Gillnet sampling 2010 |
Previous | 1 of 1 | Next |
|
small (250x250 max)
medium (500x500 max)
Large
Extra Large
large ( > 500x500)
Full Resolution
|
This page
All
|
W 2800.7 FS32r/w F-SO-R-17 2010 no.30 c.1 FINAL REPORT FEDERAL AID GRANT NO. F-50-R-17 PROJECT NO. 30 Fish Research and Surveys for Oklahoma Waters Gill net sampling: An Evaluation of Oklahoma's Standard Sampling procedures OKLAHOMA DEPARTMENT OF WILDLIFE CONSERVATION JANUARY 1, 2010 through DECEMBER 31, 2010 FINAL REPORT State: Oklahoma Grant Number: F-50-R-17 Grant Title: Fish Research for Oklahoma Waters Project Number: 30 Project Title: Gillnet sampling: An Evaluation of Oklahoma's Standard Sampling Procedures Project Leader: Greg Summers Grant Period: From: January 1,2010 to: December 31, 2010 I. Project Objective: Determine 1.) difference in catch rate and length frequency of previous and current SSP gill net configurations at historic fixed sites and 2.) difference in catch rate and length frequency of fixed and random sites using current SSP gill nets. II. Summary of Progress: A. Introduction Gill nets are a passive capture gear that catches fish by entanglement. They are underwater walls of netting usually set in a straight line (Hubert 1983; Miranda and Boxrucker 2009). Fish are caught by either being wedged-held by mesh around the body, gilled-held by mesh slipping behind the opercula, or tangled-caught by spines, teeth, or other protrusions without actually penetrating a mesh. Therefore, mesh size is an important factor influencing the size of fish captured (Reddin 1986; Hubert 1983; Miranda and Boxrucker 2009). Other factors that influence catch of gill nets include color, season, baiting (Jester 1977), net length, set time (Minns and Hurley 1988) use of mono- or multi-filament material to construct the net (Anonymous 1952; Hogman 1973; Henderson and Nepszy 1992), size of filaments (Hansen 1974), and hanging ratio (Machiels et al. 1994). Gill nets are one of the most widely used fisheries gears in the United States and Canada (Gablehouse et al. 1992); therefore many variations exist, which can make comparisons between nets difficult. Standardized sampling procedures are important in fisheries management and are used to evaluate fish populations over time and between reservoirs in a region or state (Noble 2002). In 1977 the Oklahoma Department of Wildlife Conservation (ODWC) developed "Standardized Sampling Procedures (SSP) for Reservoir and Reservoir Management Recommendations" (Erickson 1978). This protocol called for the use of gill nets 61-m long by 1.8-m deep (0.5 basis) with 7.6-m mesh panels with bar mesh sizes of 19,25,38, 51,57,76,89, and 102 mm respectively. Since that time, Oklahoma's SSP have been revised to obtain more accurate population parameters as well as maximize efficiency. In 2001, the ODWC made a change to its SSP gill net configuration. Two, 7.6-m panels were added to the nets (13- and 16-mm bar mesh) and two panels were removed (89 - and 102-mm bar mesh). These changes were made with two objectives in mind. The first was to target shad (Dorosoma spp.) that were previously not represented in gill net samples and the second was to reduce the catch of rough fish (L. Cofer, ODWC, personal communication). In 2009, the ODWC further refined their shad sampling and began using separate floating 'shad' gill nets with small bar mesh sizes specifically designed to target shad. The addition of shad nets to the ODWC's SSP made the data from the 13- mm and 16-mm mesh sizes of the SSP gill net configuration no longer necessary. The current SSP revised in 2009 specified the number of gill net samples required as a function of impoundment surface area: less than 40 ha, not more than 5 sites; 40-404 ha, 5 sites; 404-2023 ha, 10 sites; greater than 2023 ha, 15 sites (Kuklinski, ODWC, personal communication). Miranda and Boxrucker (2009) offer gill net standards for the entire southern USA with the intention of regulating gill net selectivity, deployment, effort, and timing of collection to reduce the variability that prevents adequate comparisons among agencies. They also considered sampling efficiency to account for typical time and labor restrictions. The ODWC has adopted the current standard net configuration suggested by Miranda and Boxrucker (2009) consisting of 24.4-m long by 1.8-m deep (0.5 basis) nets composed of 3.1-m panels with bar mesh sizes of38, 57, 25, 44,19,57,32, and 51 mm respectively. Oklahoma's previous gill net specifications were 61-m long by 1.8-m deep nets composed of7.6-m panels with bar mesh sizes of 13,16,19,25,38,51,57, and 76 mm. Because biologists use gill nets to look at trends in fish populations over time, it is beneficial to compare the catch rates and length frequencies of these two net designs during the same sampling season in order to understand the consequences of changing net configurations. When comparing different gear types, both gears should be used at the same time and same location. This method provides a way to convert historic data collected with one gear to data collected with another (Peterson and Paukert 2009; Noble et al. 2007). The ODWC's current SSP protocol also requires sample locations to be fixed sites selected by the biologist. Fixed sites chosen in this manner are useful for monitoring changes in the population over time but are potentially more biased than random sites with respect to abundance and length frequencies of fish they capture (Wilde and Fisher 1996). Fixed sites can produce greater precision, which allows fewer samples to achieve a similar level of precision; however, these sites may not accurately represent the entire fish population and population characteristics of these sites may change at a disproportional rate to the rest of the reservoir (Noble et al. 2007). Random sampling may give a more accurate assessment of sportfish populations and allow biologists to make more reliable comparisons between reservoirs, but at the possible cost of higher required replication to achieve the same precision. Therefore, direct comparison of fixed versus random sampling needs to be made to determine if additional sampling effort would be needed in a random sampling design. Moronids and percids are the primary target species for the ODWC's gill net SSP, specifically hybrid striped bass Morone chrysops x Morone saxatilis, white bass Morone chrysops, walleye Sander vitreus, and saugeye Sander canadense x Sander vitreus. Information gathered from gill nets about white crappie Pomoxis annularis, and channel catfish lctalurus punctatus are important to the ODWC, but are considered secondary target species. B. Methods Five Oklahoma reservoirs (Canton, Thunderbird, Kaw, Waurika, and Torn Steed) were selected for sampling in 2009 based on stockings of hybrid striped bass and either saugeye or walleye within the previous five years (Table 1). Four Oklahoma reservoirs (Ft. Cobb, Foss, Skiatook, and Torn Steed) were sampled in 2010. Gill netting took place in October and November. All sample reservoirs had natural populations of white bass and channel catfish; however, Tom Steed had supplemental stockings of channel catfish. Although Thunderbird was not stocked with hybrid striped bass it was sampled because was considered Oklahoma's best saugeye fishery. Waurika was only stocked periodically with saugeye but was sampled because it was considered Oklahoma's best hybrid striped bass fishery. Canton was considered Oklahoma's best and most productive walleye fishery. Each reservoir received a total of 45 net nights of effort (15 net nights of 61-rnnets at fixed sites, 15 net nights of 24.4-m nets at fixed sites, and 15 net nights of 24.4-m nets at random sites) for a total of 225 net nights of effort for all five reservoirs combined. Gillnet sampling began October 5t \ 2009 and continued through November 6th , 2009 in accordance with ODWC's SSP recommended time frame. Historic gillnet sites were identified at each reservoir and five 61-m long by 1.8-m deep nets composed of eight 7.6-m panels with bar mesh sizes of 13, 16, 19,25,38, 51, 57, and 76 mm were set perpendicular to the shore line and fished overnight. Nets were pulled the next day and five different fixed sites were sampled similarly each of the next two nights such that 15 different sites were sampled over three consecutive nights. Each night, another five 24.4-m long by 6-ft deep (112 basis) nets composed of eight 3.1-m panels with bar mesh sizes of38, 57,25,44, 19,57,32, and 51 mm were set perpendicular to the shoreline at the same historic sites approximately 90 meters away from the 61-m nets. Nets were set this way in order to give the same schools offish the opportunity to pass through each of the nets and become entangled. In addition, five 24.4-m nets (as described above) were set perpendicular to the shoreline at random locations each night such that 15 different random sites were sampled over three consecutive nights. Random locations were selected by placing a 274 by 274 meter grid over a map of the reservoir and grid numbers were randomly selected. Only grid sites with a depth of at least 1.8 m within approximately 46 m of the shoreline were considered to ensure that nets could hang stretched out from top to bottom. All nets were fished for a period of 17 to 24 hours. The number of fish of each species in each net was divided by the number of hours that net was fished and multiplied by 24 to give catch per 24 hours (catch/24 hours). Mean catch/24 hours was then calculated by averaging the catch/24 hours from all sites at each lake. Gillnets were pulled and fish were processed at a work station. Catch for each mesh panel was recorded separately. Total length (mm) and weight (g) were recorded for game species. Only total length was recorded for non-game species. Coefficient of variation (c.v.) was calculated for each species in each net configuration (fixed and random sites analyzed separately). The number of samples needed to detect a target c.v. of 0.25 and 0.125 were calculated using the random resampling method of Dumont and Schlechte (2004). Catch/24 hour data were log transformed (In[X+0.0004]) to correct for normality. A two-way analysis of variance (ANOVA) (net length treated as a fixed factor and reservoir treated as a blocking variable) was used to test for differences in mean catch/24 hours between 24.4-m and 61-mnets. A two-way ANOVA (site type treated as a fixed factor and lake treated as a blocking variable) was also used to test for differences in mean catch/24 hours between fixed and random sites. Linear regression was used to test for relationships between the In[X+1] mean catch/24 hours of 24.4-m and 61-mnets at fixed sites for five target species. Kolmogorov Smirnov (KS) tests were used to test for differences in length frequencies (10-mm length groupings) of 24.4-m and 61-mnets with all mesh panels included, 24.4-m and 61-mnets with no shad mesh panels (i.e., lacking 13 and 16 mm shad mesh; 61nsm), and 24.4-m nets set at fixed and random sites. The KS test only allows for comparison of two populations at a time, therefore length-frequencies of each species were tested in each reservoir. Net configurations with sample sizes ofN < 40 were omitted from length frequency analysis. All statistical tests were evaluated as significant if P :::;0.05 with the exception of the linear regression analysis, which was evaluated at P :::;0.10 due to the increased variability of catch rates from individual nets. C. Results 24.4-m vs. 61-m at Fixed Sites Catch rates The 24.4-m and 61-mnets had similar mean catch/24 hours for most of the target species (Table 2). However, channel catfish and white crappie had significantly lower mean catch/24 hours in the 24.4-m nets. Because only 6 walleye were caught at Kaw (all net types combined), the walleye catch at Kaw reservoir was excluded from all analysis. The 24.4-m nets had lower mean catch/24hours for most non-target species (by-catch; Table 2). Variability (c.v. of the mean) of the 24.4-m nets was similar to (within 5 units) or less than the 61-mnets for the target species with the exception of hybrid striped bass, which was only slightly higher. Using the 15 samples currently required by the SSP (Kuklinski, ODWC, personal communication), the 61-mnets were able to detect a 25% increase or decrease in the population (c.v. = 0.25) only for channel catfish, walleye, and white crappie; although white crappie would only require one additional sample (Table 3). With 15 samples the 24.4-m nets were able to detect a ± 25% change in the population for all target species except hybrid striped bass and channel catfish. Channel catfish would only require two additional samples, but hybrid striped bass would require considerably more effort. Neither net type was able to detect a ± 12.5% change in the population (c.v. = 0.125) for any species with 15 samples. Regressions of mean catch/24 hours explained from 60.25% (white crappie) to 90.59% (hybrid striped bass) of the variation in catch rates between 24.4-m and 61-rnnet types (Figure 1). The only target species with a slope not significantly different from zero was white crappie. Length Frequency Just over half (59%) of the KS tests comparing 24.4-m and 61-m nets were significantly different (top panels, Figures 2-6; Table 4a); however, only 29% of the test comparing 24.4-m and 61nsm-m nets were significantly different (middle panels, Figures 2-6; Table 4b). The influence of the shad mesh is especially apparent for white crappie (Figure 6) but can also be seen to a lesser degree for hybrid striped bass (Figure 3) and white bass (Figure 5). Eliminating the shad mesh in the 61nsm-m nets did not improve the length comparisons for channel catfish (Figures 2, top and middle panels). Fixed vs. Random sampling sites Catch rates The mean catch/24 hours of most target species from fixed and random sites were similar. Only white bass differed significantly (Table 2). Variability at random sites was slightly higher than that at fixed sites for all target species except saugeye and walleye, which were essentially equal (Table 2). Using either sample site strategy, it is not possible to detect a ± 12.5% change in the population for any of the target species with only 15 net nights (Table 3). Using the random site sampling strategy, it was possible to detect a ± 25% change in the population with only 15 net nights for saugeye and walleye. An additional 3 nets would make this increase or decrease detectable for white crappie. Length Frequency The length frequencies from 24.4-m nets at fixed and random sites were typically similar (Table 4). The only reservoir where the length frequency of a target species differed was Canton (channel catfish; bottom panel, Figure 2; Table 4). D. Discussion 24.4-m vs. 61-m Catch rate Because there were no significant differences in mean catch/24 hours and the 24.4-m nets are less than half the length of the 61-mnets, the new 24.4-m nets are more efficient at catching hybrid striped bass, saugeye, walleye, and white bass than the previously used nets. This should be an advantage for state agencies and researchers considering the gill net configuration offered by Miranda and Boxrucker (2009) because the 24.4-m nets were much easier and faster to deploy and retrieve. By-catch was also lower in 24.4-m nets, so total processing time of these nets is lower than for 61-m nets. Mean catch/24 hours for white crappie were significantly lower in the 24.4-m nets suggesting the 61-mnets may be better for this species. However, most biologists use trap nets as the main sampling gear for crappie (Kuklinski, ODWC, personal communication). Boxrucker and Ploskey (1988) found trap nets to have higher catch rates and less variability in catch rates and length frequency distributions of white crappie than electro fishing and gill net samples. Guy et al. (1996) recommends trap nets over gill nets for sampling crappie because trap nets catch rates give a better index of abundance than gill net catch rates, although size structure data is similar for both gears. Any significant change in crappie gill net data due to the switch to a new net configuration may be offset by the efficiency of the 24.4-m nets to catch other target species as well as the availability of other gear types to sample crappie. Channel catfish mean catch/24 hours were also significantly lower for the 24.4-m nets. Previous research comparing gears for channel catfish have produced conflicting results. Two studies addressing small impoundments recommend experimental gill nets over baited hoop nets, baited slat traps (Robinson 1999) electrofishing (AC), baited wooden and wire fish traps, and trot lines (Santucci 1999). Sullivan and Gale (1999) recommend 25.4-mm-mesh or variable-meshed baited hoop nets in a series over experimental gill nets in larger reservoirs. The conflicting results of these studies suggest there is considerable variability with respect to channel catfish sampling. If biologists continue to use gill net data to track trends in channel catfish populations, they should develop new benchmarks based on the reduced efficiency of the 24.4-m nets for this species. Catch rates of 24.4-m and 61-mnets were correlated for every target species except white crappie, but only when using a significance level of P = 0.10. Given the inherent variability of gill net catch rates, this reduced significance level is justifiable. Robson and Regier (1964) report that management studies commonly use more liberal P-values than research studies. Understanding the relationship of mean catch/24 hours between the two net types will be critical for biologist to understand when making management decisions based on either historic or current data (Peterson and Paukert 2009). One way these equations could be useful is for making adjustments to stocking criteria. According to the ODWC's stocking criteria a reservoir is considered an established hybrid striped bass fishery ifit has a gill net mean catch/24 hours of 2.4 fish using the 61-mnets. Using the equation produced from the hybrid striped bass regression analysis indicates a mean catch/24 hours of 2.4 fish from 61-mnets is equivalent to a mean catch/24 hours of 1.1 fish using the new 24.4-m nets. These types of conversions will need to be made for all species. However, the regression is based on a relatively small sample size where at times a lake had values very different than others and influenced the slope more than other points. Using either net configuration it would be impractical to set enough nets to detect a ± 12.5% change in the population for any of the target species. The 61-m nets were able to detect a ± 25% change in the population for only channel catfish and walleye with 2: 15 samples. The 24.4-m nets were able to detect a ± 25% change for all target species except hybrid striped bass and channel catfish with 15 samples. Precision of channel catfish catch rates were very close to the targeted precision, requiring only 17 samples (Table 3). The better precision of the 24.4-m nets suggests it would not be necessary to increase sample size requirements for this new net design, however we do not recommend reducing effort to less than 15 samples/reservoir as precision is still just acceptable at this level. In fact, it may be beneficial for biologist to increase sample sizes on reservoirs where hybrid striped bass are a management concern. Due to the ease of use of the 24.4-m nets, the addition of 12 net nights (4 additional nets each of the 3 sample nights) should not drastically impact the amount of time spent in the field when compared to the amount of time spent running and processing the previous 61-mnets. Length Frequency For every target species except walleye, at least one reservoir showed a difference in length frequency between the different length nets. To better understand for the mechanism underlying these differences it is necessary to examine the construction of the two nets and the purpose of each included mesh panel. It is important to consider sampling objectives when comparing gear types (Peterson and Paukert 2009). The 61-mnets have 13- and 16-mm mesh panels, which are not found in the 24.4-m nets. These panels were included in the 61-m net to specifically target small shad that are vulnerable to predation-thus making up the forage base for predators (Cofer, ODWC, personal communication). These meshes not only catch small shad but also other small fish. The information gained about small non-shad species could. be useful when looking at things such as post-stocking survival or young of year abundance (Anonymous 1958; Willis 1987). However, as a standardized gear to collect information about age and growth, length frequency, and relative abundance of predatory fish, these mesh sizes would be less useful. When assessing predatory fish stock size and size structure (e.g, for setting creel regulations, length limits, and other management decisions), generally only fish that have recruited to a size that is catchable by anglers are used. Therefore, we suggest the 24.4-m nets are more efficient for predatory fish assessment and that separate gill nets should be used when data on smaller fish is desired, such as the ODWC now does for forage assessment (Kuklinski, ODWC, personal communication). It is likely the differences in length frequencies of the two net types are largely caused by the differences in mesh sizes used; the "shad meshes" of the 61-m nets had high catch rates of juveniles such as white crappie. This influence can be seen to a lesser degree for hybrid striped bass, and white bass. Once fish caught in the shad meshes were eliminated from the data set, significant difference were not found at any reservoirs for hybrids striped bass, and were only found at 1 of 4 reservoirs for white bass (Canton). Although statistically different, the difference in length frequency of white bass was small enough to be negligible for management purposes (middle panel, Figure 5). Johnson (1999) warns against relying too heavily on significance of statistical testing when conducting biological research. Channel catfish length-frequency distributions were significantly different at both reservoirs where N > 40, even after eliminating shad meshes. One possible explanation for the difference in length frequencies could be the mechanics of capture for this species. Catfish have serrated spines, which enable them to become easily tangled in a net without actually penetrating the mesh. For most species that are "gilled" rather than entangled in gill nets, there is an optimum size of mesh for which a particular species of a given size is captured. For these species, few fish are captured with lengths that differ from the optimum by more than 20% (Hamely 1975). This makes length frequency distributions for these types of species less variable and more consistently grouped by gill net mesh size. The serrated dorsal and pectoral spines of channel catfish make almost any size fish vulnerable to capture in almost any size mesh if its spines become entangled. Buckmeier and Schlecthe (2009) found that gill nets are less efficient and produce a more biased size structure data than electrofishing and hoop-netting for both blue and channel catfish; however, Santucci et al. (1999) found that gill netting and not electrofishing sampled channel catfish in proportion to their actual abundance. Further research is needed to determine the potential bias of the 24.4-m gill net design for this species. 24.4-m Fixed sites vs. 24.4-m Random Sites Catch rates The mean catch/24 hours of fixed and random sites were similar to each other with only white bass being significantly different. Since the mean catch/24 hours for white bass only differed by 0.1, it is doubtful this statistical difference has any application to management decisions. The precision at random sites was slightly lower than that at fixed sites for all target species except saugeye and walleye, where precision was essentially equal. It would be impractical with either sampling strategy to set enough nets to detect a ± 12.5% change in the population for any of the target species. Using random site selection, it was possible to detect a ± 25% change in the population with only 15 net nights of effort for saugeye, and walleye. If 8 more randomly-sampled net nights were added (one additional day in the field or setting an additional 3 nets each night) a ± 25% change in the population would also be detectable for channel catfish, white bass, and white crappie. As with the fixed-site sampling strategy, mean catch/24 hours of hybrid striped bass were the most variable of the target species for the random sampling strategy. If a random-sampling strategy was to be included in the SSP then it would be necessary to increase sampling effort by 12 net nights for this species to achieve the desired level of precision. Length frequency The only difference in length frequency found for a target species was for channel catfish at Canton. Of the 71 KS tests performed for fixed vs. random sampling (including target species and non-target species), only 8 (11%) significant differences were found. We did not perform a Bonferroni adjustment on the KS tests because we wanted to be overly conservative. However, it is likely that many of these 8 significant differences could have occurred by chance. As mentioned earlier, channel catfish can be caught in gill nets by their serrated spines, which makes the size of channel catfish captured in each mesh size much more variable (Table 4c). Therefore, we suggest the difference we observed in Canton reservoir may not indicate a strong bias of net design, but could have just occurred by chance along. Further research would be needed to clarify this pattern. Data from 2010 is currently being entered and analyzed. E. Conclusions * Using the new 24.4-m net configuration should benefit ODWC by decreasing the time spent afield deploying, retrieving, and working up nets. * The lack of significant differences in mean catch/24 hours when comparing 24.4-m and the 61-m nets shows that the new nets catch hybrid striped bass, saugeye, walleye, and white bass more efficiently. * Since 24.4-m nets are more efficient at catching hybrid striped bass, walleye, and saugeye the reduced efficiency for sampling channel catfish and white crappie should be an acceptable trade off, especially because the 24.4-m nets catch fewer non-target species. * The regression equations developed should help biologists revise current benchmarks for evaluating sportfish populations in Oklahoma. * The variability of the new 24.4-m nets at fixed sites was essentially equal to or less than the 61-m nets with the exception of hybrid striped bass. The reduced precision for hybrid striped bass could easily be offset by adding a few more sample sites to reservoirs with hybrid striped bass fisheries. * Channel catfish appear to be the only target species that had a significant change in length frequency distributions using the new 24.4-m nets. All other target species length frequency distributions appear to be fairly similar using each net type once adjustments are made for the influence of the shad mesh sizes. * Anecdotally, 2010 results appear to be similar to 2009 data. * 2009 and 2010 data will be pooled and analyzed * Recommend 2011 sampling year be eliminated since data is consistent from first two years. III. Significant Deviations: None Prepared by: _ Ryan Ryswyk, Fisheries Biologist Date: ~ HM? \.( Approved by: G£f?j2::? C:f'J~ _ Fisheries Administration Oklahoma Department of Wildlife Conservation ~n ~ ~ ~ ~.~~-- I ohn Stafford r ~ Federal Aid Coordinator Oklahoma Department of Wildlife Conservation IV. Literature Cited Anonymous. 1952. From a report on study of nylon and linen gill nets. The Progressive Fish Culturist. 14: 18-18. Anonymous. 1958. Small-mesh gill nets used in Wyoming studies. The Progressive Fish Culturist. 20:32-32. Boxrucker, J., and G. Ploskey. 1988. Gear and seasonal biases associated with sampling crappie in Oklahoma. Proceedings of the Annual Conference Southeastern Association ofFish and Wildlife Agencies. 42:89-97. Buckmeier, D.L., and J.W. Schlechte. 2009. Capture efficiency and size selectivity of channel catfish and blue catfish sampling gears. North American Journal of Fisheries Management 29:404-416. Dumont, S.c., and W. Schlechte. 2004. Use of resamp ling to evaluate a simple random sampling design for general monitoring of fishes in Texas reservoirs. North American Journal of Fisheries Management 24:408-416. Erickson, K.E. 1978. Standardized sampling procedures for lake and reservoir management recommendations. Oklahoma Federal Aid Project F-38-R, Job 1. Performance Report. 27p. Gablehouse, D., and 20 coauthors. 1992. Fish sampling and data analysis techniques used by conservation agencies in the U.W. and Canada. American Fisheries Society, Fisheries Management Section, Bethesda, Maryland. Guy, C.S., D.W. Willis, and R.D. Shultz. 1996. Comparison of catch per unit effort and size structure of white crappies collected with trap nets and gill nets. North American Journal of Fisheries Management 16:947-951. Hamely. J.M., Review of gillnet selectivity. 1975. Journal of the Fisheries Research Board of Canada. 32: 1943-1969. Hansen, R.G., 1974. Effects of different filament diameters on the selective action of monofilament gill nets. Transactions of the American Fisheries Society. 2:386- 387. Henderson, B.A., and S.l. Nepszy. 1992. Comparison in catches of mono- and multifilament gill nets in Lake Erie. North American Journal of Fisheries Management. 12:618-624. Hogrnan, W.J., The relative efficiency of nylon gill nets after transition from cotton nets in a multispecies fishery. Transactions of the American Fisheries Society. 102:778-785. Hubert, W.A. 1983. Passive capture techniques. Pages 95-122 in Nielsen, L. A., and D.L. Johnson, editors. Fisheries Techniques. American Fisheries Society. Bethesda, Maryland. Jester, D.B., 1977. Effects of color, mesh size, fishing in seasonal concentrations, and baiting on catch rates of fishes in gill nets. Transactions of the American Fisheries Society 106:43-56. Johnson, D.H., 1999. The insignificance of statistical significance testing. The Journal of Wildlife Management 63:763-772. Machiels, M.A.H., M. Klinge, R. Lanters, and W.L.T. van Densen. 1994. Effect of snood length and hanging ratio on efficiency and selectivity of bottom-set gillnets for pikeperch, Stizostedion lucioperca, and bream, Abramis brama. Fisheries Research 19:231-239. Minns, C.K., and D.A. Hurley. 1988. Effects of net length and set time on fish catches in gill nets. North American Journal of Fisheries Management. 8:216-223. Miranda, L.E., and J. Boxrucker. 2009. Warmwater fish in standing waters. Pages 29-42 in S.A. Bonar, W.A. Hubert, and D.W. Willis, editors. Standard methods for sampling North American freshwater fishes. American Fisheries Society, Bethesda, Maryland. Noble, R.L., D.l. Austen, and M.A. Pegg. 2007. Fisheries management study design considerations. Pages 31-49 in C.S. Guy and M.L. Brown, editors. Analysis and interpretation of freshwater fisheries data. American Fisheries Society, Bethesda, Maryland. Noble, R.L. 2002. Reflections on 25 years of progress in black bass management. Pages 139-146 in D.P. Philipp and M. S. Ridgway, editors. Black bass: ecology, conservation, and management. American Fisheries Society, Symposium 31, Bethesda, Maryland. Peterson, IT., and c.P. Paukert. 2009. Converting nonstandard fish sampling data to standardized data. Pages195-215 in S.A. Bonar, W.A. Hubert, and D.W. Willis, editors. Standard methods for sampling North American freshwater fishes. American Fisheries Society, Bethesda, Maryland. Reddin, D.G. 1986. Effects of different mesh sizes on gill-net catches of Atlantic Salmon in Newfoundland. North American Journal of Fisheries Management 6:209-215. Robinson, M.S. 1999. Evaluation of three gear types for sampling channel catfish in small impoundments. Pages 265-269 in E.R Irwin, W.A. Hubert, C.F. Rabeni, H.L. Schramm Jr., T. Coon, editors. Catfish 2000: proceedings of the international Ictalurid symposium. American Fisheries Society Symposium 24, Bethesda, Maryland. Robson, D.S., and H.A. Regier. 1964. Sample size in Petersen mark-recapture experiments. Transactions of the American Fisheries Society. 93:215-226. Sullivan, K.P., and C.M. Gale. 1999. A comparison of channel catfish catch rates, size distributions, and mortalities using three different gears in a Missouri impoundment. Pages 293-300 in E.R Irwin, W.A. Hubert, C.F. Rabeni, H.L. Schramm Jr., T. Coon, editors. Catfish 2000: proceedings of the international Ictalurid symposium. American Fisheries Society Symposium 24, Bethesda, Maryland. Santucci, V.J. Jr., D.H. Wahl, and D.F. Clapp. 1999. Efficiency and selectivity of sampling methods used to collect channel catfish in impoundments. Pages 317- 328 in E.R Irwin, W.A. Hubert, C.F. Rabeni, H.L. Schramm Jr., T. Coon, editors. Catfish 2000: proceedings of the international Ictalurid symposium. American Fisheries Society Symposium 24, Bethesda, Maryland. Wilde, G.R., and W.L. Fisher. 1996. Reservoir fisheries sampling and experimental design. Pages 397-409 in L.E. Miranda and D.R. DeViries, editors. Multidimensional approaches to reservoir fisheries management. American Fisheries Society Symposium 16, Bethesda, Maryland. Willis. D.W. Use of gill net data to provide a recruitment index for walleyes. North American Journal of Fisheries Management. 7:591-592. Table 1. Number of fingerling hybrid striped bass (HSB), fingerling and fry walleye, fingerling saugeye, and 152-254 mm channel catfish stocked in sample reservoirs from 2004 to 2009. Canton Kaw Thunderbird Tom Steed Waurika Year HSB walleye HSB walleye saugeye HSB saugeye c. catfish HSB saugeye 2004 80,650 5,895,500* 34,238 64,500 67,800 33,525 106,320 2005 73,167 3,891,449* 170,000 140,850 63,800 71,000 16,345 106,700 79,200 2006 1,900 3,691,142* 168,632 180,000* 120,900 65,500 16,000 67,850 24,000 2007 70,418 1,900,000* 165,350 170,000* 56,450 64,000 21,773 107,864 2008 36,778 3,930,000* 171,600 170,000* 122860 64,716 63,750 32,475 107,352 46,150 2009 74,044 3,700,000* 44,259 180,000* 122,084 64,000 20,096 110,690 26,208 * Denotes fry stockings. Table 2. Mean catch/24 hours, coefficient of variation of the mean (c.v. mean), F statistic and P values for two net configurations: 61-m net (eight 7.6-m panels with 13,16,19,25,38,51,57, and 76 mm mesh sizes) and 24.4-m (eight 3.1-m panels with 38,57,25, 44, 19, 64, 32, and 51 mm mesh sizes) at fixed sites and or random sites. Data were from 15 net-nights effort at five Oklahoma reservoirs. Significant results (P.:::: 0.05) are bolded. 61-m Fixed 61-m vs. 24.4-m Fixed 24.4-m Fixed 24.4-m FRiaxneddovms. 24.4-m 24.4-m Random Mean Mean Mean Target Species Catch/24 c.v. mean F statistic Pvalue Catch/24 c.v. mean F statistic Pvalue Catch/24 c.v. mean hours hours hours channel cat 7.6 0.19 Fl.l44 = 14.81 0.0002 3.6 0.23 F1,144=0.12 0..7329 4.3 0.25 hybrid striped bass 6.3 0.28 F 1,115= 1.11 0.2934 4.2 0.36 F1,115= 1.22 0.2718 3.4 0.33 saugeye 3.7 0.20 F1,86= 0.061 0.8061 3.2 0.20 F1,86= 0.50 0.4817 2.6 0.20 walleye 2.7 0.17 F1,28= 3.13 0.0879 1.7 0.18 F1,28= 0.20 0.6587 1.9 0.19 white bass 17.8 0.26 F1,144= 0.030 0.8622 8.5 0.21 F1,144= 4.81 0.0299 8.6 0.27 white crappie 14.9 0.23 F1,144= 7.29 0.0078 6.6 0.21 F 1,144= 1.36 0.2459 5.3 0.24 By Catch blue catfish 3.9 0.31 F1,115= 3.90 0.0505 2.1 0.45 F1,115= 1.46 0.2288 2.2 0.32 common carp 2.3 0.32 F1,144= 9.76 0.0022 0.9 0.47 F1,144= 0.90 0.3437 0.7 0.61 drum 10.6 0.26 F 1,144= 26.49 0.0000 2.6 0.32 F1,144=0.12 0.7287 3.1 0.36 flathead catfish 0.3 0.75 F 1,144= 1.86 0.1745 0.1 0.74 F1,144= 0.00 0.9667 0.1 0.68 gizzard shad 55.6 0.32 F1,144=8.11 0.0051 16.0 0.22 F1,144= 0.67 0.4138 16.2 0.19 longnose gar 0.4 0.51 F1,144= 0.46 0.4971 0.2 0.78 F 1,144= 1.95 0.1652 0.7 0.72 river carpsucker 1.6 0.39 F1,144= 13.37 0.0004 0.5 0.49 F1,144= 0.00 0.9989 0.5 0.54 shortnose gar 0.4 0.88 F1,144= 1.07 0.3017 0.2 1.00 F1,144= 1.75 0.1885 0.4 0.84 smallmouth buffalo 1.1 0.30 F1,144= 2.03 0.1562 0.8 0.38 F1,144= 0.26 0.6079 0.8 0.45 spotted gar 0.6 0.51 F1,144= 1.70 0.1940 0.3 0.78 F 1,144= 1.96 0.1632 0.5 0.62 threadfin shad 23.7 0.48 F1.86= 11.42 0.0011 1.0 0.54 F1.86= 0.40 0.5271 1.4 0.47 Table 3. Mean number of samples required to detect a± 12.5% (c.v. = 0.125) or ± 25% (c.v. = 0.25) change in the population for target species using 61-m(eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes) and 24.4-m (eight 3.1-m panels with 38, 57, 25, 44, 19,64,32, and 51 mm mesh sizes) gill nets at fixed and random sites in 5 Oklahoma reservoirs. Mean number of samples needed to detect target c.v. of mean Fixed 61-m Fixed 24.4-m Random 24.4-m c.y. = Cc V, = c.y. = Cc V. = CvV. = Cc V. = 0.25 0.125 0.25 0.125 0.25 0.125 channel catfish 12 41 17 58 22 72 hybrid striped bass 24 83 44 151 40 137 saugeye 22 79 13 43 14 46 walleye 8 30 10 33 10 37 white bass 23 78 13 46 23 79 white crappie 16 56 14 49 18 63 Table 4. Kolmogorov-Smimoff test results (KSa and P values) for length frequency comparisons of (A) 24A-m (eight 3.1-m panels with 38, 57, 25, 44,19,64,32, and 51 mm mesh sizes) vs. 61-rnnets (eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes), (B) 24A-m vs. 61nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 mm mesh sizes) nets, and (C) fixed 24A-m vs. random 24A-m nets. Significant results (P:S 0.05) are bolded. Canton Kaw Steed Thunderbird Waurika A. 24A-m vs. 61-m KSa P value KSa P value KSa P value KSa P value KSa P value channel catfish 1.64 0.0091 1.85 0.0022 hybrid striped bass 2A7 0.0000 1.25 0.0883 0.84 OA872 saugeye 1.63 0.0096 1.00 0.2655 walleye 1.06 0.2148 white bass 4.66 0.0000 1.76 0.0041 0.94 0.3357 0.94 0.3442 1.87 0.0019 white crappie 5.69 0.0000 3A9 0.0000 3.24 0.0000 OA9 0.9702 B. 24A-m vs.61nsm-m channel catfish 1.59 0.0126 1.59 0.0125 hybrid striped bass 1.03 0.2433 1.24 0.0908 0.83 0.4944 saugeye 1.62 0.0107 1.00 0.2655 walleye 1.04 0.2254 white bass 1.43 0.0341 1.48 0.0251 OAO 0.9976 0.91 0.3757 1.05 0.2202 white crappie 1.17 0.1308 1.19 0.1176 1.24 0.0929 0.56 0.9086 C. Fixed vs. Random channel catfish 1.52 0.0198 0.63 0.8290 hybrid striped bass 1.29 0.0731 0.95 0.3307 0.61 0.8541 saugeye 1.19 0.1201 0.84 OA732 walleye 0.69 0.7301 white bass 1.32 0.0614 0.82 0.5171 0.83 OA919 1.26 0.0853 white crappie 0.87 OA378 0.63 0.8266 0.77 0.5992 0.67 0.7565 channel catfish hybrid striped bass 4 - Y= 0.8,~49x +Y.TB 4 1 y = 0.9449x + 0.5181 ew R - 0.7399 • c -' R2 = 0.9059 c 3 - P=0.0539. w 3 • • TS e~ I P = 0.0754 2 - 2 J • I< / 1 1 • I( F = 8.79 F = 11.78 0 0 Vl +-' 0 1 2 3 4 0 1 2 3 4 Cl) c E saugeye and walleye white bass IM I.D '+0- 4 - C 5 1 Vl Y = 1.0202x + 0.1765 ./ Y = 1.508x - 0.9889 .c ':-:l R' =0.84)Y.TS 4 -, R' = 0.6838?.0.c 3 - TS j P = 0.0962 • TB """ P = 0.0894 <N, 3 ..c 2 , u I< +ro-' .w 2 U crCol) 1 " F = 9.71 1 F = 5.74 ~ '+- 0 0 f---------,-- -T- ~---.- - - --- -- ----- --- 0c 0 1 2 3 4 0 1 2 3 4 -l white crappie F = 4.59 K= Kaw C = Canton 5 4 -j I[ 3 -, t 2 1 y = 0.6635x + 1. 7281 • T6 R2 = 0.602\c _ .~.------ P=O.1215 ----~ ~ ew TS = Tom Steed TB = Thunderbird 1 ., W = Waurika o 1-. -,-- -- __~ o 1 2 3 4 Ln of Mean Catch/24 hours 24.4-m nets Figure 1. Linear regression relationships of the natural log (Ln[ x+ 1]) of mean catchJ24 hours of 24.4-m (eight 3.l-m panels with 38,57,25,44,19,64,32, and 51 mm mesh sizes) nets plotted against the Ln of mean catchJ24 hours of61-m (eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes) nets for channel catfish, hybrid striped bass, saugeye, walleye, white bass, and white crappie in five Oklahoma reservoirs. Significant results (P:S 0.10) are bolded. Thunderbird Channel Catfish Canton 8% 6% 4% 2% 0% ' --2.00 N= 174 N= 81 15%, 200 N = 103 80 N = 68 10% .;. 5% v 10% o~v l-< 8% l-< ::s u u 6% 0 '+0-< 4% ~ 2% vul -< Pv-. 0% 10% 8% 6% 4% 2% 0% 00 rl 200nsill N = 155 N = 81 0% ~ TLmm Figure 2. Length frequency distributions of channel catfish caught in 61-m(eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes)' nets vs. 24.4-m (eight 3.1-m panels with 38,57,25,44,19,64,32, and 51 mm mesh sizes) nets at fixed sites (top row), 61nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 mm mesh sizes) nets vs. 24.4-m nets at fixed sites (middle row), and 24.4-m fixed site nets vs. 24.4-m random site nets (bottom row), Length frequencies with P < 0.05 are distinguished with a box around the graph. ......'''''' :.:. ••••••• 80 random N = 126 80 fixed N = 81 ••• "t --'..I~ '"--'; 0 0 0 0 0 0 0 0 0 0 N r<) <r Lf) 1.0 roo-- c00o 15% 10% 200nsm N = 95 80 N = 68 5% .~ l~' I'Ii J,"1\\ .r. II' II ~t,ll,,i.,, , , l ' ,'(, f \. • 1.,"\ O%! -.,,4 \,: " #;.~. 15% 80 R N = 56 80FN=68 5% 10%· 0% Hybrid striped bass Canton Tom Steed Waurika 20% --200 N=83 15% 200 N = 80 20% 200 N = 181 80 N = 46 80 N = 93 80 N = 92 15% 15% 10% - 10% ) 10% 5% ~_A __ )\ 5% 5% 0% ~/\ _"!\/W~,v..~lV\"_A_ O'?~ 0% I. __1'/'1.-:-... -,J <~.) 20% 15% 20% ---- 200mmN = 53 200nsm N = 77 200nsm N = 180 ~~ !-< 80 N = 46 80 N = 93 80 N = 92 . !-< 15% - ;:::l 15% <.) <.) I", 10% 0 10% " 10% 4-< I, 0 ,'. I~~ SOL , ~ . 1'1 ~~ 5% i " 1 _,r "\ ,0 1I II 5% o !-< i" - "','. ,~, 'II\! II", \ 1" ",I ~ ",.1\ j ~ 0% )'!o./',~,,\. _', f ,I' ,!, •• i" \.'~,s: 0... ' ~ .. ' •• t- - !,~:......>"'" 0% , - ~ ,',~ _·1· r -- 0% - 1- _\_~,-'- •• 15%, ! 10% ~ 50/. ,0 -" '. \ a' L...,.:,..i. ,\/,¥~\..j.. ooM 0N0 ~00 ~00 ~000000 ~ ~ .: . 20% ••••••• 80 random N = 47 80 fixed N = 46 15%, 80 R N = 45 80 F N = 93 15% " 0% :·...:·.:'·."·....1,~1r_0;·_::..•:.~. ') -• .< ,, " .' 'f .• lilt. • .g .f! ;... '},' . . ~. • '·~·,"-"--"'T ~ • 5% ....,' .M." ... ~ :::: .:: a. .,' a, a. • , • '. t. • '. t •• _" '.' '" a. ". ~ ~ :: ::: _,.,:<:,:::~_.I_~:=--~~~:j.d ~: 1-:' .: - ..--..- t~~-' __ !:_~_~_.-II..·.··--, 10% ... ..'' .', .....'"'...' 80 R N = 91 80 F N = 92 5% .. .. ''.. ""<, " ..... 10% 0% 0% o 0 ~o ~0 TLrnm Figure 3. Length frequencies of hybrid striped bass for 61-m mesh (eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes) vs. 24.4-m (eight 10-foot panels with 38,57,25,44, 19,64,32, and 51 mm mesh sizes) nets at fixed sites (top row), 61nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 rnm mesh sizes) nets vs. 24.4-m nets at fixed sites (middle row), and 24.4- m fixed site nets vs. 24.4-m random site nets (bottom row). Length frequencies with P < 0.05 are distinguished with a box around the graph. oor l oo N oor <l oo~ oo~ oo~ oo~ ooM oo N oo~ ooco oor-, Canton walleye Tom Steed saugeye Thunderbird saugeye 15% -200 N = 145 15% 10% so N = 97 200 N = 107 200 N = 77 80 N = 114 8% 80 N = 42 10% 10% f' , tjW\~~ i ~I\ 6CY~ V 4% 5% ,\ 5% J fJ ) IJ\ 2%~~I \I 0% 'J -. -- "~I;"f:-.;, 1\ ~ 0% '\ :.:.I\_-J~ -_.),.... . \.. 0% 0% 1- 15% 15%] 200nsm N = 105 80 N = 114 ___ - 200nsIll N = 144 80 N = 97 10% 10% 5% 5% 0% 12% l 10% 14% 12% 10% 8% 6% 80 R N = 66 80 F N = 114 4% 4% - J ~ I " I .:'.:.:-. .". i :.:t-: : ~! .••••••... I ••• J : -: :.,:.: ... I - - :, r-.«: •.:: '" +..::;~~.- ..;: "r· - Y'~:-,,---,---,'= ~..~ . ., ••••• SO random N = 107 :: 80 fixed N = 97 8% .:''. , : 2% 0% 6% 2% 0% 10% - 200nsm N = 77 80 N = 42 8% -I I 6% •". 11,\' '~, - ",I I , " , , .t , ~ I ; 1 1\" '\ I , ~-5., -\..,1-,- -- 4% 2% 0% 14% 12% 10% 8% 6% 80RN=60 • ~ 80 F N = 42 -... :.....,_.....-...........................•........._.. ;..:..e..."".".......... }\ /iT,~l_,lL '::~~'\:~+"".-, 4% 2% 0% 0000000 00000000000 00000000000 0000000 OLflOLflOLflOLflOLflO OLflOLflOLflOLflOLflO N ('() <::t Lfl 1O r-, CO NN('()('()<::t'<:ttJ)tJ)lOlOf'-. NN('()('()<::t'<:ttJ)tJ)lOlOf'-. TLmrn Figure 4. Length frequencies of walleye and saugeye for 61-mnets all mesh (eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mrn mesh sizes) vs. 24.4-m (eight 3.1-m panels with 38, 57,25,44, 19,64, 32, and 51 mm mesh sizes) nets at fixed sites (top row), 7.6nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 mrn mesh sizes) vs. 24.4-m nets at fixed sites (middle row), and 24.4-m fixed site nets vs. 24.4-m random site nets (bottom row). Length frequencies with P < 0.05 are distinguished with a box around the graph. Canton Kaw 25% 20% 15% , 10% 5% 0% -LOO N=403 25% 80 N = 111 20% 15% 10% 5% 0% 200 N = 53 80 N = 46 ----200nsm N = 199 80 N = 111 25% ~, I. 20% 1 "~ V '" ",," :Jt q" 15% I , , .! 1,1 10% '. I ,. I • II .~I , , .' r- I II~ , J .H. 5% i Zu 'j • J ~ , 'Ii .-\.". •.... "1 ~.. \.t~ " 0% Q) 10% () I=i Q) l-< 8% l-< ;l o 6% o0 4-< 0 4% ~ Q) 2% () l-< Q) 0... 0% 20% 15% 10% 5% 0% 200nsm N = 48 80 N = 46 ~ """,. ,I •II " , I·, /I" /', " ._ J I' ", a','" , .....U..."..~•. ••••••• 80 random 80 fixed I J I - ,. j ;; hi 1 ,':,";' •• • •••• r : ••• '-··r'--" I If - 1 aO 'J a O'J N aO 'J r<') N = 88 N=111 12% 'I 10% j Il i 8% 6% 4% 2% 0% Random omitted N <40 a O'J a O'J r<') white bass Tom Steed 10% 200 N = 187 80 N = 123 8% 10% 200nsm N =161 80 N = 123 8% ~ 6% 4% 2% , I , '\ I 0% ~ I~,_, 14% , I 80 R N = 176 80 F N = 123 12% 10% 8% 6% 4% 2% 0% ...•.. .,."''.... "'" . g.:..:;..'.t~:....~ r: ::: .... .. .::: . . +L~.: .. .., Thunderbird 20% 200 N = 111 80 N = 102 15% Waurika 15% 10% 10% . Ji~ I 5% 5% 0% ~_;''- 0% 200 N = 254 80 N = 104 20% 200nsm N = 108 15% 80 N = 102 15% -j 10% 14% 80 R N = 93 80 F N = 102 a a a a a O'J O'J O'J O'J O'J .-< N r<') <r TLmm Figure 5. Length frequencies of white bass for 61-mnets all mesh (eight 25-foot panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes) vs. 24.4-m (eight 3.1-m panels with 38,57,25,44, 19,64,32, and 51 mm mesh sizes) nets at fixed sites (top row), 61nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 mm mesh sizes) nets vs. 24.4-m nets at fixed sites (middle row), and 24.4- m fixed site nets vs. 24.4-m random site nets (bottom row). Length frequencies with P < 0.05 are distinguished with a box around the graph 12% 1 ..." ... , ",,.. -t .I. .:. :.. ~, J.'" ..••••• I•· ......•. ." . .,. ,.,."".. 1 ..•. l" I :: f= : ~ .:' 2% '" •• ,' s > : +I ::--;:--;•.•: :......:~•. 1 a a a a a O'J O'J O'J O'J O'J .-< N r<') -cr 10% 8% 6% 4% 0% 10% 5% 0% 200nsm N = 206 80 N = 104 ,,\ fi I 1 ~"I.,,, il. II,. 14%] 12% 10% 8% 80 R N = 112 80 F N = 104 6% 4% 2% 0% ," . =.:..". ::,'. ;~:.. :::::',: :: ,..::.:.'" ,:: : ::. .:....:.- ::,-:.:-;,:..", a a a a a O'J O'J O'J O'J O'J .-< N r<') '<t Kaw Tom Steed white crappie Thunderbird Waurika 25% 200 N = 385 25% 200 N = 59 80 N = 185 80 N = 76 20% 20% 15% l~\ 15% 10% 10% 5% 5% 1\- 0% +) ..;\J.._,. .." __'_~""A"~'~~"' __ 0% , 40% -200 N=195 80 N = 70 25% 200 N = 141 80 N = 43 30% . ~ 20% 10% 0%, 15% l - - - - 200nsill N = 55 1 1 :. 80 N = 70 " , I , r I ,I - ~. # " "\",. n "I':" ~J'" I :..'",I,' ,,,.I"', ,Ir. :\ -Y' -, .~I.-,--L..JI"-,-----.-- 10% 5% 0% 14% 1 ....... random 80 N = 58 12% fixed :80 N =70 10% ~ : I , " ,. 8% " ., "" . , . 6% .. : ..10,1"0 0" " ~:.~ . , ~: ' " 4% .:: '" f :: :. " , .' .... " ;..... 2% 0' .10 ••• :-::- ': ".. ," 0 0 " ,. : 0% w ..•.•.-.-....' .• r0-, N0 r0-, 0N 0r--, N0 f0'. rl rl N N rn rf) 20% 0.3 0.25 0.2 0.15 J 0.1 -i 0.05 -I , "/1 I I , I I\ 200nsm N = 257 25% 80 N = 185 20% 15% 10% 5% 0% ~~/S~\'_l_i 80 R N = 142 30% 80 F N = 185 25% 20% ~ 0, "00 , , ,,00 ,, o. 00 o , 10%' 5% 200nsm N = 55 80 N = 76 • J \ \ 0% o ~I .~~-•..••• -' ----,---- 25% 80 R N = 41 80 F N = 76 20% 200nsm N = 66 80 N = 43 20% ... 10% 15% -i .. "0' .n:;t\ "~ . 5% 0% 20% 80 R N = 52 80 F N = 43 15% 1 10% 10% 15% -i ,.0"""",,, :. - 5% 5% 10% -t 5% ! •• 0 .0'...• ·· .... • • 0 .' . " , :; ~::.:..-:\:..•• -I· . 0% +.. ~ .._:.-,-----.--=-r:::.,1:0.. ~ _ ~ • 0% ! :.: 0% ~ •••• ,,,~.-; 000 r-, N r-, N rf) rf) TL mm o 0 N r-, •..•,• rf) ro-, N0 rl ro-, N0 .-{ N or-, N Figure 6. Length frequencies of white crappie for 61-mnets all mesh (eight 7.6-m panels with 13, 16,19,25,38,51,57, and 76 mm mesh sizes) vs. 24.4-m (eight 3.1-m panels with 38,57,25,44,19,64,32, and 51 mm mesh sizes) nets at fixed sites (top row), 61nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 mm mesh sizes) nets vs. 24.4-m nets at fixed sites (middle row), and 24.4- m fixed site nets vs. 24.4-m random site nets (bottom row). Length frequencies with P < 0.05 are distinguished with a box around the graph.
Object Description
Description
Title | Gillnet sampling 2010 |
OkDocs Class# | W2800.7 F532r/w no.10 F-50-R-17 2010 |
Digital Format | PDF, Adobe Reader required |
ODL electronic copy | Deposited by agency in print; scanned by Oklahoma Department of Libraries 6/2011 |
Rights and Permissions | This Oklahoma state government publication is provided for educational purposes under U.S. copyright law. Other usage requires permission of copyright holders. |
Language | English |
Full text |
W 2800.7 FS32r/w
F-SO-R-17 2010
no.30
c.1
FINAL REPORT
FEDERAL AID GRANT NO. F-50-R-17
PROJECT NO. 30
Fish Research and Surveys for Oklahoma Waters
Gill net sampling: An Evaluation of Oklahoma's Standard
Sampling procedures
OKLAHOMA DEPARTMENT OF WILDLIFE CONSERVATION
JANUARY 1, 2010 through DECEMBER 31, 2010
FINAL REPORT
State: Oklahoma Grant Number: F-50-R-17
Grant Title: Fish Research for Oklahoma Waters
Project Number: 30
Project Title: Gillnet sampling: An Evaluation of Oklahoma's Standard Sampling
Procedures
Project Leader: Greg Summers
Grant Period: From: January 1,2010 to: December 31, 2010
I. Project Objective: Determine 1.) difference in catch rate and length frequency of previous
and current SSP gill net configurations at historic fixed sites and 2.) difference in catch
rate and length frequency of fixed and random sites using current SSP gill nets.
II. Summary of Progress:
A. Introduction
Gill nets are a passive capture gear that catches fish by entanglement. They are
underwater walls of netting usually set in a straight line (Hubert 1983; Miranda and
Boxrucker 2009). Fish are caught by either being wedged-held by mesh around the body,
gilled-held by mesh slipping behind the opercula, or tangled-caught by spines, teeth, or
other protrusions without actually penetrating a mesh. Therefore, mesh size is an
important factor influencing the size of fish captured (Reddin 1986; Hubert 1983;
Miranda and Boxrucker 2009). Other factors that influence catch of gill nets include
color, season, baiting (Jester 1977), net length, set time (Minns and Hurley 1988) use of
mono- or multi-filament material to construct the net (Anonymous 1952; Hogman 1973;
Henderson and Nepszy 1992), size of filaments (Hansen 1974), and hanging ratio
(Machiels et al. 1994). Gill nets are one of the most widely used fisheries gears in the
United States and Canada (Gablehouse et al. 1992); therefore many variations exist,
which can make comparisons between nets difficult.
Standardized sampling procedures are important in fisheries management and are used to
evaluate fish populations over time and between reservoirs in a region or state (Noble
2002). In 1977 the Oklahoma Department of Wildlife Conservation (ODWC) developed
"Standardized Sampling Procedures (SSP) for Reservoir and Reservoir Management
Recommendations" (Erickson 1978). This protocol called for the use of gill nets 61-m
long by 1.8-m deep (0.5 basis) with 7.6-m mesh panels with bar mesh sizes of 19,25,38,
51,57,76,89, and 102 mm respectively. Since that time, Oklahoma's SSP have been
revised to obtain more accurate population parameters as well as maximize efficiency.
In 2001, the ODWC made a change to its SSP gill net configuration. Two, 7.6-m panels
were added to the nets (13- and 16-mm bar mesh) and two panels were removed (89 - and
102-mm bar mesh). These changes were made with two objectives in mind. The first
was to target shad (Dorosoma spp.) that were previously not represented in gill net
samples and the second was to reduce the catch of rough fish (L. Cofer, ODWC, personal
communication). In 2009, the ODWC further refined their shad sampling and began
using separate floating 'shad' gill nets with small bar mesh sizes specifically designed to
target shad. The addition of shad nets to the ODWC's SSP made the data from the 13-
mm and 16-mm mesh sizes of the SSP gill net configuration no longer necessary. The
current SSP revised in 2009 specified the number of gill net samples required as a
function of impoundment surface area: less than 40 ha, not more than 5 sites; 40-404 ha,
5 sites; 404-2023 ha, 10 sites; greater than 2023 ha, 15 sites (Kuklinski, ODWC, personal
communication).
Miranda and Boxrucker (2009) offer gill net standards for the entire southern USA with
the intention of regulating gill net selectivity, deployment, effort, and timing of collection
to reduce the variability that prevents adequate comparisons among agencies. They also
considered sampling efficiency to account for typical time and labor restrictions. The
ODWC has adopted the current standard net configuration suggested by Miranda and
Boxrucker (2009) consisting of 24.4-m long by 1.8-m deep (0.5 basis) nets composed of
3.1-m panels with bar mesh sizes of38, 57, 25, 44,19,57,32, and 51 mm respectively.
Oklahoma's previous gill net specifications were 61-m long by 1.8-m deep nets
composed of7.6-m panels with bar mesh sizes of 13,16,19,25,38,51,57, and 76 mm.
Because biologists use gill nets to look at trends in fish populations over time, it is
beneficial to compare the catch rates and length frequencies of these two net designs
during the same sampling season in order to understand the consequences of changing net
configurations. When comparing different gear types, both gears should be used at the
same time and same location. This method provides a way to convert historic data
collected with one gear to data collected with another (Peterson and Paukert 2009; Noble
et al. 2007).
The ODWC's current SSP protocol also requires sample locations to be fixed sites
selected by the biologist. Fixed sites chosen in this manner are useful for monitoring
changes in the population over time but are potentially more biased than random sites
with respect to abundance and length frequencies of fish they capture (Wilde and Fisher
1996). Fixed sites can produce greater precision, which allows fewer samples to achieve
a similar level of precision; however, these sites may not accurately represent the entire
fish population and population characteristics of these sites may change at a
disproportional rate to the rest of the reservoir (Noble et al. 2007). Random sampling
may give a more accurate assessment of sportfish populations and allow biologists to
make more reliable comparisons between reservoirs, but at the possible cost of higher
required replication to achieve the same precision. Therefore, direct comparison of fixed
versus random sampling needs to be made to determine if additional sampling effort
would be needed in a random sampling design.
Moronids and percids are the primary target species for the ODWC's gill net SSP,
specifically hybrid striped bass Morone chrysops x Morone saxatilis, white bass Morone
chrysops, walleye Sander vitreus, and saugeye Sander canadense x Sander vitreus.
Information gathered from gill nets about white crappie Pomoxis annularis, and channel
catfish lctalurus punctatus are important to the ODWC, but are considered secondary
target species.
B. Methods
Five Oklahoma reservoirs (Canton, Thunderbird, Kaw, Waurika, and Torn Steed) were
selected for sampling in 2009 based on stockings of hybrid striped bass and either
saugeye or walleye within the previous five years (Table 1). Four Oklahoma reservoirs
(Ft. Cobb, Foss, Skiatook, and Torn Steed) were sampled in 2010. Gill netting took place
in October and November.
All sample reservoirs had natural populations of white bass and channel catfish; however,
Tom Steed had supplemental stockings of channel catfish. Although Thunderbird was
not stocked with hybrid striped bass it was sampled because was considered Oklahoma's
best saugeye fishery. Waurika was only stocked periodically with saugeye but was
sampled because it was considered Oklahoma's best hybrid striped bass fishery. Canton
was considered Oklahoma's best and most productive walleye fishery. Each reservoir
received a total of 45 net nights of effort (15 net nights of 61-rnnets at fixed sites, 15 net
nights of 24.4-m nets at fixed sites, and 15 net nights of 24.4-m nets at random sites) for a
total of 225 net nights of effort for all five reservoirs combined. Gillnet sampling began
October 5t
\ 2009 and continued through November 6th
, 2009 in accordance with
ODWC's SSP recommended time frame.
Historic gillnet sites were identified at each reservoir and five 61-m long by 1.8-m deep
nets composed of eight 7.6-m panels with bar mesh sizes of 13, 16, 19,25,38, 51, 57,
and 76 mm were set perpendicular to the shore line and fished overnight. Nets were
pulled the next day and five different fixed sites were sampled similarly each of the next
two nights such that 15 different sites were sampled over three consecutive nights. Each
night, another five 24.4-m long by 6-ft deep (112 basis) nets composed of eight 3.1-m
panels with bar mesh sizes of38, 57,25,44, 19,57,32, and 51 mm were set
perpendicular to the shoreline at the same historic sites approximately 90 meters away
from the 61-m nets. Nets were set this way in order to give the same schools offish the
opportunity to pass through each of the nets and become entangled. In addition, five
24.4-m nets (as described above) were set perpendicular to the shoreline at random
locations each night such that 15 different random sites were sampled over three
consecutive nights. Random locations were selected by placing a 274 by 274 meter grid
over a map of the reservoir and grid numbers were randomly selected. Only grid sites
with a depth of at least 1.8 m within approximately 46 m of the shoreline were considered
to ensure that nets could hang stretched out from top to bottom. All nets were fished for
a period of 17 to 24 hours. The number of fish of each species in each net was divided by
the number of hours that net was fished and multiplied by 24 to give catch per 24 hours
(catch/24 hours). Mean catch/24 hours was then calculated by averaging the catch/24
hours from all sites at each lake.
Gillnets were pulled and fish were processed at a work station. Catch for each mesh
panel was recorded separately. Total length (mm) and weight (g) were recorded for game
species. Only total length was recorded for non-game species. Coefficient of variation
(c.v.) was calculated for each species in each net configuration (fixed and random sites
analyzed separately). The number of samples needed to detect a target c.v. of 0.25 and
0.125 were calculated using the random resampling method of Dumont and Schlechte
(2004). Catch/24 hour data were log transformed (In[X+0.0004]) to correct for normality.
A two-way analysis of variance (ANOVA) (net length treated as a fixed factor and
reservoir treated as a blocking variable) was used to test for differences in mean catch/24
hours between 24.4-m and 61-mnets. A two-way ANOVA (site type treated as a fixed
factor and lake treated as a blocking variable) was also used to test for differences in
mean catch/24 hours between fixed and random sites. Linear regression was used to test
for relationships between the In[X+1] mean catch/24 hours of 24.4-m and 61-mnets at
fixed sites for five target species.
Kolmogorov Smirnov (KS) tests were used to test for differences in length frequencies
(10-mm length groupings) of 24.4-m and 61-mnets with all mesh panels included, 24.4-m
and 61-mnets with no shad mesh panels (i.e., lacking 13 and 16 mm shad mesh; 61nsm),
and 24.4-m nets set at fixed and random sites. The KS test only allows for comparison of
two populations at a time, therefore length-frequencies of each species were tested in
each reservoir. Net configurations with sample sizes ofN < 40 were omitted from length
frequency analysis. All statistical tests were evaluated as significant if P :::;0.05 with the
exception of the linear regression analysis, which was evaluated at P :::;0.10 due to the
increased variability of catch rates from individual nets.
C. Results
24.4-m vs. 61-m at Fixed Sites
Catch rates
The 24.4-m and 61-mnets had similar mean catch/24 hours for most of the target species
(Table 2). However, channel catfish and white crappie had significantly lower mean
catch/24 hours in the 24.4-m nets. Because only 6 walleye were caught at Kaw (all net
types combined), the walleye catch at Kaw reservoir was excluded from all analysis. The
24.4-m nets had lower mean catch/24hours for most non-target species (by-catch; Table
2). Variability (c.v. of the mean) of the 24.4-m nets was similar to (within 5 units) or less
than the 61-mnets for the target species with the exception of hybrid striped bass, which
was only slightly higher. Using the 15 samples currently required by the SSP (Kuklinski,
ODWC, personal communication), the 61-mnets were able to detect a 25% increase or
decrease in the population (c.v. = 0.25) only for channel catfish, walleye, and white
crappie; although white crappie would only require one additional sample (Table 3).
With 15 samples the 24.4-m nets were able to detect a ± 25% change in the population
for all target species except hybrid striped bass and channel catfish. Channel catfish
would only require two additional samples, but hybrid striped bass would require
considerably more effort. Neither net type was able to detect a ± 12.5% change in the
population (c.v. = 0.125) for any species with 15 samples.
Regressions of mean catch/24 hours explained from 60.25% (white crappie) to 90.59%
(hybrid striped bass) of the variation in catch rates between 24.4-m and 61-rnnet types
(Figure 1). The only target species with a slope not significantly different from zero was
white crappie.
Length Frequency
Just over half (59%) of the KS tests comparing 24.4-m and 61-m nets were significantly
different (top panels, Figures 2-6; Table 4a); however, only 29% of the test comparing
24.4-m and 61nsm-m nets were significantly different (middle panels, Figures 2-6; Table
4b). The influence of the shad mesh is especially apparent for white crappie (Figure 6)
but can also be seen to a lesser degree for hybrid striped bass (Figure 3) and white bass
(Figure 5). Eliminating the shad mesh in the 61nsm-m nets did not improve the length
comparisons for channel catfish (Figures 2, top and middle panels).
Fixed vs. Random sampling sites
Catch rates
The mean catch/24 hours of most target species from fixed and random sites were
similar. Only white bass differed significantly (Table 2). Variability at random sites was
slightly higher than that at fixed sites for all target species except saugeye and walleye,
which were essentially equal (Table 2). Using either sample site strategy, it is not
possible to detect a ± 12.5% change in the population for any of the target species with
only 15 net nights (Table 3). Using the random site sampling strategy, it was possible to
detect a ± 25% change in the population with only 15 net nights for saugeye and walleye.
An additional 3 nets would make this increase or decrease detectable for white crappie.
Length Frequency
The length frequencies from 24.4-m nets at fixed and random sites were typically similar
(Table 4). The only reservoir where the length frequency of a target species differed was
Canton (channel catfish; bottom panel, Figure 2; Table 4).
D. Discussion
24.4-m vs. 61-m
Catch rate
Because there were no significant differences in mean catch/24 hours and the 24.4-m nets
are less than half the length of the 61-mnets, the new 24.4-m nets are more efficient at
catching hybrid striped bass, saugeye, walleye, and white bass than the previously used
nets. This should be an advantage for state agencies and researchers considering the gill
net configuration offered by Miranda and Boxrucker (2009) because the 24.4-m nets
were much easier and faster to deploy and retrieve. By-catch was also lower in 24.4-m
nets, so total processing time of these nets is lower than for 61-m nets.
Mean catch/24 hours for white crappie were significantly lower in the 24.4-m nets
suggesting the 61-mnets may be better for this species. However, most biologists use
trap nets as the main sampling gear for crappie (Kuklinski, ODWC, personal
communication). Boxrucker and Ploskey (1988) found trap nets to have higher catch
rates and less variability in catch rates and length frequency distributions of white crappie
than electro fishing and gill net samples. Guy et al. (1996) recommends trap nets over gill
nets for sampling crappie because trap nets catch rates give a better index of abundance
than gill net catch rates, although size structure data is similar for both gears. Any
significant change in crappie gill net data due to the switch to a new net configuration
may be offset by the efficiency of the 24.4-m nets to catch other target species as well as
the availability of other gear types to sample crappie.
Channel catfish mean catch/24 hours were also significantly lower for the 24.4-m nets.
Previous research comparing gears for channel catfish have produced conflicting results.
Two studies addressing small impoundments recommend experimental gill nets over
baited hoop nets, baited slat traps (Robinson 1999) electrofishing (AC), baited wooden
and wire fish traps, and trot lines (Santucci 1999). Sullivan and Gale (1999) recommend
25.4-mm-mesh or variable-meshed baited hoop nets in a series over experimental gill
nets in larger reservoirs. The conflicting results of these studies suggest there is
considerable variability with respect to channel catfish sampling. If biologists continue
to use gill net data to track trends in channel catfish populations, they should develop new
benchmarks based on the reduced efficiency of the 24.4-m nets for this species.
Catch rates of 24.4-m and 61-mnets were correlated for every target species except white
crappie, but only when using a significance level of P = 0.10. Given the inherent
variability of gill net catch rates, this reduced significance level is justifiable. Robson
and Regier (1964) report that management studies commonly use more liberal P-values
than research studies. Understanding the relationship of mean catch/24 hours between
the two net types will be critical for biologist to understand when making management
decisions based on either historic or current data (Peterson and Paukert 2009). One way
these equations could be useful is for making adjustments to stocking criteria. According
to the ODWC's stocking criteria a reservoir is considered an established hybrid striped
bass fishery ifit has a gill net mean catch/24 hours of 2.4 fish using the 61-mnets. Using
the equation produced from the hybrid striped bass regression analysis indicates a mean
catch/24 hours of 2.4 fish from 61-mnets is equivalent to a mean catch/24 hours of 1.1
fish using the new 24.4-m nets. These types of conversions will need to be made for all
species. However, the regression is based on a relatively small sample size where at times
a lake had values very different than others and influenced the slope more than other
points.
Using either net configuration it would be impractical to set enough nets to detect a ±
12.5% change in the population for any of the target species. The 61-m nets were able to
detect a ± 25% change in the population for only channel catfish and walleye with 2: 15
samples. The 24.4-m nets were able to detect a ± 25% change for all target species
except hybrid striped bass and channel catfish with 15 samples. Precision of channel
catfish catch rates were very close to the targeted precision, requiring only 17 samples
(Table 3). The better precision of the 24.4-m nets suggests it would not be necessary to
increase sample size requirements for this new net design, however we do not
recommend reducing effort to less than 15 samples/reservoir as precision is still just
acceptable at this level. In fact, it may be beneficial for biologist to increase sample sizes
on reservoirs where hybrid striped bass are a management concern. Due to the ease of
use of the 24.4-m nets, the addition of 12 net nights (4 additional nets each of the 3
sample nights) should not drastically impact the amount of time spent in the field when
compared to the amount of time spent running and processing the previous 61-mnets.
Length Frequency
For every target species except walleye, at least one reservoir showed a difference in
length frequency between the different length nets. To better understand for the
mechanism underlying these differences it is necessary to examine the construction of the
two nets and the purpose of each included mesh panel.
It is important to consider sampling objectives when comparing gear types (Peterson and
Paukert 2009). The 61-mnets have 13- and 16-mm mesh panels, which are not found in
the 24.4-m nets. These panels were included in the 61-m net to specifically target small
shad that are vulnerable to predation-thus making up the forage base for predators (Cofer,
ODWC, personal communication). These meshes not only catch small shad but also
other small fish. The information gained about small non-shad species could. be useful
when looking at things such as post-stocking survival or young of year abundance
(Anonymous 1958; Willis 1987). However, as a standardized gear to collect information
about age and growth, length frequency, and relative abundance of predatory fish, these
mesh sizes would be less useful. When assessing predatory fish stock size and size
structure (e.g, for setting creel regulations, length limits, and other management
decisions), generally only fish that have recruited to a size that is catchable by anglers are
used. Therefore, we suggest the 24.4-m nets are more efficient for predatory fish
assessment and that separate gill nets should be used when data on smaller fish is desired,
such as the ODWC now does for forage assessment (Kuklinski, ODWC, personal
communication).
It is likely the differences in length frequencies of the two net types are largely caused by
the differences in mesh sizes used; the "shad meshes" of the 61-m nets had high catch
rates of juveniles such as white crappie. This influence can be seen to a lesser degree for
hybrid striped bass, and white bass. Once fish caught in the shad meshes were eliminated
from the data set, significant difference were not found at any reservoirs for hybrids
striped bass, and were only found at 1 of 4 reservoirs for white bass (Canton). Although
statistically different, the difference in length frequency of white bass was small enough
to be negligible for management purposes (middle panel, Figure 5). Johnson (1999)
warns against relying too heavily on significance of statistical testing when conducting
biological research.
Channel catfish length-frequency distributions were significantly different at both
reservoirs where N > 40, even after eliminating shad meshes. One possible explanation
for the difference in length frequencies could be the mechanics of capture for this species.
Catfish have serrated spines, which enable them to become easily tangled in a net without
actually penetrating the mesh. For most species that are "gilled" rather than entangled in
gill nets, there is an optimum size of mesh for which a particular species of a given size is
captured. For these species, few fish are captured with lengths that differ from the
optimum by more than 20% (Hamely 1975). This makes length frequency distributions
for these types of species less variable and more consistently grouped by gill net mesh
size. The serrated dorsal and pectoral spines of channel catfish make almost any size fish
vulnerable to capture in almost any size mesh if its spines become entangled. Buckmeier
and Schlecthe (2009) found that gill nets are less efficient and produce a more biased size
structure data than electrofishing and hoop-netting for both blue and channel catfish;
however, Santucci et al. (1999) found that gill netting and not electrofishing sampled
channel catfish in proportion to their actual abundance. Further research is needed to
determine the potential bias of the 24.4-m gill net design for this species.
24.4-m Fixed sites vs. 24.4-m Random Sites
Catch rates
The mean catch/24 hours of fixed and random sites were similar to each other with only
white bass being significantly different. Since the mean catch/24 hours for white bass
only differed by 0.1, it is doubtful this statistical difference has any application to
management decisions. The precision at random sites was slightly lower than that at
fixed sites for all target species except saugeye and walleye, where precision was
essentially equal. It would be impractical with either sampling strategy to set enough
nets to detect a ± 12.5% change in the population for any of the target species. Using
random site selection, it was possible to detect a ± 25% change in the population with
only 15 net nights of effort for saugeye, and walleye. If 8 more randomly-sampled net
nights were added (one additional day in the field or setting an additional 3 nets each
night) a ± 25% change in the population would also be detectable for channel catfish,
white bass, and white crappie. As with the fixed-site sampling strategy, mean catch/24
hours of hybrid striped bass were the most variable of the target species for the random
sampling strategy. If a random-sampling strategy was to be included in the SSP then it
would be necessary to increase sampling effort by 12 net nights for this species to
achieve the desired level of precision.
Length frequency
The only difference in length frequency found for a target species was for channel catfish
at Canton. Of the 71 KS tests performed for fixed vs. random sampling (including target
species and non-target species), only 8 (11%) significant differences were found. We did
not perform a Bonferroni adjustment on the KS tests because we wanted to be overly
conservative. However, it is likely that many of these 8 significant differences could
have occurred by chance. As mentioned earlier, channel catfish can be caught in gill nets
by their serrated spines, which makes the size of channel catfish captured in each mesh
size much more variable (Table 4c). Therefore, we suggest the difference we observed in
Canton reservoir may not indicate a strong bias of net design, but could have just
occurred by chance along. Further research would be needed to clarify this pattern.
Data from 2010 is currently being entered and analyzed.
E. Conclusions
* Using the new 24.4-m net configuration should benefit ODWC by decreasing the time
spent afield deploying, retrieving, and working up nets.
* The lack of significant differences in mean catch/24 hours when comparing 24.4-m
and the 61-m nets shows that the new nets catch hybrid striped bass, saugeye, walleye,
and white bass more efficiently.
* Since 24.4-m nets are more efficient at catching hybrid striped bass, walleye, and
saugeye the reduced efficiency for sampling channel catfish and white crappie should be
an acceptable trade off, especially because the 24.4-m nets catch fewer non-target
species.
* The regression equations developed should help biologists revise current benchmarks
for evaluating sportfish populations in Oklahoma.
* The variability of the new 24.4-m nets at fixed sites was essentially equal to or less
than the 61-m nets with the exception of hybrid striped bass. The reduced precision for
hybrid striped bass could easily be offset by adding a few more sample sites to reservoirs
with hybrid striped bass fisheries.
* Channel catfish appear to be the only target species that had a significant change in
length frequency distributions using the new 24.4-m nets. All other target species length
frequency distributions appear to be fairly similar using each net type once adjustments
are made for the influence of the shad mesh sizes.
* Anecdotally, 2010 results appear to be similar to 2009 data.
* 2009 and 2010 data will be pooled and analyzed
* Recommend 2011 sampling year be eliminated since data is consistent from first two
years.
III. Significant Deviations: None
Prepared by: _
Ryan Ryswyk, Fisheries Biologist
Date: ~ HM? \.(
Approved by: G£f?j2::? C:f'J~ _
Fisheries Administration
Oklahoma Department of Wildlife Conservation
~n ~ ~ ~
~.~~-- I ohn Stafford r ~
Federal Aid Coordinator
Oklahoma Department of Wildlife Conservation
IV. Literature Cited
Anonymous. 1952. From a report on study of nylon and linen gill nets. The Progressive
Fish Culturist. 14: 18-18.
Anonymous. 1958. Small-mesh gill nets used in Wyoming studies. The Progressive Fish
Culturist. 20:32-32.
Boxrucker, J., and G. Ploskey. 1988. Gear and seasonal biases associated with sampling
crappie in Oklahoma. Proceedings of the Annual Conference Southeastern
Association ofFish and Wildlife Agencies. 42:89-97.
Buckmeier, D.L., and J.W. Schlechte. 2009. Capture efficiency and size selectivity of
channel catfish and blue catfish sampling gears. North American Journal of
Fisheries Management 29:404-416.
Dumont, S.c., and W. Schlechte. 2004. Use of resamp ling to evaluate a simple random
sampling design for general monitoring of fishes in Texas reservoirs. North
American Journal of Fisheries Management 24:408-416.
Erickson, K.E. 1978. Standardized sampling procedures for lake and reservoir
management recommendations. Oklahoma Federal Aid Project F-38-R, Job 1.
Performance Report. 27p.
Gablehouse, D., and 20 coauthors. 1992. Fish sampling and data analysis techniques used
by conservation agencies in the U.W. and Canada. American Fisheries Society,
Fisheries Management Section, Bethesda, Maryland.
Guy, C.S., D.W. Willis, and R.D. Shultz. 1996. Comparison of catch per unit effort and
size structure of white crappies collected with trap nets and gill nets. North
American Journal of Fisheries Management 16:947-951.
Hamely. J.M., Review of gillnet selectivity. 1975. Journal of the Fisheries Research
Board of Canada. 32: 1943-1969.
Hansen, R.G., 1974. Effects of different filament diameters on the selective action of
monofilament gill nets. Transactions of the American Fisheries Society. 2:386-
387.
Henderson, B.A., and S.l. Nepszy. 1992. Comparison in catches of mono- and
multifilament gill nets in Lake Erie. North American Journal of Fisheries
Management. 12:618-624.
Hogrnan, W.J., The relative efficiency of nylon gill nets after transition from cotton nets
in a multispecies fishery. Transactions of the American Fisheries Society.
102:778-785.
Hubert, W.A. 1983. Passive capture techniques. Pages 95-122 in Nielsen, L. A., and
D.L. Johnson, editors. Fisheries Techniques. American Fisheries Society.
Bethesda, Maryland.
Jester, D.B., 1977. Effects of color, mesh size, fishing in seasonal concentrations, and
baiting on catch rates of fishes in gill nets. Transactions of the American Fisheries
Society 106:43-56.
Johnson, D.H., 1999. The insignificance of statistical significance testing. The Journal of
Wildlife Management 63:763-772.
Machiels, M.A.H., M. Klinge, R. Lanters, and W.L.T. van Densen. 1994. Effect of snood
length and hanging ratio on efficiency and selectivity of bottom-set gillnets for
pikeperch, Stizostedion lucioperca, and bream, Abramis brama. Fisheries
Research 19:231-239.
Minns, C.K., and D.A. Hurley. 1988. Effects of net length and set time on fish catches in
gill nets. North American Journal of Fisheries Management. 8:216-223.
Miranda, L.E., and J. Boxrucker. 2009. Warmwater fish in standing waters. Pages 29-42
in S.A. Bonar, W.A. Hubert, and D.W. Willis, editors. Standard methods for
sampling North American freshwater fishes. American Fisheries Society,
Bethesda, Maryland.
Noble, R.L., D.l. Austen, and M.A. Pegg. 2007. Fisheries management study design
considerations. Pages 31-49 in C.S. Guy and M.L. Brown, editors. Analysis and
interpretation of freshwater fisheries data. American Fisheries Society, Bethesda,
Maryland.
Noble, R.L. 2002. Reflections on 25 years of progress in black bass management. Pages
139-146 in D.P. Philipp and M. S. Ridgway, editors. Black bass: ecology,
conservation, and management. American Fisheries Society, Symposium 31,
Bethesda, Maryland.
Peterson, IT., and c.P. Paukert. 2009. Converting nonstandard fish sampling data to
standardized data. Pages195-215 in S.A. Bonar, W.A. Hubert, and D.W. Willis,
editors. Standard methods for sampling North American freshwater fishes.
American Fisheries Society, Bethesda, Maryland.
Reddin, D.G. 1986. Effects of different mesh sizes on gill-net catches of Atlantic Salmon
in Newfoundland. North American Journal of Fisheries Management 6:209-215.
Robinson, M.S. 1999. Evaluation of three gear types for sampling channel catfish in
small impoundments. Pages 265-269 in E.R Irwin, W.A. Hubert, C.F. Rabeni,
H.L. Schramm Jr., T. Coon, editors. Catfish 2000: proceedings of the international
Ictalurid symposium. American Fisheries Society Symposium 24, Bethesda,
Maryland.
Robson, D.S., and H.A. Regier. 1964. Sample size in Petersen mark-recapture
experiments. Transactions of the American Fisheries Society. 93:215-226.
Sullivan, K.P., and C.M. Gale. 1999. A comparison of channel catfish catch rates, size
distributions, and mortalities using three different gears in a Missouri
impoundment. Pages 293-300 in E.R Irwin, W.A. Hubert, C.F. Rabeni, H.L.
Schramm Jr., T. Coon, editors. Catfish 2000: proceedings of the international
Ictalurid symposium. American Fisheries Society Symposium 24, Bethesda,
Maryland.
Santucci, V.J. Jr., D.H. Wahl, and D.F. Clapp. 1999. Efficiency and selectivity of
sampling methods used to collect channel catfish in impoundments. Pages 317-
328 in E.R Irwin, W.A. Hubert, C.F. Rabeni, H.L. Schramm Jr., T. Coon, editors.
Catfish 2000: proceedings of the international Ictalurid symposium. American
Fisheries Society Symposium 24, Bethesda, Maryland.
Wilde, G.R., and W.L. Fisher. 1996. Reservoir fisheries sampling and experimental
design. Pages 397-409 in L.E. Miranda and D.R. DeViries, editors.
Multidimensional approaches to reservoir fisheries management. American
Fisheries Society Symposium 16, Bethesda, Maryland.
Willis. D.W. Use of gill net data to provide a recruitment index for walleyes. North
American Journal of Fisheries Management. 7:591-592.
Table 1. Number of fingerling hybrid striped bass (HSB), fingerling and fry walleye, fingerling saugeye, and 152-254 mm channel
catfish stocked in sample reservoirs from 2004 to 2009.
Canton Kaw Thunderbird Tom Steed Waurika
Year HSB walleye HSB walleye saugeye HSB saugeye c. catfish HSB saugeye
2004 80,650 5,895,500* 34,238 64,500 67,800 33,525 106,320
2005 73,167 3,891,449* 170,000 140,850 63,800 71,000 16,345 106,700 79,200
2006 1,900 3,691,142* 168,632 180,000* 120,900 65,500 16,000 67,850
24,000
2007 70,418 1,900,000* 165,350 170,000* 56,450 64,000 21,773 107,864
2008 36,778 3,930,000* 171,600 170,000* 122860 64,716 63,750 32,475 107,352
46,150
2009 74,044 3,700,000* 44,259 180,000* 122,084 64,000 20,096 110,690
26,208
* Denotes fry stockings.
Table 2. Mean catch/24 hours, coefficient of variation of the mean (c.v. mean), F statistic and P values for two net configurations: 61-m
net (eight 7.6-m panels with 13,16,19,25,38,51,57, and 76 mm mesh sizes) and 24.4-m (eight 3.1-m panels with 38,57,25,
44, 19, 64, 32, and 51 mm mesh sizes) at fixed sites and or random sites. Data were from 15 net-nights effort at five Oklahoma
reservoirs. Significant results (P.:::: 0.05) are bolded.
61-m Fixed 61-m vs. 24.4-m Fixed 24.4-m Fixed 24.4-m FRiaxneddovms. 24.4-m 24.4-m Random
Mean Mean Mean
Target Species Catch/24 c.v. mean F statistic Pvalue Catch/24 c.v. mean F statistic Pvalue Catch/24 c.v. mean
hours hours hours
channel cat 7.6 0.19 Fl.l44 = 14.81 0.0002 3.6 0.23 F1,144=0.12 0..7329 4.3 0.25
hybrid striped
bass 6.3 0.28 F 1,115= 1.11 0.2934 4.2 0.36 F1,115= 1.22 0.2718 3.4 0.33
saugeye 3.7 0.20 F1,86= 0.061 0.8061 3.2 0.20 F1,86= 0.50 0.4817 2.6 0.20
walleye 2.7 0.17 F1,28= 3.13 0.0879 1.7 0.18 F1,28= 0.20 0.6587 1.9 0.19
white bass 17.8 0.26 F1,144= 0.030 0.8622 8.5 0.21 F1,144= 4.81 0.0299 8.6 0.27
white crappie 14.9 0.23 F1,144= 7.29 0.0078 6.6 0.21 F 1,144= 1.36 0.2459 5.3 0.24
By Catch
blue catfish 3.9 0.31 F1,115= 3.90 0.0505 2.1 0.45 F1,115= 1.46 0.2288 2.2 0.32
common carp 2.3 0.32 F1,144= 9.76 0.0022 0.9 0.47 F1,144= 0.90 0.3437 0.7 0.61
drum 10.6 0.26 F 1,144= 26.49 0.0000 2.6 0.32 F1,144=0.12 0.7287 3.1 0.36
flathead catfish 0.3 0.75 F 1,144= 1.86 0.1745 0.1 0.74 F1,144= 0.00 0.9667 0.1 0.68
gizzard shad 55.6 0.32 F1,144=8.11 0.0051 16.0 0.22 F1,144= 0.67 0.4138 16.2 0.19
longnose gar 0.4 0.51 F1,144= 0.46 0.4971 0.2 0.78 F 1,144= 1.95 0.1652 0.7 0.72
river carpsucker 1.6 0.39 F1,144= 13.37 0.0004 0.5 0.49 F1,144= 0.00 0.9989 0.5 0.54
shortnose gar 0.4 0.88 F1,144= 1.07 0.3017 0.2 1.00 F1,144= 1.75 0.1885 0.4 0.84
smallmouth
buffalo 1.1 0.30 F1,144= 2.03 0.1562 0.8 0.38 F1,144= 0.26 0.6079 0.8 0.45
spotted gar 0.6 0.51 F1,144= 1.70 0.1940 0.3 0.78 F 1,144= 1.96 0.1632 0.5 0.62
threadfin shad 23.7 0.48 F1.86= 11.42 0.0011 1.0 0.54 F1.86= 0.40 0.5271 1.4 0.47
Table 3. Mean number of samples required to detect a± 12.5% (c.v. = 0.125) or ± 25% (c.v. = 0.25) change in the population for
target species using 61-m(eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm mesh sizes) and 24.4-m (eight 3.1-m panels
with 38, 57, 25, 44, 19,64,32, and 51 mm mesh sizes) gill nets at fixed and random sites in 5 Oklahoma reservoirs.
Mean number of samples needed to detect target c.v. of mean
Fixed 61-m Fixed 24.4-m Random 24.4-m
c.y. = Cc V, = c.y. = Cc V. = CvV. = Cc V. =
0.25 0.125 0.25 0.125 0.25 0.125
channel catfish 12 41 17 58 22 72
hybrid striped bass 24 83 44 151 40 137
saugeye 22 79 13 43 14 46
walleye 8 30 10 33 10 37
white bass 23 78 13 46 23 79
white crappie 16 56 14 49 18 63
Table 4. Kolmogorov-Smimoff test results (KSa and P values) for length frequency comparisons of (A) 24A-m (eight 3.1-m panels
with 38, 57, 25, 44,19,64,32, and 51 mm mesh sizes) vs. 61-rnnets (eight 7.6-m panels with 13, 16, 19,25,38,51,57, and 76 mm
mesh sizes), (B) 24A-m vs. 61nsm-m (six 7.6-m panels with 19,25,38,51,57, and 76 mm mesh sizes) nets, and (C) fixed 24A-m vs.
random 24A-m nets. Significant results (P:S 0.05) are bolded.
Canton Kaw Steed Thunderbird Waurika
A. 24A-m vs. 61-m KSa P value KSa P value KSa P value KSa P value KSa P value
channel catfish 1.64 0.0091 1.85 0.0022
hybrid striped bass 2A7 0.0000 1.25 0.0883 0.84 OA872
saugeye 1.63 0.0096 1.00 0.2655
walleye 1.06 0.2148
white bass 4.66 0.0000 1.76 0.0041 0.94 0.3357 0.94 0.3442 1.87 0.0019
white crappie 5.69 0.0000 3A9 0.0000 3.24 0.0000 OA9 0.9702
B. 24A-m vs.61nsm-m
channel catfish 1.59 0.0126 1.59 0.0125
hybrid striped bass 1.03 0.2433 1.24 0.0908 0.83 0.4944
saugeye 1.62 0.0107 1.00 0.2655
walleye 1.04 0.2254
white bass 1.43 0.0341 1.48 0.0251 OAO 0.9976 0.91 0.3757 1.05 0.2202
white crappie 1.17 0.1308 1.19 0.1176 1.24 0.0929 0.56 0.9086
C. Fixed vs. Random
channel catfish 1.52 0.0198 0.63 0.8290
hybrid striped bass 1.29 0.0731 0.95 0.3307 0.61 0.8541
saugeye 1.19 0.1201 0.84 OA732
walleye 0.69 0.7301
white bass 1.32 0.0614 0.82 0.5171 0.83 OA919 1.26 0.0853
white crappie 0.87 OA378 0.63 0.8266 0.77 0.5992 0.67 0.7565
channel catfish hybrid striped bass
4 -
Y= 0.8,~49x +Y.TB
4 1 y = 0.9449x + 0.5181 ew R - 0.7399 • c -' R2 = 0.9059 c
3 - P=0.0539. w 3 • • TS
e~
I P = 0.0754
2 - 2 J
• I< / 1 1 • I(
F = 8.79 F = 11.78
0 0
Vl
+-' 0 1 2 3 4 0 1 2 3 4 Cl) c
E
saugeye and walleye white bass IM
I.D
'+0- 4 - C 5 1
Vl Y = 1.0202x + 0.1765 ./ Y = 1.508x - 0.9889 .c ':-:l R' =0.84)Y.TS 4 -, R' = 0.6838?.0.c 3 - TS j P = 0.0962 • TB
""" P = 0.0894 |
Date created | 2011-06-09 |
Date modified | 2011-10-27 |