Fish samples collected during different times of the year can be subject to various biases, but the influence of sampling during different seasons on population dynamics and yield metrics in large reservoirs is not well reported. This study compared the age structure, growth, mortality, and yield estimates of Channel Catfish Ictalurus punctatus and Walleye Sander vitreus collected during spring and fall with standardized gill netting in a large Nebraska, USA, reservoir. We sampled fish by using the Nebraska Game and Parks Commission standardized gill net survey methodology. We estimated ages from pectoral spines of Channel Catfish and from sagittal otoliths of Walleye, derived age-length keys, and compared mean spring and fall ages with t-tests. We compared spring and fall von Bertalanffy growth curves with likelihood-ratio tests and mortality estimates from weighted catch curves with an analysis of variance. We visually compared spring and fall yield estimates derived from yield-per-recruit models to assess the impact of variable population dynamics estimates. Estimates of mean age, growth coefficient, mean asymptotic length, total annual mortality, and yield of Channel Catfish did not differ between spring and fall. Conversely, older age structure of Walleye in spring resulted in lower estimates of total annual mortality and higher yield than in fall. Estimates of mean asymptotic length and growth coefficient differed between spring and fall for female Walleye, and mean asymptotic length, growth coefficient, and theoretical time at age 0 estimates varied between spring and fall for male Walleye. Fall yield estimates were substantially lower than spring estimates for both male and female Walleye. These results demonstrate that the sample collection season can impact population dynamics estimates for certain species, whereas others remain relatively unaffected.
Population dynamics estimates are important tools for managing sportfish populations. An understanding of age, growth, and mortality is essential for evaluating regulations (Holley et al. 2009; Schoenebeck and Brown 2011), implementing stocking strategies (Nate et al. 2000; Perrion et al. 2020), determining the effect of various regulations on harvest and return to angler via yield modeling (Doll et al. 2017; Schall and Lucchesi, in press), and making comparisons among populations (Hubert 1999; Quist et al. 2003; Schall et al. 2016). However, fish sampling is inherently biased, and the time of year samples are collected can influence sampling results. A multitude of fish metrics vary between seasons, including fish condition (Blackwell et al. 2000; Hansen and Nate 2005), size structure (Pope and Willis 1996; Fischer and Quist 2014; Schall et al. 2019a), catchability (Schoenebeck and Hansen 2005), catch rate (Schall et al. 2019a), and age structure (Hofmann and Fischer 2001). However, seasonal trends are not consistent among species.
A short sampling timeframe is common in population dynamics studies to minimize long-term influences and avoid changes in population rate functions (Hilling et al. 2016; Seibert et al. 2018; Schall et al. 2019b). However, short-timeframe studies can be biased by the seasonal differences in condition, size and age structure, and catch rates or by additional biotic and abiotic factors such as prey availability and consumption (Quist et al. 2002; Bowlby and Hoyle 2011), environmental effects (Gordoa et al. 2000) and angler harvest dynamics (Schoenebeck et al. 2010). Factors that can influence population dynamics require careful consideration to reduce bias within estimates. Previous studies indicate changes in population dynamics estimates can occur due to the influence of sampling gear (Colombo et al. 2008; Porreca et al. 2013) and aging structure (Colombo et al. 2010; Koenigs et al. 2013; Porta et al. 2018), but a greater understanding of seasonal sampling influence on population dynamics estimates is needed.
Two species that receive considerable management attention in the Great Plains are Channel Catfish Ictalurus punctatus and Walleye Sander vitreus (Bauer 2002; Quist et al. 2002). Standardized sampling of Channel Catfish and Walleye in large Great Plains reservoirs typically occurs in the fall (Zuerlein and Taylor 1985; Miranda and Boxrucker 2009; Koch et al. 2014). However, Channel Catfish sampling continues to require further study, including consistency between seasonal sampling events (Bodine et al. 2013), and Walleye population dynamics trends have been evaluated using spring samples in other areas (Kocovsky and Carline 2001; Pedersen et al. 2018). In addition, size structure of Channel Catfish and Walleye from fall samples is smaller than that from spring samples (Schall et al. 2019a), which could have implications for estimating population dynamics. Therefore, the similarity of spring and fall population dynamics estimates of Channel Catfish and Walleye in large reservoirs warrants further examination. The objective of this study was to quantify seasonal bias in estimates of age structure, growth, mortality, and yield for Channel Catfish and Walleye in a large Great Plains reservoir.
Lake McConaughy is an irrigation, flood control, and hydroelectric reservoir located on the North Platte River in western Nebraska, USA. At full pool, the reservoir covers 14,164 ha, has a mean depth of 22 m, a maximum depth of 53 m, and extends approximately 35 km (Taylor and Hams 1981). Introduction of Channel Catfish and Walleye into Lake McConaughy occurred shortly after construction of the reservoir (McCarraher et al. 1971), with supplemental stocking of Walleye since the early 1990s. Walleye is the species predominantly sought by anglers in Lake McConaughy (Porath et al. 2003), and anglers identified Channel Catfish as the second most sought species (Chizinski et al. 2014).
We collected Channel Catfish and Walleye in 2015 and 2016 by using monofilament gill nets measuring 45.7 m in length and 1.8 m in height. Gill nets were composed of six 7.6-m-long panels with bar mesh sizes arranged in the following order: 19.1, 25.4, 31.8, 38.1, 50.8, and 76.2 mm. We set gill nets perpendicular to shore with the small-mesh end set closest to shore at a standardized depth of 2.0–3.0 m and the large-mesh end farthest from shore in water depths ranging from 2.3 to 8.4 m. We deployed gill nets on the bottom of the reservoir near sunset and retrieved them beginning at sunrise the next morning.
Gill net sampling effort was distributed across the reservoir during each sampling season. We collected samples in spring 2015 (May) and fall 2015 (September) and fall 2016. We divided the reservoir longitudinally into three zones approximately 8 km in length, and we further divided each reservoir zone into 0.5-km subunits along the north and south shorelines. Gill net sampling occurred at randomly selected sites, without replacement each season, from the 0.5-km subunits on each shoreline in each reservoir zone. During each sampling season, we deployed a total of 36 gill nets evenly among the three reservoir zones such that 12 nets were set per zone each season, with 6 nets deployed on each shoreline (Figure 1).
We measured all captured fish for total length (TL; mm) and weight (g), and sex was determined by expression of gametes or observation of gonads on mortalities for Walleye (Data S1, Supplemental Material). Because of the low incidence of Channel Catfish mortalities, we collected pectoral spines from up to ten live individuals per centimeter length interval in spring 2015 and up to five individuals per centimeter length interval in fall 2016. We collected sagittal otoliths on Walleye mortalities up to ten per centimeter length interval in spring and fall 2015 and five per centimeter length interval in fall 2016. We did not collect aging structures in spring 2016. Channel Catfish pectoral spines have been validated for ages 1–4 y (Buckmeier et al. 2002), and back-calculation of pectoral spines and otoliths provides similar length-at-age estimates outside the validated age range (Michaletz et al. 2009). Koenigs et al. (2015) validated Walleye sagittal otoliths to at least age 10 y, and the estimates appear valid to older ages.
Age and growth
Pectoral spines were boiled to remove excess flesh, allowed to air dry for at least 1 wk, and cut at the distal end of the basal recess into 0.76-mm-thick sections by using a low-speed isomet saw. We cleaned otoliths and allowed them to dry for at least 1 wk. After drying, we mounted the otoliths in epoxy, transversely sectioned (0.76 mm) them, and sanded them. We placed aging structure sections in immersion oil and took images of otoliths and spines at ×40 and ×20 magnification, respectively, by using a Canon EOS Rebel T6i camera mounted to a Motic BA410E compound microscope. Two readers independently aged each structure and read in concert any structure with an age disagreement until a consensus age was agreed upon (Quist et al. 2012), or we removed the structure from analysis.
We compared the length distribution of live and dead Walleye to evaluate the influence of collecting age structures from mortalities. All Walleye less than 300 mm (except one individual in fall 2015) were dead when we retrieved the nets. We used Kolmogorov–Smirnov tests to compare TL distributions of live and dead Walleye greater than or equal to 300 mm in spring and fall. An a priori significance level of 0.05 was used for all statistical tests.
We developed age-length keys for each sample by using a semirandom age assignment within the FSA package in program R (Ogle et al. 2018), where the expected number of fish within a given length interval of a given age was assigned that age, with exceptions for fractionality when the observed distributions of ages within a length interval cannot be evenly applied to unaged fish (Isermann and Knight 2005; Data S1, Supplemental Materials). We used two-sample t-tests to determine whether mean ages derived from age-length keys differed between Walleye sampled in fall 2015 and fall 2016. We compared species-specific mean ages between spring and fall with two-sample t-tests using all available ages derived from age-length keys. We used Channel Catfish greater than or equal to 180 mm and Walleye greater than or equal to 200 mm to compare mean ages to avoid sizes not fully recruited to the gear based on selectivity curves for the smallest gill net mesh size deployed (Shoup and Ryswyk 2016).
We estimated instantaneous natural mortality (M) and conditional natural mortality (cm) from spring and fall samples for each species. We developed length–weight regressions of Channel Catfish and Walleye from spring and fall data to produce slope (b) and intercept (a) values for use in natural mortality estimation. We derived estimates of M and cm from spring and fall data by using seven natural mortality estimators and estimated mean values for use in yield-per-recruit (YPR) models (Hoenig 1983; Chen and Watanabe 1989; Djabali et al. 1993; Jensen 1996; Lorenzen 1996; Cubillos et al. 1999; Quinn and Deriso 1999; Slipke and Maceina 2014). Spring and fall von Bertalanffy growth parameter estimates, maximum observed ages, proportion surviving to maximum age, and the estimated maximum theoretical weight are input variables necessary to calculate M and cm by using the Fishery Analysis and Modeling Simulator software (Slipke and Maceina 2014).
We assessed the effect of seasonal population dynamics estimates on estimated yield by using a YPR model in Fishery Analysis and Modeling Simulator software. We used spring and fall von Bertalanffy growth parameter estimates, maximum observed ages, slope and intercept values from length–weight regressions, and cm estimates as input parameters for a YPR model beginning with 1,000 recruits. We used a minimum length limit model for both species, with the minimum length set to 381 mm for Walleye based on the current regulation in Nebraska and 305 mm for Channel Catfish based on the minimum suggested harvestable size for other North American freshwater catfish species (Travnichek 2011). We estimated yield (kg) per 1,000 recruits across a range of conditional fishing mortality (cf) values (0.00–0.95).
We collected a total of 498 and 202 individuals in spring 2015 and fall 2016, respectively. We collected pectoral spines from 104 individuals in the spring and 97 in the fall and assigned ages to a total of 494 fish in the spring and 199 fish in the fall by using age-length keys. Ages ranged from 3 to 18 y in spring and from 1 to 17 y in fall, and the maximum observed length was 769 mm in spring and 724 mm in fall.
The mean age estimates differed between seasons, but growth estimates were similar between spring and fall samples. Mean age was greater in the spring than in the fall (t = 7.265, P < 0.001), despite a similar maximum age between seasons (Table 1). Estimates of L∞ and K were similar between seasons, but t0 was lower in spring than in fall (χ2 = 26.807, P < 0.001; Figure 2).
We derived A estimates by using ages 6–18 y in the spring (n = 423) and ages 5–17 y in the fall (n = 143). Spring and fall A estimates were similar (F = 0.004, P = 0.951). We estimated total annual mortality to be 29% (95% confidence interval [CI]: 12–43%) in spring and 30% (95% CI: 19–39%) in fall (Figure 3). Estimates of M (mean ± SE) were 0.169 ± 0.037 in the spring and 0.155 ± 0.027 in the fall.
Yield was similar between spring and fall models. Mean cm estimates used to model yield were nearly identical between spring and fall (Table 2). Maximum yield was at a cf of 0.21 during both seasons, producing an estimated yield of 182 kg in spring and 176 kg in fall (Figure 4).
We collected a total of 330, 639, and 710 individuals in spring 2015, fall 2015, and fall 2016, respectively. We aged otoliths from 213 fish in spring, 293 fish in fall 2015, and 173 fish in fall 2016 (466 individuals total in fall) and assigned ages by using age-length keys to 329 fish in the spring, 634 fish in fall 2015, and 702 fish in fall 2016 (1,336 total in the fall). Ages ranged from 1 to 18 y in spring and from 0 to 10 y in fall, and the maximum observed length was 753 mm in spring and 676 mm in fall. Of the 330 individuals collected in spring 2015, 95 (29%) were alive and 89 (7%) of 1,349 collected in the fall were alive. The length frequency distributions of live and dead fish greater than or equal to 300 mm were similar in spring (D = 0.09, P = 0.585) and fall (D = 0.139, P = 0.088).
The mean age and von Bertalanffy growth parameter estimates differed between the spring and fall. Mean age was similar between 2015 and 2016 fall samples (t = 0.504, P = 0.615); thus, we pooled fall age data to compare mean ages against spring data. Mean age was nearly three times higher in the spring than in the fall (t = 19.483, P < 0.001; Table 1). The maximum observed age was nearly twice as high in spring (18 y) than in the fall (10 y). However, seasonal differences in maximum observed age were more pronounced in males (spring = 18 y, fall = 6 y) than females (spring = 13 y, fall = 10 y; Table 2). Von Bertalanffy growth analysis included a total of 110 males and 60 females in spring and 159 males and 219 females in fall. Spring and fall estimates of L∞ and K differed for females (χ2 = 34.882, P < 0.001; Figure 2), and all three growth parameters (L∞, K, and t0) differed between spring and fall for males (χ2 = 31.906, P < 0.001; Figure 2).
The A estimates also varied by season due to the truncated age structure in the fall samples. We used ages 2–18 y in the spring mortality analysis (n = 293) and ages 1–10 y in the fall mortality analysis (n = 1,217). The weighted spring A estimate of 20% (95% CI: 8–31%) was significantly lower than the fall A estimate of 53% (95% CI: 39–64%; F = 14.283, P = 0.001; Figure 3). The M estimates (mean ± SE) of females were 0.324 ± 0.012 in the spring and 0.450 ± 0.040 in the fall. The M estimates of males were 0.342 ± 0.028 in the spring and 0.655 ± 0.061 in the fall.
Yield estimates were substantially higher in the spring than the fall for both sexes (Figure 4). Estimated cm values used for the YPR model input were higher in the fall than in the spring for both females and males (Table 2). We estimated maximum spring yield of females to be 291 kg at a cf of 0.53, but yield only reached a maximum of 186 kg in the fall assuming a cf of 0.87. For males, yield reached a maximum of 222 kg in the spring at a cf of 0.77 and 104 at a cf of 0.95 in the fall.
Population dynamics estimates of Channel Catfish were generally similar between spring and fall samples. Differences in mean age and t0 did not have meaningful impact on mortality or growth analysis, and t0 is a modeling artifact of the von Bertalanffy growth equation with little biological relevance (Ogle 2016). The similarity in population dynamics estimates can be attributed to similar size and age structure collected in spring and fall, which may be a result of thermal preference and inshore nocturnal movements to forage. Mean littoral water temperatures were below optimal temperatures for Channel Catfish growth in spring (11.4°C in 2015 and 11.7°C in 2016) and fall (21.7°C in 2015 and 20.1°C in 2016) at Lake McConaughy (McMahon and Terrell 1982; Schall 2016), and historic temperature profiles indicate temperature rapidly declines with depth (McCarraher et al. 1971). Channel Catfish are known to occupy a broad range of depths in some reservoirs (Hubert and O'Shea 1992) and more inshore distributions in others (Edds et al. 2002), but warmer nearshore temperatures may have been preferred to deep offshore locations. Channel Catfish are also known to move inshore at night to forage (McMahon and Terrell 1982), likely making them susceptible to littoral gill nets. Because spawning migrations occur in late spring or early summer (McCarraher et al. 1971; Duncan and Myers 1978) and large movements away from the littoral zone are not known to occur, the impact to spring or fall population dynamics estimates is likely to be limited.
Conversely, age structure, growth, and mortality estimates varied between seasons for Walleye, likely as a result of seasonal migratory behavior. The seasonal variability in growth and mortality rates could be attributed to the lack of old, large individuals (>600 mm) in the fall samples compared with the spring samples, where Walleye greater than 760 mm were observed. The spring total annual mortality estimate for Walleye was similar to those of populations experiencing low exploitation (Kocovsky and Carline 2001; Koenigs et al. 2013). Adult Walleye move into shallow littoral areas to spawn in the spring (Colby et al. 1979), which may have overrepresented old age classes (Schoenebeck and Hansen 2005). The lack of older individuals in fall samples resulted in estimates similar to highly exploited systems (Quist et al. 2004, 2010), and limiting analysis to fall samples may lead to estimating higher mortality than actually occurred. In the summer and fall, large adult Walleye move offshore into deeper cool water to presumably avoid high littoral water temperatures (Johnson 1969; Colby et al. 1979; Mero and Willis 1992) and seek concentrated pelagic prey species (Rawson 1957; Wang et al. 2007; Bowlby and Hoyle 2011), thereby reducing the likelihood that they would be captured in littoral sampling gear. Fall littoral water temperatures at Lake McConaughy were within the preferred range of 20–23°C (Schall 2016), but higher water temperatures in summer may have caused Walleye to move offshore before the fall sampling timeframe. Also, Alewife Alosa pseudoharengus and Gizzard Shad Dorosoma cepedianum are among the most commonly consumed prey in Lake McConaughy (Porath and Peters 1997; Perrion 2016), and the lack of large Walleye in fall samples at Lake McConaughy may reflect movement offshore to follow schools of these pelagic fish. Additional research is needed to identify sampling timeframes and methods that most accurately reflect the true population dynamics of the Walleye population in large reservoirs and to investigate the effects of how additional factors, such as sex- and size-specific Walleye migration timelines (Rawson 1957; Johnson 1969), impact variability in population dynamics estimates. We postulate similar movements of other species, such as off-shore movements in search of thermal refuge or to use pelagic prey, may also result in seasonally underrepresented samples in traditional near-shore sampling gear and therefore biased representation of the whole population.
The variability in seasonal Walleye population dynamics estimates can result in different management implications. In addition to estimating yield, YPR models are used to evaluate the effects of regulations and the impact of varying levels of exploitation on fish populations (Holley et al. 2009; Quist et al. 2010; Eder et al. 2016; Stewart et al. 2016; Doll et al. 2017). The YPR models rely on accurate estimates of population dynamics parameters (Slipke and Maceina 2014), and differences in input parameters can produce results with contrasting management interpretations. Yield estimates for Walleye have been shown to improve with increasingly restrictive regulations in heavily exploited systems (Quist et al. 2010). Walleye yield estimates were lower in the fall at Lake McConaughy due to the lack of large, older individuals in fall samples, and interpretation of YPR modeling by using only the fall sampling data may provide support for regulations that are not necessary. Sensitivity of modeled results to seasonal biases from populations that exhibit highly variable seasonal behavior could be considered before implementation of management actions.
Results from this study suggest that species-specific limitations of population dynamics estimates derived from data collected during standardized sampling timeframes are important to consider. Standardized sampling of Channel Catfish and Walleye in Great Plains reservoirs typically occurs in the fall (Zuerlein and Taylor 1985; Miranda and Boxrucker 2009; Koch et al. 2014), and this study demonstrated that population dynamics estimates derived from spring and fall Channel Catfish gill net samples were consistent and should allow some flexibility in timing of Channel Catfish population dynamics assessments. Conversely, greater consideration of the limitations of using fall samples to estimate Walleye population dynamics is warranted. Reduced catchability of large Walleye by using littoral gill nets sets in the fall should be expected based on the previously observed movement patterns. Although standardized spring fyke netting is used by some fisheries agencies to collect adult Walleye in large lentic systems (Miranda and Boxrucker 2009; Pedersen et al. 2018), the accuracy of spring-collected population dynamics data in large Great Plains reservoirs requires further study. Therefore, special consideration may be important when estimating population dynamics from species such as Walleye that exhibit variable seasonal movements, because estimates derived using standardized sampling timeframes could be misleading.
Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.
Data S1. Channel Catfish Ictalurus punctatus (CCF) and Walleye Sander vitreus (WAE) sampling data, including total length (mm), weight (g), sex (F = female, M = male, and U = unknown), year and month sampled, and age (y). Back-calculated total length at age data for all aged fish are included, but growth estimates derived with the von Bertalanffy growth curve were limited to the most recent fully formed annulus.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S1 (79 KB CSV).
Reference S1. Chizinski CJ, Martin DR, Pope KL. 2014. Angler behavior in response to management actions on Nebraska reservoirs. Lincoln, Nebraska: Nebraska Cooperative Fish and Wildlife Research Unit and Nebraska Game and Parks Commission. Federal Aid in Sportfish Restoration Performance Report Project F-182-R.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S2 (1.56 MB PDF).
Reference S2. Colby PJ, McNicol RE, Ryder RA. 1979. Synopsis of biological data on the Walleye Stizostedion v. vitreum (Mitchill 1818). Food and Agriculture Organization of the United Nations Fisheries Synopsis 119.
Reference S3. McMahon TE, Terrell JW. 1982. Habitat suitability index models: Channel catfish. Washington, D.C.: U.S. Department of the Interior, Fish and Wildlife Service. Report FWS/OBS-82/l0.2.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S4 (2.25 MB PDF); also available at https://pubs.er.usgs.gov/publication/fwsobs82_10_2.
Reference S4. Perrion MA. 2016. Early life history characteristics of juvenile fishes in Lake McConaughy, Nebraska: an assessment of stock contribution and food habits. Master's thesis. Kearney, Nebraska: University of Nebraska at Kearney.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S5 (1.94 MB PDF).
Reference S5. Schall BJ. 2016. Spatial distribution of fishes and population dynamics of sportfish in Lake McConaughy, Nebraska. Master's thesis. Kearney, Nebraska: University of Nebraska at Kearney.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S6 (4.68 MB PDF).
Reference S6. Taylor MW, Hams KM. 1981. The physical and chemical limnology of Lake McConaughy with reference to fisheries management. Lincoln: Nebraska Game and Parks Commission. Nebraska Technical Series No. 9.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S7 (2.8 MB PDF); also available at https://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1032&context=nebgamepub.
Reference S7. Zuerlein GJ, Taylor MW. 1985. Standard survey guidelines for sampling lake fishery resources. Lincoln: Nebraska Game and Parks Commission.
Found at DOI: https://doi.org/10.3996/JFWM-20-027.S8 (4.91 MB PDF).
We thank Matthew Perrion, Josh Kreitman, Cale Hadan, Tyler Jackson, Mark Staab, Nicole Pauley, and volunteers from the University of Nebraska–Kearney and the Nebraska Game and Park Commission for field sampling and laboratory processing assistance. We also thank the Associate Editor and two anonymous reviewers for their feedback that greatly improved this manuscript. This project was funded by Federal aid in Sport Fish Restoration funds (Project F-196-R) administered through the Nebraska Game and Parks Commission and the University of Nebraska–Kearney.
Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.
The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.
Citation: Schall BJ, Schoenebeck CW, Koupal KD. 2021. Seasonal sampling influence on population dynamics and yield of Channel Catfish and Walleye in a large Great Plains reservoir. Journal of Fish and Wildlife Management 12(1):223–233; e1944-687X. https://doi.org/10.3996/JFWM-20-027