Bigmouth and Smallmouth Buffalo in Wisconsin: Population Characteristics and Contributions to Bowfishing Harvest

Bigmouth buffalo (Ictiobus cyprinellus) and smallmouth buffalo (Ictiobus bubalus) are native to the midwestern USA and receive little management attention due to their status as non-game fishes. As a result, little is known about their population dynamics and characteristics across the landscape. Interest in buffalo and other native non-game fishes has increased as recently reported maximum ages of many non-game species are greater than previously thought. Recent interest in the conservation of these species has coincided with the increased popularity of bowfishing over the last decade. Bow fishers harvest invasive cyprinids and native non-game species, but the contribution of buffalo to bowfishing harvest is largely unknown. Consequently, my research objectives were to 1) describe age composition, reproductive traits (fecundity and maturation schedules), and population dynamics (growth, mortality, and recruitment variability) for bigmouth and smallmouth buffalo populations in Wisconsin; 2) determine the minimum number of fish needed to accurately represent age composition and variation in population metrics; 3) assess the relative resiliency of Wisconsin buffalo populations to harvest using population metrics and spawning potential ratio (SPR), and 4) determine the contribution of buffalo to bowfishing harvest in tournament and non-tournament settings. Over a two-year period in 2023 and 2024, I collected buffalo from 14 water bodies across Wisconsin. Buffalo were collected from Wisconsin Department of Natural Resources (WDNR) sampling using electrofishing, fyke nets, seines, and gillnets. Buffalo were also obtained from 10 Wisconsin Bowfishing Association (WBA) tournaments. Buffalo were identified to species, measured to the nearest millimeter (total length; TL), weighed to the nearest 0.01 kilogram, and sex and maturity were determined from visual inspection of the gonads. Otoliths were used for age estimation. A modified age-length approach was used to select buffalo for age estimation, where the target was to collect 10 fish under 254 mm TL, 5 fish each per 50.8 mm TL bin for fish between 254 mm - 660 mm TL, and 10 fish over 660 mm TL. Heads were stored with identification tags in a freezer until otoliths could be removed in the laboratory following the protocol described by Lackmann et al. (2019). Otoliths were thin sectioned using a Buehler IsoMetTM 1000 Precision Cutter, viewed under a Nikon 1500 SMZ stereomicroscope, and images were broadcast to a 65 × 40 cm flat-screen monitor from which two independent readers enumerated annuli. Age length keys were created for each population and were used to assign ages to all fish in a sample. Age composition of each population was described using Shannon diversity index (H) and number of age classes present. Age (A50) and TL (L50) at which probability of maturity was 50% were determined for each population using logistic regression. Mean fecundity at median TL of all mature females in the sample (663 mm) was predicted using a power function. Growth trajectories were quantified using a von Bertalanffy growth function (VBGF) and buffalo growth was described using predicted TLs-at-age 3 and age 11, and time in years to reach the TL at which buffalo were estimated to become vulnerable to bowfishing harvest (500 mm) and median observed TL observed across all populations (585 mm). Catch curves were used to estimate total mortality (Z) of each population using the Chapman-Robson method. Variation in recruitment was described using recruitment variation index. To determine the minimum number of fish needed to accurately represent age composition and variation in population metrics, a minimum of 90 fish were collected from Yellowstone Lake and the Lower Wisconsin River. “Oversampled” populations were randomly subsampled in increments of 10 and this subsampling was performed 1,000 times at each increment (i.e., 10 fish were randomly sampled 1,000 times, 20 fish were randomly sampled 1,000 times and so on). The minimum number of fish needed to accurately estimate the number of age classes present, mean A50, and mean L50 was determined for each population as the increment at which the calculated metric was within 10% of the metric calculated for the entire sample. Relative resiliency to harvest was assessed using two different approaches. First, SPR was used to predict the effects of varying levels of instantaneous fishing mortality (F) on the reproductive potential of each buffalo population. Instantaneous natural mortality rates of female buffalo were determined using six different estimators available in Fishery Analysis and Modeling Simulator (FAMS), and F was simulated from 0 to 1.0 in increments of 0.05. The F value required to reduce SPR below 0.3 (F0.3) was used as a comparison metric among populations. Additionally, population-specific estimates of metrics outlined in the first objective were compared to respective median values of those metrics obtained from all populations. Populations were considered relatively more resilient if they grew quicker, matured earlier, and exhibited a greater F0.3. Contribution of buffalo to tournament harvest was determined from 10 WBA tournaments held in 2023 and 2024. Harvested fish were enumerated and identified to species or species groups. A random sample of harvested buffalo at each tournament was measured for TL (mm). Contribution of buffalo to non-tournament harvest was determined using a bow fisher diary which disseminated to six WBA volunteers. Diary entries consisted of date, location, number of individuals in the party, target species, number of each species harvested, and trip duration. Number of age classes present ranged from 12 to 20 (median = 15.5) in bigmouth buffalo populations and from 15 to 30 (median = 20) in smallmouth buffalo populations. Maximum age ranged from 14 to 71 years (median = 32.5 years) in bigmouth buffalo populations and from 37 to 58 years (median = 48 years) in smallmouth buffalo populations. Age at 50% maturity ranged from 2.0 to 5.0 years (median = 2.99 years) in bigmouth buffalo populations and from 3.41 to 8 years (median = 4.39 years) in smallmouth buffalo populations. Time to reach 500 mm TL ranged from 3.4 to 6.1 years (median = 4.7 years) in bigmouth buffalo populations and from 4.5 to 7.0 years (median = 6.0 years). Predicted TLs at age 3 ranged from 295 to 489 mm (median = 435 mm) in bigmouth buffalo populations and from 417 to 475 mm (median = 433 mm) in smallmouth buffalo populations. Recruitment variation indices ranged from -1.79 to -0.34 (median = -0.55) in bigmouth buffalo populations and from -1.51 to 0.02 (median = -0.25) in smallmouth buffalo populations. Total mortality estimates ranged from 0.14 to 0.65 (median = 0.30) in bigmouth buffalo populations and from 0.18 to 0.23 (median = 0.19) in smallmouth buffalo populations. A sample of 70 fish with otolith-based age estimates was needed in both oversampled populations to accurately estimate the number of age classes present in the entire sample. Maturation schedules were generally able to be accurately estimated with as few as 10 fish, given that at least 1 immature fish was present in the sample. Samples of 50 and 100 fish were needed to accurately estimate H and RVI, respectively. The rate of F required to reduce SPR to the threshold value ranged from 0.12 to 0.23 (median = 0.15) in bigmouth buffalo populations and from 0.08 to 0.09 (median = 0.09) in smallmouth buffalo populations. Populations considered relatively resilient to harvest included bigmouth buffalo from the Rock and Crawfish rivers, while populations considered less resilient to harvest included smallmouth buffalo from Lake St. Croix and the Lower Wisconsin River. Total contribution of buffalo to bowfishing tournament harvest was 6.2% in 2023 and 32.3% in 2024, with an overall contribution of 17.4% across both years. In total, 555 buffalo were harvested in 2023 and 2,192 were harvested in 2024, for a total of 2,747 across both years. Most of these fish (2,717) were bigmouth buffalo and only 30 were smallmouth buffalo. Mean TL of harvested buffalo across all tournaments was 653 mm (SD = 124). Non-tournament bowfishing trips mostly occurred in May (36%) and relatively least trips occurred in July (12%). Common carp constituted the greatest proportion of total harvest (31.8%), followed by gar (11.7%) and buffalo (10.6%). My results suggest that the characteristics of buffalo populations vary substantially across Wisconsin. While previous studies have suggested buffalo may attain ages of 100 years or more, my assessment suggests this does not occur in all populations. Buffalo recruitment appears to occur more frequently in my Wisconsin study populations than in populations previously studied in Minnesota and Saskatchewan. Bigmouth buffalo populations appeared to be more resilient to harvest than smallmouth buffalo populations, and populations inhabiting eutrophic lakes (Yellowstone and Beaver Dam lakes) and small rivers (Crawfish and Rock rivers) tended to be more resilient than those which inhabited large river systems (Mississippi and Lower Wisconsin rivers). The variation in the characteristics of buffalo populations across the landscape suggests that a single approach to buffalo management in Wisconsin may be inadequate. The contribution of buffalo to tournament harvest varied considerably with location and time of year, with the greatest contributions occurring in the spring when buffalo were actively spawning. My results suggest that bowfishing harvest may not be an important driver of population dynamics in many Wisconsin populations where recruitment is observed with some regularity or in populations that support considerable commercial harvest (which is likely far greater in extent than bowfishing harvest). However, harvest may be more likely to affect populations which recruit sporadically, exhibit relatively long average lifespans, and are not c

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.