in the combined July-November time period,
211 shown are boxplots of residence times for receivers 14- 17. The Y axis is residence time (h); the
212 X axis is cluster number. These data are means for all individual fish within a cluster.
213
214 Chapter 2 Appendix Figure 7. For the clusters in the combined July-November, shown are
215 boxplots of residence times for receivers 18 and 19. The Y axis is residence time (h); the X axis
216 is cluster number. These data are means for all individual fish in a cluster.
217
218 Chapter 2 Appendix Figure 8. Shown is a silhouette plot identifying clusters based on residence
219 time (h) for July. Identity and Jaccard bootstrap values for all clusters are indicated. Appendix
220 Figures 7-11 depict a single cluster analysis.
221
222 Chapter 2 Appendix Figure 9. For the clusters in July, shown are boxplots of residence times for
223 receivers 2- 5. The Y axis is residence time (h); the X axis is cluster number. These data are
224 means for all individual fish in each cluster.
225
226 Chapter 2 Appendix Figure 10. For the clusters in July, shown are boxplots of residence times
227 for receivers 6, 9, 10, 12. The Y axis is residence time (h); the X axis is cluster number. These
228 data are means for all individual fish within a cluster.
Chapter 2 – Distribution and Egress - Figure Captions
57
229
Chapter 2 Appendix Figure 11. For the clusters in July, shown are 230 boxplots of residence times
231 for receivers 14- 17. The Y axis is residence time (h); the X axis is cluster number. These data
232 are means for all individual fish within a cluster.
233
234 Chapter 2 Appendix Figure 12. For the clusters in July, shown are boxplots of residence times
235 for receivers 18 and 19. The Y axis is residence time (h); the X axis is cluster number. These
236 data are means for all individual fish in a cluster.
237
238 Chapter 2 Appendix Figure 13. Shown is a silhouette plot identifying clusters based on
239 residence time for August. Identity and Jaccard bootstrap values for all clusters are indicated.
240 Appendix Figures 12-16 depict a single cluster analysis.
241
242 Chapter 2 Appendix Figure 14. For the clusters in August, shown are boxplots of residence
243 times for receivers 2- 5. The Y axis is residence time (h); the X axis is cluster number. These
244 data are means for all individual fish in each cluster.
245
246 Chapter 2 Appendix Figure 15. For the clusters in August, shown are boxplots of residence times
247 for receivers 6, 9, 10, 12. The Y axis is residence time (h); the X axis is cluster number. These
248 data are means for all individual fish within a cluster.
249
Chapter 2 – Distribution and Egress - Figure Captions
58
Chapter 2 Appendix Figure 16. For the clusters in August, s 250 hown are boxplots of residence
251 times for receivers 14- 17. The Y axis is residence time (h); the X axis is cluster number. These
252 data are means for all individual fish within a cluster.
253
254 Chapter 2 Appendix Figure 17. For the clusters in August, shown are boxplots of residence times
255 for receivers 18 and 19. The Y axis is residence time (h); the X axis is cluster number. These
256 data are means for all individual fish in a cluster.
257
258 Chapter 2 Appendix Figure 18. Shown is a silhouette plot identifying clusters based on residence
259 time for September. Identity and Jaccard bootstrap values for all clusters are indicated.
260 Appendix Figures 17-21 depict a single cluster analysis.
261
262 Chapter 2 Appendix Figure 19. For the clusters in September, shown are boxplots of residence
263 times for receivers 2- 5. The Y axis is residence time (h); the X a