 |
||||
|
||||
|
|
|
 |
||
|
||||
      |
Modeling Historic and Spatial Variation of Great Lakes Fish Maturation Schedules
This project is no longer current. Please see the Research Programs page for a list of current research projects.
Tomas
Hook
Cooperative Institute for Limnology and Ecosystems Research (CILER
web site)
Collaborators
Scott Peacor , Great
Lakes Environmental Research Laboratory
Hui-Yu Wang, Cooperative Institute for Limnology and Ecosystems Research
Executive Summary
Harvesting of fish stocks removes individuals based upon traits including size, behavior, and location. Selective harvesting could therefore alter genetic composition, and negatively affect growth, maturation schedules, and whole stock productive capacity. Great Lakes fishes are selectively harvested based on size. This research will evaluate if maturation schedules of Great Lakes fishes have changed over time, estimate genetically-based effects of selective harvesting, and provide recommendations for minimizing these effects.
2006 Plans
- Assemble historic data on Great Lakes whitefish and walleye maturation schedules.
- Analyze maturation data (fit spatially and temporally specific reaction norms; see below).
- Develop a generalized individual-based model (with an inheritance sub-routine) to simulate long-term population genetic composition and stock productivity.
- Use individual-based model to explore the consequences of various harvesting scenarios on long-term population genetic composition and stock productivity.
Project Rationale
There
is increasing evidence that genetic selection occurs at a time scale relevant
to population dynamics and fisheries management. Further, ecological processes
(e.g. intra-specific competition, interacting species dynamics) can influence
genetic selection and phenotypic expression within a population. These
linkages between ecology and evolution are particularly relevant for fisheries
management. Fish are generally selectively harvested based upon individual
traits such as size, behavior, and location. Such selective harvesting
can alter the genetic composition of the population and affect growth
rates, maturation schedules, and productive capacity of the exploited
fish stock. In order to adequately mange fish populations for sustainable
yields and genetic diversity, it is necessary to simultaneously consider
ecological interactions and genetic selection.
The
selective harvest of commercially and recreationally important Great Lakes
fish stocks could lead to the evolution of altered growth rates and maturation
schedules. Specifically, harvest of large individuals should favor slower
growth and maturation at earlier age and smaller size. This possibility
should be assessed as such harvest-induced evolution could decrease the
productive capacity of fish stocks, and may (as suggested for Atlantic
cod) be a harbinger of stock and recruitment over-fishing. However, as
growth and age/size at maturity are both genetically and environmentally
determined it is not straightforward to attribute trends in these measures
to genetic selection alone. In addition, there is currently not a defined
modeling framework to link stock harvest with genetic selection and future
stock productivity.
This project will evaluate the hypothesis that genetically-determined maturation schedules of Great Lakes fishes vary temporally and spatially, and will provide recommendations for how fish may be harvested in a manner to minimize fishery-induced changes to maturation schedules. We will compare probabilities of maturity among Great Lakes whitefish and walleye stocks (i.e., spatially-distinct populations experiencing different harvest dynamics and abiotic/biotic environments). In order to account for environmental effects (e.g. temperature, food supply) on size/age at maturity, we will fit mean reaction norm parameters (see below) for specific sub-stocks and time periods, based on data on size at age and size/age at maturity. We will compare these parameters among locations and time periods to test if genetically-defined reaction norms dictating size/age at maturity vary spatially and temporally. Finally, we will develop a generalized individual-based genetic-ecological model, and we will use this simulation model to predict how maturation schedules and stock productivity may change under various harvesting and heritability scenarios.
We will assemble historic and present-day data on individual size, age, sex, and maturity status for whitefish and walleye collected during late summer and fall (the time period when maturity can be assessed). These data are available through historic and on-going efforts by the state, provincial and tribal natural resource agencies and include fishery independent and dependent surveys. Most data are already in electronic form and can therefore be easily assembled. However, some historic data exist only as hard copies and will be manually integrated with other data.
Upon assembling these data, we will calculate age and size specific probabilities of maturity for discrete time periods and locations. These metrics are used extensively in fisheries management models, but are of little use for assessing population-level genetic changes over time. Individual phenotypic expression (e.g., age or size at maturity) is a function of both individual genetics and environmental (both biotic and abiotic) interactions. A reaction norm (RN) is a genetic algorithm which dictates how individual phenotypic expression changes with environmental experiences. A probabilistic maturity reaction norm is typically represented as a response curve based on age (x-axis) and size (y-axis). Parameters describing RNs are essentially heritable, and selection will alter the frequencies of such parameters. We will estimate mean population-level RNs based on historic and current data on size, age, sex, and maturity status, using established statistical approaches (Heino et al. 2002a; Heino et al. 2002b). For a specific location or time-period, we will group individuals by age and fit age-specific logistic regression equations to estimate size at 50% maturity. We will then fit a curve representing the 50% maturity RN through the collection of logistic regression estimates for size (y-axis) at maturity, given age (x-axis). Finally, we will statistically compare RNs among locations and time periods, thereby allowing us to test the hypothesis that genetically-determined lake whitefish maturation schedules vary temporally and spatially.
Figure 1: Two hypothetical probabilistic reaction norms (RN1 and RN2; bold solid curves represent reaction norm medians, i.e., 50% maturation probability; thin dashed lines represent 25% and 75% maturation probabilities). Gray lines represent two potential growth trajectories (A and B). Maturation occurs (based on defined probabilities) when a growth trajectory crosses the maturation reaction norm. Thus, for either of these two reaction norms maturation occurs earlier and at a larger size for the faster growth trajectory (A). Also, note that for either growth trajectory maturation occurs earlier and at smaller size based on RN2 compared to RN1. Thus, a shift from RN1 towards RN2 may be expected as a result of intensive size-selective fishery harvest.
We will use an individual-based simulation model (IBM) and evolving reaction norms (i.e., include an inheritance subroutine allowing individuals to pass on adaptive traits to subsequent generations) to predict how maturation schedules may change under various harvesting and heritability scenarios. Initial model reaction norm parameters will be derived from statistical analyses described above. Our approach involves modelling the experiences and subsequent modifications of multiple individuals and integrating individual-level information to consider population-level traits. This approach acknowledges that ecological interactions occur at the individual-level, and is particularly useful when modelling populations with highly size-dependent interactions as is the case for most fish populations. For similar reasons, this modelling approach should ultimately prove advantageous for integrating evolutionary processes into population models. Evolutionary selective forces operate at the individual level and many such forces are size-dependent (e.g., model harvest of whitefish will be size-dependent). We will use the model to examine how past and potential future fishing practices may alter maturation schedules, and how such changes may impact the productive capacity of the stock.
