EVALUATING WALLEYE (SANDER VITREUS) THERMAL AND OPTICAL HABITAT OCCUPANCY IN NORTHERN WISCONSIN LAKES USING TWO FORMS OF TECHNOLOGY

File(s)
Date
2024-04Author
Vasquez, Benjamin
Publisher
College of Natural Resources, University of Wisconsin-Stevens Point
Advisor(s)
Isermann, Daniel A.
Metadata
Show full item recordAbstract
Previous research suggests that walleye (Sander vitreus) fishery production is related to lake temperature and clarity. Based on this understanding, thermal (11–25°C) and optical (8–68 lumens/m2, or lux) conditions presumed to be preferred by walleye were combined into lake area estimates of thermal-optical habitat area (TOHA; Lester et al. 2004), which served as a conceptualization of optimal habitat for walleye growth. There is substantial evidence that TOHA has decreased in some northern Midwest lakes, and that trends in walleye production and yield may be related to these shifts. However, no rigorous evaluation of walleye thermal-optical habitat occupancy exists and therefore it is unclear if walleye occupy optimal habitat as defined in the conceptual model. To assess the validity of TOHA, I tagged walleye with acoustic transmitters and archival tags that recorded high-frequency temperature and depth data. My objectives were to: 1) determine if walleye thermal and optical habitat occupancy varies among three northern Wisconsin lakes in relation to season and total length (TL); 2) determine if thermal and optical habitat boundaries employed in previous research are occupied by walleye; 3) redefine optimal walleye habitat using direct measurements of temperature, light, and dissolved oxygen (DO) occupancy; and 4) determine if data resolution and cost-effectiveness differs between two tags used to monitor walleye temperature occupancy.
In 2022, I surgically implanted acoustic transmitters and archival tags into walleye collected from Escanaba (n = 84), McDermott (n = 16; archival tags only), and Sparkling (n = 64) lakes and distributed these tags across three TL categories: small, 310–380 mm (n = 54); medium, 381–456 mm (n = 41); and large, ≥ 457 mm (n = 69). Each lake varied in regards to walleye recruitment status, stocking, morphometry, and temperature profiles. Most walleye were tagged during May 2022 during adult population surveys using fyke netting and nearshore night electrofishing. Tags were inserted into the body cavity of walleye ≥ 310 mm with additional external tags to help identify tagged fish if recaptured or harvested. After initial tagging, acoustic receivers were deployed in Escanaba (n = 8) and Sparkling (n = 5) lakes to record observations from acoustic transmitters. During 2022, I opportunistically tagged additional walleye in June (McDermott Lake, n = 5) and September (Escanaba Lake, n = 7) to replace fish that were harvested by recreational anglers or tribal spearers from Escanaba Lake.
Temperature and light profiles used to characterize thermal and optical habitat were measured with strands of HOBO loggers deployed in the pelagic zone of each lake. These loggers recorded temperature and light intensity values every 10 min throughout the study. Weekly limnological measurements of water clarity and dissolved oxygen were also collected to complement HOBO logger data. I used these pelagic strands of loggers to determine the light experienced by a walleye at a specific depth through linear interpolation with walleye depth values. Additionally, walleye DO concentration at depth was determined with weekly profiles of dissolved oxygen and linear interpolation. As a post-hoc supplement to HOBO loggers in the pelagic zone of each lake, HOBO loggers were deployed in the littoral and pelagic zone of each lake during 2023 to address potential differences in thermal and optical conditions between these zones.
To address my objectives, I first tested for differences in temperature, depth, and light values of walleye across lakes and walleye TL categories using generalized additive mixed-effects models (GAMMs). I also determined if observed walleye temperature and light values aligned with thermal and optical habitat model ranges used in previous evaluations of TOHA by calculating occupancy within thermal and optical habitat when those habitats were available in each lake. Occupancy was calculated as the percent of observations that occurred within each habitat range divided by the total number of observations. Additionally, occupancy rates were calculated by month (June, July, August, September, October) and time of day (dawn, day, dusk, night) to assess if occupancy varied in relation to these temporal categorizations. As an additional analysis, I calculated proportional benthic area (used in previous TOHA models) and average depth of thermal and optical habitat to examine trends in habitat availability across lakes. To redefine optimal walleye habitat, I calculated summary statistics, ecological niches, and habitat selection ratios for all walleye temperature, light, and DO observations. Last, to evaluate the cost efficiency and data resolution of both tag types deployed in Escanaba and Sparkling lakes, I created a cost-benefit analysis that included: 1) the number of individual fish providing data relative to number of tags released; 2) the total number of temperature observations obtained from each fish by tag type; 3) the average time between observations, 4) the total effort expended to collect and recover data, and 5) the total cost including effort-based costs. Data resolution metrics were only calculated for temperature sensors present in each tag type to reduce observation redundancy.
In 2023, walleye were recaptured during adult population estimates with fyke nets and electrofishing. In McDermott and Sparkling lakes, walleye with archival tags had tags surgically removed and were released alive. Walleye recaptured in Sparkling Lake with acoustic transmitters were released. All internally tagged walleye encountered on Escanaba Lake were euthanized for an ongoing exploitation experiment. Acoustic receiver detections were offloaded after walleye population estimates concluded. During this study, 176 individual walleye were tagged and viable data were recovered from 88 of these fish (50%). Analyses were conducted using dates ranging from 01 June to 31 October 2022 (i.e., days when thermal and optical habitat existed across study lakes). During this period, temperature and light profiles varied across lakes. Additionally, when examining data from summer 2023, I found that temperatures in the littoral zone of each lake were comparable to those recorded in the pelagic zone but the pelagic zone was much brighter than the littoral zone.
GAMMs predicted that daily walleye temperature occupancy varied by lake and TL category. Walleye in Escanaba Lake (x̄ = 19.4°C; N = 4,679) occupied slightly cooler water than McDermott Lake (x̄ = 19.6°C; N = 739; p < 0.01), but walleye in Escanaba Lake and McDermott Lake occupied warmer water than Sparkling Lake (x̄ = 18.2°C; N = 4,485; p < 0.01). Additionally, small (x̄ = 19.2°C; N = 2,843) and medium (x̄ = 19.4°C; N = 2,723) walleye used similar temperatures but large walleye were in slightly cooler water than small walleye (x̄ = 18.4°C; N = 4,337; p = 0.02). Daily averages predicted that walleye in Sparkling Lake (x̄ = 4.59 m; N = 4,434) occupied deeper water than walleye in Escanaba (x̄ = 2.01 m; N = 4,355) and McDermott lakes (x̄ = 2.44 m; N = 739, p < 0.01). In contrast to the depth results, mean daily light (excluding night and square-root transformed for residuals to meet Gaussian distribution) predicted that lake and TL category were not sufficient at describing interpolated walleye light occupancy. When evaluating occupancy in thermal habitat, I found that walleye were generally within 11–25°C (95.6% of observations) as long as it was available. In stark contrast, I found that walleye were rarely within optical habitat (6.4% of observations) and most observations occurred at thousands of lux higher than the optical habitat range (8–68 lux). Occupancy in 8–68 lux only increased when optical habitat availability increased in shallower water (i.e., during dawn and dusk). The metrics I used to redefine the thermal boundary of TOHA suggested that 20–23°C are the most preferred temperatures when available. However, these metrics were unable provide agreement on a preferred light range for light values; regardless, all metrics suggested that 8–68 lux was too limited of a range to serve as a landscape-scale characterization of preferred walleye habitat. To improve the TOHA framework, I also calculated DO occupancy rates and found that concentrations ≥ 7 mg/L were preferred. My tag cost-benefit analysis suggested that archival tags collected more data at a lower cost than acoustic transmitters, largely because of high recapture rates across all three study lakes.
These results suggest that walleye ≥ 310 mm prefer similar temperatures and dissolved oxygen concentrations among lakes. Because I found no evidence of selection for a certain light range, I suggest removing light from estimates of TOHA and instead using Secchi depth (1–3 m Secchi depth, reported across multiple studies as being optimal for walleye populations) as a landscape-scale predictor. Additionally, truncating thermal habitat from 11–25°C to 20–23°C and adding dissolved oxygen concentrations ≥ 7 mg/L may explain variation in walleye production or abundance better than previous TOHA models. Furthermore, my tag analysis suggested that archival tags and acoustic transmitters are both adequate at collecting temperature data on fish in inland lakes, but cost efficiency is highly dependent on fish recovery rate. Improvement of the TOHA conceptual model could aid walleye management, and my cost-benefit analysis of tag types may help agency personnel choose a tagging technology for future research projects.
Subject
acoustic telemetry
archival tag
dissolved oxygen
temperature
TOHA
walleye
Permanent Link
http://digital.library.wisc.edu/1793/85172Type
Thesis