Healthy salmon populations have complex population structures characterized by the return of multiple age classes and by milted mixing between neighboring populations. This variety in life history patterns allows them to persist through periodic disasters and unfavorable environmental conditions and to colonize newly available areas.
The complex structure of salmon populations most likely functions as a metapopulation. A metapopulation is a subdivided population, with each subdivision having its own internal dynamics, yet connected to other subdivisions by dispersal (Hanski and Gilpin 1991). The ability of naturally reproducing populations to persist and flourish in a fluctuating environment depends on demographic factors, genetic factors, and interactions between local populations. Properly describing this population structure and the interrelationships is a complex task based upon examining the statistical properties of a number of phenotypic, genetic, and behavioral information (NRC 1995). While a multi-attribute approach to describing stock structure is a difficult task, it has the advantage of avoiding the shortcomings of any single approach and more accurately describes the important traits which have allowed these fish to persist for thousands of years.
The lack of data necessary to describe metapopulation structure has led to management uncertainty. This has been particularly true when developing management strategies to conserve small populations. Current strategies often rely heavily on one aspect such as genetic factors, while not accounting for extreme demographic risk.
Lack of information has also led to speculation and disagreement about the impacts of using hatchery technology as a tool to increase natural salmon populations. Certainly, traditional production hatchery practices have resulted in behavioral and genetic changes in the cultured fish compared to the original population. Radically different hatchery practices, as proposed by the tribes to deal with the immediate risks of further extirpations, will be designed to minimize potential genetic effects. On the other hand, the use of captive breeding practices using all of the few remaining adults from a population might keep the populations from being extirpated, but would almost certainly have a greater genetic and behavioral impact than earlier intervention using a low-impact supplementation approach. In either case, there is a need to track the medium- and long-term genetic and behavioral effects of hatchery technology on naturally spawning populations. The information obtained over time will help describe the population structure for salmon and evaluate the effects of hatchery techniques, so that natural and hatchery production can be wisely managed over the long term.