Stock-recruitment Relationship Research Articles

Recruitment and survival rate variability in fish populations: density-dependent regulation or further evidence of environmental determinants?

Recently, Minto et al. (2008) , based on a fishery data set including marine, estuarine, and freshwater fishes, described higher variability in the survival rates of juveniles at low rather than at high parental density in an inversely density-dependent fashion and suggested density-dependent mechanisms underpinning those patterns. This study, based on a long-term study of brown trout (Salmo trutta; a species and habitat not included in the Minto et al. (2008) analysis), documents that survival rates in these stream-living populations exhibit a pattern that matches exactly those reported by Minto et al. (2008) . Nevertheless, hypothesis testing rejected the occurrence of stock–recruitment relationships and the operation of density-dependent recruitment regulation. The patterns elucidated for these brown trout populations can be entirely explained by the operation of two single environmental factors, namely, stream discharge in March determining annual survival rates across streams and sites and site-specific depth determining site-specific survival rates. It is open to question that exactly the same patterns can be generated by two sets of opposing factors, density-dependent (i.e., Minto et al. 2008 ) and environmental factors (i.e., this study). The consistency of this pattern suggests that survival rates and recruitment are probably determined by environmental factors across fish populations and habitats.

Model for stock-recruitment dynamics of the Peruvian anchoveta (<i>Eugraulis ringens</i>) off Peru

This paper was aimed at re-examining the validity of the results from Cahuin et al. (Estuar. Coast. Shelf Sci. 84, 2009) and identifying a model to describe the stock-recruitment relationship of the Peruvian anchoveta (Eugraulis ringens). Regression analysis was used to determine if density-dependent effects were present. The analysis did not show the existence of any densitydependent effects. It is important to use environmental factors and take observational and process errors into account when attempting to identify density-dependent effects in fish populations. Sea surface temperature (SST) and Southern Oscillation Index (SOI) were used as independent variables to fit the recruitment dynamics of the anchoveta. Both SST and SOI were found to be significantly important parameters in structuring anchoveta dynamics according to Akaike Information Criterion (AIC) and R2 values. The results of this study do not correlate with the findings of Cahuin et al., (2009), where density-dependent effects and the presence of regimes were detected. In conclusion, the recruitment Rt is essentially determined in proportion to spawning stock biomass St, and then environmental factors in year t further change the recruitments. This mechanism is completely same with that for Japanese sardine proposed by Sakuramoto (The Open Fish. Sci. 5, 2012).

Modeling the Effect of Environmental Factors on the Ricker Stock-Recruitment Relationship for North Pacific Albacore Using Generalized Additive Models

This study investigates how the inclusion of El Nino episodes affects the modeling of recruitment in North Pacific albacore. The relationship between spawning stock size and recruitment for the 1976 - 2004 period, including environmental variables (Southern Oscillation Index and sea surface temperature anomalies), was conducted by general linear and generalized additive models (GAM). The results indicate that the Southern Oscillation Index and the interaction between spawning stock size and sea surface temperature anomalies have significant effects on the recruitment of North Pacific albacore. GAM fit the original data better than general linear models, according to the Akaike information criterion. The recruitment declined when El Nino episodes occurred, but increased when La Nino episodes occurred. The recruitment also increased when the spawning stock size density was above 11.21 million tons with a -0.3°C sea surface temperature anomaly. El Nino was observed to have both long-period (above 10 - 15 years) and short-period (1 - 2 years) effects on North Pacific recruitment.

Relating a recruitment shift of Patagonian grenadier (Macruronus magellanicus Lönnberg) to large-scale environmental changes off Southern Chile

The SE Pacific stock of Patagonian grenadier (Macruronus magellanicus) showed evidence of an abrupt reduction in recruitment after 2000. This drop exceeded expectations from changes in the spawning stock biomass (SSB), indicating a change in the stock-recruitment relationship (S-R). We evaluated whether variability in recruitment could be explained by concurrent changes in three environmental indices: sea-surface temperature anomaly (SSTA); southern oscillation index (SOI); and latitudinal position of the west wind drift bifurcation (WWDL). Continuous and discrete effects of these indices were tested as covariates in linear log-log and non-linear Ricker's S-R models. Discrete effects represented regime shifts detected in SSTA (1998), SOI (1998) and WWDL (1995). While SSTA was the only continuous variable found to be significant, the discrete 1998 regime shift supported the most informative model. The best Ricker's model considered a discrete intercept change in the same year: 1998. Although a spurious correlation between SSTA and S-R changes is possible, SSTA may be reflecting major physical or biological changes relevant to M. magellanicus juveniles in the SE Pacific.

Biotic and Abiotic Factors Influencing Cisco Recruitment Dynamics in Lake Superior during 1978–2007

AbstractHistorically, the Cisco Coregonus artedi was abundant throughout the Great Lakes basin, but anthropogenic influences caused the collapse of stocks during the mid‐1900s, and fishery managers are currently exploring options for restoration. To increase understanding of biotic and abiotic factors influencing Cisco recruitment dynamics throughout the Great Lakes, we used Ricker stock–recruitment models with rearing‐habitat‐weighted indices of recruitment and adult spawning stock size in Lake Superior to identify and quantify (1) the appropriate spatial scale for modeling age‐1 recruitment dynamics and (2) the effects of biotic and abiotic factors on age‐1 recruitment dynamics within the regions identified for modeling. Cisco recruitment variation in Lake Superior was best described by a regional model with separate stock–recruitment relationships for western, southern, eastern, and northern regions. The spatial scale for modeling was approximately 260 km. Age‐1 recruitment was negatively correlated with adult spawning stock size in all four regions. Multifactor models suggested (1) a positive correlation between age‐1 recruitment and the interaction between wind speed and air temperature on a lakewide scale and (2) a negative correlation between age‐1 recruitment and the biomass of Rainbow Smelt Osmerus mordax on a regional scale. Large‐scale abiotic factors are beyond the control of fishery managers, so harvest of adult Ciscoes and potential predators and competitors should be carefully managed to achieve desired goals in Lake Superior. Although our study was limited to Cisco stocks in Lake Superior, we believe that our general findings can be more broadly applied (albeit with caution) to the restoration and management of remnant stocks throughout the Great Lakes basin.Received June 4, 2013; accepted August 16, 2013

Comparisons of meta-analytic methods for deriving a probability distribution for the steepness of the stock–recruitment relationship

The steepness parameter of the stock–recruitment relationship (the proportion of unfished recruitment when spawning biomass is reduced to 20% of its unfished level) is a key parameter in stock assessment models, and hence in the provision of scientific management advice for many fisheries. Prior probability distributions for steepness have been used when conducting assessments of US west coast groundfish in the absence of data to estimate steepness reliably. These priors have been developed by applying meta-analytic methods to the results from stock assessments, but the performances of these methods have not been evaluated. Three potential methods for applying meta-analysis to construct steepness priors are available: non-linear mixed models, Bayesian hierarchical methods, and a novel method which approximates marginal likelihoods using likelihood profiles. These methods are evaluated using simulation. The profile method is found to perform best. Estimates of the parameters which define the steepness prior are uncertain owing primarily to uncertainty associated with the results of the stock assessments which provide the input for the meta-analysis methods, and because of the small number of stocks available for inclusion in the meta-analysis.

Reference point based management of Norwegian Atlantic salmon populations

SUMMARYWhile management according to biological reference points is well established for many commercial marine fisheries, similar systems for more leisure based fisheries for freshwater fishes are less common. This paper describes the scientific foundation for management according to conservation limits and management targets for Norwegian populations of Atlantic salmon, a highly valued and heavily exploited anadromous fish species. Based on stock recruitment relationships during the freshwater phase, the biomass of females necessary to attain the carrying capacity (yielding average maximum recruitment) has been established as conservation limits for each of the 439 Norwegian populations. Using a simulation model based on reported catch and estimates of exploitation rates, the probability and percentage attainment of the conservation limits have been assessed annually since 2008, and exploitation advice provided for 176 of the largest populations. The number of populations that attained their conservation limits increased substantially after the new management scheme was introduced, despite that the number of returning salmon remained at historical low levels. Overall the populations evaluated in 2011 were at 95% of their conservation limits compared to 91% in 2008 and 85% in 2005. The improvement could largely be attributed to reduced exploitation rates, due to new restrictions in both the marine and river fisheries. The new management scheme also improved the catch statistics and stimulated data acquisition for management. Implementation of management according to conservation limits has been a success in terms of attaining the main management goal of protecting the Atlantic salmon populations by ensuring that an increasing number of the populations likely are at their maximum reproductive capacity. Long-term increases in fisheries yield, the secondary management goal, are likely to be attained, but remain to be documented. Reference point based management of Atlantic salmon exemplifies management within the intersection of fisheries management and conservation biology, borrowing principles from both sides.

Cost–Benefit Analysis of Alternative Techniques for Rehabilitating Abalone Reefs Depleted by Abalone Viral Ganglioneuritis

The relative cost–benefit of alternative methods for rehabilitating blacklip abalone (Haliotis rubra) reefs depleted by viral ganglioneuritis was evaluated using quantitative population and bio-economic modeling. Fishery assessment models were adapted to describe population dynamics and economics of reseeding, translocation, and natural rebuilding. Past and current literature on reseeding, translocation, and mortality rates informed a stakeholders’ workshop that agreed on the parameters for the analysis. Comparative economic performance was quantified as the impaired value of the individual transferable quota from the area until the stock recovered to the level of maximum sustainable yield, as it was prior to the abalone viral ganglioneuritis impact. When the cost of capital is considered, none of the intervention scenarios produced any cost–benefit above that estimated to accrue from an unfished natural recovery. The translocation of adults and reseeding of juveniles were found to be similarly effective largely due to the productivity of the assumed stock–recruitment relationship at low stock densities. Across the scenarios, translocation was estimated to always at least pay for direct costs but rarely covered capital costs while reseeding only covered direct costs if the price of abalone was >$25/kg and seed was <$0.50 per individual. When a depensatory form of stock–recruitment relationship, related to adult biomass, was assumed, translocation was estimated to be an order of magnitude more effective due to the time required to augment the breeding biomass with reseeded juveniles.

Bio-economic management strategy evaluation of deepwater stocks using the FLBEIA model

Deepwater fish are characterized by long lifespans, late maturity and low productivity. This implies slow recovery from low biomass levels and it is, therefore, important to manage these stocks correctly to avoid overfishing. However, these stocks are generally data poor, which it makes difficult to apply quantitative assessment models on which to base their management. The management strategy evaluation (MSE) approach consists in evaluating the performance of management strategies by simulation before their implementation. In the evaluation, the main sources of uncertainty in a fishery system and its management process should be taken into account in order to find robust management strategies. Thus, the MSE approach is relevant to the management of data-poor stocks. We used the FLBEIA software, an FLR library coded in the R statistical language, to conduct a bio-economic impact assessment of fisheries management strategies. We evaluated the performance of several management strategies in three different deepwater case studies using the MSE approach: beaked redfish (Sebastes mentella) in the Norwegian Sea, blackspot seabream (Pagellus bogaraveo) in the Strait of Gibraltar, and French mixed deepwater fisheries in the North East Atlantic. In conclusion, we found that the S. mentella stock is very sensitive to the stock recruitment relationship chosen. For P. bogaraveo, simulations showed that the present constant catch management is as good as more sophisticated strategies. Finally, in mixed fisheries, case study fleet dynamics were key for the success of management strategies

Stock–recruitment relations controlled by feeding interactions alone

Stock–recruitment (S–R) relationships are the centrepiece of fisheries management aimed at achieving maximum sustainable yield. Here we consider the possibility that the density dependence evident in S–R relations is controlled by feeding interactions alone. We simulate a food-web model with dynamic representations of intra- and inter-specific size structure and a linear relation between food intake and hatchling production of adults. Population sizes of individual stocks are modified by imposing additional mortality. The predominant functional forms and the steepness of resulting S–R relationships agree well with observations. We conclude that recruitment is plausibly regulated by feeding interactions alone.

Development of a Stock–Recruitment Model and Assessment of Biological Reference Points for the Lake Erie Walleye Fishery

AbstractWe developed an updated stock–recruitment relationship for Lake Erie Walleye Sander vitreus using the Akaike information criterion model selection approach. Our best stock–recruitment relationship was a Ricker spawner–recruit function to which spring warming rate was added as an environmental variable, and this regression model explained 39% of the variability in Walleye recruitment over the 1978 through 2006 year‐classes. Thus, most of the variability in Lake Erie Walleye recruitment appeared to be attributable to factors other than spawning stock size and spring warming rate. The abundance of age‐0 Gizzard Shad Dorosoma cepedianum, which was an important term in previous models, may still be an important factor for Walleye recruitment, but poorer ability to monitor Gizzard Shad since the late 1990s could have led to that term failing to appear in our best model. Secondly, we used numerical simulation to demonstrate how to use the stock recruitment relationship to characterize the population dynamics (such as stable age structure, carrying capacity, and maximum sustainable yield) and some biological reference points (such as fishing rates at different important biomass or harvest levels) for an age‐structured population in a deterministic way.Received September 21, 2012; accepted July 2, 2013

A perspective on steepness, reference points, and stock assessment

We provide a perspective on steepness, reference points for fishery management, and stock assessment. We first review published data and give new results showing that key reference points are fixed when steepness and other life history parameters are fixed in stock assessments using a Beverton–Holt stock–recruitment relationship. We use both production and age-structured models to explore these patterns. For the production model, we derive explicit relationships for steepness and life history parameters and then for steepness and major reference points. For the age-structured model, we are required to generally use numerical computation, and so we provide an example that complements the analytical results of the production model. We discuss what it means to set steepness equal to 1 and how to construct a prior for steepness. Ways out of the difficult situation raised by fixing steepness and life history parameters include not fixing them, using a more complicated stock–recruitment relationship, and being more explicit about the information content of the data and what that means for policy makers. We discuss the strengths and limitations of each approach.

A semiparametric Bayesian approach to estimating maximum reproductive rates at low population sizes

The maximum annual reproductive rate (i.e., the slope at the origin in a stock-recruitment relationship) is one of the most important biological reference points in fisheries; it sets the upper limit to sustainable fishing mortality. Estimating the maximum reproductive rate by fitting parametric models to stock-recruitment data may not be a robust approach because two statistically indistinguishable models can generate radically different estimates. To mitigate this issue, we developed a flexible, semiparametric Bayesian approach based on a conditional Gaussian process prior specifically designed to estimate the maximum annual reproductive rate, and applied it to analyze simulated stock-recruitment data sets. Compared with results based on other Gaussian process priors, we found that the conditional Gaussian process prior provided superior results: the accuracy and precision of estimates were enhanced without increasing model complexity. Moreover, compared with parametric alternatives, performance of the conditional Gaussian process prior was comparable to that of the data-generating model and always better than the wrong model.

Demographic effects on the use of genetic options for the control of mosquitofish, Gambusia holbrooki

This study tests the sensitivity of genetically based pest control options based on sex ratio distortion to intra- and intersexual aggressive interactions that affect male and female survival and fitness. Data on these interactions and their impacts were gathered for the mosquitofish Gambusia holbrooki (Poeciliidae), a promiscuous species with a strongly male-biased operational sex ratio and well-documented male harassment of females. The experimental design consisted of an orthogonal combination of two population densities and three sex ratios, ranging from strongly male-biased to strongly female-biased, and long-term observations of laboratory populations. Contrary to expectations, the number of males in a population had little evident effect on population demographics. Rather, the density of adult females determined population fecundity (as a result of a stock-recruitment relationship involving females, but not males), constrained male densities (apparently as a result of cannibalism or intersexual aggression), and regulated itself (most likely through effects of intrasexual aggression on female recruitment). The principal effect of males was to constrain their own densities via effects of male-male aggression on adult male mortality rates. Through use of a realistically parameterized genetic/demographic model, we show that of three different genetic options applied to control G. holbrooki, one based on recombinant sex ratio distortion (release of Female Lethal carriers) is marginally more efficient than a sterile male release program, and both outperform an option based on chromosomal sex ratio distortion (Trojan W). Nonlinear dependence of reproductive rate on female density reduces the efficacy of all three approaches. The major effect of intra- and intersexual aggression is mediated through females, whose interactions reduce female numbers and increase the efficacy of a control program based on sex ratio. Socially mediated male mortality has a small impact on control programs due to operational sex ratios that are heavily male-biased. The sensitivity of sex ratio-based control options to social factors will depend on the mating system of the targeted pest, but evidence of widespread density-dependent population regulation suggests that, for most species, the effects of elevated adult mortality (due to intra- and intersexual aggression) on control programs are likely to be slight.

A stock–recruitment model for highly fecund species based on temporal and spatial extent of spawning

A stock–recruitment model is described for highly fecund species based on the contraction of the spatial and temporal extent of spawning when a population is reduced in size: R=αK(1−exp[−S/(mK)]), where S is the number of spawners, K is the carrying capacity in units of the number of habitat patches that can produce recruits, α is the average number of recruits per unit of habitat, and m is the number of spawners that group together to spawn. The model is based on three simplifying assumptions: (1) the environment is divided into K units; (2) the presence of one spawner provides sufficient eggs to fill the capacity of that unit, any additional spawners in that unit will not increase recruitment; and (3) groups of fish are randomly distributed over the environment. The model allows for a flat top curve, which is consistent with highly fecund species that are continuous spawners and do not aggregate to spawn. It also allows for a strong relationship between spawners and recruits, which is more consistent with species that aggregate in time and space to spawn. This stock–recruitment model can be approximated in terms of parameters commonly used in contemporary stock assessment models (virgin recruitment, R0, and steepness of the stock–recruitment relationship, h, virgin spawning biomass, S0): R=R0(1−exp(5ln(1−h)S/S0)). The functional form is compared with the Beverton–Holt stock–recruitment model.

Recruitment in the Barents Sea, Icelandic, and eastern Newfoundland/Labrador capelin (Mallotus villosus) stocks

Capelin (Mallotus villosus) is a short-lived, coldwater, pelagic species that occurs in large populations in the Barents Sea, around Iceland and in the Newfoundland–Labrador area in the North Atlantic. Most individuals spawn only once at age three or four and die shortly after spawning. The commercial fisheries for capelin in the three areas are prosecuted on the pre-spawning and spawning age groups and knowledge of recruitment to the spawning stock and factors affecting recruitment are important in the management of the fisheries. Herein, we review the state of knowledge regarding recruitment in the three areas and factors that affect recruitment.Year class strength of capelin is fixed fairly early in life; at 0-group stage in the Barents Sea, by age 1 or earlier for Icelandic capelin, and within 2weeks of hatching for the beach spawners in the Newfoundland area. No obvious single factor, common to the three stocks, has been proven to exert a significant influence on recruitment. In the Barents Sea, juvenile herring prey heavily on capelin larvae and when the estimates of the abundance of these juvenile herring were accounted for in the capelin stock–recruitment analysis, the fit in the stock–recruitment relationship improved. In Iceland, there is no reason to implicate predation as a dominant factor, although detailed studies have not been carried out. In the Newfoundland and Labrador area earlier studies suggested two factors were important for recruitment success: the frequency of onshore winds at a critical period during the residence of newly hatched larvae in the beach gravel and the presence of warm waters during the early larval phase in the pelagic environment. With further testing using more data, the temperature factor was found not to be significant but the frequency of onshore winds still explained a significant portion of the variation in year class strength. There appears to be no significant biological influence, such as predation, that has influenced recruitment in this area.

Maximum survival of eggs as the key parameter of stock–recruit meta-analysis: accounting for parameter and structural uncertainty

Despite their name, hierarchical stock–recruit meta-analyses are often parameterized in terms of steepness, which depends not only on the assumed stock–recruitment relationship but also on the recruit–spawner relationship. This parameterization requires assumptions about the reproductive potential of the recruit that are not desirable if the focus of the study is limited to the spawning–recruitment phase instead of the full life cycle. Thus, usage of steepness should be avoided in studies that aim to produce informative priors for the stock–recruit relationship for use in studies of other salmon stocks. An alternative key parameter for stock–recruit models is the maximum survival of eggs, which is the slope at the origin of the stock–recruitment curve when spawning stock size is defined in terms of the number of eggs. Furthermore, the current widely used practices in stock–recruit modeling could be improved by taking into account the stock-specific model uncertainty. We use the method of Bayesian model averaging to build a hierarchical stock–recruit model that allows stock-specific model structures with Beverton–Holt, Ricker, and hockey stick models as alternatives, all of which can be parameterized with the maximum survival of eggs. We illustrate our approach by analyzing nine previously published datasets for Atlantic salmon (Salmo salar).

A recruitment forecasting model for the Pacific stock of the Japanese sardine (&amp;lt;i&amp;gt;Sardinops melanostictus&amp;lt;/i&amp;gt;) that does not assume density-dependent effects

This study developed a recruitment forecasting model based on a new concept of the stock recruitment relationship. No density-dependent effect in the relationship was assumed in the model, which showed that fluctuations in recruitment and spawning stock biomass of Japanese sardine in the northwestern Pacific can be explained mainly by environmental factors and the effects of fishing. The February Arctic Oscillation (AO) and sea surface temperature over the southern area of the Kuroshio Extension (30 - 35°N and 145 - 180°E; KEST) were used as the environmental factors. The recruitment forecasting model is proposed: The values for recruitment (), spawning stock biomass, (), in year t, forecast by this model accurately reproduced those estimated by tuning virtual population analysis (VPA), and the pattern of variability in the stock recruitment relationship was also reproduced well. In conclusion, a density-dependent effect does not necessarily have to be included to explain the large variations in recruitment and the spawning stock biomass of the Japanese sardine.

Construction of biological reference points for management of the Dungeness crab, Cancer magister, fishery in the Fraser River Delta, British Columbia, Canada

The objective of this study was to define two biological reference points, the Upper Stock Reference Point (USR) and the Limit Reference Point (LRP), to facilitate management of the Fraser River delta commercial Dungeness crab (Cancer magister) fishery in a precautionary manner. The USR is the stock level below which the removal rate is progressively reduced to avoid reaching the LRP. The LRP is the stock level below which productivity is sufficiently impaired to cause serious harm. Currently, the commercial Dungeness crab fishery in the Fraser River delta is managed with a minimum size limit (165mm carapace width point-to-point), non-retention of females and soft-shell crabs, and a restricted fishing season (June to November) to protect large males during the winter/spring moult period. We evaluated variations in legal male, sublegal male, and female Dungeness crab abundances using fishery-independent trap surveys from 1988 to 2010. We also examined proportions of soft-shell legal male and female crabs using survey data and information collected from commercial fishing. A relationship existed between female abundance and corresponding legal male abundance five years later, which is well described by the Beverton–Holt stock–recruitment model. We conducted simulations to assess possible adverse impacts of fishing on the crab population. We assumed the population would stabilize at the existing abundance upon suspension of commercial harvesting if abundance has been steadily declining. Using the stock–recruitment relationship along with information on commercial catch compositions and soft-shell proportions, we assessed the potential reduction in legal male and female crab abundances, given persistent fishing, before the population reaches a lower equilibrium. The amount of reduction to legal male and female crabs increases with decreasing population size. If legal male crab abundance is 50% and 20% of maximum possible recruitment, then fishing would reduce female abundance by 20% and 48%, respectively, before the population is stabilized. We propose these two female abundance levels as the USR and LRP for managing the Fraser River commercial Dungeness crab fishery.

Exploring the Influence of Stock–Recruitment Relationships and Environmental Variables on Black Bass and Crappie Recruitment Dynamics in Missouri Reservoirs

AbstractLarge reservoirs provide important sport fisheries, but managing these fisheries is difficult because of the multiple biological and environmental variables that interact to shape them. Understanding how sport fish recruitment responds to parental stock and environmental influences would improve our management capabilities. We compiled long‐term datasets for 15 Missouri reservoirs and used them to examine the influence of parental stock abundance and a suite of environmental variables on the recruitment of Largemouth Bass Micropterus salmoides, Spotted Bass M. punctulatus, White Crappie Pomoxis annularis, and Black Crappie P. nigromaculatus. Comparisons of log‐linear, Ricker, and Beverton–Holt stock–recruit models revealed that the log‐linear model was the most parsimonious for Largemouth Bass, White Crappie, and Black Crappie, but the Ricker model was best for Spotted Bass. However, stock–recruit models alone explained less than 13% of the variation in Largemouth Bass, Spotted Bass, and White Crappie recruitment while explaining about 33% of the variability in recruitment of Black Crappie. Largemouth Bass recruitment was positively related to spring rises in water levels, summer water levels, and average summer air temperatures, whereas Spotted Bass recruitment was positively related to spring air temperatures. White Crappie recruitment was positively related to total phosphorus levels and drops in spring water levels. Black Crappie recruitment was best explained by the model positively relating recruitment to spawning stock abundance, spring water levels, and summer water levels, while negatively relating recruitment to rapid drops in spring water levels. With the exception of models for Black Crappie, the best models did not include spawning stock abundance, and environmental variables appeared more important in explaining variation in recruitment. However, environmental variables in these models explained only a modest amount of variation (27–44%) in recruitment, revealing the complexity of recruitment processes in large reservoirs.