Compensatory response of invasive common carp Cyprinus carpio to harvest

Fisheries Research 179 (2016) 168–178

Contents lists available at ScienceDirect

Fisheries Research journal homepage: www.elsevier.com/locate/fishres

Compensatory response of invasive common carp Cyprinus carpio to harvest Michael J. Weber a,∗ , Matthew J. Hennen a,1 , Michael L. Brown a , David O. Lucchesi b , Todd R. St. Sauver b a b

Department of Natural Resource Management, South Dakota State University, Brookings, SD 57007, USA South Dakota Department of Game, Fish and Parks, 4500 South Oxbow Avenue, Sioux Falls, South Dakota 57106, USA

a r t i c l e

i n f o

Article history: Received 4 January 2016 Received in revised form 25 February 2016 Accepted 26 February 2016 Keywords: Density-dependence Exploitation Invasion Survival Population dynamics Harvest

a b s t r a c t Invasive species are often mechanically removed to reduce or eliminate their populations. However, removal may release survivors from density-dependent mechanisms resulting in stable or increasing population abundance through compensatory processes. Additionally, immigration of new individuals into systems where removal is occurring may negate efforts to control population abundance. Thus, understanding population-level responses to removal and immigration rates are essential aspects of invasive species management. We evaluated how common carp Cyprinus carpio populations respond to removal through commercial harvest in three interconnected lakes over five years. Nearly 230,000 common carp (up to 55 fish/ha/year) were removed and exploitation rates ranged from <1 to 43% across three lakes over four years. Despite high removal rates in some years, carp population abundance, recruitment, and growth remained stable. Carp survival ranged between 54–79% and was inversely related to removal rate. However, survival only decreased by 25% at 43% exploitation, suggesting a partial compensatory rather than additive response. Emigration among lakes was low (<1%; >2000 carp), but varied among years in response to water level fluctuations. Our results indicate that carp control is difficult in large interconnected systems due to compensatory mortality and interbasin movement patterns, limiting the ability of removal-based management practices alone to control these invasive populations. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Successful invasions of non-native species are a primary source of irreversible change in ecological processes (Pimentel et al., 2000; Lodge et al., 2006) and present a formidable challenge to resource managers (Vitousek et al., 1996; Byers et al., 2002). The invasion and spread of non-native species into previously uninhabited environments can disrupt ecosystem structure and function at multiple levels by altering food webs (Hrabik et al., 1998; Weber and Brown, 2009), resulting in the loss of native species and biodiversity (Allan and Flecker, 1993; Weber and Brown, 2011a) and a homogenization of freshwater fauna (Rahel, 2002). Due to the deleterious effects of invasive species, there is a need to develop

∗ Corresponding author. Current address: Department of Natural Resource Ecology and Management, Iowa State University, 339 Science Hall II, Ames, IA 50011, USA. Fax: +515 294 2995. E-mail address: [email protected] (M.J. Weber). 1 Current address: Minnesota Department of Natural Resources, 650 Highway 169, Tower, MN 55790, USA. http://dx.doi.org/10.1016/j.fishres.2016.02.024 0165-7836/© 2016 Elsevier B.V. All rights reserved.

management strategies to reduce their adverse impacts. Harvest management (i.e., mechanical removal of individuals to suppress or eliminate a population) has been conducted on a wide range of taxa including aquatic invertebrates (Hein et al., 2006), fish (Peterson et al., 2004; Coggins et al., 2011), birds (Frederiksen et al., 2001), and mammals (Howald et al., 2007). Successful eradication or suppression of invasive species may reduce their ecological effects and serve as a restoration strategy in disrupted ecosystems (Hein et al., 2006; Weidel et al., 2007; Coggins et al., 2011). However, formal evaluations quantifying the success of eradication projects are rare, often only summarized in technical reports (Meronek et al., 1996; Simberloff, 2001), or been conducted on small, closed systems where eradication is likely to be successful (e.g., Peterson et al., 2004; Weidel et al., 2007; Bajer et al., 2011). In contrast, little is known concerning how populations respond to eradication projects in larger interconnected systems where animals are abundant, immigration and emigration occurs, and eradication is unlikely. The goal of invasive species harvest management is population eradication or reduction to a level that minimizes the ecological impacts of the targeted species on invaded ecosystems. How-

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

ever, removal of individuals provides additional resources (e.g., food, habitat) to those that remain, potentially releasing them from density-dependent processes. If populations are regulated by density-dependent processes, survivors may compensate with increased levels of recruitment and growth or decreased natural mortality, resulting in stable (compensatory) or increasing (overcompensatory) population size (Brooks, 2002; Matsuda and Abrams, 2004; De Roos and Schellekens, 2007; Zipkin et al., 2008). Several theoretical models predict compensatory responses from populations subjected to harvest (Zipkin et al., 2008; Abrams, 2009; Colvin et al., 2012). However, empirical evidence of compensatory responses in natural populations subjected to harvest is rare (Abrams, 2009), but is needed to identify conditions under which compensatory responses occur. Common carp Cyprinus carpio is a widespread invader worldwide that has been described as an ecosystem engineer due to its large-scale ecosystem disturbances (Koehn, 2004; Weber and Brown, 2009). Common carp are associated with physical, chemical, and biological alterations to shallow aquatic systems (Weber and Brown, 2009, 2011a). Direct effects of common carp (e.g., decreased water clarity and aquatic macrophyte coverage, increased suspended nutrient availability) can result in shifts in aquatic ecosystems between alternative equilibria from a heterogeneous, clear macrophyte-dominated state to a homogenous turbid, plankton-dominated state (Scheffer et al., 2001; Weber and Brown, 2009), which may indirectly result in changes in native fish assemblages and food web dynamics (Jackson et al., 2010; Weber and Brown, 2011a; Letvin, 2013). A variety of techniques have been applied to control common carp populations, including mechanical removal (e.g., netting and trapping), toxicants, water level manipulation, fish barriers, biotechnology, and immunological methods (Brown and Walker, 2004; Weber and Brown, 2009; Brown and Gilligan, 2014). Wholelake chemical piscicide applications have been successful at reducing or eliminating invasive fish populations in small systems, but their use is limited by lake size, costs, and public concerns about effects on non-target fishes and risks to human health (Meronek et al., 1996; Harig and Bain, 1998). In contrast, mechanical removal of invasive species provides a species-specific alternative to chemical reclamation (Peterson et al., 2004; Hein et al., 2006; Zipkin et al., 2008), but formal holistic evaluations of its effectiveness at controlling invasive populations are scarce. The most widely attempted mechanical method to control common carp populations is commercial fishing using large seine nets (e.g., Weber and Brown, 2009; Bajer et al., 2011; Colvin et al., 2012). However, supporting population dynamic data for common carp is lacking and population-level effects of mechanical removal are unknown. Controlling invasive species is a population-level phenomenon, but is rarely studied as such (Parker, 2000). The success of invasive populations depends on population growth, which fluctuates with recruitment, growth, mortality (natural and harvest), and immigration and emigration (Vermeij, 1996; Gotelli, 2001). Understanding population level mechanisms resulting in the success of invasive species is necessary to formulate control and recovery strategies (Parker et al., 1999; Sakai et al., 2001). Given the deleterious effects of common carp on aquatic ecosystems, there may be substantial ecological benefits to reducing their abundance in invaded lakes. However, very few manipulative approaches have been implemented to evaluate the effects of removal through harvest on populations. Here, we conducted a five-year manipulative experiment in three interconnected lakes to evaluate the effect of harvest on common carp abundance and population dynamics. Success of invasive species control depends upon the feasibility of population reduction, the potential for immigration, and the susceptibility to control measures (Myers et al., 2000). We assessed common carp population estimates and relative abundance, survival, inter-

169

basin movement, size structure, growth and recruitment and how they changed in relation to varying levels of removal. Our work was based on the premise of harvest management that predicts removing individuals from a population results in reduced survival and population size. Thus, we hypothesized that common carp survival and abundance would decline following harvest management. Alternatively, survival may decline at a lower rate than exploitation or may not be related to harvest and instead be constant through time or simply vary among lakes and years. If common carp survival and abundance were reduced following harvest, we hypothesized that recruitment and growth rates would increase and size structure would shift toward larger bodied individuals. 2. Methods 2.1. Study site Common carp populations were sampled in an interconnected chain of natural lakes in southeastern South Dakota, USA (Fig. 1). Lake Herman forms the headwaters of the chain of lakes and connects to lakes Madison and Brant via Silver Creek. Lake Herman is intermediate in size (521 ha) compared to the other lakes, has a mean depth of 1.4 m and a maximum depth of 3.9 m. Lake Madison is the largest (1069 ha, mean depth = 2.4 m, maximum depth = 4.9 m) of the four lakes. Brant Lake, last in the chain, is 421 ha with a mean depth of 2.9 m and a maximum depth of 4.7 m and separated from Lake Madison by Round Lake (76 ha), which, due to its unimpeded connection with Brant Lake, was not considered a separate system in this study. A 1.5 m elevation change exists at the outlet of Lake Madison, prohibiting upstream movement from Lake Brant. All lakes are predominately silt bottoms with limited submerged macrophyte coverage. 2.2. Relative abundance The relative abundance of adult and age-1 common carp was indexed using trap nets from 2004 through 2012 in lakes Herman, Madison, and Brant and in two control lakes where harvest did not occur, lakes Thompson and Sinai. Control lakes were selected because they are the only lakes comparable in size with similar fish communities that are located in close proximity (<30 km) to the study lakes. Carp catches in trap nets were positively correlated with population estimates from lakes Herman, Madison, and Brant from 2007 to 2010 (n = 10 lake/years; r = 0.58), indicating that trap nets provided a useful index of carp relative abundance. Trap nets were constructed with 19-mm bar mesh netting, frames were 0.9 m high by 1.5 m wide, and had 18.3-m-long leads. Experimental gill nets were 45.7 m long by 1.8 m deep with one 7.6-m panel each of 13-, 19-, 25-, 32-, 38- and 51-mm bar mesh monofilament netting. Effort deployed for each gear varied by lake as a function of lake surface area, but all locations were randomly selected and nets were soaked for 24 h. Common carp were counted in each gear and measured (nearest 1 mm). Von Bertalanffy growth curves indicate that age-1 common carp in these systems are 150–325 mm in length whereas those >350 mm are age-2 or older (Weber et al., 2011). Thus, individuals were categorized as age-1 or adult based on their length, and the relative abundance of each age group (age1 and adult) captured in each gear was indexed as catch per unit effort (CPUE; number of fish per net night). We used a multiple site and year, before and after control and impact (BACI) design to test for differences in adult and age-1 common carp abundance in removal lakes following harvest. This approach compares differences in catch rates measured before and after a treatment in ‘impacted’ (lakes Herman, Madison, and Brant) and ‘control’ (lakes Thompson and Sinai) systems. BACI designs can be used to differ-

170

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

Fig. 1. Map of the Madison chain of lakes (star) in eastern South Dakota. Lakes Thompson and Sinai that served as reference lakes are denoted on the map inset.

Fig. 2. Multistate design illustrating the four states (Herman (H), Madison (M), Brant (B), tag loss(T)), apparent survival (Si ), transition probability (ϕ), detection probability (pi ), and reporting rate (ri ).

entiate natural variability in fish abundances from variability likely caused by harvest (Schmitt and Osenberg, 1996). Individual nets were subreplicates and lakes were nested within treatment for each sampling period. Samples taken each year were treated as replicates in a two-way ANOVA with treatment (harvest versus control lake) and period (before versus after harvest) as the main effects. A significant interaction indicated that the treatment (harvest) had an influence on population abundance.

2.3. Tagging and recapture In May and June 2007, common carp were collected for tagging in lakes Brant, Madison, and Herman using boat DC-pulsed electrofishing. Due to a limited number of fish captured and tagged during 2007 (n = 3,354), commercial seining (bag seine, 1,200 m long, 5-cm bar mesh) was used from 2008 through 2010 to capture a larger number of individuals. All common carp collected were marked with individually numbered T-bar anchor tags (FD94; Floy® ) and gender was determined via manual extrusion of eggs

100

Herman population Herman harvest

80 60 40 20 0

Adult carp CPUE (number/net night)

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

Madison population Madison harvest

300 200 100 0 300 250

Age-1 carp CPUE (number/net night)

Density (number/ha)

400

30

Treatment= 0.67 Period = 0.82 T*P = 0.35

25 20 15 10 5 0 5 4

Treatment= 0.41 Period = 0.84 T*P = 0.59

3 2 1 0 2004

Brant population Brant harvest

171

2006

2008

2010

2012

Fig. 4. Relative abundance of adult (top) and age-1 (bottom) common carp in control lakes (dark circles, n = 2) and in lakes where adults were harvested (open circles, n = 3) before (2004–2006) and during (2007–2012) removals. The vertical line between 2006 and 2007 denotes when harvest began in harvest lakes. Values in the figures indicate P-values from the BACI design for each main factor and their interaction which would indicate a decline in abundance as a result of carp removal.

200 150 100 50 0 2007

2008

2009

2010

Year Fig. 3. Density of common carp populations (black circles ± 95% confidence intervals) and harvest (number per hectare) in lakes Herman, Madison, and Brant during 2007–2010.

or milt. Total length (nearest mm) was measured and a pectoral (2007) or dorsal (2008–2010) fin was removed for aging from a subset of individuals captured. Although different structures were used for aging, both pectoral rays and dorsal spines provide comparable age estimates (Weber and Brown, 2011b). Each fish was aged by two independent readers experienced with aging carp without prior knowledge of fish size. If the two age estimates differed, the structure was aged again and a consensus on the age estimate determined. A von Bertalanffy growth function was used to describe the growth of common carp Lt = L∞ × 1-e

were counted, and then released. Between five and ten capture and recapture events occurred on each system each year, which were incorporated in modified Schnabel’s population estimates for multiple capture-recapture events using the number of marked and unmarked individuals captured during each sampling event and the number of marked individuals at large (Van Den Avyle and Hayward, 1999). Commercial fishermen harvested common carp from all three systems during fall and winter 2007 through 2011 using a large bag seine described above. Common carp were harvested during these periods because they tend to form large aggregations then and are more vulnerable to harvest (Bajer et al., 2011). All individuals captured were examined for tags and enumerated to document harvest (number of common carp removed per hectare) and exploitation (number of tagged fish harvested divided by the total number of tagged individuals released that year corrected for tag loss (10.8%; see Section 3)). 2.4. Capture-recapture analysis

−K(t−t0)

where Lt = length at time t, L∞ = the theoretical maximum length of common carp in the population, K = the growth coefficient, and t0 = time when length would theoretically equal 0 mm. Due to the low numbers of fish captured <3 years of age (n = 15), initial fitting of von Bertalanffy models to the data resulted in illogical intercept (t0 ) estimates. Consequently, t0 was fixed at zero when deriving k and L∞ from the von Bertalanffy models (Weber et al., 2011). Common carp generally mature by age-3. Thus, the length at age-3 and k were used as growth indices and correlated with carp density in each of the lakes to evaluate potential density-dependent growth rates. Approximately one month after the conclusion of tagging, boat electrofishing and commercial seining was conducted to recapture individuals. All recaptured individuals were checked for presence of a tag that was subsequently read and recorded, fish

Individual encounter histories of tagged carp were analyzed in program MARK (White and Burnham, 1999) using the live-dead capture multistate model, which is an extension of the CormackJolly-Seber live recapture model extended for more than one area, allowing for transitions among areas (White et al., 2006). Using both live recaptures and dead (i.e., harvest) recoveries simultaneously increases precision and reduces bias of model estimates (Lebreton et al., 1999). Models assume that tagged individuals are representative of the population, number of tagged individuals is known, tagging does not influence survival, tags are not lost, all tags are correctly recorded, releases and recaptures occur over brief periods compared to time between tagging, recapture does not affect survival or future recapture, individual fates within and among cohorts are independent, individuals within a cohort have the same survival and recapture probability for each time interval, and parameter

172

7 Herman Madison Brant

6 5 4 3 2 1 0 0

50

100

150

200

250

300

100

120

Adult carp density (#/ha) Age-1 carp CPUE (number/net night)

estimates are conditional upon the model used (Burnham et al., 1987). Parameters of interest in the live-dead multistate model include apparent survival (S; probability a carp is alive, remains within the chain of lakes, and is available for recapture), detection probability (p), transition probability (ϕ), and reporting rate (probability that a carp is harvested and the tag is reported) of dead individuals (r; White et al., 2006). The multistate model uses individual tagged fish encounter history data to form the log of the likelihood function. For instance, transitioning from one state to another over time implies that the fish survived over that time interval. Similarly, not capturing an individual for multiple sampling events before capturing it again at a later date implies that the individual had previously survived but was not detected. A more detailed explanation of the encounter histories and how they are used in the multistate model to estimate survival, detection probability, and transition probability can be found in White and Burnham (1999) and White et al. (2006). There were three states that common carp could transition among: Herman, Madison, and Brant. Fish could also transition to a ‘tag loss’ state, which allowed us to estimate the rate of tag loss while adjusting other parameter estimates of interest (S, p, ϕ, r) accordingly. Transition probability from the ‘tag loss’ state back to one of the three lakes was fixed to 0 as individuals that shed tags are not able to reacquire the tag. Tagged and released individuals survive during time interval i with probability Si , are faithful to the state or transition to another state (ϕi ), and are recaptured during the sampling session k with probability pk (Fig. 2). We developed a set of 46 models to evaluate the effects of lake (L), year (t), gender (g), stream flow, and harvest on apparent survival, detection, and transition probabilities in addition to the effect of tag age on the probability of tag loss. The global model {S(L*g*t) p(L*g*t) ϕ(L*g*t) r(L*g*t) tag loss(L*g*t)} estimated the probability of annual survival (S), detection (p), interbasin movement among Herman, Madison, and Brant (ϕ), recovery during commercial fishing conditional on mortality (r), and tag loss by lake, gender, and year. Additional parameters evaluated were the effects of number of fish harvested (harvest) on S, average daily flow rate from Silver Creek (average flow) and flow during 2010 versus other years (flood) on ϕ, and effects of tag age (first year of tagging versus subsequent years) on tag loss. Our primary hypothesis was that common carp survival was inversely related to harvest, and that transition probability would be related to flow rate in Silver Creek. Alternatively, survival, transition probability, detection, and reporting rate could be unrelated to harvest or flow rate, but differ among

Age-1 carp CPUE (number/net night)

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

7 6 5 4 3 2 1 0 0

20

40

60

80

Adult carp harvest (#/ha) Fig. 5. Relationships between common carp density (number/hectare) and abundance of age-1 carp (number/net night; top pannel) and common carp harvest (number/hectare) and abundance of age-1 carp (number/net night; bottom panel) in lakes Herman, Madison, and Brant from 2007 through 2010.

the three lakes (L), between genders (g), among years (t), be constant (.) across lakes, genders, or years, and various combinations therein. Hypotheses were stated in model form in program MARK and compared using Akaike’s Information Criterion corrected for small sample size (AICc ; Burnham and Anderson, 1998). Because a reliable goodness-of-fit statistic does not exist for multistate models, model selection was determined with AICc rather than QAICc (Conn et al., 2004). Models with lower AICc values were considered more parsimonious and closer to the unknown ‘truth’. Akaike weights (wi ) were also calculated to address potential uncertainty concerning the selection of the best model (Burnham and Anderson,

Table 1 Number common carp tagged and recaptured within lake (Herman (H), Madison (M), and Brant (B)) and year from 2007–2011. Recapture location 2007

2008

Year

Tagging location

Number tagged

H

M

B

H

2007

Herman (H) Madison (M) Brant (B)

889 1538 927

215 0 1

0 67 0

0 0 0

60 1 0

2008

Herman (H) Madison (M) Brant (B)

2282 5490 2955

2009

Herman (H) Madison (M) Brant (B)

2098 6391 4007

2010

Herman (H) Madison (M) Brant (B)

991 1412 983

Total

Herman (H) Madison (M) Brant (B)

6260 14831 8872

215 0 1

0 67 0

2009 M

B

4 350 0

0 0 109

272 2 4 1396 0 0

0 0 104

0 332 6 0 5 1746 0 0 0

0 0 213

H 28 1 0

2010 M

B

H

2011 M

B

H

Total M

B

2 246 0

0 0 310

23 1 0

0 18 0

0 2 62

16 4 0

0 10 0

0 0 8

107 11 1 1309 0 0

0 1 1064

155 2 0

2 121 0

0 3 259

158 4 0

0 101 0

2 13 0 2200 0 0

0 0 1546

167 6 0

4 251 0

1 6 349

150 8 0

20 2 0

2 60 0

0 4 219

365 11 0

8 450 0

1 15 889

137 26 2 3755 0 0

0 1 2920

H 342 7 1

M

B

6 691 0

0 2 489

0 0 12

692 15 11 2927 0 0

0 4 1439

4 126 0

1 2 17

319 21 14 2577 0 0

2 8 1912

147 1 0

2 45 0

0 0 7

471 17 0

6 282 0

167 3 0

4 105 0

0 4 226

1 1520 46 2 35 6300 44 1 0

2 18 4066

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

Herman

173

Madison

Brant

15 2007

2007

2007

2008

2008

2008

2009

2009

2009

2010

2010

2010

12 9 6 3 0 15 12 9

Frequency (%)

6 3 0 15 12 9 6 3 0 15 12 9 6 3 0 200

400

600

800

1000

200

400

600

800

1000

200

400

600

800

1000

Length (mm) Fig. 6. Length-frequency distributions of common carp collected in seines from lakes Herman, Madison, and Brant during 2007 through 2010.

1998). After running all candidate models, model averaging was used to obtain robust estimates of daily apparent survival, detection probability, and escapement probability (Burnham and Anderson, 1998). Model averaging allows parameters that appear in all candidate models to be estimated based on Akaike weights, therefore allowing models with greater weight to provide more information in predicting parameter values (Burnham and Anderson, 1998).

3. Results 3.1. Population estimates and harvest rates A total of 29,963 carp were captured, tagged, and released during 2007–2010 with 11,988 subsequent recaptures of tagged individuals through 2011 (Table 1). Common carp population density ranged from 35.1 to 83.0 fish/ha in Lake Herman, 62.6 to 255.3 fish/ha in Lake Madison, and 108.6 to 210.7 fish/ha in Brant Lake (Fig. 3). From fall 2007 through fall 2011, 46,665 carp were harvested from Lake Herman, 133,427 were harvested from Lake Madison, and 76,488 were harvested from Brant Lake. Exploitation

on any given lake and year was highly variable ranging from 0.1 to 42.7% (Table 2). Carp harvest in Lake Herman was highest during 2007 (55.3 fish/ha), but considerably lower from 2008–2010 (mean = 3.1 fish/ha). During this time, common carp density in Lake Herman showed a decreasing trend (Fig. 3). In Lake Madison, the largest harvest occurred during 2008 (45.8 fish/ha) and 2009 (39.5 fish/ha) with lower harvest observed in 2007 (11.3 fish/ha) and 2010 (2.1 fish/ha). Carp density in Lake Madison was stable throughout the study (mean = 116.8 fish/ha (± 46.3 SE)) except during 2008 when density peaked at 255.3 fish/ha. The largest harvest in Brant Lake occurred during 2009 when 105.3 fish/ha were removed and the smallest occurred in 2007 when 0 fish/ha were removed. Carp density in Lake Brant increased from 108.6 fish/ha in 2007 to 210.7 fish/ha in 2010 (Fig. 3) despite a general trend of increasing harvest from 2007–2010,

3.2. Relative abundance, recruitment, and growth Mean relative abundance of adult common carp during 2004–2007 prior to harvest was 6.9 (±4.0 SE) fish per trap net night

174

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

Table 2 Number of common carp tagged, tagged fish harvested, and annual exploitation (%; corrected for 10.8% tag loss) of tagged fish in lakes Herman, Madison, and Brant from 2007–2010. Total tagged and harvested represent the summation of carp tagged and harvested across lakes within a year whereas total exploitation is the number of fish harvested divided by the number of fish tagged across lakes within a year corrected for tag loss. 2007

2008

2009

2010

Total

Tagged Harvested Exploitation Tagged Harvested Exploitation Tagged Harvested Exploitation Tagged Harvested Exploitation Tagged Harvested Exploitation 26.8% 4.8% 0.1% 9.3%

2282 275 5490 1400 2955 104 10727 1779

13.4% 28.3% 3.9% 18.4%

2098 15 6391 2200 4007 1546 12496 3761

in harvest lakes compared to 1.7 (±0.5 SE) fish per trap net night in control lakes (Fig. 4). After harvest began in fall 2007, relative abundance of adult common carp declined to 4.0 (±1.0 SE) fish per trap net night in harvest lakes but doubled to 3.4 (±0.7 SE) fish per trap net night in control lakes (Fig. 4). However, trends in adult carp relative abundance were similar between harvest and control lakes and the effects of treatment, period, and their interaction were insignificant (Fig. 4). Abundance of age-1 carp during 2004–2007 was 0.6 (±0.3 SE) fish per trap net night in harvest lakes compared to 0.1 (±0.1 SE) fish per trap net night in control lakes (Fig. 4). After harvest started in fall 2007, relative abundance of age-1 common carp was 0.6 (±0.4 SE) fish per trap net night in harvest lakes compared to 0.3 (±0.1 SE) fish per trap net night in control lakes (Fig. 4). Similar to abundance of adult carp, trends in relative abundance of age-1 carp were comparable between harvest and control lakes and the effects of treatment, period, and their interaction on were insignificant (Fig. 4). Finally, adult common carp density and harvest rates were not related to abundance of age-1 carp in lakes Herman, Madison, or Brant (Fig. 5). The size distribution of common carp collected in the seine was similar during 2007–2010 in all three lakes (Fig. 6). Size distributions of carp collected using electrofishing and in seines did not differ significantly (Kolmogorov-Smirnov test P > 0.05). Most fish collected were between 400–800 mm, although a few smaller and larger individuals were captured in some instances. Growth rates, as estimated with k, ranged between 0.21 and 0.48 year−1 , and mean total length at age-3 varied between 424 mm to 569 mm (Fig. 7). However, neither k or total length at age-3 was related to carp density in any of the lakes (P > 0.05).

3.3. Survival, interbasin movement, detection, and reporting rate The global model with common carp survival, detection probability, interbasin movement, reporting rate, and tag loss dependent upon lake, gender, and time was not supported (model rank: 46 of 46). Of the 46 models evaluated, two models received substantial support. Both of the best two models had harvestdependent survival (S), lake- and time-dependent detection probability (p), lake and time dependent reporting probability (r), and age dependent tag loss. Differences between best top two models occurred within the interbasin movement parameter: the best model contained annual variation in interbasin movement (ϕ; AICc = 0, Wi = 0.56) and second ranked model contained flood-dependent interbasin movement (ϕ; AICc = 0.5, Wi = 0.44; Table 3). Tag loss was estimated at 10.8% the first year following tagging (95% CI = 6.0 − 18.6%) but low thereafter (<0.1%). There was little support for gender-dependent models, indicating similar survival, detection probability, interbasin movement, and reporting rate for male and female carp. During years when no harvest occurred, carp survival was 79% (95% confidence interval = 75–82%). As harvest increased, carp survival declined (Fig. 8A). However, the slope of the relationship between survival and exploitation was less than one (Fig. 8B). At the highest level of exploitation (42.7%), com-

0.8% 38.1% 42.7% 33.3%

991 1412 983 3386

22 66 218 306

2.5% 5.2% 24.6% 10.0%

6260 527 14831 3733 8872 1869 29963 6129

9.3% 27.9% 23.3% 22.7%

0.50 0.45 Growth rate (k)

215 67 1 283

0.40 0.35 0.30 0.25 0.20 0.15 600 550

Age-3 TL (mm)

Herman 889 Madison 1538 Brant 927 3354 Total

500 450 Herman Madison Brant

400 0

50

100

150

200

250

300

Adult carp density (#/ha) Fig. 7. Relationship between common carp density (number/hectare) and growth rate (top) and mean length at age-3 (bottom) in lakes Herman, Madison, and Brant from 2007 through 2010.

mon carp survival declined 25% to an annual apparent survival rate of 54% (95% CI = 51–58%). Common carp interbasin movement was low, with annual probability of movement typically <0.2%. The most supported model indicated that interbasin movement was similar from Herman to Madison, Madison to Herman, and Madison to Brant but movement rates were variable from 2007–2011 (Table 3). Patterns of annual interbasin movement closely matched patterns in mean summer flow rate (Fig. 9). However, there was little support that annual variation in summer flow rates in Silver Creek was related to interbasin movement rates of common carp (model 6, Table 3). The highest interbasin movement occurred during 2010 (mean = 0.7%, 95% CI = 0.3–1.7%), which corresponded with the highest observed flow rates in Silver Creek and there was substantial evidence that 2010 flooding resulted in higher movement rates compared to other nonflood years (Table 3, Model 2). Detection probability and reporting rate depended on both lake and time. Detection probability was lowest in 2011, but similar from 2008–2010 (Fig. 10). Common carp in Brant Lake had the highest detection probability in 2008 and 2009 whereas carp in Lake Herman had the highest detection probability in 2010. Mean reporting rate was 37% (±8% SE) and remained stable from 2007 through 2011 but varied greatly among lakes (Fig. 10). The Lake Herman population had the highest reporting rate during 2007 and

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

1.2

P e rc e n t ( %)

80

Exploitation Survival

60 40 20 Exploitation = 0.28*harvest + 2.95 r2 = 0.83, P < 0.0001

0 0

20

40

60

80

100

120

Survival (%)

60 50 40 20

30

40

200

0.6

150

0.4

100

0.2 50

2009

2010

2011

Fig. 9. Common carp interbasin movement probability and spring-summer flow rates in Silver Creek (±1 standard error) from 2008–2011.

B

70

10

250

0.8

2008

80

0

1.0

0

140

100 Survival = -0.60*Exploitation + 79.36 r2 = 0.81, P < 0.0001

Interbasin movement Discharge

0.0

Carp harvested (#/ha)

90

300

A

Flow rate (m3/minute)

Survival = -0.21*Harvest + 79.04 r2 = 0.99, P < 0.0001

Interbasin movement (%)

100

175

50

Exploitation (%) Fig. 8. Relationships between number of carp harvested and exploitation and survival (A) and between exploitation and survival (B) in lakes Herman, Madison, and Brant from 2007–2010. In panel B, the solid represents a 1:1 line, the long dash represents the survival regression, and the short dash represents 95% confidence intervals for the survival regression.

2011, Lake Madison had the highest reporting rate in 2008, and Lake Brant had the highest reporting rate in 2009 and 2010. 4. Discussion Our work indicates that although survival of common carp can be reduced through exploitation, population abundance may not be reduced, potentially due to compensatory processes. Consequently, high exploitation rates (>50%) sustained over long time periods are needed if harvest management is to be successful at suppressing population abundance. In other studies, abundance of invaders were also not reduced following harvest management directed at population eradication or sustained reduction. For example, biomass of common carp in Clear Lake, Iowa increased following consecutive years of intensive harvest (Colvin et al., 2012). Lake trout Salvelinus namaycush abundance increased in Yellowstone

Lake following more than a decade of intensive harvest (Syslo et al., 2011). Smallmouth bass Micropterus dolomieu populations also increased following seven years of intensive harvest (Weidel et al., 2007). Because successful invaders typically have few co-evolved predators or competitors, they are often present at high densities near carrying capacity, which are difficult to control compared to lower-density populations (Bartmann et al., 1992). Population densities of common carp in this study are high compared to many other areas (35–255 individuals/ha, 175–1276 kg/ha) likely making it more difficult to control their populations. Population abundance can remain stable or increase only if one or more density-dependent functions (i.e., recruitment or survival) are compensatory. Theoretical models indicate that recruitment is the primary parameter responsible for overcompensation and that density-dependent survival alone is insufficient to produce a compensatory response (Zipkin et al., 2009). However, our results indicate that density-dependent survival may have produced stable common carp population abundance, as survival did not decline proportionally with exploitation. The partial compensatory mortality hypothesis predicts that harvest mortality may be partially additive, but that compensatory survival is exhibited at low levels of harvest (Nichols et al., 1984). Because carp survival did not decline proportionally with harvest or exploitation, our results suggest that common carp can exhibit partial compensatory mortality. One issue with trying to evaluate the potential for compensatory survival is the difficulty in accurately measuring survival rates, which may explain why it has rarely been used to explain compensatory population responses. Multistate capture-recapture models have rarely been used to evaluate survival and movement parameters of fishes (but see Massicotte et al., 2008; Frank et al., 2012) but here provided a useful approach to evaluate these important dynamics and population responses to exploitation.

Table 3 Model selection results from Program MARK evaluating factors influencing survival (S), live encounter probability (p), interbasin movement (ϕ), reporting probability of a dead animal (r), and tag loss for common carp in lakes Herman, Madison, and Brant from 2007–2011. Only the top 10 of 46 models run are presented. Model likelihood is calculated as exp(−0.5*AICc) . Model

AICc a

Wi

Model Likelihood

Deviance

S(harvest) p(L*t) ϕ (t) r(L*t) tag loss(age) S(harvest) p(L*t) ϕ (Flood) r(L*t) tag loss(age) S(L*g*t) p(t) ϕ (t) r(L*g) tag loss(age) S(L*g*t) p(t) ϕ (t) r(L) tag loss(age) S(L*t) p(t) ϕ (t) r(L*g) tag loss(age) S(harvest) p(L*t) ϕ (avg flow) r(L*t) tag loss(age) S(L*g*t) p(t) ϕ (t) r(.) tag loss(age) S(L*g*t) p(t) ϕ (t) r(t) tag loss(age) S(L*g*t) p(t) ϕ (L) r(.) tag loss(age) S(L*g*t) p(t) ϕ (2007–2009 vs 2010–2011) r(.) tag loss(age)

0.0 0.5 26.7 28.7 32.9 35.2 43.2 62.0 70.1 71.1

0.56 0.44 0 0 0 0 0 0 0 0

1 0.77 0 0 0 0 0 0 0 0

63366.3 63368.8 63363.0 63371.0 63399.3 63405.5 63389.5 63400.2 63418.4 63421.4

a

AICc for the top model was 63428.4.

176

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

16 Herman Madison Brant

Detec tio n p ro b ab ility (%)

14 12 10 8 6 4 2

NA

0 2007

2008

2009

2010

2011

2007

2008

2009

2010

2011

100

R ep o rting rate (%)

80

60

40

20

0 Year Fig. 10. Detection probability and reporting rate (±1 standard error) for common carp in lakes Herman, Madison, and Brant from 2007–2011 estimated from Program MARK. Detection probability was not estimable in 2007 (NA).

Heterogeneity and seasonal timing of harvest can be two factors influencing whether harvest mortality is additive or compensatory. First, survival and catchability heterogeneity between genders can result in a compensatory response (Lebreton, 2005). We tested for variation in survival, detection, and reporting rates between males and females but these models received little support. Thus, it is unlikely that gender heterogeneity produced a compensatory response in these study populations. Second, seasonal timing of harvest can mediate its impact on survival. Harvest is more likely to induce additive mortality when it follows or overlaps periods of high natural mortality, but will be compensatory if harvest precedes seasonal periods of high natural mortality (Hudson et al., 1997; Boyce et al., 1999; Ratikainen et al., 2008). The sources of natural mortality in these carp populations are not known. However, severe winter conditions (e.g., prolonged ice and snow cover, low dissolved oxygen) can induce mass mortality of carp populations in shallow lakes (Weber and Brown, 2016). Although mass-mortality events were not observed in these systems during the study, dead individuals were periodically observed in spring following ice-off. Thus, we suspect that stressful overwinter conditions followed by outbreaks of spring viremia induced most natural mortality observed in these populations. Spring viremia, a contagious viral (a rhabdovirus) disease, occurs annually in eastern South Dakota lakes, including lakes Herman, Madison and Brant, with variable virulence (R. Neiger, Animal Disease Research and Diagnostic Lab, SDSU, personal communication). In our study, common carp were harvested in the fall or winter, before or during the suspected period of high natural mortality. Thus, timing of harvest may have added to

the probability that it would be compensatory rather than additive mortality. Harvest mortality is more likely to be additive for species with high survival and low fecundity because they have a lower capacity to compensate natural mortality with harvest whereas highly fecund species with low survival often produce a surplus of offspring allowing for them to compensate for adult harvest (Coulson et al., 2000; Sedinger et al., 2007). Populations experiencing high exploitation may respond by increasing fecundity (Baccante and Reid, 1988) and recruitment (Ricker 1954), resulting in increases in population abundance (Weidel et al., 2007). Compensation is common in populations where maximum per capita fecundity is large (≥4) resulting in stable or increasing population abundance (Zipkin et al., 2009) and the likelihood of overcompensation increases when survival is high during juvenile and adult life stages and when only adults are harvested (Zipkin et al., 2009; Weber et al., 2011). Common carp are highly fecund (Weber and Brown, 2012), age-0 common carp abundance can be high and regulated by densitydependent processes (Weber and Brown, 2013), and mechanical removal efforts are highly size-selective, targeting large individuals (Weber et al., 2011). Thus, great potential exists for common carp populations to exhibit compensatory recruitment in response to harvest. However, we found no evidence indicating that recruitment (abundance of carp <325 mm) increased in these systems following exploitation, potentially because adult densities were not sufficiently reduced in these systems. Artificial size-selective fishing mortality can have a significant influence on individual life histories. Body size is an important and practical response variable in natural resource management (Hoxmeier et al., 2009). We hypothesized compensatory responses of common carp to include increased growth rates and size structure as a result of decreases in intraspecific competition. Carp populations in South Dakota appear to exhibit density-dependent growth (Weber et al., 2010; Weber and Brown, 2013). However, we found no relationship between carp density and growth rates during this study. Growth rates of carp in these populations were high prior to harvest compared to other areas of invasion (Weber et al., 2010). Additionally, despite high harvest rates, carp abundance was not reduced, which may not have released populations from density-dependent processes. Emigration and immigration may also affect the outcomes of harvest-oriented management. Dispersal of individuals can result in compensation through demographic rescue and source-sink dynamics (Cooley et al., 2009) whereas harvest is more likely to be additive for populations with limited dispersal ability (Guthery et al., 2000). Less than 1% of the population displayed interbasin movement among these systems on an annual basis, but immigration was related to environmental conditions that increased connectivity of these systems (i.e., flood events). The probability of fish immigrating was highest during a period of high water. In highly-interconnected lakes, common carp may display spawning movements to adjacent water bodies but display site fidelity to their lake of origin (Bajer and Sorensen, 2010). In contrast, systems evaluated here are further apart, less connected to one another, and none of the fish that that displayed interbasin movements returned to their lake of origin (although our interbasin movement and detection probability results indicate the chances of this occurring would be quite low). Although immigration was only 0.7% in 2010, we estimate it resulted in more than 2000 carp moving among systems. Immigration of only a few individuals can result in rapid population recovery (Cooley et al., 2009), especially for highly fecund species including carp (Weber and Brown, 2012). Thus, even limited movement between basins must be stopped to control populations and prevent recolonization. A major challenge of harvest-oriented management is that harvest rates are rarely high enough for long enough periods of time to

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

allow for successful control (Wydoski and Wiley, 1999; Colvin et al., 2012). Common carp control specifically is challenging and multiple removal attempts are likely necessary before reductions are achieved (Schrage and Downing, 2004; Weber and Brown, 2009; Colvin et al., 2012). Models indicate that long-term harvest rates between 20–40% exploitation may reduce common carp size structure, abundance, and spawning potential (Weber et al., 2011; Colvin et al., 2012), but ignore potential for density-dependent processes. Additionally, commercial fisheries that remove non-native fishes such as carp are highly size-selective, targeting large-bodied individuals (Weber et al., 2011). These types of adult-only removal strategies make it difficult to reduce population abundance because they do not account for recruitment of new individuals into the population. Even when removals are successful at reducing abundance, success is often short lived and population abundance rapidly increases if persistent control strategies are not maintained (Meijer et al., 1990; Colvin et al., 2012). An added complication is that commercial fishers have a vested interest in sustainable harvest to ensure they do not collapse populations. Thus, fisheries managers may need to provide commercial fishers financial incentives to achieve higher levels of exploitation over consecutive years that may be more likely to control populations. Large-scale experiments evaluating population-level responses of invasive species to harvest-oriented management are rare but needed to design improved control measures. Our evaluation of common carp harvest-oriented management provided key insights into the management of invasive species. Previously, empirical evidence of invaders exhibiting a compensatory response to exploitation was rare. However, by experimentally manipulating exploitation over three systems and five years, our results reveal that common carp may respond to harvest with partial compensatory survival. Evidence for partial compensatory survival by common carp is consistent with four of the five factors that determine population responses to harvest: (1) common carp are highly fecund, (2) population densities are high, (3) harvest was seasonal and likely occurred prior to periods of high natural mortality, and (4) movement among populations occurs. Yield models that predict population responses to harvest assume additive mortality (Colvin et al., 2012) but may underestimate the amount of harvest required to control populations when harvest is compensatory. Our results indicate controlling invasive species through harvest-oriented management must take into account population-level responses to harvest and potential for densitydependent responses. Due to the increasing number of detrimental invaders and the need to control these populations, the likelihood for compensatory mortality to affect harvest management of other invaders warrants further investigation. Acknowledgments We thank Dave Raw Fish, Inc., for collecting carp for tagging and conducting carp removals. We also thank the numerous technicians and students at South Dakota State University and South Dakota Game, Fish and Parks that assisted with tagging and data collection for this project. Finally, we thank G. White and W. Kendall at Colorado State University for assistance with Program MARK analyses. Partial funding for this project was provided through the Federal Aid in Sport Fish Restoration Act Study 1513 (Project F-15-R-42) administered through South Dakota Department of Game, Fish and Parks and the South Dakota Agricultural Experiment Station. References Abrams, P.A., 2009. When does greater mortality increase population size?: The long history and diverse mechanisms underlying the hydra effect. Ecol. Lett. 12, 462–474.

177

Allan, J.D., Flecker, A.S., 1993. Biodiversity conservation in running waters: identifying the major factors that threaten destruction of riverine species and ecosystems. BioScience 43, 32–43. Baccante, D.A., Reid, D.M., 1988. Fecundity changes in two exploited walleye populations. N. Am J. Fish. Manage. 8, 199–209. Bajer, P.G., Sorensen, P.W., 2010. Recruitment and abundance of an invasive fish the common carp, is driven by its propensity to invade and reproduce in basins that experience winter-time hypoxia in interconnected lakes. Biol. Invasions 12, 1101–1112. Bajer, P.G., Chizinski, C.J., Sorensen, P.W., 2011. Using the Judas technique to locate and remove wintertime aggregations of invasive common carp. Fish. Manage. Ecol. 18, 497–505. Bartmann, R.M., White, G.C., Carpenter, L.H., 1992. Compensatory mortality in a Colorado mule deer population. Wildl. Monogr. 121, 3–39. Boyce, M.S., Sinclair, A.R.E., White, G.C., 1999. Seasonal compensation of predation and harvesting. Oikos 87, 419–426. Brooks, E.N., 2002. Using reproductive values to define optimal harvesting for multisite density-dependent populations: example with a marine reserve. Can. J. Fish. Aquat. Sci. 59, 875–885. Brown, P., Gilligan, D., 2014. Optimizing an integrated pest-management strategy for a spatially structured population of common carp (Cyprinus carpio) using meta-population modelling. Mar. Freshwater Res. 65, 538–550. Brown, P., Walker, T.I., 2004. CARPSIM: stochastic simulation modelling of wild carp (Cyprinus carpio L.) population dynamics: with applications to pest control. Ecol. Modell. 176, 83–97. Burnham, K.P., Anderson, D.A., 1998. Model Selection and Inference. Springer-Verlag, New York, NY. Burnham, K.P., Anderson, D.R., White, G.C., Brownie, C., Pollock, K.H., 1987. Design and Analysis Methods for Fish Survival Experiments Based on Release-Recapture. American Fisheries Society, Monograph 5, Bethesda, Maryland. Byers, J.E., Reichard, S., Randall, J.M., Parker, I.M., Smith, C.S., Lonsdale, W.M., Atkinson, I.A.E., Seastedt, T.R., Williamson, M., Chornesky, E., Hayes, D., 2002. Directing research to reduce impacts of nonindigenous species. Conserv. Biol. 16, 630–640. Coggins, L.G., Yard, M.D., Pine, W.E., 2011. Nonnative fish control in the Colorado River in Grand Canyon, Arizona: an effective program or serendipitous timing? Trans. Am. Fish. Soc. 140, 456–470. Colvin, M.E., Pierce, C.L., Stewart, T.W., Grummer, S.E., 2012. Strategies to control a common carp population by pulsed commercial harvest. N. Am. J. Fish. Manage. 32, 1251–1264. Conn, P.B., Kendall, W.L., Samuel, M.D., 2004. A general model for the analysis of mark-resight, mark-recapture, and band-recovery data under tag loss. Biometrics 60, 900–909. Cooley, H.S., Wielgus, R.B., Koehler, G.M., Robinson, H.W., Maletzke, B.T., 2009. Does hunting regulate cougar populations: a test of the compensatory mortality hypothesis. Ecology 90, 2913–2921. Coulson, T., Milner-Gulland, E.J., Clutton-Brock, T., 2000. The relative roles of density and climatic variation on population dynamics and fecundity rates in three contrasting ungulate species. Pro. R. Soc. Lond. B 267, 1771–1779. De Roos, A.M., Schellekens, T., van Kooten, T., Van De Wolfsharr, K., Claessen, D., Persson, L., 2007. Food-dependent growth leads to overcompensation in stage-specific biomass when mortality increases: the influence of maturation versus reproduction regulation. Am. Nat. 170, 59–76. Frank, B.M., Gimenez, O., Baret, P.V., 2012. Assessing brown trout (Salmo trutta) spawning movements with multistate capture-recapture models: a case study in a fully controlled Belgian brook. Can. J. Fish. Aquat. Sci. 69, 1091–1104. Frederiksen, M., Lebreton, J.D., Bregnballe, T., 2001. The interplay between culling and density-dependence in the great cormorant: a modeling approach. J. Appl. Ecol. 38, 617–627. Gotelli, N.J., 2001. A Primer of Ecology, third edition. Sinauer Associates, Sunderland, MA, USA. Guthery, F.S., Peterson, M.J., George, R.R., 2000. Viability of northern bobwhite populations. J. Wildl. Manage. 64, 656–662. Harig, A.L., Bain, M.B., 1998. Defining and restoring biological integrity in wilderness lakes. Ecol. Appl. 8, 71–87. Hein, C.L., Roth, B.M., Ives, A.R., Vander Zanden, M.J., 2006. Fish predation and trapping for rusty crayfish (Orconectes rusticus) control: a whole-lake experiment. Can. J. Fish. Aquat. Sci. 63, 383–393. Howald, G., Donlan, D.J., Galvan, J.P., Russell, J.C., Parkes, J., Samaniego, A., Wang, Y., Veitch, D., Genovesi, P., Pascal, M., Saunders, A., Tershy, B., 2007. Invasive rodent eradication on islands. Conserv. Biol. 21, 1258–1268. Hoxmeier, J., Aday, D.D., Wahl, D.H., 2009. Examining interpopulation variation in bluegill growth rates and size structure: effects of harvest, maturation, and environmental variables. Trans. Am. Fish. Soc. 138, 423–432. Hrabik, T.R., Magnuson, J.J., McLain, A.S., 1998. Predicting the effects of rainbow smelt on native fishes in small lakes: evidence from long-term research on two lakes. Can. J. Fish. Aquat. Sci. 55, 1364–1371. Hudson, P.J., Newborn, D., Robertson, P.A., 1997. Geographical and seasonal patterns of mortality in red grouse Lagopus lagopus scoticus populations. Wildl. Biol. 3, 79–87. Jackson, Z.J., Quist, M.C., Downing, J.A., Larscheid, J.G., 2010. Common carp (Cyprinus carpio), sport fishes, and water quality: Ecological thresholds in agriculturally eutrophic lakes. Lake Res. Manage. 26, 14–22. Koehn, J.D., 2004. Carp (Cyprinus carpio) as a powerful invader in Australian waterways. Fresh. Biol. 49, 882–894.

178

M.J. Weber et al. / Fisheries Research 179 (2016) 168–178

Lebreton, J.D., 2005. Dynamical and statistical models for exploited populations. Aust. N. Z. J. Stat. 47, 49–63. Lebreton, J.D., Almeras, T., Pradel, R., 1999. Competing events, mixture of information and multistratum recapture models. Bird Study 46S, S32–38. Letvin, A., 2013. Changes in food web dynamics of shallow lakes as a result of common carp perturbation. In: Master’s Thesis. State University, South Dakota. Lodge, D.M., Williams, S., MacIsaac, H.J., Hayes, K.R., Leung, B., Reichard, S., Mack, R.N., Moyle, P.B., Smith, M., Andow, D.A., Carlton, J.T., McMichael, A., 2006. Biological invasions: recommendations for US policy and management. Ecol. Appl. 16, 2035–2054. Massicotte, R., Magnan, P., Angers, B., 2008. Intralacustrine site fidelity and nonrandom mating in the littoral-spawning northern redbelly dace (Phoxinus eos). Can. J. Fish. Aquat. Sci. 65, 2016–2025. Matsuda, H., Abrams, P.A., 2004. Effects of predator-prey interactions and adaptive change on sustainable yield. Can. J. Fish. Aquat. Sci. 61, 175–184. Meijer, M.-L., de Hann, M.W., Breukelaar, A.W., Buiteveld, H., 1990. Is reduction of benthivorous fish an important cause of high transparency following biomanipulation in shallow lakes? Hydrobiologia 200/201, 303–315. Meronek, T.G., Bouchard, P.M., Buckner, E.R., Burri, T.M., Demmerly, K.K., Hatleli, D.C., Klumb, R.A., Schmidt, S.H., Coble, D.W., 1996. A review of fish control projects. N. Am. J. Fish. Manage. 16, 63–74. Myers, J.H., Simberloff, D., Kuris, A.M., Carey, J.R., 2000. Eradication revisited: dealing with exotic species. Trends Ecol. Evol. 15, 316–320. Nichols, J.D., Conroy, M.J., Anderson, D.R., Burnham, K.P., 1984. Compensatory mortality in waterfowl populations: a review of the evidence and implications for research and management. Trans. N. Am. Wildl. Conf. 49, 535–554. Parker, I.M., 2000. Invasion dynamics of Cytisus scoparius: a matrix model approach. Ecol. Appl. 10, 726–743. Parker, I.M., Simberloff, D., Londsdale, W.M., Goodell, K., Wonham, M., Kareiva, P.M., Williamson, M.H., Von Holle, B., Moyle, P.B., Byers, J.E., Goldwasser, L., 1999. Impact: toward a framework for understanding the ecological effects of invaders. Biol. Invasions 1, 3–19. Peterson, D.P., Fausch, K.D., White, G.C., 2004. Population ecology of an invasion: effects of brook trout on native cutthroat trout. Ecol. Appl. 14, 754–772. Pimentel, D., Lach, L., Zuniga, R., Morrison, D., 2000. Environmental and economic costs associated with non-indigenous species in the United States. BioScience 50, 53–65. Rahel, F.J., 2002. Homogenization of freshwater faunas. Annu. Rev. Ecol. Syst. 33, 291–315. Ratikainen, I.I., Gill, J.A., Gunnarsson, T.G., Sutherland, W.J., Kokko, H., 2008. When density dependence is not instantaneous: theoretical developments and management implications. Ecol. Lett. 11, 184–198. Ricker, W.E., 1954. Stock and recruitment. J. Fish. Res. Board Can. 11, 559–623. Sakai, A.K., Allendorf, F.W., Holt, J.S., Lodge, D.M., Molofsky, J., With, K.A., Baughman, S., Cabin, R.J., Cohen, J.E., Ellstrand, N.C., McCauley, D.E., O’Neil, P., Parker, I.M., Thompson, J.N., Weller, S.G., 2001. The population biology of invasive species. Annu. Rev. Ecol. Syst. 32, 305–332. Scheffer, M., Carpenter, S., Foley, J.A., Folke, C., Walker, B., 2001. Catastrophic shifts in ecosystems. Nature 413, 591–596. Schmitt, R.J., Osenberg, C.W., 1996. Detecting Ecological Impacts. Academic Press, Oxford, United Kingdom. Schrage, L.J., Downing, J.A., 2004. Pathways of increased water clarity after fish removal from Ventura Marsh; a shallow, eutrophic wetland. Hydrobiologia 511, 215–231.

Sedinger, J.S., Nicolai, C.A., Lensink, C.J., Wentworth, C., Conant, B., 2007. Black brant harvest, density dependence, and survival: a record of population dynamics. J. Wild. Manage. 71, 496–506. Simberloff, D., 2001. Eradication of island invasives: practical actions and results achieved. Trends Ecol. Evol. 16, 273–274. Syslo, J.M., Guy, C.S., Bigelow, P.E., Doepke, P.D., Ertel, B.D., Koel, T.M., 2011. Response of non-native lake trout (Salvelinus namaycush) to 15 years of harvest in Yellowstone Lake, Yellowstone National Park. Can. J. Fish. Aquat. Sci. 68, 2132–2145. Van Den Avyle, M.J., Hayward, R.H., 1999. Dynamics of exploited fish populations. In: Kohler, C.C., Hubert, W.A. (Eds.), Inland Fisheries Management in North America. , 2nd edition. American Fisheries Society, Bethesda, MD, pp. 127–163. Vermeij, G.J., 1996. An agenda for invasion biology. Biol. Conserv. 78, 3–9. Vitousek, P.M., D’Antonio, C.M., Loope, L.L., Westbrooks, R., 1996. Biological invasions as global environmental change. Am. Sci. 84, 468–477. Weber, M.J., Brown, M.L., 2016. Effects of resource pulses on nutrient availability, ecosystem productivity, and temporal variability following a stochastic disturbance in eutrophic glacial lakes. Hydrobiologia, http://dx.doi.org/10. 1007/s10750-015-2628-z. Weber, M.J., Brown, M.L., 2013. Density-dependence and environmental conditions regulate recruitment and first year-growth of common carp in shallow lakes. Trans. Am. Fish. Soc. 142, 471–482. Weber, M.J., Brown, M.L., 2012. Maternal effect of common carp on egg quantity and quality. J. Fresh. Ecol. 27, 409–417. Weber, M.J., Brown, M.L., 2011a. Relationships among invasive common carp: native fishes and physicochemical characteristics in upper Midwest (USA) lakes. Ecol. Freshwater Fish 20, 270–278. Weber, M.J., Brown, M.L., 2011b. Comparison of common carp (Cyprinus carpio) age estimates derived from dorsal fin spines and pectoral fin rays. J. Freshwater Ecol. 26, 195–202. Weber, M.J., Hennen, M.J., Brown, M.L., 2011. Simulated population responses of common carp to commercial exploitation. N. Am J. Fish. Manage. 31, 269–279. Weber, M.J., Brown, M.L., Willis, D.W., 2010. Spatial variability of common carp populations in relation to lake morphology and physicochemical parameters in the upper Midwest United States. Ecol. Freshwater Fish 19, 555–565. Weber, M.J., Brown, M.L., 2009. Effects of common carp on aquatic ecosystems 80 years after ‘Carp as a dominant’: ecological insights for fisheries management. Rev. Fish. Sci. 17, 524–537. Weidel, B.C., Josephson, D.C., Kraft, C.E., 2007. Littoral fish community response to smallmouth bass removal from an Adirondack lake. Trans. Am. Fish. Soc. 136, 778–789. White, G.C., Burnham, K.P., 1999. Program MARK: survival estimation from populations of marked animals. Bird Study (Suppl 46), 120–138. White, G.C., Kendall, W.L., Barker, R.J., 2006. Multistate survival models and their extensions in program MARK. J. Wildl. Manage. 70, 1521–1529. Wydoski, R.S., Wiley, R.W., 1999. Management of undesirable fish species. In: Kohler, C.C., Hubert, W.A. (Eds.), Inland Fisheries Management in North America. American Fisheries Society, Bethesda, Maryland, pp. 403–430. Zipkin, E.F., Kraft, C.E., Cooch, E.G., Sullivan, P.J., 2009. When can efforts to control nuisance and invasive species backfire? Ecol. Appl. 19, 1585–1595. Zipkin, E.F., Sullivan, P.J., Cooch, E.G., Kraft, C.E., Shuter, B.J., Weidel, B.C., 2008. Overcompensatory response of a smallmouth bass (Micropterus dolomieu) population to harvest: release from competition? Can. J. Fish. Aquat. Sci. 65, 2279–2292.