Knowledge of the distribution of ages of fish within a stock, and subsequently individual growth rates, allows managers the ability to calculate key metrics (i.e., recruitment, mortality, and stock growth rate) that greatly improve stock assessment models. Two remnant stocks of Lake Sturgeon Acipenser fulvescens exist near and within the Niagara River, one primarily occupying the headwaters and the other primarily occupying the mouth. Though initial efforts in the late 1990s collected data on the lower Niagara River stock, a long-term comprehensive examination of age and growth is lacking and the age structure of the stock found at the headwaters has yet to be formally described to our knowledge. To ascertain the current age structure of these two stocks we sampled Lake Sturgeon in the lower Niagara River and at the headwaters of the Niagara River between 2012 and 2017 and took a portion of the leading pectoral fin spine of captured Lake Sturgeon for age estimation. Ages ranged between 4 and 42 y, with females generally being older and larger than males. The median age appeared to increase from 14 to 18 y throughout our study in both stocks. Lengths at age of both stocks were larger than those reported in other systems and growth rates appear to have increased over the past decade in the lower Niagara River. Despite efforts to improve age estimation accuracy, age estimates from fish whose ages were partly known (derived from multiple age estimates from fish that were captured multiple times) demonstrated that assigned ages may have greater error than expected. Additionally, a lack of young individuals confounded growth analyses. Although there was uncertainty in the assigned ages, this study still provides evidence of consistent recruitment in both stocks and, to our knowledge, the first characterization of the age structure of the Lake Sturgeon stock occupying the headwaters of the Niagara River.
Lake Sturgeon Acipenser fulvescens are an ancient, keystone species that have undergone dramatic declines throughout their distribution (Tody 1974; Auer 1999). Since their decline throughout the Laurentian Great Lakes in the late 19th and early 20th centuries, Lake Sturgeon stock status and recovery have been difficult to assess because of their rarity and biology. As the largest and longest-lived freshwater fish species endemic to North America (Scott and Crossman 1973), Lake Sturgeon exhibit late age at maturity (Scott and Crossman 1973; Winemiller and Rose 1992), slow growth, and skip-spawning, which predisposes them to overharvest and prolonged recovery rates (Haxton et al. 2014; Sweka et al. 2018). Like other long-lived, cartilaginous species, assigning accurate ages to Lake Sturgeon is difficult. However, age estimates of individuals within a stock greatly improve stock assessment models. As the age structure of a stock is identified, researchers and managers can begin to quantify key metrics, such as recruitment, individual growth rate, and total mortality, and how these metrics vary temporally.
Since ages of wild Lake Sturgeon are not known, researchers and managers rely on osseochronometry, or the interpretation of fast and slow calcium carbonate deposition on calcified accretions to indirectly estimate growth and assign an estimated age to a fish (Casselman 1987). The accuracy of the age estimate is predicated on the choice of calcified structure, preparation method, and interpretation and is thus subject to error (Casselman 1983). Consequently, Beamish and McFarlane (1983) and Campana (2001) strongly recommend the accuracy and precision of age estimates be validated when possible.
Lake Sturgeon are currently listed as a threatened species in New York State and are considered a priority species for recovery (NYSDEC 2018). Two known stocks of Lake Sturgeon exist in and around the Niagara River; one at the headwaters (HNR), upstream of Niagara Falls, and the other at the mouth of the river, downstream of Niagara Falls (LNR; Figure 1; Hughes et al. 2005; Neuenhoff et al. 2018). Though genetically similar (M. Bartron, US Fish and Wildlife Service, unpublished data), these two groups of Lake Sturgeon are spatially separated by Niagara Falls and differ demographically (present study). Given the spatial separation and demographic differences, these two groups of Lake Sturgeon are considered two distinct stocks and are managed as such (Wells and Richmond 1995; NYSDEC 2018). Describing the age composition of these stocks is important to managers because it provides a means for assessing recruitment and cohort strength over time, as well as age at sexual maturity. The New York State Lake Sturgeon recovery plan requires two criteria to be met within six of seven management units before initiating delisting: 1) evidence of at least 750 sexually mature Lake Sturgeon and 2) evidence of natural recruitment in at least 3 y of a 5-y period within the last 20 calendar years (NYSDEC 2018). The LNR and HNR stocks represent two of the seven management units. Age estimates can provide managers with age at maturity and recruitment, contributing to the assessment of both delisting criteria within management units. In the late 1990s, efforts were made to assess the stock age structure of Lake Sturgeon in the LNR (Hughes et al. 2005); however, no studies we are aware of have re-examined the age structure of this stock and no studies have documented the age structure of Lake Sturgeon occupying the HNR. To facilitate future stock assessment, we sought to characterize the age distributions and average individual growth rates of these two Lake Sturgeon stocks.
The Niagara River is approximately 58 km long and is the major connecting waterway between Lake Erie and Lake Ontario (Figure 1). The river forms the international border of Canada and the United States beginning at the northeast end of Lake Erie and flowing north into the main stem of the river. The majority of Lake Erie's discharge flows into the headwaters of the river, near Buffalo, New York and Fort Erie, Ontario, and begins its way through the 35-km reach of the upper Niagara River.
The Niagara River can be described by its three main sections, the upper Niagara River and two sections in the LNR: the Niagara Gorge and the lower portion of the Niagara River. The upper Niagara is relatively shallow (7 m) compared with the LNR and is characterized by variable water velocities depending on river width, ranging from 0.6 m/s to 3.7 m/s (INWC 2016). Once the river reaches Grand Island, it splits into two channels before reuniting at the northwestern end of Grand Island. The river then enters the Niagara Cascade for about 1 km before dropping 57 m over Niagara Falls and entering the LNR. The first portion of the LNR, known as the Niagara Gorge, is a 7-km gorge comprised of a deep pool just below the falls followed by deep rapids with velocities measured up to 9.0 m/s (INWC 2016). After the gorge, the river slows and traverses approximately 15 km before reaching the mouth where it meets Lake Ontario.
We sampled in the HNR, proximate to documented Lake Sturgeon egg deposition (Neuenhoff et al. 2018), south-southeast of Bird Island reef and in the LNR, downstream of the gorge and around the mouth of the river. We used experimental gill nets to sample Lake Sturgeon at the HNR (see Neuenhoff et al. 2017 for sampling method specifics) throughout May and we used baited setlines in the LNR (see Bruestle et al. 2018 for sampling method specifics) between March and October from 2012 to 2017. We measured and tagged captured Lake Sturgeon with passive integrated transponders and FLOY T-bar anchor tags (Floy Tag & Mfg. Inc., Seattle, WA) to identify recaptured individuals. We assigned sex when possible by gamete expression or gonad assessment via visual observation through a small incision. We took a small (1.5 to 2.5 cm) section of the left leading pectoral fin spine roughly 5 cm from the point of articulation for age estimation. If an individual was recaptured in a later year of sampling, we removed the right leading pectoral fin spine for partial age validation. We stored spine samples in envelopes and allowed them to air dry for a minimum of 48 h before processing.
Processing, reader selection, and age estimation
Once dried, we took a thin (0.5 mm) transverse section from the basal end of the spine sample using a low-speed diamond-bladed sectioning saw. The transverse section was allowed to dry before fixing it to microscope slides using a two-part epoxy. Once the epoxy hardened, we ground and polished the sample to preferred thicknesses (translucency).
Since age assessment is subjective and can have low reader accuracy despite high reader precision, there can be large amounts of error in age estimates. To select a reader to estimate ages, we had three individuals estimate ages of known-age Lake Sturgeon spine sections collected from Lake Winnebago provided by the Wisconsin Department of Natural Resources. Readers defined an annulus as a translucent check following opaque zonation such that the area from the distal edge of one translucent check to the distal edge of the consecutive translucent zone would denote 1 calendar year. Beginning at the focus, or origin, readers would identify the first annulus and progressively denote annuli while moving toward the edge of the spine. Though the sample size was relatively small within year classes and only contained individuals up to age 14, assessing the accuracy of each reader with this index was believed to be valuable enough to identify a single best reader. Once we determined a single best reader from accuracy of age estimates on the known-age spines from Lake Winnebago, that individual proceeded to estimate the ages of spines collected from Lake Sturgeon from the HNR and LNR in the same manner (see Data S1, Supplemental Material, for age estimates).
Our age estimator viewed each spine section under a compound microscope (Zeiss Axioskop 2 plus) using transmitted light and a magnification between ×12.5 and ×100. Our estimator identified the focus of each sample and identified each annulus, considered to be one complete opaque zone—relating to faster growth during the growing season—and one translucent zone—pertaining to the slower growth during winter months, toward the outer edge. Our estimator was given the month of capture to help determine whether marginal increment growth near the outer edge should be classified as a complete annulus. In addition to providing an age estimate, our estimator assigned a confidence rating to each section where samples with zones that were indistinct, variable in appearance, or coalesced received a lower rating. To reduce error, we removed sections that received very low confidence ratings from further analyses.
, where I is the age at the inflection point; and the logistic equation:
, where g∞ is the instantaneous growth rate at infinity.
We compared stock age structures and lengths between stocks using a Mann–Whitney U test since data were not normally distributed (using Shapiro–Wilk tests on LNR ages: W = 0.98, P < 0.001; LNR total lengths: W = 0.99, P < 0.001; HNR ages: W = 0.81, P < 0.001; HNR total lengths: W = 0.96, P < 0.001). We also compared differences in lengths and age structure between sexes within and between stocks using Mann–Whitney U tests. Finally, we used a Kruskal–Wallis test to test for differences in age by year within each stock.
We examined recaptured Lake Sturgeon spines as a form of partial age validation. We took the age estimate from the first capture as the “true” age. Our expected age at time of recapture was calculated as the true age estimate from the time of first capture plus time at liberty. We then compared this expected age estimate with the age estimate assigned to the spine taken during recapture.
Of the 668 Lake Sturgeon we captured in the LNR between 2012 and 2017, we deemed that 618 had age estimates we considered trustworthy. Of the 186 Lake Sturgeon we captured in the HNR between 2012 and 2017, we deemed that 164 had age estimates we considered trustworthy. Age estimates ranged between 4 and 30 y in the LNR (median = 15 y) and 9 and 42 y at the HNR (median = 16 y). Lengths ranged from 81.8 to 181.9 cm in the LNR (median = 140 cm) and from 117.8 to 187.4 cm at the HNR (median = 144.7 cm).
We found that age distributions and sizes differed by sex and stock. Despite the age distributions appearing similar between stocks, with the majority of captured fish having ages between 14 and 19 y, HNR Lake Sturgeon were older (W = 45282, P = 0.035). Overall sizes at the HNR were also slightly larger than those present in the LNR (W = 38020, P < 0.001). Female Lake Sturgeon sampled in the LNR were longer (W = 915, P < 0.001) and estimated to be older (W = 1178, P < 0.001) than males, with females estimated to be 11 to 30 y old (median = 18 y) and ranginged in size from 123.1 to 181.9 cm (median = 149.4 cm), whereas males were estimated to be 9 to 26 y old (median = 16 y) and ranged in size from 115.3 to 170.7 cm (median = 137.5 cm). Similarly, females at the HNR were longer (W = 181, P = 0.004) and older (W = 176, P = 0.003) than males, with females estimated to be between 14 to 42 y old (median = 23 y) and ranged in size from 131.1 to 187.4 cm (median = 165.5 cm); males were between 10 and 33 y old (median = 16 y) and ranged in size from 117.8 to 178.6 cm (median = 143.3 cm). Age distributions varied between years, with median ages increasing from 14 in 2012 to 18 in 2017 in both stocks (Figures 2 and 3; LNR: χ2 = 63.01, df = 5, P < 0.001; HNR: χ2 = 19.69, df = 5, P = 0.001). Cohorts ranged from 1983 to 2013 in the LNR and from 1974 to 2006 at the HNR (Figure 4). The strongest year classes were 1998 in the LNR, with the majority of fish coming from 1997 to 2001, and 1999 at the HNR, with the majority of fish coming from 1996 to 2000.
Our partial age validation, by comparing recaptured fish's age estimates with their age estimate at first capture, showed age estimates of recaptures generally overestimated the expected age (Figure 5). Linear regressions of estimated age at recapture vs. expected age given the assigned age at original capture showed slopes differing from 1 and intercepts differing from 0 (parameter estimate ± SE; slope: 0.84 ± 0.13, intercept: 3.78 ± 2.22). This indicated that there was an inconsistent bias in age estimation: ages of younger Lake Sturgeon tended to be overestimated more than those of older Lake Sturgeon.
We calculated average individual growth rates for Lake Sturgeon captured in the LNR using a von Bertalanffy growth curve (Figure 6). We were not able to fit growth curves to the length-at-age data from those sampled at the HNR (Figure 7). We made attempts to fit a growth curve to the data by fixing t0 and constraining L∞ in a von Bertalanffy model as well as fitting Gomperz and logistic models; however, an absence of young individuals likely caused data to appear more linear and prevented data from fitting these growth models. The lack of young individuals produced unrealistic length at age-0 estimates; therefore we did not pursue further considerations about the von Bertalanffy growth curve.
We found differences in sizes and ages of Lake Sturgeon between sexes within both the LNR and the HNR stocks. Since sampling of both stocks occurred near known spawning sites during spawning seasons, when sex is more confidently assigned, individuals whose sex was assigned were more likely to be sexually mature. Given that females sexually mature at older ages and larger sizes than males (Pikitch et al. 2005; Bruch 2008), it is not surprising that the median age of sampled females was older than that of males.
We also observed differences in lengths and ages between the two stocks. Observed differences could be driven by disparities in resources or habitat. Though both systems provide heavy flow and preferred spawning substrate, warmer temperatures and higher productivity in Lake Erie may facilitate faster growth and larger sizes than in Lake Ontario. Alternatively, differences may be an artifact of using different sampling gears between stocks (gill nets in HNR vs. setlines in LNR).
In addition to having similar age distributions, both stocks showed a lack of young, small individuals and old, large individuals. Median ages in both stocks increased from 14 to 18 y between 2012 and 2017, suggesting some regional effect that produced a strong year class in 1998. The lack of young, small individuals is in part due to the selectivity of the gear used to catch Lake Sturgeon. The lack of old individuals could be due to low year class strength before 1998 and also the filtering of age data with low reader confidence (generally spines with very high numbers of compacted growth increments or where resorption may have occurred).
We found that mean lengths and lengths at age in the LNR and HNR were larger than those reported in most other systems, whereas mean ages were similar to many other systems. We found that mean total lengths of Lake Sturgeon from the LNR and HNR were larger than mean total lengths of Lake Sturgeon from Lake St. Clair River, St. Marys River, Lake St. Clair, Lake Huron in Michigan waters, and the North Channel and Georgian Bay of Lake Huron; comparable with mean total lengths of Lake Sturgeon from the Winnebago system; and smaller than mean total lengths of Lake Sturgeon from the Detroit River (Table 1). Mean ages of Lake Sturgeon from the LNR and HNR were slightly less than mean ages reported from the Winnebago system, Lake St. Clair River, and Lake St. Clair but greater than mean ages reported in Lake Huron in Michigan waters and the North Channel and Georgian Bay of Lake Huron (Table 1). Both LNR and HNR Lake Sturgeon length at age 10 were predicted to be larger than Lake Sturgeon at age 10 in the St. Marys River (Table 1). Lake Sturgeon from the LNR were predicted to be larger at age 10 than the previous length-at-age prediction (Table 1).
Although gear-based size selection tended toward larger individuals, size at age appears to have increased over time. Hughes et al. (2005) captured slightly younger, smaller individuals using the same sampling methods in the late 1990s and early 2000s. The increase in mean total lengths and length at age in the LNR could signify a shift in growth rate over time. Jacobs et al. (2017) proposed that young Lake Sturgeon may be transitioning from foraging on lower trophic prey items to higher trophic prey items, such as round goby (Neogobius melanostomus), whose higher energy content could result in increased Lake Sturgeon growth at younger ages. This may partly explain the faster growth rate we observed than what Hughes et al. (2005) reported.
It is important to note that length-at-age von Bertalanffy models are sensitive to age estimation error. Both the LNR and HNR lack data from young individuals, which are important to developing a growth model, and though we attempted to quantify estimation errors, no consistent bias was detected; therefore we could not correct data for these errors. Given that many of the captured individuals were estimated to be greater than 14 y, age estimation errors are to be expected given the nature of the spine structure used to estimate ages of these fish (Bruch et al. 2009). Future studies on the age structure of these stocks would benefit from the collection of younger (less than 14 y) Lake Sturgeon. Furthermore, despite removing low-quality spines for age estimation, our age estimates appear somewhat poor given partial age validation through ageing individuals captured multiple times. The lack of agreement between estimates from partially validated ages demonstrates that errors likely occurred during processing or interpretation. Though we attempted to train readers on known-age spines, there was a lack of samples available within and across age classes. Future studies may benefit from training readers on inventoried, known-age spines (n > 100 of varying ages) to gain experience in interpretation before attempting to assign ages. Although we used known-age samples for reader evaluation and training, sample sizes within each cohort were limited and many ages were not present. Establishing a long-term index of known-age samples with more than five samples in each cohort would be a valued resource in age interpretation training. Additionally, the use of a fluorochrome label has been shown to assist in Lake Sturgeon pectoral fin spine interpretation by providing readers with spatial reference points in the form of fluorochrome marks within calcified structures that relate to time of previous capture (Rossiter et al. 1995).
Although we recognize that there is error in our estimates of Lake Sturgeon age, these data still aid in determining the status of the HNR and LNR stocks. One of the delisting criteria of New York State's Lake Sturgeon recovery plan (see NYSDEC 2018) is evidence of natural recruitment in at least 3 y of a 5-y period within the last 20 calendar years. Even with the inaccuracies of ageing Lake Sturgeon spines, our data demonstrate that this criterion has been met given the number of consecutive cohorts within each stock. Error in age estimates could present problems when using these data in other stock assessment models such as catch curve analysis to estimate mortality or when conducting virtual stock analysis or catch-at-age modeling. In such models, aging error can lead to inappropriate estimates of recruitment time series, yield predictions, and ultimately lead to overfishing (Tyler et al. 1989; Bradford 1991). Thus, additional efforts to increase the accuracy of age estimation in Lake Sturgeon are warranted, especially if such data are used to manage a recovered fishery with some level of allowable exploitation.
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. Raw age and length data of captured and recaptured Lake Sturgeon Acipenser fulvescens collected from headwaters of the Niagara River and the lower Niagara River from spring sampling between 2012 and 2017.
Found at DOI: https://doi.org/10.3996/102019-JFWM-085.S1 (51 KB CSV).
Reference S1.Casselman JM. 1983. Age and growth assessment of fish from their calcified structures—techniques and tools. Pages 1–17 in Prince ED, Pulos LM, editors. Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks. Miami, Florida: National Oceanic Atmospheric Administration Technical Report National Marine Fisheries Service.
Found at DOI: https://doi.org/10.3996/102019-JFWM-085.S2 (749 KB PDF); also available at https://spo.nmfs.noaa.gov/Technical%20Report/tr8opt.pdf.
Reference S2.Hill TD, McClain JR, editors. 2002. Activities of the Central Great Lakes Binational Lake Sturgeon Group. Alpena, Michigan: U.S. Fish and Wildlife Service, Alpena Fishery Resources Office.
Found at DOI: https://doi.org/10.3996/102019-JFWM-085.S3 (945 KB PDF); also available at https://www.fws.gov/midwest/sturgeon/documents/rpt-stnib02.pdf.
The authors thank staff at the Northeast Fishery Center, the New York State Department of Environmental Conservation, and the Lower Great Lakes Fish and Wildlife Conservation Office for their assistance in capturing and marking Lake Sturgeon throughout this study. Thanks and recognition are given to Dr. Mark Clapsadl and the Great Lakes Science Center as well as Chief Timothy Rafter and the Aids to Navigation unit of the Buffalo, New York U.S. Coast Guard for equipment storage and providing facility space. Two anonymous reviewers and the Associate Editor provided comments that improved an earlier version of this manuscript. This study was funded by the Great Lakes Restoration Initiative.
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.
Citation: Withers JL, Gorsky D, Biesinger Z, Einhouse D, Clancy M, Davis L, Karboski C, Legard C, Bruestle E, Markley N, Roth R, Zimar R, Sweka JA. 2020. Age and growth of Niagara River Lake Sturgeon. Journal of Fish and Wildlife Management 11(2):634–643; e1944-687X. https://doi.org/10.3996/102019-JFWM-085
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.