Evaluation of Daily Creel and Minimum Length Limits for Black Crappie and Yellow Perch in Wisconsin

Crappies (both Black Crappie Pomoxis nigromaculatus and White Crappie Pomoxis annularis) and Yellow Perch Perca flavescens support popular, harvest-oriented recreational fisheries across much of their range, including Wisconsin. Angler exploitation can reduce both abundance and size structure in crappie and Yellow Perch populations (Goedde and Coble 1981; Webb and Ott 1991; Isermann et al. 2005). Specifically, anglers selectively remove relatively large individuals from panfish populations (Goedde and Coble 1981; Colvin 1991; Beard and Kampa 1999; Boxrucker 2002). In most states and provinces, harvest of crappies and Yellow Perch is regulated by daily creel limits and minimum length limits (MLLs) (Boxrucker and Irwin 2002; Isermann et al. 2005). Daily creel limits restrict the number of fish that can be harvested by an individual angler in a single day. Minimum length limits prohibit harvest of fish less than a specified length. Both types of harvest regulations are implemented under the assumption that individual anglers will release fish accordingly (i.e., most will not retain illegally). These regulations have been implemented to reduce harvest (Colvin 1991; Hale et al. 1999; Isermann et al. 2007), improve catch rates, and size structure (Hale et al. 1999; Bister et al. 2002; Boxrucker 2002) or to distribute harvest among anglers over longer periods (Carlton 1975; Fox 1975; Cook et al. 2001; Hurley and Jackson 2002).

Previous evaluations have suggested that creel limits for crappies and Yellow Perch affect only a small percentage of anglers because most anglers harvest few or no fish and few anglers harvest a full daily limit (Snow 1982; Webb and Ott 1991; Baccante 1995; Cook et al. 2001). Conversely, if angler harvest represents a significant source of mortality, reduced creel limits could improve size structure of crappies and Yellow Perch (Colvin 1991; Allen and Miranda 1995; Isermann et al. 2007), provided anglers comply with the regulations (Isermann and Carlson 2009). Reducing harvest in crappie and Yellow Perch fisheries by ≥25% may require daily creel limits of <10 fish/ angler (Radomski 2003; Isermann et al. 2007), which may not be socially acceptable to some anglers (Hale et al. 1999; Reed and Parsons 1999; Boxrucker 2002).

Minimum length limits have reduced crappie harvest, while increasing abundance, alleviating growth overfishing (i.e., removal of fish at a rate and size that does result in maximum yield), improving size structure, and maintaining yield during times of increased fishing pressure (Webb and Ott 1991; Colvin 1991; Boxrucker 2002), but such benefits may not be universally effective for all crappie fisheries (Hale et al. 1999; Bister et al. 2002; Hurley and Jackson 2002). Minimum length limits may not be effective in achieving management objectives based on yield or size structure when natural mortality rates are high (Larson et al. 1991; Reed and Davies 1991) or if reductions in fishing mortality have little effect on total mortality (Allen et al. 1998; Hansen et al. 2011). Additionally, crappie growth rates have decreased after implementation of MLLs (Hale et al. 1999; Bister et al. 2002; Hurley and Jackson 2002) and reductions in harvest may not be popular among harvest-oriented anglers (Reed and Parsons 1999; Boxrucker 2002; Hurley and Jackson 2002).

Modeling has suggested that MLLs can increase abundance or size structure in crappie and Yellow Perch populations (Allen and Miranda 1995; Maceina et al. 1998; Isermann et al. 2002, 2007) and can reduce the potential for recruitment overfishing in certain situations (Allen et al. 2013). Minimum length limits are predicted to improve yield in crappie fisheries, only if growth is relatively fast and conditional natural mortality rates are <30% (Allen and Miranda 1995). A 254-mm MLL for crappies in Weiss Lake, Alabama, was predicted to increase yield because conditional natural mortality was low (<35%), but harvest would be reduced by 23% (Maceina et al. 1998). Minimum length limits (i.e., 229 mm and 254 mm) were predicted to increase age and size structure for Yellow Perch in South Dakota lakes (Isermann et al. 2007); the largest improvements were observed where growth was relatively fast and fishing mortality represented the majority of total mortality occurring within a population.

Based on previous studies, rates of natural mortality and growth are important factors that determine whether MLLs meet management objectives (Reed and Davies 1991; Allen and Miranda 1995; Isermann et al. 2007). Crappies and Yellow Perch typically exhibit high natural mortality (Goedde and Coble 1981; Larson et al. 1991; Reed and Davies 1991), which might negate benefits from a MLL if few fish reach the MLL. Slow growth might prevent increases in size structure and yield expected from a MLL if most fish are lost to natural mortality before they reach the minimum length (Larson et al. 1991; Reed and Davies 1991; Hale et al. 1999). While previous studies offered insight regarding the expected response of crappies and Yellow Perch to MLLs, rates of growth and natural mortality probably vary due to differences in latitude, fish assemblages, lake productivity and morphometry, and fishing mortality. Specifically, most of the previous work related to MLLs for crappies has focused on populations occurring in relatively large reservoirs in the southern USA, where crappie population dynamics could differ considerably compared with smaller lakes in the upper Midwest. Based on growth and natural mortality trends observed for other species (Pauly 1980; Beamesderfer and North 1995), crappies in northern waters may grow slower and exhibit lower rates of natural mortality than in southern portions of their range, but no previous study has simulated the potential effects of MLLs for a broad range of crappie populations in northern lakes and previous model-based assessments of MLLs for Yellow Perch have included but six populations (Boe 1984; Lucchesi 1988; Isermann et al. 2007).

Crappies and Yellow Perch represent two of the most harvested and popular fish among recreational anglers in the state of Wisconsin (McClanahan and Hansen 2003). Based on an angler survey conducted for 2006 (B. Weigel, Wisconsin Department of Natural Resources [WDNR], unpublished data), panfish (e.g., Bluegills Lepomis macrochirus, Yellow Perch, and crappie) constitute the most sought-after group of fish among Wisconsin anglers, accounting for 45% of all angling trips. Panfish also represented the majority of fish harvested by Wisconsin anglers (25.7 million fish; 78% of all fish harvested), and the panfish harvest rate (number harvested/number caught) was 45% and was second only to the estimated harvest rate for catfish (69%). A previous evaluation demonstrated that size structure of Yellow Perch harvested by anglers in Wisconsin declined during 1967–1991 (Beard and Kampa 1999), but this trend was not apparent for Black Crappie. A more recent analysis conducted by the WDNR indicates that size structure of both Yellow Perch and Black Crappie has declined (A. Rypel, WDNR, unpublished data).

In 2012, WDNR began to develop a new statewide management plan for panfish, and the revised plan could include changes to harvest regulations with the intent of reducing angler harvest to improve crappie and Yellow Perch size structure. As part of this management process, WDNR conducted surveys of anglers that attended various sport fishing shows throughout the state. They found that 47% of respondents believed the statewide aggregate daily creel limit of 25 panfish/angler should be maintained and 47% believed it should not (WDNR, unpublished data). Furthermore, 61% of survey respondents indicated that they would definitely or probably prefer to catch and keep fewer panfish provided they were larger in size. Fishery biologists in the state of Wisconsin routinely deal with angler requests to implement more restrictive harvest regulations for panfish because anglers perceive that these changes will improve size structure. Fishery managers continue to propose changes to panfish harvest regulations for individual waters, but some of these changes may not be warranted because they are unlikely to achieve management objectives.

Harvest regulations for crappies and Yellow Perch in Wisconsin have varied widely since the first panfish harvest regulation was implemented in 1925, which restricted angler harvest to 20 crappies ≥152 mm (6 in), 30 sunfish, and 30 Yellow Perch per day (Becker 1983). Since the implementation of the current statewide aggregate daily creel limit of 25 panfish per angler in 1998, the number of harvest regulations for panfish in individual bodies of water has increased and these regulations have included reduced daily creel limits, no daily creel limits, and MLLs of 203 and 254 mm TL. Reasons for implementing harvest regulations that deviate from the statewide regulation were not well documented and the effectiveness of these regulations was not thoroughly evaluated. Consequently, the potential effects of reduced daily creel limits and increased MLLs on these fisheries are not known. Furthermore, increasing complexity of harvest regulations is a concern among Wisconsin anglers and WDNR fishery managers; that is, a complex array of different harvest regulations may not be warranted for crappies and Yellow Perch in Wisconsin. Proliferation of panfish regulations for individual waters suggests that fishery managers have developed criteria for implementation, but no rationale has been defined for selecting specific harvest regulations for crappie and Yellow Perch fisheries in Wisconsin.

Direct assessment of harvest regulations for any species in Wisconsin would require a substantial investment of agency resources at a large spatial scale because of the diverse array of waters within the state. While this type of assessment may occur in the future, it will be many years before the information would be available for use in making decisions regarding the management of crappies and Yellow Perch. Simulating the potential effects of harvest regulations in relation to growth and mortality would provide some insight as to when and where harvest regulations may improve crappie and Yellow Perch fisheries and this information could be used in development of panfish management plans and in designing future assessments of harvest regulations. Therefore, our objectives were to determine whether (1) transitioning from an aggregate statewide daily creel limit of 25 panfish to species-specific daily creel limits <25 fish per angler or implementing statewide MLLs would reduce harvest of Black Crappie and Yellow Perch in Wisconsin by ≥25%, and (2) MLLs would improve yield by ≥10% and increase mean TL of harvested fish by ≥25 mm in Wisconsin fisheries.

METHODS

Creel and length-limit reductions.—We used angler harvest data from 263 WDNR creel surveys conducted on 186 waters during 1998–2008 to quantify the effects of species-specific reductions in daily creel limits and statewide MLLs on angler harvest of Black Crappie and Yellow Perch. Most of these surveys were conducted on lakes in the northern third of the state (Mosel 2012). We did not use creel data from before 1998 because the daily creel limit was reduced to 25 panfish in 1998. Creel surveys were conducted during the angling season from the first Saturday in May through March 1 of the following year using a random stratified roving-access design (Rasmussen et al. 1998). Some creel surveys (i.e., 17 for Black Crappie and 16 for Yellow Perch) covered the entire month of March or a full calendar year. Only creel surveys with at least 30 angling parties interviewed and with ≥50 harvested Black Crappie or Yellow Perch measured for TL by creel clerks were used in analyses.

For each angling party, we determined how many Black Crappie or Yellow Perch would have been legally harvested if the daily creel limit for each species had been <25 fish/d or if a statewide MLL had been in effect, assuming complete angler compliance with regulations. The overall percent reduction in harvest (Pr) achieved under each daily creel limit reduction or statewide MLL for all parties was calculated as

where H₁ is fish harvested under a reduced daily creel or MLL (i.e., the test condition), and H₂ is fish harvested under a 25-fish creel limit with no MLL, which represents the observed number of fish harvested under the current statewide harvest regulation. Reduced daily creel limits of interest were 20, 15, 10, or 5 fish/d, and MLLs of interest were 203, 229, 254, and 279 mm TL. As in most creel surveys, harvest rates (fish/h) of anglers interviewed by creel clerks were assumed to represent harvest rates of all anglers on a particular lake, and observed harvest rates were multiplied by estimated angler effort to estimate total harvest. Consequently, our estimates of harvest reduction provided by the previous equation reflect the extent to which total harvest estimates for Black Crappie and Yellow Perch would have been reduced under each daily creel limit or statewide MLL.

Growth and natural mortality.—Black Crappie and Yellow Perch length and age data were collected during 1990–2010 from individual Wisconsin waters using fyke-nets (1.2 × 1.8-m frames with a 1.2-m-diameter cod end; leads were 23 m long by 1.2 m high; 19.1-mm bar mesh) set during March to June. Ages were usually estimated using scales, but otoliths were used for some populations. Only surveys with ≥50 Black Crappie or Yellow Perch measured were used to estimate mean lengths at ages 3 through 9. Age 3 represented the age at which both species appeared to fully recruit to fyke nets, and few fish older than age 9 were collected. Mean lengths at age and von Bertalanffy growth models (equation 9.9 in Ricker 1975) were used to estimate asymptotic maximum TLs (L_∞) and Brody–Bertalanffy growth coefficients (K) for each population. Growth models were forced through the origin (t₀ = 0) because fyke nets did not effectively capture fish younger than age 3 resulting in obvious negative bias in estimates of t₀. Although setting t₀ = 0 could result in biased estimates of L_∞ and K (Gwinn et al. 2010), this approach generally provided more realistic estimates of the time required for fish to reach specified lengths. Populations were excluded from further analyses if L_∞ exceeded the state record length for each species by 10% or more.

We developed three growth categories for both Black Crappie and Yellow Perch via linear regressions of L_∞ against K for each species. The bivariate distribution of L_∞ and K was divided perpendicularly to the linear regression line at the 33rd percentile of K and the 66th percentile of L_∞ to define slow growth (Mosel 2012). Fast growth was defined using the 66th percentile of K and the 33rd percentile of L_∞. Average growth was defined using all data points between the 33^rd and 66th percentiles. Growth trajectories for populations displaying slow, average, and fast growth were estimated by calculating the mean L_∞ and K for all populations within each growth category. Mean asymptotic maximum weight (W_∞; g) was calculated for each growth category by using mean L_∞ in the log₁₀ weight–log₁₀ TL model for each species:

Parameters of log₁₀ weight–log₁₀ TL models were estimated using all fully recruited Black Crappies (TL ≥ 100 mm; N = 7,679) or Yellow Perch (TL ≥ 127 mm; N = 2,719) weighed during fyke-netting. Weight–length models were not fitted for individual growth categories because weight data were not available for some categories.

Instantaneous natural mortality rate (M) was estimated for individual populations using the equation proposed by Pauly (1980):

where T = mean annual water temperature. Mean annual air temperatures approximately correspond to mean annual water temperatures (Pauly 1980; Shuter et al. 1983) and were used because mean annual water temperatures were not available. Mean annual air temperature data were acquired from the National Climatic Data Center for a 30-year period during 1961–1990 for individual populations (NOAA 2005). Estimates of M for each population were averaged within each growth category for modeling purposes.

Modeling simulations.—We used a modeling approach similar to that of Allen and Miranda (1995) to evaluate effects of MLLs on yield, harvest, and mean TL of harvested fish in Wisconsin Black Crappie and Yellow Perch fisheries. We used a Beverton–Holt yield per recruit model (equation 10.21 in Ricker 1975) to simulate how Black Crappie and Yellow Perch yield (Y) per 100 initial recruits (N₀) would be affected by implementing different MLLs for each growth category:

where F = instantaneous fishing mortality rate, t_r = time in years to recruit to a specified MLL, and Z = total instantaneous mortality rate (i.e., Z = F + M). For each growth category, MLL simulations were run over the range of F equivalent to a 0–60% range of annual exploitation rates (u). Exploitation rates were calculated as u = F(A/Z), where total annual mortality rate (A) is 1 − e^−Z.

Black Crappie and Yellow Perch were assumed to fully recruit to harvest at 203 mm TL under no MLL, which approximated the 50th percentile of the cumulative frequency distribution of lengths for all harvested fish measured in creel surveys (Figure 1). Based on length distributions of harvested fish observed in creel surveys and MLLs used in previous modeling exercises (Allen and Miranda 1995; Maceina et al. 1998; Isermann et al. 2002), we simulated the effects of four MLLs (203 [i.e., no length limit], 229, 254, and 279 mm) on yield, although MLLs of 254 and 279 mm were not reasonable for some Wisconsin Yellow Perch populations based on estimates of L_∞. We rearranged the von Bertalanffy equation to estimate the time in years required to reach each MLL for each growth category. These values were used for t_r in the Beverton–Holt model.

We calculated the predicted number of fish harvested for each MLL simulation as , where and represents the number of fish recruiting to the fishery at t_r. Yield was divided by the number of fish harvested to obtain mean weight (g) of harvested fish; mean weights were converted to mean TLs using the log₁₀ weight–log₁₀ TL model for each species.

RESULTS

Black Crappie

Creel and length limits.—Predicted reduction in harvest was 7% under a statewide MLL of 203 mm, but predicted harvest reductions were >10% at statewide MLLs of 229 mm (35% reduction), 254 mm (69% reduction), and 279 mm (89% reduction; Figure 1). Only 5% (10,644 of 234,076) of anglers reported harvesting at least one Black Crappie and only 0.04% (95 of 234,076) of anglers harvested a daily creel limit of 25 fish, so the current daily creel limit affected few anglers (Figure 2). Predicted reductions in harvest were <10% at reduced creel limits of 20 (2% reduction) and 15 (5% reduction) fish/d and were >10% at reduced creel limits of 10 (13% reduction) and 5 fish/d (33% reduction).

Modeling simulations.—Growth and natural mortality parameters used in MLL simulations for Black Crappie appear in Table 1. Minimum length limits were always predicted to reduce harvest (22–93% reductions) and increase mean TL of harvested fish (13–61 mm increases) compared with no MLL, while providing little to no improvement in yield (largest increase = 2%; Figure 3). For all growth scenarios, 254-mm and 279-mm MLLs were predicted to improve the mean lengths of harvested fish by >25 mm TL, but predicted reductions in harvest exceeded 40%. Predicted increases in the mean lengths of harvested fish under a 229-mm MLL were always <25 mm TL, but reductions in harvest were less severe (reductions between 21% and 35%). Despite projected improvements in mean lengths of harvested fish, yield was usually highest with no MLL in place. When growth was slow and u was >45%; a 229-mm MLL provided slightly higher yields (<5% higher) than no MLL, but 22% fewer fish were harvested with the 229-mm MLL in place (Figure 3).

Table 1. Asymptotic maximum TLs (L_∞; mm), Brody–Bertalanffy growth coefficients (K), mean instantaneous natural mortality rates (M), time in years to reach 203 (t₂₀₃), 229 (t₂₂₉), 254 (t₂₅₄), and 279 mm TL (t₂₇₉), and asymptotic maximum weights (W_∞; g), for Wisconsin Black Crappie and Yellow Perch populations displaying slow, average, or fast growth. Fast-growing Yellow Perch populations were not predicted to reach 254 mm and 279 mm TL, based on estimates of L_∞.

Species	Growth category	L∞	K	M	t203	t229	t254	t279	W∞
Black Crappie	Slow	362	0.23	0.33	3.55	4.30	5.20	6.36	803
	Average	324	0.28	0.36	3.48	4.32	5.41	6.99	561
	Fast	289	0.40	0.48	3.05	3.92	5.28	8.45	388
Yellow Perch	Slow	336	0.20	0.15	4.59	5.64	6.97	8.80	590
	Average	300	0.24	0.17	4.47	5.69	7.47	10.84	399
	Fast	253	0.36	0.23	5.10	7.84			222

Yellow Perch

Creel and length limits.—Predicted harvest reductions were ≥25% at statewide MLLs of 203 mm (25% reduction), 229 mm (75% reduction), 254 mm (90% reduction), and 279 mm (97% reduction; Figure 1). Only 5% (12,112 of 234,067) of anglers reported harvesting at least one Yellow Perch and only 0.1% (280 of 234,067) of anglers harvested a daily creel limit of 25 Yellow Perch, so the current daily creel limit affected few anglers (Figure 2). Predicted reductions in harvest were <10% at reduced creel limits of 20 (3% reduction) and 15 (9% reduction) fish/d and were >10% at reduced creel limits of 10 (19% reduction) and 5 (40% reduction) fish/d (Figure 2).

Modeling simulations.—We did not simulate 254-mm and 279-mm MLLs for fast-growing Yellow Perch populations because mean L_∞ for this growth category was 253 mm (Table 1). Minimum length limits were always predicted to reduce harvest (15–66% reductions) and increase mean TL of harvested fish (10–58 mm increases) compared with no MLL (Figure 4). Improvements in yield ≥10% were only observed under MLLs when growth was slow and u exceeded 45% (largest predicted yield increase = 17%). For slow-growth and average-growth scenarios, 254-mm and 279-mm MLLs were predicted to improve the mean TL of harvested fish by >25 mm, but predicted reductions in harvest were between 30% and 66%. Predicted increases in the mean length of harvested fish under a 229-mm MLL were always <25 mm TL, but reductions in harvest were less severe (reductions between 15% and 36%). Despite projected improvements in mean length of harvested fish, when growth was average yield under MLLs was usually lower or within 6% of predicted yields with no MLL in place. When growth was slow and u was >45%, 229-mm and 254-mm MLLs provided yields that were 8–17% higher than no MLL, but 15–31% fewer fish were harvested with the MLLs in place (Figure 4).

DISCUSSION

The current daily creel limit of 25 panfish/angler does little to reduce harvest of Black Crappie and Yellow Perch in Wisconsin and reducing harvest by ≥25% would require daily creel limits of <10 fish/angler. Our findings are consistent with results from previous evaluations of daily creel limits for crappie and Yellow Perch populations (Cook et al. 2001; Radomski 2003; Isermann et al. 2007), which also indicated that creel limits would need to be ≤10 fish/d to effectively reduce harvest. While daily creel limits of ≤10 fish/angler could reduce harvest and possibly improve size structure, these low creel limits may be socially unacceptable (Reed and Parsons 1999; Cook et al. 2001; Radomski 2003; Edison et al. 2006). For example, most Minnesota anglers (53%) believed that a daily limit of 30 Bluegill Lepomis macrochirus was adequate for proper management, and they opposed reducing the creel limit to 15 (78% opposition), 10 (96% opposition), or 5 (100% opposition) Bluegill/d (Reed and Parsons 1999). Similarly, resort owners on Lake Winnibigoshish, Minnesota, were opposed to reducing the daily creel and total possession limit of 100 Yellow Perch/angler to a 20-fish daily creel limit and 30-fish possession limit because they believed this would negatively affect their businesses (Radomski 2003). Conversely, Illinois anglers favored a 10-fish daily creel limit for Bluegill over a 25-fish daily creel limit (Edison et al. 2006). Alternatively, daily creel limits could be set at a level where 10% of the anglers would be expected to harvest a daily creel limit (Cook et al. 2001). Using this strategy, current creel limits would need to be adjusted to <10 fish/d for both Black Crappie and Yellow Perch in Wisconsin.

Implementing statewide MLLs >229 mm for Black Crappie and >203 mm for Yellow Perch could reduce harvest at a statewide level by 25% or more, but predicted responses to MLLs varied among simulated populations in relation to growth and mortality. Our model results indicate that MLLs can improve the mean lengths of Black Crappies and Yellow Perch harvested by anglers, but these improvements did not typically coincide with measurable increases (≥10%) in yield because of corresponding reductions in the number of fish harvested. Specifically, increases in the mean lengths of harvested fish of ≥25 mm TL were only observed under 254-mm and 279-mm MLLs, but our models suggest that anglers would have to accept large reductions in harvest (≥30%) to achieve these improvements. Given the harvest-oriented nature of crappie and Yellow Perch fisheries, a 229-mm MLL offered a more equitable tradeoff between improvements in the mean lengths of harvested fish (11–21 mm improvements) and reductions in harvest (22–37% reductions). While anglers targeting panfish are often harvest-oriented to some degree (Allen and Miranda 1996; Hurley and Jackson 2002; Beardmore et al. 2011), unlike commercial fishers, they typically do not assess fishing satisfaction in terms of yield. However, yield provides a single metric useful for assessing the tradeoff between the number and size of harvested fish, which do represent two metrics often used to evaluate fishing satisfaction or angler motivations (Allen and Miranda 1996; Isermann et al. 2005; Edison et al. 2006). While previous studies have indicated that yield in crappie and Yellow Perch fisheries could be improved if growth was above average and natural mortality was low, improvements in yield ≥10% under MLLs were only observed in general for slow-growing Yellow Perch at relatively high exploitation rates. Consequently, while MLLs could improve the size of fish harvested by anglers, in most scenarios MLLs would not maximize the weight of edible flesh obtained by anglers from an individual fishery.

Several previous modeling evaluations have demonstrated that high M could negate potential improvements in crappie and Yellow Perch yield resulting from implementation of MLLs (Bronte et al. 1993; Allen and Miranda 1995; Isermann et al. 2007). Our estimates of M for Black Crappie populations in Wisconsin were similar to rates observed within some Minnesota populations (M = 0.21–0.43; Parsons and Reed 1998) but were generally lower than rates estimated for crappie populations in Alabama, Oklahoma, Missouri, Kansas, Nebraska, and Ohio (M = 0.08–0.94, Ellison 1984; Mosher 1985; Angyal et al. 1987; Colvin 1991; Hammers and Miranda 1991; Miller 1991; Reed and Davies 1991; Zale and Stubbs 1991; Brock 1994). However, given the growth trajectories we observed, the estimates of M we used for Black Crappies in Wisconsin were sufficiently high to negate potential improvements in crappie yields that might result from implementing MLLs, especially in fast-growth simulations where the lowest M observed for any single population is 0.38. Previous studies predicted that yield improvements under MLLs were more likely to occur when M was <0.40 and growth was relatively fast (Allen and Miranda 1995; Maceina et al. 1998; Isermann et al. 2002), but these conditions may not exist for most fast-growing Black Crappie populations in Wisconsin. Yellow Perch populations in Wisconsin had lower rates of M than those reported for one South Dakota population (M = 0.54; Isermann et al. 2005) and for Chequamegon Bay, Lake Superior (M = 0.51; Bronte et al. 1993), but this did not translate into improved yields in most model simulations.

We used the model provided by Pauly (1980) to estimate M because estimates of natural mortality rates are not available for Black Crappie and Yellow Perch populations in Wisconsin, and are available for only a few populations in the northern USA (Bronte et al. 1993; Parsons and Reed 1998; Isermann et al. 2005; Schoenebeck and Brown 2011). We also used the Pauly (1980) equation because we assumed that M was positively related to growth rate of crappies and Yellow Perch and negatively related to latitude. While this assumption is probably correct in a general sense, the relationships among M, growth rates, and latitude have not been explicitly defined for crappies and Yellow Perch. The rates of M we used for each growth scenario may not have been accurate if species-specific relationships among these variables deviate significantly from the Pauly (1980) model. A better understanding of these relationships would be useful for fishery managers and would improve future efforts to assess the potential effects of harvest regulations.

Our modeling approach was less complex than some previous efforts to simulate the responses of crappie populations to angler harvest and MLLs (Miranda and Allen 2000; Allen et al. 2013), but we wanted our modeling framework to accurately reflect the extent of data available to most fishery managers in North America. Information on recruitment dynamics of crappies and Yellow Perch is lacking for most populations in the upper Midwestern USA because abundances of adult and juvenile fish are not evaluated on a meaningful temporal scale, largely because of the substantial number of populations to be sampled and intensive sampling efforts for other species. Consistent with previous modeling exercises (Allen and Miranda 1995; Maceina et al. 1998; Isermann et al. 2007), we did not account for population responses that could occur after changes in harvest regulations, such as changes in natural mortality rates (Allen et al. 1998; Boxrucker 2002), growth rates (Hurley and Jackson 2002), longevity (Newman and Hoff 2000), and fishing mortality for fish longer than the MLL (Larscheid and Hawkins 2005). For example, Yellow Perch and Black Crappie growth is often density dependent (Guy and Willis 1995; Staggs and Otis 1996; Pierce et al. 2006), so reductions in harvest resulting from changes in regulations could reduce growth (Serns 1978; Carline et al. 1984; Munger and Kraai 1997). We did not account for these potential responses to MLLs because these responses have not been clearly defined. The effectiveness of more stringent harvest regulations in meeting management objectives could be negatively affected by relatively high postrelease mortality (Coggins et al. 2007), which we did not account for in our modeling. For White Crappies in Columbus and Aliceville reservoirs, Mississippi, delayed mortality was low (3%) if crappies were caught from depths <10 m, but delayed mortality was higher at depths of 13 (29% mortality) and 16 m (67% mortality; Hubbard and Miranda 1991), so release mortality could be a concern in some Wisconsin lakes where crappies are caught from deeper water.

We may have overestimated rates of growth and M because we relied on scales as the primary structure for age estimation. Previous studies have demonstrated that scale ages underestimate the age of older fish compared with otoliths for both crappies and Yellow Perch (Ross et al. 2005; Maceina and Sammons 2006; Vandergoot et al. 2008; Isermann et al. 2010). Consequently, fish may have required more time to reach some MLLs, extending the period of time where only M was occurring, leading us to overestimate yield and mean length of harvested fish. Conversely, if growth of Black Crappie and Yellow Perch was slower, our estimates of M would have also been lower, potentially offsetting some of the bias in estimating yield and mean length of harvested fish. While otoliths could provide more accurate age estimates (Ross et al. 2005), Black Crappies and Yellow Perch are not routinely sacrificed during WDNR surveys and, to our knowledge, most resource agencies in the upper Midwest continue to rely on scales for estimating the age of crappies and Yellow Perch.

Lastly, changes in harvest regulations could attract or deter anglers depending on their motivations and perceptions of fishing opportunities (Johnson and Carpenter 1994; Aas et al. 2000; Beard et al. 2003). For example, angler effort in Wisconsin for Walleye Sander vitreus was higher on lakes with higher daily creel limits, despite the fact lakes with smaller creel limits had higher catch rates (Beard et al. 2003). Conversely, harvest regulations could increase angler use if the regulation creates an opportunity for quality fishing (Cox and Walters 2002) or if these harvest regulations are implemented where quality fishing opportunities already exist, allowing anglers to more easily identify these opportunities. Anglers in Minnesota indicated that they would increase the number of fishing trips they took if Bluegill size structure was higher (Reed and Parsons 1999). Consequently, sufficient levels of angler effort could negate potential fishery improvements that might be achieved under more restrictive harvest regulations. For example, under a daily creel limit of 10 fish/angler, which was enacted to protect a population of relatively large Yellow Perch in Pelican Lake, South Dakota, estimated u exceeded 60%, despite the fact that only a small percentage (5%) of anglers harvested a full daily creel limit (Isermann et al. 2005). Lastly, catch-at-age data indicate that Wisconsin Black Crappie and Yellow Perch populations exhibit high variation in recruitment, which often results in pulses of harvestable fish that may attract high angler effort. Effort at these times could be high enough to negate the predicted benefits of low creel limits or MLLs.

MANAGEMENT IMPLICATIONS

Recruitment variation in Black Crappie and Yellow Perch populations usually results in inconsistent fisheries, but the primary management concern related to recreational fisheries for these species is not recruitment overfishing and stock sustainability, but growth and quality overfishing. Growth overfishing is often defined as removal of fish at a rate and size that does result in maximum yield for a specific fishery (Radomski et al. 2001; Pereira and Hansen 2003). Our modeling results suggests that growth overfishing of this nature is not a widespread problem for Black Crappie and Yellow Perch fisheries in Wisconsin because increasing MLLs generally did not increase yield compared with no MLL. Consequently, if the management objective is to prevent growth overfishing (i.e., improve yield) for these fisheries, then our analyses indicate that managers do not need to implement more restrictive MLLs.

Defining quality overfishing is more difficult because definitions of fishery quality can vary among both anglers and fisheries (Hahn 1991; Allen and Miranda 1996; Radomski et al. 2001). One component of quality overfishing may be harvest-related reductions in population size structure that result in low angler satisfaction with regard to the average size of fish they catch or harvest (Pereira and Hansen 2003). Reducing the effects of size-based quality overfishing might be achieved by reducing daily creel limits, but this seems unlikely to occur unless creel limits are reduced to <10 fish/ angler. Additionally, implementing MLLs could reduce the effects of size-based quality overfishing, but substantive gains (i.e., ≥25 mm increases in the mean TL of harvested fish) would probably require substantial reductions in harvest (≥30%) and would not typically improve yield. However, anglers may also consider the number of fish harvested when defining fishery quality (Boxrucker 2002; Hurley and Jackson 2002; Pereira and Hansen 2003). Consequently, implementing MLLs may merely improve one aspect of the fishery (i.e., size structure), while reducing quality in terms of the opportunity to harvest fish. If the primary management objective is to achieve a relative balance between reductions in harvest and improvements in the size of fish available to anglers, a 229-mm MLL offers the most reasonable option for both species.

While our modeling provides a framework for the WDNR and other resource agencies to make more informed decisions regarding the implementation of harvest regulations, understanding angler motivations is critical to defining management objectives and selecting appropriate management actions. This can be difficult because opinions regarding tradeoffs between number of fish harvested and the average size of fish available to them will vary among anglers. Some anglers may be content to harvest larger numbers of relatively small fish, while other anglers would be willing to harvest fewer fish if reduced harvest improved size structure. Additionally, angler opinions regarding a certain fishery can change after new regulations are implemented. For example, abundance of quality and preferred-length crappies and angler catch rates increased under a 254-mm MLL was implemented on Ft. Supply Reservoir, Oklahoma, which seemed to meet angler preferences documented before the MLL was implemented; however, angler dissatisfaction regarding reduced harvest led to removal of the length limit (Boxrucker 2002). Making anglers aware of the potential tradeoffs between number of fish harvested and possible improvements in the size of the harvested fish should be an important outreach component to increase support for changes in harvest regulations. Our results offer fishery managers a means to evaluate these potential tradeoffs for Black Crappies and Yellow Perch in northern waters.

A better understanding regarding sources of mortality in Black Crappie and Yellow Perch populations would aid managers in determining whether changes in harvest regulations could improve individual fisheries. The large number of fishable bodies of water in the upper Midwest and the time dedicated to the management of other prominent sport fish such as Walleye and Muskellunge Esox masquinongy make it unlikely that state resource agencies in the midwestern USA and portions of Canada will devote substantial resources to increased assessment of Black Crappie and Yellow Perch populations and fisheries. However, estimating u for a suite of fisheries over several years would be helpful in determining if u is really an important factor affecting the quality of these fisheries, or whether M and recruitment variability are more important in regulating population size structure and the availability of harvestable size fish.

Lastly, we suggest that well-designed experiments conducted within an adaptive management framework are needed to actually determine whether changes in harvest regulations will achieve desired management objectives. These experiments will require careful planning and intensive monitoring of both treatment and reference lakes during preregulation and postregulation periods. Our work provides biologists with important guidance regarding which harvest regulations and populations are worthy of consideration in these experiments because inclusion of too many treatment levels can lead to poor replication and inconclusive results (Carlson and Isermann 2009). Our results indicate that (1) daily creel limits of >10 fish/d could be excluded from future experiments, (2) MLLs ≥254 mm TL may not be of interest unless anglers are willing to accept large reductions in harvest, (3) if the number of populations that can be sampled is limited, moderate-growing and slow-growing populations may represent better experimental units than fast-growing populations, and (4) if only a single MLL can be evaluated, a 229-mm MLL may be more palatable to anglers because it offers the most equitable tradeoff between reductions in harvest and improvements in mean length of harvested fish. Future evaluations should also account for angler behavior in response to regulation changes to determine whether more restrictive regulations were more or less effective merely because of changes in effort.