Cooper and Mangel: Metapopulation structure in the conservation of salmonids 



221 



Distance 



Figure 5 



Average year in which sinks attained their maximum population size as a function of the distance to the nearest 

 source. The year in which a sink population attains its maximum size depends upon the distance to the closest 

 source, but there is considerable variation in dependence. Error bars represent one standard deviation. 



initially increase but then decrease as habitat deg- 

 radation progresses. Some demes in sink habitats, 

 such as demes 1 and 4, were relatively stable if not 

 increasing over a fair portion of the simulation pe- 

 riod before declining towards extinction. 



Discussion 



Undetected metapopulation structure in salmonid 

 populations may obscure the signals that managers 

 use to determine the need for convervation action. 

 Abundance trends, either absolute numbers or simple 

 indices of abundance, constitute the primary input 

 into analyses used for fisheries management and 

 decision-making (Hilborn and Walters, 1992). The 

 majority of these techniques assume that the popu- 

 lation in question is a closed system, that any immi- 

 gration or emigration can be considered negligible. 

 In California, Oregon, and Washington, estimates 

 of population or run size for most salmonid species 

 are commonly based on number of spawning fish or 

 redd counts taken from index reaches or streams 

 ( WDFW, 1994; ODFW, 1995; Weitkamp et al., 1995). 



The absolute numbers of spawning fish or redds 

 counted in these index streams, which incorporate 

 only a minute portion of the available spawning habi- 

 tat within the watershed, are then used to extrapo- 

 late watershed- or basin-level abundance estimates, 

 the very estimates upon which managers base their 

 decisions. In fact, in their discussion on the data 

 available for assessing the population size and risk 

 of extinction for coho salmon along the west coast of 

 the United States, Weitkamp et al. (1995, p. 106) 

 stated ". . . where [stream] surveys were conducted, 

 they are the best local indication we have of popula- 

 tion abundance trends." Index reaches and streams 

 such as these are precisely the situation modeled in 

 the specific examples discussed in this paper; stray- 

 ing between these reaches is likely and differential 

 reproductive success between strays and natal fish 

 may not occur. The NMFS's analysis of coho salmon 

 populations considered short- and long-term trends 

 in abundance to be the main indicators for the risk of 

 extinction but avoided using estimates based on index 

 streams (Weitkamp et al., 1995). Using the trends in 

 abundance for an entire ESU, Weitkamp et al. (1995) 

 attempted to avoid the problems associated with the 



