Evaluation of Precision and Sample Sizes Using Standardized Sampling in Kansas Reservoirs

Fisheries managers desire efficient sampling methods with sufficient precision for monitoring ictalurid catfish populations. For this study, I conducted single unbaited hoop‐net surveys in the Minnesota River to evaluate the efficacy of simultaneously assessing stock‐size (≥280‐mm) Channel Catfish Ictalurus punctatus and quality‐size (≥510‐mm) Flathead Catfish Pylodictis olivaris populations. I compared catch rates and length‐frequency distributions among months (April–September) and used resampling procedures to estimate the number of hoop‐net samples needed to achieve desirable precision of catch rate estimates and adequate statistical power. Catch rates (mean ± SE) were 2.1 ± 0.3 fish/net‐night for Channel Catfish and 0.8 ± 0.1 fish/net‐night for Flathead Catfish. Resampling procedures demonstrated that desirable precision (i.e., relative SE ≤ 25) and adequate statistical power (≥0.80 with α = 0.10) to detect 50% reductions in mean catch rates with two‐tailed t‐tests could be achieved for both species with as few as 68 hoop‐net samples during August. Unbaited hoop nets are effective for simultaneously assessing Channel Catfish and Flathead Catfish populations in the Minnesota River, and conducting surveys during August provides the best balance between catch rate and precision while minimizing perceived environmental and behavioral biases. However, the amount of sampling effort needed to detect small population changes (<50%) and to collect sufficient sample sizes for age and growth analyses may be logistically prohibitive. This study emphasizes the importance of evaluating precision and statistical power of sampling methods so that fisheries managers can identify the most suitable methods for their specific objectives.

Section: Discussionmentioning

confidence: 99%

“…Lower RSEs indicate greater precision, with a minimum goal of RSE ≤ 25 and a preferred goal of RSE ≤ 15 (Dumont and Schlechte ; Koch et al. ). Statistical power generally increases with precision, with a common power goal of at least 0.80.…”

Section: Methodsmentioning

confidence: 99%

Evaluation of Unbaited Hoop Nets for Simultaneously Assessing Channel Catfish and Flathead Catfish populations in the Minnesota River

Sindt

2018

“…Rigorous assessments, in terms of sampling effort (see Koch et al. ) and temporal duration (Wagner et al. ), provide fisheries managers with the information needed to determine applicability of MLLs as tools for meeting fisheries objectives (Isermann and Paukert ).…”

mentioning

confidence: 99%

Standard Assessments Reveal Context‐Dependent Responses of Crappie Populations to a Length‐Based Regulation in Ohio Reservoirs

Pritt

Conroy

Page

et al. 2019

Fisheries managers implement minimum length limits (MLLs) to improve the size structure of populations of crappie Pomoxis spp. throughout Midwestern and southeastern U.S. reservoirs. The success of these regulations has been mixed, with several implementations resulting in undesirable outcomes, including slow growth and stunting of crappies, and ultimately regulation removal. Consequently, it is unclear whether and where MLLs should be used to improve crappie size structure. Beginning in 2003, Ohio instituted a statewide standard sampling protocol to monitor reservoir crappie populations, and between 2003 and 2010 a 229‐mm MLL was implemented at over 40 reservoirs throughout the state. Using these spatially and temporally extensive crappie population data, we sought to (1) test for the response of crappie population adult abundance, growth, and size structure to the MLL, (2) test whether reservoir surface area and/or productivity mediated this response, and (3) model fisheries outcomes at different reservoir sizes and productivities to predict where application of a MLL would be most appropriate. Using linear mixed‐effects models and an information theoretic approach, we found that crappie population abundances generally increased, whereas growth decreased in small, unproductive reservoirs but increased in large, productive reservoirs. Overall, the MLL failed to produce increases in size structure, except in large reservoirs. Yield‐per‐recruit models predicted a decrease in angler yield in response to a 229‐mm MLL in small reservoirs, whereas in large, productive reservoirs this regulation was predicted to increase angler yield. Consequently, the MLL could improve crappie fisheries in reservoirs larger than 1,000 ha with high productivity (total phosphorus concentrations > 50 μg/L) but could be counterproductive in small, unproductive reservoirs. Finally, our approach highlights the benefits of standard sampling as we were able to integrate data across years and reservoirs to make comparisons on a statewide scale.

“…A minimum of 2 and maximum of 100 gill‐net deployments were utilized for both CPUE and length‐frequency analysis to reflect practical sampling effort (Dumont and Schlechte ; Koch et al. ). The mean seasonal sampling efforts for CPUE and length‐frequency data were reported from the 2015 and 2016 sample estimates.…”

Section: Methodsmentioning

confidence: 99%

“…The sample sizes used in standardized fish sampling assessments in large reservoirs are often lower than the required effort needed to achieve suitable confidence in sampling results (Koch et al. ; Miller et al. ).…”

mentioning

confidence: 99%

Spatial and Temporal Variability in a Large‐Reservoir Fish Assessment and Application of a Stratified Random Sampling Approach

Schall

Schoenebeck

Koupal

2019

Sampling is often standardized with regard to time and space to monitor population characteristics. Effective sampling of large reservoirs can be complicated by the inherit bias of habitat variability along the reservoir gradient. Our objective was to assess spatial and seasonal distribution patterns of target and nontarget fish species in a large reservoir to identify potential sources of bias. A multitude of fish community and habitat characteristics were measured during seasonal (spring, summer, and fall) gill‐net surveys in 2015 and 2016 at Lake McConaughy, Nebraska. Spatial assessments along the longitudinal gradient (up‐reservoir to down‐reservoir) and between opposing shorelines were made for fish metrics from the seven most common species, and associations between species‐specific relative abundance and habitat data were examined. Spatial differences in relative abundance and size structure occurred for nearly all species, with the greatest spatial variability occurring between the upper and lower reservoir zones, but few strong correlations were observed between species‐specific relative abundances and habitat data. Seasonal differences occurred in relative abundance and size structure for most species, and size structure was lowest in the fall for most species. To reduce the impact of spatial variability, we determined the minimum required sampling effort to achieve precise estimates of mean catch per net‐night using two resampling strategies, with 16 or fewer gill‐net sets required for five of seven species and 14 or fewer required if a stratified random approach is utilized. This research quantifies the extent of spatial and temporal variability in standardized sampling on a large reservoir and provides novel insight for managers of similar systems considering a stratified random approach to reduce sampling effort while maintaining desired levels of precision.