Centrarchid population responses to intensive removal in a northern Wisconsin lake with an assessment of otolith-based back-calculation methods for estimating growth
File(s)
Date
2024-08Author
Henningsen, Rebecca Marie
Publisher
College of Natural Resources, University of Wisconsin-Stevens Point
Advisor(s)
Isermann, Daniel A.
Metadata
Show full item recordAbstract
Several species in the Centrarchidae family support popular recreational fisheries within their respective native ranges in North America. Once established, members of the Centrarchidae family (referred to as centrarchids herein) are prolific, which can lead to “stunting.” Stunting is a widespread condition characterized by slow growth due to density dependence, with few fish reaching sizes desirable to anglers. In addition, climate patterns in the northern USA and Canada over the past several decades have been more favorable to warmwater species, including centrarchids. Increased centrarchid abundance could influence abundance of other species through predation or competition for prey resources. Centrarchids introduced outside of their native ranges can produce negative effects through these same mechanisms.
Whether the management concern is stunting or interactions with other species, fishery managers often address these concerns using direct or indirect methods of centrarchid removal. However, little is known regarding what level of removal is sufficient to elicit population-level responses. This study focused on McDermott Lake, a northern Wisconsin lake with a history of stunted centrarchid populations, to gain further insights in these areas. Approximately 280,000 centrarchids were removed from the lake in 2018-2021, which provided a unique opportunity to examine how centrarchid populations respond to persistent large-scale removal. Therefore, the objective of Chapter 1 of this study was to determine the effects of the intensive removal on the population dynamics and demographics of black crappie Pomoxis nigromaculatus, bluegill Lepomis macrochirus, and largemouth bass Micropterus salmoides relative to observations from a nearby reference lake.
McDermott Lake and the reference lake, Sandy Beach Lake are both located in Iron County in northern Wisconsin. Sampling began on both lakes in April 2017 and continued through August 2023. All centrarchids collected during 2018-2021 were removed from McDermott Lake. Sampling began immediately after ice-out with the deployment of fyke nets Fish were collected via AC boat electrofishing during May and June and with mini-fyke nets and cloverleaf traps during July and August. All centrarchids collected in these gears were counted and subsamples of each species were measured (total length [TL] in mm) and length-stratified subsamples of measured fish were sacrificed for otolith removal and age estimation.
Trends in relative abundance were assessed using catch-per-unit effort (CPUE) from fyke netting (fish/net night) for black crappie and bluegill and from electrofishing (fish/h) for bluegill and largemouth bass. Catch-per-unit effort of age-0 and age-1 fish in August cloverleaf traps were used to index bluegill recruitment and CPUE of age-0 fish in mini-fyke nets was used to index largemouth bass recruitment. Size structure was described using mean TL (mm) and Proportional Size Distribution (PSD). Mean ages were used to assess trends in age structure and mean TLs at ages 3 and 5 were used to describe growth trends. Body condition was evaluated using mean relative weight (Wr). Two-way t-tests (α= 0.10) were used to determine the difference between metric estimates in pre-removal years (2017-2018) versus post-removal years (2019-2023) within a single lake. When possible, an analysis of variance (α= 0.10) using a before-after-control-impact (BACI) framework was used to test for differences in annual estimates of each metric in pre-removal years versus post-removal years between lakes.
Relative abundance of black crappie declined significantly in McDermott Lake after the removal, but the other metrics showed limited responses. There was no clear evidence the removal affected the bluegill population in McDermott Lake. Additionally, significant improvements in PSD-Q and Wr of stock-length bluegill were observed in the control lake but not in the treatment lake. The number of largemouth bass removed from McDermott declined over time, suggesting declines in abundance. However, relative abundance of largemouth bass > 200 mm did not significantly change. While not significant, substantial increases in relative abundance of age-0 largemouth bass and of bass < 200 mm TL were observed. This suggests a compensatory recruitment response to the removal may have occurred in McDermott Lake which was further supported by significant declines in largemouth bass mean TL, PSD, PSD-P, and mean age.
Given the observed responses of the centrarchid populations within McDermott Lake relative to sampling effort expended to remove fish, it is unlikely manual removal is a viable management tool for lakes of this type where centrarchid populations are established and prolific. Manual removal may be a useful, although inefficient, tool to reduce black crappie abundances. However, my study suggests this strategy would not be effective for reducing abundance of species with more consistent reproductive success, such as largemouth bass and bluegill. Furthermore, even when a reduction in abundance of a centrarchid species is achieved, this change may not always result in improvements in growth and size structure.
Chapter 2 of this study focused on using otolith-based back calculation to estimate previous mean lengths at age of centrarchids, which would provide a means to supplement pre-treatment growth data available in before-and-after comparison studies. In terms of both accuracy and precision, otoliths provide the best approach for estimating the age of centrarchids; however, little research has been done on otolith-based back calculation for these species. Therefore, my objectives for Chapter 2 were to: 1) assess the validity of otolith-based back-calculations for estimating previously observed mean lengths at age for specific cohorts of black crappie, bluegill, and largemouth bass; 2) determine the necessity of incorporating biological intercepts into otolith-based back-calculation methods for largemouth bass; and 3) to determine the extent of the variation in back-calculated TL at age of black crappie, bluegill, and largemouth bass when computed using otolith radius and annuli transect measurements made by two independent observers. The fish used in used in my Chapter 2 analyses were collected from McDermott Lake and Sandy Beach Lake and assigned ages using the methods previously described for Chapter 1. Back-calculations were completed using sectioned sagittal otoliths. Back-calculated estimates of TL at age were calculated using the Dahl-Lea formula. All back-calculations were completed in using the RFishBC package in the programming language R.
To address my first objective, mean back-calculated TLs at age were compared against observed mean TLs at age for specific cohorts within each species using t-tests ( = 0.05). For my second objective, biological intercepts were determined using post-larval largemouth bass. Paired t-tests ( = 0.05) were used to compare mean back-calculated TLs at age of largemouth bass calculated using biological intercepts to those back-calculated for the same fish using the standard Dahl-Lea formula. Additionally, intercept-corrected mean back-calculated TLs at age of largemouth bass were compared to observed mean TLs at age for cohorts of bass using t-tests. Finally, for my third objective, back-calculated TLs determined by two independent readers were compared for each species at ages 1 through 5 using paired t-tests ( = 0.05).
A total of 31 pairings were used for comparisons between back-calculated and observed mean TL for the three centrarchid species. Back-calculated TLs at age tended to overestimate observed TLs at age by 1 to 20 mm, with an average of 6 mm. Only 10 observed differences were ≥ 10 mm and 20 of 31 comparisons indicated no significant difference between observed and back-calculated mean TLs at age. The difference between mean TLs for largemouth bass back-calculated using the two approaches (biological intercept vs. no intercept) was significantly different from zero for all seven comparisons. However, absolute differences in mean TL at age between the two back-calculation methods ranged from 1 to 3 mm. Intercept-corrected estimates more accurately reflected observed estimates for five of the seven comparisons; however, the increase in accuracy was slight (mean improvement = 3 mm). Between-reader differences in individual back-calculated TLs at age ranged from 0 to 56 mm (mean = 6 mm) among species and ages and 17% of the observed differences were ≥ 10 mm. Mean back-calculated TL only differed significantly between observers for age 1 black crappie. Despite the lack of differences, for 12 of the 15 comparisons, mean back-calculated TLs at age for reader 1 were lower than those for reader 2. The largest discrepancies between readers occurred for largemouth bass at all ages.
Despite occasional differences between back-calculated and observed TL distributions, back-calculated mean total lengths at age were generally similar to observed values for black crappie, bluegill, and largemouth bass in McDermott Lake and Sandy Beach Lake and most of differences were < 10 mm. As such, the discrepancies between back-calculated and observed values were small enough that they are unlikely to substantially alter the conclusions of growth-based analyses or influence management decisions. Similarly, the differences in back-calculated TLs between intercept-corrected and no-intercept models (all <5 mm) would likely be considered negligible within a management context. While the magnitude of these discrepancies between readers was generally not large enough to cause significant differences in our study, the consistent relative over-measuring indicates it is possible for reader identity to produce biased back-calculated lengths. Overall, this study demonstrates that otolith-based back-calculation can provide accurate estimates of observed mean TLs at age for black crappie, bluegill, and largemouth bass growth and may not require use of biological intercepts.
Subject
Back-calculation
Bass
Centrarchidae
Fish Removal
Otolith
Panfish
Permanent Link
http://digital.library.wisc.edu/1793/85743Type
Thesis